music-composition

어떤 스킬인가요?

Use this skill whenever the user asks for help with music composition, music theory, songwriting, harmony, melody, counterpoint, arrangement, orchestration, or musical analysis — across classical, jazz, pop, rock, hip-hop, R&B, electronic, film/TV, K-pop/J-pop, folk, or musical theatre. Triggers include explicit theory questions ("what scale fits over Cm7", "explain ii-V-I", "modulate from C to E♭"), craft questions ("write a chord progression for a melancholy bridge", "make this melody more memorable", "voice this chord on piano"), and vague creative problems that map to compositional techniques ("the chorus feels weak", "this transition is awkward", "my song sounds generic"). Use this skill even when the user describes the goal in non-technical terms — translating fuzzy creative problems into concrete compositional moves is exactly what this skill is for. Do NOT use for DAW operation, MIDI file generation, audio engineering, mixing/mastering, sound design, or notation software UI.

설치 방법

프로젝트 폴더에서 아래 명령어를 실행하고, 설치할 에이전트를 선택하세요.

설치 명령을 직접 실행해야 적용됩니다. 지원 에이전트와 필요한 권한·라이선스는 제작자의 안내를 확인하세요.

스킬 원문 파일

# Python
__pycache__/
*.py[cod]
*$py.class

# OS
.DS_Store
Thumbs.db
desktop.ini

# Editors / IDEs
.vscode/
.idea/
*.swp
*~

# Benchmark raw outputs (timestamped per run; canonical eval lives in v1.0-eval.md)
benchmarks/results/

# Local env
.env
.env.local

# License

This repository uses a split license.

## 1. Documentation and skill content

Unless otherwise noted, the Markdown documents, reference materials, skill instructions, assets, release notes, validation reports, and other non-code written materials in this repository are licensed under:

**Creative Commons Attribution 4.0 International (CC BY 4.0)**  
SPDX-License-Identifier: `CC-BY-4.0`

Copyright (c) 2026 SJY051 and contributors.

You are free to:

- share — copy and redistribute the material in any medium or format;
- adapt — remix, transform, and build upon the material for any purpose, including commercial purposes;

provided that you give appropriate credit, provide a link to the license, and indicate if changes were made.

Full license text: https://creativecommons.org/licenses/by/4.0/legalcode

Recommended attribution format:

```text
Based on “Music Composition Agent Skill”, licensed under CC BY 4.0.
```

Or, with version information:

```text
Based on “Music Composition Agent Skill v1.0” by the Music Composition Agent Skill contributors, licensed under CC BY 4.0.
```

## 2. Code

Code files in `scripts/` are licensed under the MIT License.

SPDX-License-Identifier: `MIT`

MIT License

Copyright (c) 2026 SJY051 and contributors

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the “Software”), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

## 3. Third-party material

References to artists, songs, genres, platforms, products, institutions, charts, music services, and other third-party names are for identification, commentary, analysis, and research-routing purposes only. They do not imply endorsement, affiliation, or ownership.

This repository does not grant rights to any third-party copyrighted music, lyrics, recordings, samples, trademarks, logos, or protected expressive material.

## 4. Style and reference use

The skill is intended to help users create original music and analyze compositional craft. It should not be used to copy protected expression from existing songs or artists, including but not limited to melodies, lyrics, distinctive riffs, samples, signature producer tags, or highly specific vocal identities.

## 5. No legal advice

This license file and the repository documentation are provided for general project governance and do not constitute legal advice. For legal questions about copyright, licensing, music publishing, sampling, or commercial release, consult a qualified legal professional.

# Known Limitations for First Release

This file records what the first release of the music-composition skill is **not** trying to solve yet. It should be shipped with release candidates so maintainers and evaluators can distinguish genuine bugs from intentionally deferred scope.

## First-release boundary

The first release is designed to be a strong composition assistant: it can help users brainstorm, draft, revise, diagnose, analyze, learn, and prepare reference-informed musical ideas. It is not a complete music encyclopedia, a DAW manual, a legal service, or a replacement for expert performers and tradition-bearers.

Use this file with `RELEASE-ROADMAP.md`, `references/validation/first-release-readiness.md`, and `references/validation/phase-c-smoke-test-results.md` when preparing a release candidate.

## Known limitations

| Area | Limitation | Expected behavior |
|---|---|---|
| Current trends | Static genre files cannot stay current on charts, platform norms, viral scenes, or artist activity. | Use `references/research/web-music-trend-research.md` and active web verification for current claims. |
| User listening data | The skill cannot access private streaming libraries or account data by itself. | Ask only for safe user-provided evidence: public playlist links, track lists, Replay/Wrapped summaries, or screenshots. Never request credentials. |
| Copyright / style | The skill cannot guarantee legal clearance or produce legal advice. | Avoid copying melody, lyrics, riffs, samples, vocal identity, signature tags, or protected expression. Extract craft variables instead. |
| Cultural / traditional music | Starter guides are not substitutes for practitioner knowledge, language expertise, or tradition-specific study. | Name the tradition, scene, language, rhythm, ornament, function, and collaboration need instead of using broad labels. |
| Audio analysis | The skill does not transcribe audio directly, detect exact chords from recordings, or verify every famous-song progression. | Treat user-provided charts as approximate unless a reliable source is supplied; flag uncertainty. |
| MIDI / notation generation | This skill is text guidance, not a MIDI or score-generation engine. | Provide chord charts, scale degrees, form maps, rhythm grids, or notation-like sketches. |
| DAW / production operation | It does not teach DAW UI, plugin operation, detailed synthesis, mixing, or mastering. | Stay composition-facing: arrangement density, register, energy, and pre-production decisions. |
| Performance pedagogy | Instrument-idiom files are about writing playable parts, not training performers. | Give playability constraints and idiomatic writing moves, not technique-practice regimens. |
| Full genre coverage | Many scenes remain covered only by starter or dispatch documents. | Work from available genre files plus research workflows; say when a dedicated file does not exist. |
| Automated validation | The sanity checker catches regressions and broken links, not all musical truth. | Use manual review for nuanced harmony, genre, cultural, and currentness claims. |

## Non-blocking gaps for RC1

The following are useful but should not block a first release:

- dedicated deep guides for regional Mexican, Punjabi pop, Turkish arabesk-pop, Maghreb / raï, gqom, singeli, kuduro, sertanejo, pagode, arrocha, piseiro, and related local scenes;
- additional instrument idiom files for harp, mallet percussion, organ, accordion, synth performance idioms, and DJ / turntable-informed writing;
- a larger prompt-evaluation harness with automatically recorded routing traces;
- a deeper symbolic music-theory parser covering every chord symbol, Nashville number, and enharmonic edge case;
- comprehensive source citations inside every reference file;
- notation-software, DAW, audio-transcription, and MIDI-export workflows.

## When a limitation becomes a release blocker

A limitation becomes a blocker if it causes the agent to answer with false confidence rather than scoped uncertainty.

| Situation | Treat as blocker? | Why |
|---|---|---|
| A missing dedicated genre file, but the agent says it has one | Yes | Misrepresents coverage. |
| A current trend answer without web verification | Yes, when currentness matters | Static files should not freeze living scenes. |
| A cultural shortcut such as generic “world” or broad culture-as-color guidance | Yes | Violates the cultural-specificity principle. |
| A stable-theory answer that asks for web research unnecessarily | Usually no | Annoying behavior, but not false; improve routing if frequent. |
| A reference-style answer that copies a melody/riff/lyric/sample | Yes | Violates originality and safety guardrails. |
| A validation report claims unrun tests passed | Yes | Breaks release trust. |

## Recommended wording when a gap appears

Use a compact disclosure rather than a long apology:

```md
This skill does not have a dedicated guide for [scene] yet. I’ll treat it as a current scene question, verify recent references, and translate what I find into tempo, groove, harmony, topline, form, and arrangement variables.
```

For tradition-specific requests:

```md
I can sketch a respectful starting point, but this is not a substitute for a practitioner or native-language collaborator. I’ll name which musical variables I’m borrowing and avoid treating the tradition as generic color.
```

For exact transcription requests:

```md
I can analyze the chart or notes you provide, but I won’t claim this is the exact recording transcription unless you provide a reliable score or source.
```

## Cross-references

- Release roadmap → `RELEASE-ROADMAP.md`
- First-release readiness → `references/validation/first-release-readiness.md`
- Phase C results → `references/validation/phase-c-smoke-test-results.md`
- Web trend research → `references/research/web-music-trend-research.md`
- User listening context → `references/research/user-listening-context-and-streaming-services.md`
- Style/reference guardrails → `references/research/style-reference-and-copyright.md`
- Source bibliography → `references/source-bibliography.md`

# RC1 → v1.0 / RC2 Decision Memo

Date: 2026-04-27  
Package reviewed: `music-composition-RC1-2026-04-27.zip`

## Decision

**Promote RC1 to v1.0. Do not create RC2 unless a new release blocker appears during final v1.0 packaging.**

RC2 is not recommended right now. The current remaining work is mostly expansion, deeper evaluation, or tooling polish, not first-release blocking correction.

## Evidence reviewed

- `VALIDATION-RC1.md`
- `RELEASE-NOTES-RC1.md`
- `RELEASE-ROADMAP.md`
- `KNOWN-LIMITATIONS.md`
- `references/validation/phase-c-smoke-test-results.md`
- RC1 package file list and validation structure
- Local sanity-check rerun from extracted RC1 package

Local sanity-check rerun:

```text
music_theory_sanity_check: PASS
Checked 112 markdown files.
Elapsed: 0:25.98
```

The checker runtime is somewhat noticeable, but acceptable for RC1/v1.0 validation. Optimizing it can be treated as post-v1.0 tooling polish unless it becomes disruptive in routine use.

## Why v1.0 is the better next step

RC1 has already cleared the gates that matter for a first release:

| Gate | Status | Decision impact |
|---|---|---|
| Phase A content lock | Complete | No RC2 needed |
| Phase B correctness pass | Complete | No RC2 needed |
| Phase C prompt smoke tests | Passed by user report | No RC2 needed |
| Phase D packaging | Complete | No RC2 needed |
| Sanity checker | Re-ran and passed | No RC2 needed |
| ZIP/package structure | Already validated in RC1 | No RC2 needed |
| Known limitations | Explicitly shipped | No RC2 needed |

The release roadmap already says to hold after RC1 only if a release blocker appears. No such blocker is currently visible.

## Why RC2 is not recommended now

A new RC should fix a meaningful release-risk problem. At this point, the likely RC2 candidates would be:

- more regional or microgenre guides;
- a deeper chord-symbol / scale-degree parser;
- faster validation tooling;
- broader prompt-evaluation harnesses;
- more source mapping and bibliography detail;
- additional instrument-idiom files.

All of these are useful, but they are better as v1.1 / v1.2 work. Adding them before v1.0 would increase churn without reducing a known first-release risk.

## Final v1.0 packaging scope

The v1.0 pass should be **metadata-only / release-finalization only**:

- add or rename final release notes as `RELEASE-NOTES-v1.0.md`;
- add `VALIDATION-v1.0.md`;
- update `RELEASE-ROADMAP.md` from “RC1 packaged” to “v1.0 shipped”;
- update `MAINTENANCE.md` release artifact list;
- optionally keep RC1 notes for traceability or replace them with v1.0 notes, but avoid duplicate/confusing release-state language;
- generate `music-composition-v1.0-2026-04-27.zip`, patch, and SHA256 file;
- do not add new runtime content during v1.0 packaging.

## RC2 trigger conditions

Create RC2 instead of v1.0 only if one of these appears during the final packaging pass:

1. sanity checker fails;
2. ZIP integrity or hygiene fails;
3. internal `.md` links break;
4. a newly found P0/P1 music-theory error appears in high-use assets;
5. a privacy/currentness/style/copyright/cultural-specificity blocker appears;
6. a serious mismatch appears between `SKILL.md`, `references/00-navigation.md`, and actual files;
7. a validation report claims a check passed when it was not actually run or cannot be reproduced;
8. a Phase C failure pattern reappears and affects core routing or answer shape.

## Post-v1.0 backlog

Treat these as v1.1 / v1.2 candidates:

- dedicated deep guides for regional Mexican, Punjabi pop, Turkish arabesk-pop, Maghreb/raï, gqom, singeli, kuduro, sertanejo/pagode/arrocha/piseiro;
- additional instrument idiom files for harp, mallet percussion, organ, accordion, synth performance idioms, and DJ/turntable-informed writing;
- larger prompt-evaluation harness with routing traces;
- deeper symbolic music-theory parser for chord symbols, Nashville numbers, altered dominants, modal mixture, and scale-degree spelling;
- validation script performance optimization;
- more source citations and bibliography mapping for high-risk cultural/currentness files;
- accessibility and wellness-oriented composition guidance.

## Final recommendation

**Proceed directly from RC1 to v1.0 packaging.**

Hold RC2 in reserve only for blocker fixes discovered during the final packaging pass.

# Release Notes — v1.0

Release: **music-composition v1.0**  
Release date: **2026-04-27**  
Base package: `music-composition-RC1-2026-04-27.zip`  
Release type: **first stable release**

## Release decision

v1.0 promotes RC1 to the first stable release. No RC2 was created because Phase B correctness work is complete, Phase C prompt smoke testing passed by user report, and the release-candidate packaging checks did not reveal a release blocker.

This release is intended to be a stable, practical music-composition skill for brainstorming, drafting, analysis, revision, teaching, reference research, and user-agent composition sessions. It is not intended to be a complete encyclopedia of every style, region, instrument, platform, or legal edge case.

## What changed from RC1

v1.0 does **not** add new runtime music content beyond RC1. The finalization pass adds release-final artifacts and updates release status wording only.

Added:

- `RELEASE-NOTES-v1.0.md`
- `VALIDATION-v1.0.md`
- `RC1-v1.0-decision-memo-2026-04-27.md`

Updated:

- `MAINTENANCE.md`
- `RELEASE-ROADMAP.md`
- `references/00-navigation.md`
- `references/validation/first-release-readiness.md`
- `scripts/music_theory_sanity_check.py`

The updates are release-finalization / validation updates, not new composition behavior.

## What is included

v1.0 includes:

- `SKILL.md` as the runtime entry point;
- `references/00-navigation.md` as the routing guide;
- short, quotable assets in `assets/`;
- longer theory, genre, workflow, research, orchestration, instrument-idiom, and validation references in `references/`;
- music-specific brainstorming, collaboration, revision, answer-calibration, and session-log workflows;
- web research, reference-track digging, user listening context, regional trend analysis, microgenre dispatch, and style/originality guardrails;
- genre and instrument-idiom starter guides;
- `KNOWN-LIMITATIONS.md`;
- `scripts/music_theory_sanity_check.py`.

## Release-readiness summary

| Area | v1.0 status |
|---|---|
| Phase A content lock | Complete |
| Phase B correctness pass | Complete |
| Phase C prompt smoke tests | Pass result recorded |
| Phase D RC packaging | Complete |
| v1.0 finalization | Complete |
| Automated sanity checker | Pass |
| Internal Markdown links | Pass |
| ZIP hygiene | Pass |
| Known limitations | Included |

## How to validate locally

From the repository root:

```bash
python3 scripts/music_theory_sanity_check.py
```

Expected output for this package:

```text
music_theory_sanity_check: PASS
Checked 115 markdown files.
```

If a future maintainer adds or removes Markdown files, the count may change; the validation report should then be updated.

## What to read first

For maintainers:

1. `RELEASE-ROADMAP.md`
2. `VALIDATION-v1.0.md`
3. `KNOWN-LIMITATIONS.md`
4. `references/validation/first-release-readiness.md`
5. `references/validation/phase-c-smoke-test-results.md`

For runtime behavior:

1. `SKILL.md`
2. `references/00-navigation.md`
3. the relevant files selected by the navigation guide

## Known limitations

See `KNOWN-LIMITATIONS.md`.

The most important caveat is that current music trends, artist activity, chart behavior, platform surfaces, streaming-service behavior, AI tools, copyright status, and living scenes must still be handled through active research workflows rather than frozen static claims.

## Post-release direction

After v1.0, treat new work as update patches rather than first-release blockers. Good next updates include dedicated regional guides, deeper parser-like validation, larger prompt-evaluation harnesses, more instrument idiom files, and additional currentness/research refinements.

## Cross-references

- Entry point → `SKILL.md`
- Maintenance guide → `MAINTENANCE.md`
- Release roadmap → `RELEASE-ROADMAP.md`
- v1.0 validation report → `VALIDATION-v1.0.md`
- RC1 → v1.0 decision memo → `RC1-v1.0-decision-memo-2026-04-27.md`
- RC1 validation report → `VALIDATION-RC1.md`
- Known limitations → `KNOWN-LIMITATIONS.md`
- Phase C results → `references/validation/phase-c-smoke-test-results.md`
- First-release readiness → `references/validation/first-release-readiness.md`

# Release Notes — RC1

Release candidate: **music-composition RC1**  
Snapshot date: **2026-04-27**  
Base package: `music-composition-patched-2026-04-27-v8.zip`

## Release decision

RC1 is ready for review. Phase B correctness work is recorded, Phase C prompt smoke testing has a pass result recorded, and the final packaging sanity checks pass.

This release candidate is intended to be a stable first release of the music-composition skill, not a final encyclopedia of every style, region, instrument, or workflow.

## What is included

RC1 includes:

- a runtime entry point in `SKILL.md`;
- routing in `references/00-navigation.md`;
- short quotable assets in `assets/`;
- longer composition references in `references/`;
- user-agent workflow guides for brainstorming, collaboration, revision, answer calibration, and session logging;
- research workflows for web trend research, reference-track digging, user listening context, regional trend analysis, microgenre dispatch, and style/originality boundaries;
- genre and instrument-idiom starter guides;
- release-readiness and validation documentation;
- `KNOWN-LIMITATIONS.md`;
- `scripts/music_theory_sanity_check.py`.

## Release-readiness summary

| Area | RC1 status |
|---|---|
| Phase A content lock | Complete for RC1 |
| Phase B correctness pass | Complete |
| Phase C prompt smoke tests | Pass result recorded |
| Phase D packaging | Complete |
| Automated sanity checker | Pass |
| Internal Markdown links | Pass |
| ZIP hygiene | Pass |
| Known limitations | Included |

## Changes from v8

RC1 does not add major new runtime capability. The v8 content is preserved and packaged as a release candidate with:

- this release-notes file;
- `VALIDATION-RC1.md`;
- roadmap wording updated from “ready for RC packaging” to “RC1 packaged”;
- maintenance documentation updated to list the RC release artifacts.

## How to validate locally

From the repository root:

```bash
python3 scripts/music_theory_sanity_check.py
```

Expected output:

```text
music_theory_sanity_check: PASS
Checked 112 markdown files.
```

The exact Markdown count should remain stable unless a maintainer intentionally adds or removes files and updates the validation report.

## What to read first

For maintainers:

1. `RELEASE-ROADMAP.md`
2. `VALIDATION-RC1.md`
3. `KNOWN-LIMITATIONS.md`
4. `references/validation/first-release-readiness.md`
5. `references/validation/phase-c-smoke-test-results.md`

For runtime behavior:

1. `SKILL.md`
2. `references/00-navigation.md`
3. the relevant files selected by the navigation guide

## Release caveat

This RC is a strong first-release candidate, but it still relies on active web research for living scenes, charts, platform behavior, AI/copyright status, current artist activity, and rapidly changing music trends.

## Cross-references

- Entry point → `SKILL.md`
- Maintenance guide → `MAINTENANCE.md`
- Release roadmap → `RELEASE-ROADMAP.md`
- Validation report → `VALIDATION-RC1.md`
- Known limitations → `KNOWN-LIMITATIONS.md`
- RC packaging checklist → `references/validation/rc1-packaging-checklist.md`
- Phase C results → `references/validation/phase-c-smoke-test-results.md`

[English](./README.md) | [한국어](./README.ko.md)

# Music Composition Agent Skill

![Version](https://img.shields.io/badge/version-1.0-blue) ![Docs License](https://img.shields.io/badge/docs-CC%20BY%204.0-green) ![Code License](https://img.shields.io/badge/code-MIT-blue) ![Compatibility](https://img.shields.io/badge/compatibility-any%20SKILL.md%20harness-orange)

A modular knowledge and workflow skill for an AI music-composition assistant.

This repository is designed to help an agent turn vague creative requests — “make the chorus hit harder,” “give me a darker K-pop bridge,” “reharmonize this like jazz but not too dense,” “find reference tracks for my playlist” — into concrete musical decisions: chords, melody shapes, grooves, forms, arrangement moves, diagnostic feedback, and research workflows.

**Current release:** `v1.0`  
**Release date:** `2026-04-27`

---

## Installation

The simplest way is via the [`skills` CLI](https://www.npmjs.com/package/skills) (requires Node.js):

```bash
npx skills add SJY051/music-composition
```

This installs to your current project's agent directory (e.g., `.claude/skills/`). Add `-g` for user-level (global) install, or `--all -y` to skip prompts.

For a manual install, clone the repository and point your agent at it (or copy `SKILL.md` and the relevant subdirectories into your agent's skill folder, such as `~/.claude/skills/` or `.claude/skills/` in a project):

```bash
git clone https://github.com/SJY051/music-composition.git
```

### Compatibility

The skill follows the open **Agent Skills standard** — the SKILL.md + YAML frontmatter format used by both [Claude](https://code.claude.com/docs/en/skills.md) (Claude Code, Claude Agent SDK, Claude.ai skills) and [OpenAI](https://developers.openai.com/api/docs/guides/tools-skills) (Responses API `tools[].environment.skills`). It works on any harness that reads this format.

---

## Does it work?

Two benchmarks ship with v1.0:

**Approach 1 (system-prompt injection, GLM-5.1, blind eval).** With `SKILL.md` prepended to the system prompt, the skill version wins **6 of 8** representative composition prompts in an author-bias-controlled blind comparison. The strongest signal is the boundary test (P8) — the skill version explicitly declines an out-of-scope mixing question while the baseline answers freely.

![GLM-5.1 blind evaluation — wins per condition](./benchmarks/charts/chart_glm_winners.png)

**Approach 2 (native lazy loading, Claude Opus 4.7).** With the skill installed via Claude Desktop's Skills feature, the model reads references on demand. Most striking: pansori OST advice cites *yoseong*, *toeseong*, *chuseong* with cents-and-milliseconds implementation values that exist only inside `references/genres/korean-traditional.md` — direct evidence the model used the skill's full content, not just `SKILL.md`.

![P8 boundary respect across both benchmarks](./benchmarks/charts/chart_p8_boundary.png)

Together: **the skill content shapes behavior even on weaker setups, and content depth unlocks on stronger setups.** Full methodology and per-prompt analysis: [`benchmarks/v1.0-eval.md`](./benchmarks/v1.0-eval.md), [`benchmarks/v1.0-claude.md`](./benchmarks/v1.0-claude.md).

---

## What this skill is for

Use this skill when an AI agent needs to help with:

- composition brainstorming
- chord progressions and harmonic direction
- melody, motif, phrase, and topline writing
- song forms, transitions, and narrative structure
- rhythm, groove, meter, and feel
- orchestration, arrangement, voicing, and instrument idiom
- genre-aware writing across pop, jazz, classical, game, media, regional, traditional, and hybrid styles
- critique and revision of a work in progress
- teaching music concepts in a practical composition-first way
- current music trend research and reference-track digging
- user-agent collaborative workflows across multiple turns

The skill is not a DAW, audio engine, MIDI generator, transcription engine, legal service, or substitute for a human specialist in culturally specific traditions. It is a structured reference and workflow layer for an AI assistant.

---

## Core design philosophy

The skill is built around a few practical principles:

1. **Composition is decision-making, not rule-following.**  
   Music theory is treated as a map of possible effects, not a list of commandments.

2. **Give playable, actionable output.**  
   Prefer concrete examples such as chord charts, scale-degree melodies, groove sketches, voicing suggestions, and revision blocks.

3. **Preserve user agency.**  
   Offer options, explain tradeoffs, and let the user choose a direction.

4. **Use only the references needed.**  
   The agent should usually start with `SKILL.md`, route through `references/00-navigation.md`, and load only the most relevant 1–3 files.

5. **Treat current music trends as research tasks.**  
   Static genre files provide stable craft conventions. Recent artists, charts, platform norms, and scene-specific references should be checked through active web research.

6. **Avoid shallow cultural shortcuts.**  
   Tradition-specific and region-specific music should not be reduced to generic “flavor.” The skill encourages naming the source tradition, function, language, rhythm system, instrumentation, and collaboration needs.

---

## Repository structure

```text
.
├── SKILL.md
├── MAINTENANCE.md
├── RELEASE-ROADMAP.md
├── KNOWN-LIMITATIONS.md
├── RELEASE-NOTES-v1.0.md
├── VALIDATION-v1.0.md
├── assets/
├── references/
│   ├── 00-navigation.md
│   ├── creative-workflows/
│   ├── form/
│   ├── fundamentals/
│   ├── genres/
│   ├── harmony/
│   ├── instrument-idiom/
│   ├── melody/
│   ├── orchestration/
│   ├── production-aware/
│   ├── research/
│   ├── rhythm-groove/
│   ├── songwriting/
│   ├── techniques/
│   └── validation/
└── scripts/
    └── music_theory_sanity_check.py
```

### Important files

| File | Purpose |
|---|---|
| `SKILL.md` | Root operating instructions for the agent. Start here. |
| `references/00-navigation.md` | Routing map for choosing the right reference files. |
| `assets/` | Short cheat sheets and response templates for quick use. |
| `references/creative-workflows/` | Brainstorming, revision, answer calibration, and user-agent collaboration. |
| `references/research/` | Web trend research, reference-track digging, streaming-service-aware research, and style/copyright guardrails. |
| `references/genres/` | Genre and regional starter guides. |
| `references/instrument-idiom/` | Composition-facing playability and idiomatic writing guides. |
| `references/validation/` | Release readiness, smoke tests, correctness passes, and packaging checklists. |
| `scripts/music_theory_sanity_check.py` | Regression checker for links, known theory errors, risky terminology, snapshot notes, and required files. |

---

## Suggested agent workflow

For most user requests:

```text
1. Read SKILL.md.
2. Use references/00-navigation.md to identify the smallest useful document set.
3. Load only the relevant reference files, usually 1–3.
4. Generate concrete musical output.
5. Explain why the choices work.
6. Offer a small number of next-step options.
```

For example:

| User request | Likely routing |
|---|---|
| “Give me a sad but not cliché progression.” | `harmony/functional-harmony.md`, `harmony/modal-harmony.md`, `assets/progressions-catalog.md` |
| “The chorus does not hit.” | `assets/diagnostic-checklists.md`, `songwriting/hooks-and-memorability.md`, `production-aware/energy-and-dynamics.md` |
| “Make this more playable on guitar.” | `instrument-idiom/guitar.md`, plus the relevant harmony or arrangement file |
| “Brainstorm three directions.” | `creative-workflows/musical-brainstorming.md`, `assets/musical-brainstorming-cards.md` |
| “Find current J-pop references.” | `research/web-music-trend-research.md`, `research/reference-track-digging.md`, `assets/web-search-cheatsheet.md` |
| “Use my playlist as a reference.” | `research/user-listening-context-and-streaming-services.md`, `assets/trend-and-reference-matrices.md` |

---

## Currentness policy

Some files include a snapshot note. These files contain information that can become outdated, especially around:

- current genre trends
- charts and streaming-platform behavior
- artist references
- regional scenes
- AI-assisted music tools
- copyright and policy developments

The snapshot note does not mean the file is permanently current. It means the file was last reviewed as of that snapshot date. For fresh claims, use the research documents and verify current sources.

Useful research files:

```text
references/research/web-music-trend-research.md
references/research/reference-track-digging.md
references/research/user-listening-context-and-streaming-services.md
references/research/regional-trend-evolution-analysis.md
references/research/style-reference-and-copyright.md
assets/web-search-cheatsheet.md
```

---

## Style, originality, and copyright guardrails

This skill is intended to help generate original musical ideas. It should not be used to copy protected expression from existing songs or artists.

When a user asks for a specific artist or song style, the agent should translate the reference into craft variables such as:

- tempo range
- groove type
- harmonic density
- form
- vocal range and phrasing style
- arrangement density
- instrumentation
- energy curve
- production-era cues

The agent should avoid reproducing or imitating protected elements such as:

- melody
- lyrics
- distinctive riffs
- samples
- signature producer tags
- highly specific vocal identity
- recognizable arrangement sequences from a particular recording

See:

```text
references/research/style-reference-and-copyright.md
KNOWN-LIMITATIONS.md
```

---

## Cultural and regional specificity

The skill contains guides for traditional, regional, and hybrid styles. These are composition-facing starter references, not replacements for cultural expertise, native-language knowledge, or practitioner collaboration.

When working with culturally specific material, prefer:

```text
specific tradition / region / language / function / rhythm / instrument idiom / performance context
```

Instead of vague labels such as:

```text
generic world flavor / ethnic color / Asian flavor / Middle Eastern flavor
```

Relevant files include:

```text
references/genres/korean-traditional.md
references/genres/folk-roots-and-traditions.md
references/genres/mena-pop.md
references/genres/south-asian-film-pop.md
references/genres/regional-scene-starters.md
references/research/regional-trend-evolution-analysis.md
```

---

## Validation

The v1.0 package passed the release validation recorded in:

```text
VALIDATION-v1.0.md
references/validation/first-release-readiness.md
references/validation/phase-b-correctness-pass.md
references/validation/phase-c-smoke-test-results.md
```

To run the sanity checker:

```bash
python scripts/music_theory_sanity_check.py
```

The checker looks for issues such as:

- broken internal Markdown references
- known high-risk music-theory regressions
- risky cultural shortcut terminology
- missing snapshot notes in current-sensitive files
- required release and validation files
- selected interval, mode, chord-symbol, and altered-dominant sanity checks

This checker is a regression tool, not a complete proof of musical correctness.

---

## Maintenance

Before changing the repository:

1. Read `MAINTENANCE.md`.
2. Update `references/00-navigation.md` when adding or moving files.
3. Keep high-use assets concise and accurate.
4. Use snapshot notes for current-sensitive files.
5. Run `scripts/music_theory_sanity_check.py`.
6. Record substantial validation work in the relevant validation document.

Recommended versioning:

| Change type | Suggested version |
|---|---|
| Typo, link, or small validation-script fix | `v1.0.x` |
| P0/P1 correctness hotfix | `v1.0.x` hotfix |
| New genre, workflow, or research documents | `v1.1` |
| Larger parser/evaluation harness improvements | `v1.1` or `v1.2` |
| Major structure change or file reorganization | `v2.0` candidate |

---

## Known limitations

See `KNOWN-LIMITATIONS.md` for the full list.

In brief, this skill does not directly provide:

- audio transcription
- DAW control
- MIDI export
- audio mixing or mastering
- legal advice
- guaranteed current trend knowledge without web research
- exhaustive expertise in every local or traditional music practice

---

## License

This repository uses a split license:

- Skill documents, references, assets, release notes, and validation documents are licensed under **Creative Commons Attribution 4.0 International (CC BY 4.0)**.
- Code in `scripts/` is licensed under the **MIT License**.

See `LICENSE.md` for details.

# First Release Roadmap

This roadmap defines what “release-ready” means for the music-composition skill. It is intentionally practical: the goal is not to finish every possible genre, technique, or culture area, but to ship a stable first release that gives accurate, useful, safe, and extensible composition help.

## Release target

**First release goal:** a composition assistant skill that can reliably help users brainstorm, draft, analyze, revise, and learn music across common Western theory, popular songwriting, selected global genre frames, current-reference workflows, and user-agent collaboration sessions.

The first release should feel like a strong working composer / arranger / songwriting coach. It does not need to feel like a complete music encyclopedia.

## Release principles

1. **Accuracy before breadth.** A smaller correct skill beats a larger skill that repeats wrong theory confidently.
2. **Workflow before trivia.** The assistant should help the user make musical decisions, not just explain concepts.
3. **Currentness through process.** Living scenes, charts, streaming services, platform norms, AI tools, and copyright status should be handled through active research workflows, not frozen claims.
4. **Cultural specificity over broad labels.** Avoid continent-scale or “world music” shorthand. Name the tradition, scene, language, rhythm, instrument idiom, or source context.
5. **Templates are part of the product.** Output formats, brainstorming cards, diagnostic checklists, and decision logs make the skill reliable in real sessions.
6. **Automated checks are regression guards, not truth machines.** Scripts should catch broken links and known high-risk mistakes, but expert review still matters.

## Minimum viable first release

The first release is ready when the following are true.

### 1. Navigation and structure

- `SKILL.md` accurately describes the folder structure.
- `references/00-navigation.md` routes all major user request types to 1–3 relevant files.
- Every major new folder has at least one overview or routing file.
- All internal `.md` links resolve.
- Deprecated or renamed files are either removed cleanly or kept as redirect stubs.

### 2. Core music theory correctness

- High-use assets are clean: chord symbols, intervals, modes, cadences, jazz voicings, progressions.
- Known corrected regressions remain fixed: `C(no5)`, A5/m6 inversion, E7alt tension labels, relative-vs-parallel mode tables, augmented sixth enharmonic claims, maqam/makam reduction, 12-TET overgeneralization.
- Chord-symbol ambiguity guidance exists for common user notation variants.
- Scale-degree spelling guidance exists for Roman numerals, modal mixture, secondary functions, and altered tensions.

### 3. User-agent workflow

- The skill supports brainstorming, drafting, critique, revision, selection, and handoff.
- The assistant can keep a project card, decision log, branch tracker, do-not-change list, and before/after revision block.
- Vague user feedback can be translated into musical variables.
- The assistant can offer divergent options before converging.

### 4. Research and currentness

- Current-sensitive genre, AI, platform, and trend files have snapshot notes.
- The skill includes a general web trend research workflow.
- The skill includes a fast web-search cheatsheet.
- The skill includes reference-track digging, style/copyright safety, user listening context, streaming-service surfaces, regional trend evolution, and microgenre dispatch.
- The assistant never asks for passwords, tokens, or private account access to use streaming-service context.

### 5. Genre and idiom coverage

The first release should cover, at a practical starter level:

- Classical period conventions and common forms
- Jazz harmony and selected jazz styles
- Pop, rock, indie, hip-hop, R&B, electronic, film/TV, game, media/commercial, musical theatre
- K-pop/J-pop, C-pop / Southeast Asian pop
- Latin pop/reggaeton, Brazilian pop/funk, Afrobeats/amapiano, South Asian film-pop, MENA pop
- Korean traditional and K-trad fusion
- Country/Americana, metal/punk/hardcore, gospel/CCM
- Folk/roots/tradition-specific writing
- Regional scene starters, microgenres, and hybrid genres
- Instrument-idiom basics for piano, guitar, bass, drums/percussion, strings, winds, brass, and vocals

The first release does **not** require every subgenre to have a dedicated file. Microgenre and regional-scene hubs are acceptable when they route the assistant to research workflows and major-genre craft variables.

### 6. Safety and originality

- Style-reference requests are converted into craft variables.
- The assistant avoids copying protected melody, lyrics, riffs, samples, signature tags, or vocal identity.
- Traditional or regional material is treated with specificity, context, and collaboration cautions where appropriate.
- AI-assisted composition guidance clearly separates ideation, generation, selection, editing, ownership, and current legal uncertainty.

### 7. Validation

- `scripts/music_theory_sanity_check.py` passes.
- A first-release readiness checklist exists.
- A prompt smoke-test suite exists.
- Validation reports summarize file counts, broken-link status, snapshot coverage, known-regression checks, and remaining limitations.

## Suggested release phases

### Phase A — content lock

Finish all first-release structural content:

- release roadmap
- first-release readiness checklist
- prompt smoke-test suite
- chord-symbol ambiguity / parsing asset
- scale-degree spelling asset
- revision-loop workflow
- audit rubric

Exit criteria:

- No planned first-release file is missing.
- Navigation and `SKILL.md` are updated.
- Sanity checker passes.

### Phase B — correctness pass

Focus on correctness rather than expansion. Use `references/validation/phase-b-correctness-pass.md` as the maintainer-facing checklist for this phase.

Review order:

1. `assets/` high-use cheat sheets
2. `references/harmony/`
3. `references/fundamentals/`
4. `references/genres/korean-traditional.md`, `references/genres/mena-pop.md`, `references/genres/south-asian-film-pop.md`, `references/genres/folk-roots-and-traditions.md`
5. `references/research/` currentness and privacy boundaries
6. `references/creative-workflows/` output behavior

Exit criteria:

- No known P0/P1 theory errors remain.
- Culture-language shortcuts are absent from active guidance.
- Currentness claims are either snapshot-labeled or moved to active research workflows.

### Phase C — prompt simulation pass

Run the prompt smoke tests in `references/validation/prompt-smoke-tests.md`.

The assistant should be tested on:

- chord progression generation
- melody revision
- weak chorus diagnosis
- jazz reharmonization
- K-pop/J-pop reference request
- user streaming-service reference workflow
- microgenre request
- regional trend comparison
- traditional-fusion request
- style/copyright-safe request
- multi-turn revision loop
- teaching-mode explanation

Exit criteria:

- The assistant routes to the intended files.
- The answer is concrete, musical, and genre-aware.
- The answer does not over-research when a stable theory answer is enough.
- The answer does use active web research when currentness matters.
- The answer offers choices and preserves user agency.

### Phase D — release candidate

Prepare the first release package.

Required artifacts:

- final ZIP
- unified diff from previous release candidate
- validation report
- known limitations section or `KNOWN-LIMITATIONS.md`
- Phase C smoke-test result record
- changelog summary

Exit criteria:

- ZIP integrity passes.
- No `__pycache__`, temporary scratch folders, generated logs, or obsolete validation artifacts are included.
- File counts and expected-files list match the validation report.

## What not to build before first release

These are useful, but they should not block first release.

- A complete world-genre encyclopedia
- Full historical musicology coverage
- DAW operation, plugin manuals, sound-design recipes, mix/mastering instruction
- MIDI file generation or notation-software UI automation
- Exhaustive artist-by-artist style guides
- Every possible microgenre as a dedicated file
- A perfect music-theory parser for every notation edge case

## Post-release backlog

Good post-release expansions:

- Dedicated regional guides: regional Mexican, Punjabi pop, Turkish arabesk-pop, Maghreb/raï, gqom, singeli, kuduro, sertanejo/pagode/arrocha/piseiro
- More instrument-idiom files: harp, mallet percussion, accordion, organ, synth performance idioms, DJ/turntable-informed writing
- Deeper advanced harmony checks: chord-scale mapping, altered dominant spelling, secondary-function consistency, modal-mixture spelling, Nashville number parsing
- Larger prompt-evaluation harness with expected routing and expected answer shape
- Accessibility and wellness-oriented composition guidance
- Collaboration handoff docs for human co-writers, producers, arrangers, and performers

## Release status labels

Use these labels in validation reports and maintenance notes:

| Label | Meaning |
|---|---|
| `draft` | Useful but not reviewed enough for release reliance |
| `reviewed` | Human-reviewed for obvious errors and structure |
| `snapshot-current` | Current-sensitive claims checked at the named snapshot date |
| `stable-reference` | Mostly durable theory/craft reference |
| `needs-web-check` | Must be verified when answering current scene/platform/law/tool questions |
| `release-blocker` | Must be fixed before first release |
| `post-release` | Valuable, but not required for first release |

## Current first-release assessment

As of v1.0 packaging on 2026-04-27, the first-release phases are complete:

| Phase | Status |
|---|---|
| Phase A — content lock | Complete |
| Phase B — correctness pass | Complete |
| Phase C — prompt simulation pass | User-run pass recorded |
| Phase D — release-candidate packaging | Complete |
| v1.0 finalization | Complete |

RC1 was promoted to v1.0 without creating RC2 because no release blocker was found during the RC1 → v1.0 decision pass. Remaining desirable work is treated as post-release update/backlog material.

v1.0 packaging is complete when the following artifacts are present:

1. `RELEASE-NOTES-v1.0.md`,
2. `VALIDATION-v1.0.md`,
3. `KNOWN-LIMITATIONS.md`,
4. a clean v1.0 release ZIP,
5. a packaging diff from RC1,
6. a SHA256 checksum file,
7. a passing `scripts/music_theory_sanity_check.py` result,
8. a passing ZIP integrity check.

After v1.0, use the post-release backlog above to plan update patches. New features should be treated as v1.x additions unless they fix a release-blocking defect.

[English](./README.md) | [한국어](./README.ko.md)

# Music Composition Agent Skill

![Version](https://img.shields.io/badge/version-1.0-blue) ![Docs License](https://img.shields.io/badge/docs-CC%20BY%204.0-green) ![Code License](https://img.shields.io/badge/code-MIT-blue) ![Compatibility](https://img.shields.io/badge/compatibility-any%20SKILL.md%20harness-orange)

AI 작곡 어시스턴트를 위한 모듈식 지식 / 워크플로 스킬.

이 저장소는 에이전트가 모호한 창작 요청 — "후렴이 더 강하게 꽂혔으면 좋겠어", "K-pop 브릿지를 좀 더 어둡게", "재즈처럼 리하모니제이션하되 너무 빽빽하지 않게", "내 플레이리스트 기준으로 레퍼런스 트랙 찾아줘" — 를 구체적인 음악적 결정으로 번역하도록 돕는다: 화성, 멜로디 형태, 그루브, 형식, 편곡 무브, 진단 피드백, 리서치 워크플로 등.

**현재 릴리스:** `v1.0`
**릴리스 날짜:** `2026-04-27`

---

## 설치

가장 간단한 방법은 [`skills` CLI](https://www.npmjs.com/package/skills) 사용 (Node.js 필요):

```bash
npx skills add SJY051/music-composition
```

현재 프로젝트의 에이전트 디렉토리(예: `.claude/skills/`)에 설치된다. 사용자 전역 설치는 `-g`, 프롬프트 없이 진행하려면 `--all -y`를 추가.

수동 설치 시에는 저장소를 클론한 뒤 에이전트가 참조하게 하거나, `SKILL.md`와 필요한 하위 디렉토리를 에이전트 스킬 폴더(예: `~/.claude/skills/` 또는 프로젝트의 `.claude/skills/`)로 복사한다:

```bash
git clone https://github.com/SJY051/music-composition.git
```

### 호환성

이 스킬은 open **Agent Skills 표준**을 따른다 — [Claude](https://code.claude.com/docs/en/skills.md)(Claude Code, Claude Agent SDK, Claude.ai skills 기능)와 [OpenAI](https://developers.openai.com/api/docs/guides/tools-skills)(Responses API `tools[].environment.skills`) 모두 공통으로 사용하는 SKILL.md + YAML frontmatter 포맷이다. 이 포맷을 읽는 모든 하니스에서 작동한다.

---

## 실제로 작동하나요

v1.0과 함께 두 가지 벤치마크가 들어가 있다:

**Approach 1 (시스템 프롬프트 주입, GLM-5.1, 블라인드 평가).** `SKILL.md`를 시스템 프롬프트에 주입한 조건이 작가 편향을 통제한 블라인드 비교에서 8개 프롬프트 중 **6개**를 이김. 가장 강한 신호는 boundary 테스트(P8) — 스킬 적용 응답은 범위 밖 믹싱 질문을 명시적으로 거절했고, 베이스라인은 그냥 답함.

![GLM-5.1 블라인드 평가 — 조건별 승수](./benchmarks/charts/chart_glm_winners.png)

**Approach 2 (네이티브 lazy loading, Claude Opus 4.7).** Claude Desktop의 Skills 기능으로 스킬을 설치한 조건. 모델이 참조 파일을 필요 시 읽음. 가장 인상적인 부분 — 판소리 OST 조언에서 요성·퇴성·추성을 cents·ms 단위 구현값과 함께 인용. 이 값들은 `SKILL.md`나 `00-navigation.md`엔 없고 `references/genres/korean-traditional.md`에만 있는 정보라, 모델이 스킬 전체 콘텐츠를 *실제로* 활용했다는 직접 증거다.

![두 벤치마크에서의 P8 boundary respect](./benchmarks/charts/chart_p8_boundary.png)

종합하자면 — **약한 모델 / 약한 메커니즘에서도 스킬 콘텐츠가 행동을 형성하고, 강한 모델 / 강한 메커니즘에서는 콘텐츠 깊이까지 잠금 해제된다.** 메소돌로지와 프롬프트별 분석: [`benchmarks/v1.0-eval.md`](./benchmarks/v1.0-eval.md), [`benchmarks/v1.0-claude.md`](./benchmarks/v1.0-claude.md).

---

## 이 스킬은 어디에 쓰는가

AI 에이전트가 다음과 같은 작업을 도와야 할 때 사용한다:

- 작곡 브레인스토밍
- 코드 진행과 화성 방향 잡기
- 멜로디, 모티프, 프레이즈, 탑라인 작업
- 곡 형식, 전환, 서사 구조
- 리듬, 그루브, 박자, 필
- 오케스트레이션, 편곡, 보이싱, 악기 이디엄
- 팝 / 재즈 / 클래식 / 게임 / 미디어 / 지역 / 전통 / 하이브리드 스타일 전반의 장르 의식 작업
- 작업 중인 곡에 대한 비평과 수정
- 작곡 우선 관점의 음악 이론 교육
- 최신 음악 트렌드 리서치와 레퍼런스 트랙 디깅
- 여러 턴에 걸친 사용자–에이전트 협업 워크플로

이 스킬은 DAW, 오디오 엔진, MIDI 생성기, 트랜스크립션 엔진, 법률 서비스, 또는 특정 전통에 특화된 인간 전문가를 대체하는 도구가 **아니다**. 어디까지나 AI 어시스턴트를 위한 **구조화된 레퍼런스 / 워크플로 레이어**다.

---

## 핵심 설계 철학

이 스킬은 몇 가지 실용적 원칙 위에 세워졌다.

1. **작곡은 규칙 따르기가 아니라 의사결정이다.**
   음악 이론은 명령 목록이 아니라 가능한 효과의 지도로 다룬다.

2. **연주 가능하고 실행 가능한 출력.**
   코드 차트, 도수 멜로디, 그루브 스케치, 보이싱 제안, 수정 블록처럼 구체적인 예시를 우선한다.

3. **사용자의 주도권 보존.**
   여러 옵션을 제시하고, 트레이드오프를 설명하고, 방향 선택은 사용자에게 맡긴다.

4. **필요한 레퍼런스만 로드.**
   에이전트는 기본적으로 `SKILL.md`에서 시작 → `references/00-navigation.md`에서 라우팅 → 가장 관련성 높은 1–3개 파일만 로드한다.

5. **최신 음악 트렌드는 리서치 작업으로 다룬다.**
   정적 장르 파일은 안정적인 작곡 컨벤션을 제공한다. 최근 아티스트, 차트, 플랫폼 동향, 씬별 레퍼런스는 능동적 웹 리서치로 확인한다.

6. **얕은 문화적 단축은 피한다.**
   전통과 지역에 특화된 음악을 단순 "플레이버"로 환원하지 않는다. 전통, 기능, 언어, 리듬 시스템, 악기 편성, 협업 필요성을 명확히 명명하도록 유도한다.

---

## 저장소 구조

```text
.
├── SKILL.md
├── MAINTENANCE.md
├── RELEASE-ROADMAP.md
├── KNOWN-LIMITATIONS.md
├── RELEASE-NOTES-v1.0.md
├── VALIDATION-v1.0.md
├── assets/
├── references/
│   ├── 00-navigation.md
│   ├── creative-workflows/
│   ├── form/
│   ├── fundamentals/
│   ├── genres/
│   ├── harmony/
│   ├── instrument-idiom/
│   ├── melody/
│   ├── orchestration/
│   ├── production-aware/
│   ├── research/
│   ├── rhythm-groove/
│   ├── songwriting/
│   ├── techniques/
│   └── validation/
└── scripts/
    └── music_theory_sanity_check.py
```

### 주요 파일

| 파일 | 용도 |
|---|---|
| `SKILL.md` | 에이전트의 최상위 운영 지침. 여기서부터 시작. |
| `references/00-navigation.md` | 적절한 레퍼런스 파일을 고르기 위한 라우팅 맵. |
| `assets/` | 즉시 사용 가능한 짧은 치트시트와 응답 템플릿. |
| `references/creative-workflows/` | 브레인스토밍, 수정, 답변 캘리브레이션, 사용자–에이전트 협업. |
| `references/research/` | 웹 트렌드 리서치, 레퍼런스 트랙 디깅, 스트리밍 서비스 인지 리서치, 스타일/저작권 가드레일. |
| `references/genres/` | 장르 / 지역별 스타터 가이드. |
| `references/instrument-idiom/` | 작곡 관점에서의 연주 가능성과 이디엄 가이드. |
| `references/validation/` | 릴리스 준비도, 스모크 테스트, 정합성 패스, 패키징 체크리스트. |
| `scripts/music_theory_sanity_check.py` | 링크, 알려진 이론 오류, 위험 용어, 스냅샷 노트, 필수 파일에 대한 회귀 검사기. |

---

## 권장 에이전트 워크플로

대부분의 사용자 요청에 대해:

```text
1. SKILL.md를 읽는다.
2. references/00-navigation.md를 사용해 가장 작은 유용한 문서 집합을 식별한다.
3. 관련 레퍼런스 파일만 로드한다(보통 1–3개).
4. 구체적인 음악적 출력을 생성한다.
5. 왜 그 선택이 통하는지 설명한다.
6. 적은 수의 다음 옵션을 제시한다.
```

예를 들어:

| 사용자 요청 | 라우팅 |
|---|---|
| "슬프지만 진부하지 않은 진행 줘." | `harmony/functional-harmony.md`, `harmony/modal-harmony.md`, `assets/progressions-catalog.md` |
| "후렴이 안 꽂혀." | `assets/diagnostic-checklists.md`, `songwriting/hooks-and-memorability.md`, `production-aware/energy-and-dynamics.md` |
| "이 부분 기타로 더 치기 쉽게 해 줘." | `instrument-idiom/guitar.md` + 관련 화성 / 편곡 파일 |
| "방향성 세 가지 브레인스토밍해 줘." | `creative-workflows/musical-brainstorming.md`, `assets/musical-brainstorming-cards.md` |
| "최근 J-pop 레퍼런스 찾아 줘." | `research/web-music-trend-research.md`, `research/reference-track-digging.md`, `assets/web-search-cheatsheet.md` |
| "내 플레이리스트를 레퍼런스로 써 줘." | `research/user-listening-context-and-streaming-services.md`, `assets/trend-and-reference-matrices.md` |

---

## 최신성 정책

일부 파일에는 스냅샷 노트가 포함되어 있다. 이런 파일은 시간이 지나면서 낡을 수 있는 정보를 담고 있다 — 특히:

- 현재 장르 트렌드
- 차트와 스트리밍 플랫폼 동향
- 아티스트 레퍼런스
- 지역 씬
- AI 보조 음악 도구
- 저작권 / 정책 동향

스냅샷 노트는 "이 파일이 영구적으로 최신"이라는 의미가 아니라, 해당 날짜 기준으로 마지막 검토가 이루어졌다는 의미다. 새로운 클레임에는 리서치 문서를 활용하고 출처를 검증한다.

유용한 리서치 파일:

```text
references/research/web-music-trend-research.md
references/research/reference-track-digging.md
references/research/user-listening-context-and-streaming-services.md
references/research/regional-trend-evolution-analysis.md
references/research/style-reference-and-copyright.md
assets/web-search-cheatsheet.md
```

---

## 스타일, 오리지널리티, 저작권 가드레일

이 스킬은 **오리지널 음악 아이디어 생성**을 돕기 위한 것이다. 기존 곡이나 아티스트의 보호받는 표현을 복제하는 데 써서는 안 된다.

사용자가 특정 아티스트 / 곡 스타일을 요청하면, 에이전트는 그 레퍼런스를 다음과 같은 작업 변수로 번역한다:

- 템포 범위
- 그루브 타입
- 화성 밀도
- 형식
- 보컬 음역과 프레이징 스타일
- 편곡 밀도
- 악기 편성
- 에너지 곡선
- 프로덕션 시대 단서

다음과 같은 보호받는 요소는 복제 / 모방하지 않는다:

- 멜로디
- 가사
- 특징적인 리프
- 샘플
- 시그니처 프로듀서 태그
- 매우 구체적인 보컬 정체성
- 특정 녹음의 식별 가능한 편곡 시퀀스

참고:

```text
references/research/style-reference-and-copyright.md
KNOWN-LIMITATIONS.md
```

---

## 문화적 / 지역적 특수성

이 스킬에는 전통, 지역, 하이브리드 스타일에 대한 가이드가 포함되어 있다. 이들은 **작곡 관점의 스타터 레퍼런스**이지, 문화적 전문성, 모국어 지식, 실천자 협업을 대체하지 않는다.

문화적으로 특수한 자료를 다룰 때는 다음과 같이 명시한다:

```text
구체적 전통 / 지역 / 언어 / 기능 / 리듬 / 악기 이디엄 / 연주 맥락
```

다음과 같은 모호한 라벨을 피한다:

```text
포괄적 월드 플레이버 / 민족적 색깔 / 아시아 플레이버 / 중동 플레이버
```

관련 파일:

```text
references/genres/korean-traditional.md
references/genres/folk-roots-and-traditions.md
references/genres/mena-pop.md
references/genres/south-asian-film-pop.md
references/genres/regional-scene-starters.md
references/research/regional-trend-evolution-analysis.md
```

---

## 검증

v1.0 패키지는 다음 문서에 기록된 릴리스 검증을 통과했다:

```text
VALIDATION-v1.0.md
references/validation/first-release-readiness.md
references/validation/phase-b-correctness-pass.md
references/validation/phase-c-smoke-test-results.md
```

정합성 검사 실행:

```bash
python scripts/music_theory_sanity_check.py
```

검사기는 다음과 같은 문제를 잡는다:

- 끊어진 내부 마크다운 레퍼런스
- 알려진 고위험 음악 이론 회귀
- 위험한 문화적 단축 용어
- 최신성에 민감한 파일에서 누락된 스냅샷 노트
- 필수 릴리스 / 검증 파일 존재 여부
- 선택된 음정, 모드, 코드 표기, 알터드 도미넌트 정합성

이 검사기는 회귀 방지 도구이지, 음악적 정확성에 대한 완전한 증명이 아니다.

---

## 유지보수

저장소를 변경하기 전에:

1. `MAINTENANCE.md`를 읽는다.
2. 파일을 추가하거나 옮길 때 `references/00-navigation.md`를 갱신한다.
3. 사용 빈도가 높은 에셋은 간결하고 정확하게 유지한다.
4. 최신성에 민감한 파일에는 스냅샷 노트를 사용한다.
5. `scripts/music_theory_sanity_check.py`를 실행한다.
6. 의미 있는 검증 작업은 해당 검증 문서에 기록한다.

권장 버전 정책:

| 변경 유형 | 권장 버전 |
|---|---|
| 오타, 링크, 작은 검증 스크립트 수정 | `v1.0.x` |
| P0/P1 정합성 핫픽스 | `v1.0.x` 핫픽스 |
| 새 장르, 워크플로, 리서치 문서 | `v1.1` |
| 더 큰 파서 / 평가 하니스 개선 | `v1.1` 또는 `v1.2` |
| 주요 구조 변경 또는 파일 재구성 | `v2.0` 후보 |

---

## 알려진 한계

전체 목록은 `KNOWN-LIMITATIONS.md` 참조.

요약하자면, 이 스킬은 다음을 직접 제공하지 않는다:

- 오디오 트랜스크립션
- DAW 제어
- MIDI 익스포트
- 오디오 믹싱 / 마스터링
- 법률 자문
- 웹 리서치 없는 보장된 최신 트렌드 지식
- 모든 지역 / 전통 음악 실천에 대한 완전한 전문성

---

## 라이선스

이 저장소는 분리된 라이선스를 사용한다:

- 스킬 문서, 레퍼런스, 에셋, 릴리스 노트, 검증 문서는 **Creative Commons Attribution 4.0 International (CC BY 4.0)** 라이선스로 제공된다.
- `scripts/`의 코드는 **MIT License**로 제공된다.

자세한 내용은 `LICENSE.md` 참조.

---
name: music-composition
description: Use this skill whenever the user asks for help with music composition, music theory, songwriting, harmony, melody, counterpoint, arrangement, orchestration, or musical analysis — across classical, jazz, pop, rock, hip-hop, R&B, electronic, film/TV, K-pop/J-pop, folk, or musical theatre. Triggers include explicit theory questions ("what scale fits over Cm7", "explain ii-V-I", "modulate from C to E♭"), craft questions ("write a chord progression for a melancholy bridge", "make this melody more memorable", "voice this chord on piano"), and vague creative problems that map to compositional techniques ("the chorus feels weak", "this transition is awkward", "my song sounds generic"). Use this skill even when the user describes the goal in non-technical terms — translating fuzzy creative problems into concrete compositional moves is exactly what this skill is for. Do NOT use for DAW operation, MIDI file generation, audio engineering, mixing/mastering, sound design, or notation software UI.
---

# Music Composition

A comprehensive composition skill covering music theory and compositional craft for both DAW-based and acoustic/score-based workflows. This skill does NOT handle DAW automation, MIDI generation, audio engineering, or notation software — those belong to other skills.

## Core philosophy

Composition is a series of decisions, not a list of rules to follow. When advising:

- **Frame techniques as options, not commandments.** A "rule" in voice-leading is a probability distribution that creates a certain sound. Breaking it makes a different sound, which may be exactly what's wanted. Always know which sound the user is after.
- **Explain *why* a technique produces its effect.** Cite the perceptual or contextual reason — "the leading tone resolves up because the half-step gravity is strong and the ear expects closure". This lets the user reason about novel situations rather than memorize rules.
- **When the user describes a feeling or goal, give multiple options.** "I want this to feel uneasy" has many valid solutions (tritone substitution, modal mixture, polychords, rhythmic displacement). Offer 2–4 with their trade-offs rather than picking one.
- **Be concrete.** "Try a iv chord" beats "try modal mixture." "Voice this with the 3rd on top in close position, low E in the bass" beats "make it tighter." Always give the user something they can play.
- **Respect genre conventions but don't be enslaved by them.** A "wrong" choice in jazz might be perfect in indie rock. Always know which genre frame the user is operating in.
- **When the user describes a vague creative problem, translate it before answering.** "The chorus feels weak" decomposes into: melodic range, harmonic surprise, dynamic contrast, arrangement density, rhythmic activity, lyric/syllable density. Diagnose, then prescribe.
- **Don't moralize about technique.** Parallel fifths aren't immoral. They're a sound. Bach avoided them; Debussy used them; the user's track might want them.

## What this skill does NOT cover

- **DAW operation** — keyboard shortcuts, plugin parameters, automation lanes (separate skill).
- **MIDI file generation or direct manipulation** — handled elsewhere.
- **Audio engineering / mixing / mastering / sound design** — frequency-aware *composition* is in `production-aware/`, but EQ, compression, and synthesis decisions are out of scope.
- **Music history trivia** unconnected to composition technique. (Stylistic conventions of a period ARE in scope.)
- **Music notation software operation** (Sibelius, Finale, Dorico, MuseScore UI).
- **Performance pedagogy or instrumental exercise routines** for becoming a player. Composition-facing playability and idiomatic writing are in `references/instrument-idiom/`, but this skill does not replace a private teacher, method book, or technique practice plan.

## Navigation

The bulk of this skill lives in `references/`. **Always start by reading `references/00-navigation.md`** — it maps user requests to the specific reference files you need.

Don't load every reference. Load only what's directly relevant to the current question. For most requests, 1–3 reference files is enough. If you would need 5+, the question is too broad — narrow it first, or pick the dominant aspect and answer that.

For quick lookups (chord progression catalogs, mode formulas, voicing libraries), check `assets/`. Asset files are short and can be quoted directly. Reference files should be synthesized, not quoted.

## Top-level structure

```
references/
├── 00-navigation.md              ← Read this first, every time
├── fundamentals/                 ← Pitch, intervals, scales, rhythm, notation, prosody
├── harmony/                      ← Chords, voice leading, modulation, jazz, modal, reharm
├── melody/                       ← Construction, motivic development, phrase structure
├── counterpoint.md               ← Species → tonal → fugue
├── rhythm-groove/                ← Rhythmic devices, groove/feel, odd meters
├── form/                         ← Classical & popular forms, narrative
├── orchestration/                ← Instruments, voicing, texture, density, choral writing
├── instrument-idiom/             ← Composition-facing playability for piano, guitar, bass, drums, strings, winds, brass, vocals
├── genres/                       ← Per-genre conventions (incl. Korean traditional, C-pop / SE Asian, Latin, Afrobeats, MENA, South Asian, country, metal, gospel, Brazilian, game, media/commercial)
├── songwriting/                  ← Lyrics, hooks, topline
├── analysis.md                   ← Roman numeral, form, Schenker, set theory
├── techniques/                   ← Theme & variation, 20th C., constraint-based, microtonal, algorithmic / AI-assisted
├── production-aware/             ← Pre-production, composition-mix interface, energy
├── research/                     ← Web trends, user listening context, regional trend evolution, reference digging, style/copyright guardrails
├── creative-workflows/           ← Musical brainstorming, answer calibration, revision loops, and user-agent collaboration patterns
├── validation/                   ← Release readiness, Phase B/C records, RC packaging, and prompt smoke tests
├── source-bibliography.md        ← Maintainer-facing source map and verification guide
├── workflow.md                   ← Starting, revising, hybrid-genre advising
├── critique-and-feedback.md      ← Evaluating user's existing work (critique mode)
└── teaching-composition.md       ← Pedagogy mode, learning paths, exercises

scripts/
└── music_theory_sanity_check.py   ← Optional validation and regression helper for maintainers

assets/
├── progressions-catalog.md
├── response-templates.md
├── diagnostic-checklists.md
├── trend-and-reference-matrices.md
├── musical-brainstorming-cards.md
├── session-brief-and-decision-log.md
├── web-search-cheatsheet.md
├── chord-symbol-ambiguity-and-parsing.md
├── scale-degree-spelling-cheatsheet.md
├── music-theory-audit-rubric.md
├── cadence-reference.md
├── modes-cheatsheet.md
├── intervals-and-scale-formulas.md
├── jazz-voicings.md
├── form-templates.md
└── chord-symbol-conventions.md
```

## Typical response workflow

1. **Identify the user's actual ask.** Surface goal (e.g., "write a sad chord progression in C") and underlying goal (e.g., "I want emotional weight without sounding cliché"). They're often different.
2. **Identify the genre frame.** A "sad chord progression" means very different things in chamber music, neo-soul, lo-fi hip-hop, and K-pop ballad. If unstated and ambiguous, ask. If implied, work with it but name the assumption.
3. **Read `references/00-navigation.md`** to find the relevant reference files.
4. **Load the minimal set** of reference files needed.
5. **Translate theory into concrete options.** Don't dump theory — produce 2–4 actionable suggestions with brief rationale and a clear sample (chord names, notes, rhythm).
6. **Note adjacent considerations** the user may not have thought of (e.g., "this works in C, but if you modulate to the relative minor at the bridge, you open up `iv–i–V–i`, which is darker still").
7. **Stop before over-explaining.** A concise, useful answer beats a comprehensive lecture.

## Notation conventions

Use these consistently throughout the skill:

- **Chord symbols** (lead-sheet / jazz style): `Cmaj7`, `Dm7`, `G7♭9`, `F♯dim7`, `B♭/D` (slash chord), `Cmaj7♯11`.
- **Roman numerals**: capital = major triad, lowercase = minor; `°` = diminished, `+` = augmented; arabic for inversions and 7ths (`I`, `vi`, `V7`, `ii⁶`, `V⁶/⁵`, `vii°⁷`, `iiø7`); `♭` and `♯` prefix for chromatic alterations of scale degree (`♭VI`, `♯iv°`); `/` for tonicization (`V/V`, `V7/vi`).
- **Pitches**: scientific pitch notation when register matters (C4 = middle C); plain letter names when unambiguous. Sharp = `♯`, flat = `♭`, double-sharp = `𝄪`, double-flat = `𝄫`.
- **Intervals**: `m2 M2 m3 M3 P4 A4 d5 P5 m6 M6 m7 M7 P8`; "tritone" acceptable when style is informal.
- **Scale degrees**: `^1 ^2 ^3 ...` when concise; "tonic / supertonic / mediant / subdominant / dominant / submediant / leading tone" in prose.
- **Tempo**: BPM (`120 BPM`); `♩=120` when discussing notation.

## Genre framing — always required

The same musical question has different answers in different genres. Before applying any technique, identify the genre frame:

- Classical (which period?)
- Jazz (which sub-style?)
- Pop / rock / indie
- Hip-hop / R&B / neo-soul
- Electronic (which sub-style?)
- Film / TV (what dramatic context?)
- Game / interactive media (what platform / interactive system?)
- Advertising / podcast / branded content / trailer / library music
- K-pop / J-pop
- C-pop / Cantopop / T-pop / V-pop / OPM / Indo-pop
- Latin pop / reggaeton / urbano
- Brazilian pop / MPB / samba-pop / funk carioca
- Afrobeats / amapiano
- South Asian film-pop / raga-derived pop
- MENA pop / maqam-adjacent fusion
- Korean traditional (*gugak*) and K-trad fusion
- Country / Americana / roots
- Metal / punk / hardcore
- Gospel / CCM / worship
- Folk / roots / tradition-specific writing
- Regional scene starters and local-pop idioms
- Minor / microgenre / hybrid scenes
- Musical theatre

If the user hasn't named one and context doesn't make it obvious, ask. Don't default silently — you'll give wrong-feeling advice.

## Extensibility

This skill is designed to grow. To add new content:

1. **New genre or sub-genre**: add `references/genres/<name>.md`, update `00-navigation.md`'s genre table.
2. **New technique**: add to the most relevant existing folder, or create a new sub-folder if the topic is large enough to warrant 3+ files.
3. **New cheatsheet / matrix asset**: add to `assets/`, update `00-navigation.md`'s asset map.
4. **New instrument idiom**: add `references/instrument-idiom/<instrument>.md`, update `instrument-idiom/overview.md` and `00-navigation.md`.
5. **New current-research workflow**: add to `references/research/`, include a snapshot note, update `00-navigation.md`, and link from related genre files.
6. **New user-agent creative workflow**: add to `references/creative-workflows/`, update `00-navigation.md`, and add quick templates in `assets/` if the workflow benefits from reusable cards. Use `creative-workflows/answer-calibration.md` for bounded variation, answer length, and loading discipline.
7. **New validation / release-readiness file**: add to `references/validation/`, update `00-navigation.md`, `RELEASE-ROADMAP.md` if relevant, and add expected-file checks when it becomes release-critical. Ship `KNOWN-LIMITATIONS.md` or include its caveats in release validation reports.

When adding files, keep them under ~600 lines (split if longer), one topic per file, descriptive names. Always update `00-navigation.md` so the agent can find new content.

Recent expansion includes game/interactive scoring, microtonal/xenharmonic systems, choral writing, music for media, Korean traditional / K-trad fusion, C-pop / Southeast Asian pop, prosody, and algorithmic / AI-assisted composition. Recent expansion also includes response templates, diagnostic checklists, trend/reference matrices, musical brainstorming cards, session/decision-log templates, a web-search cheatsheet, style/copyright guardrails, user streaming-service listening context, regional trend-evolution analysis, regional scene starters, minor/microgenre handling, genre idiom files for Latin pop/reggaeton, Afrobeats/amapiano, South Asian film-pop, MENA pop, country/Americana, metal/punk/hardcore, gospel/CCM, and Brazilian pop/funk, plus composition-facing `instrument-idiom/`, `creative-workflows/`, and `validation/` folders, release-readiness docs, Phase C smoke-test result recording, RC1 packaging guidance, known-limitations documentation, answer-calibration workflow guidance, chord-symbol/scale-degree audit assets, and a maintainer-facing source bibliography. Future additions should prioritize prompt-smoke-test runs, deeper regional genre idioms, instrument-family expansions where needed, style-specific reference libraries, and stronger automated validation; the Phase B correctness-prep checklist now lives in `references/validation/phase-b-correctness-pass.md`.

## Honesty about gaps

If a user asks about a topic this skill doesn't yet cover, say so explicitly: "This skill doesn't have a dedicated reference for [X] yet — I'll work from general principles and flag what I'm uncertain about." Then do that. Don't fabricate confident answers from gaps.

# Maintenance Guide

This file is for whoever is *maintaining* this skill — not for the agent using it. The agent doesn't need to read this. It documents the skill's internal conventions, file-link structure, and the procedures for adding, renaming, or removing content without breaking anything.

For the skill's user-facing philosophy, see `SKILL.md`. For navigation logic the agent uses at runtime, see `references/00-navigation.md`. For the first stable release artifacts, see `RELEASE-NOTES-v1.0.md` and `VALIDATION-v1.0.md`; RC1 records remain in `RELEASE-NOTES-RC1.md` and `VALIDATION-RC1.md`.

## What lives where

```
SKILL.md                          — entry point for the agent; philosophy + structure
MAINTENANCE.md                    — this file; for the maintainer only
RELEASE-ROADMAP.md                — first-release phase plan and post-release backlog
RELEASE-NOTES-v1.0.md             — first stable release notes
VALIDATION-v1.0.md                — first stable release validation report
RC1-v1.0-decision-memo-2026-04-27.md — RC1 promotion decision record
RELEASE-NOTES-RC1.md              — release-candidate notes retained for audit trail
VALIDATION-RC1.md                 — release-candidate validation report retained for audit trail
KNOWN-LIMITATIONS.md              — first-release scope boundaries and caveats
assets/                           — short, quotable cheat-sheets
references/
├── 00-navigation.md              — routing table the agent reads first
├── workflow.md                   — meta-file: how the agent advises
├── analysis.md                   — methodology for analyzing existing music
├── counterpoint.md               — single-file topic
├── fundamentals/                 — pitch, rhythm, notation
├── harmony/                      — chord vocabulary and motion
├── melody/                       — line writing and development
├── rhythm-groove/                — meter manipulation and feel
├── form/                         — structural templates
├── orchestration/                — instrumentation, ensemble texture, density, choral writing
├── instrument-idiom/             — composition-facing playability and instrument-specific writing
├── songwriting/                  — lyric, hook, topline craft
├── techniques/                   — variation, 20th-C., constraint-based
├── production-aware/             — composition decisions touching the mix
├── research/                     — web trend research, reference digging, style/copyright guardrails
├── creative-workflows/           — brainstorming, answer calibration, and user-agent collaboration workflows
├── validation/                   — release readiness, Phase B/C records, RC packaging, smoke tests
├── source-bibliography.md        — maintainer-facing source map and verification guide
└── genres/                       — per-genre conventions
```

`assets/` files are short and can be quoted directly by the agent. `references/` files are longer and are meant to be synthesized, not quoted.

## File conventions

Every reference file (`references/**/*.md`) should have:

1. **Title** (`# File Title`) matching the topic
2. **Short opening paragraph** defining the topic in 2–4 sentences
3. **Cross-link paragraph** pointing to adjacent files for related topics
4. **`## When to consult this file`** section — bullet list of triggers
5. **Body** organized with `##` headers
6. **`## Common pitfalls when advising`** section near the end (where applicable)
7. **`## Quick decision matrix`** table near the end (where applicable)
8. **`## Cross-references`** section at the very end — bullet list of links

Asset files are looser but should still have a title, a one-paragraph framing, and at minimum a cross-references section at the end.

The `## When to consult this file` section is what the agent uses to decide whether to load the file. Write triggers as concrete user-asks, not abstract topic descriptions.

## Tone conventions

All files are written in the same advisory voice. Some patterns to keep consistent:

- **Direct second person ("you") is rare.** The file talks about the user as "the user" and the writer as "the composer"; second person sneaks in occasionally for instructional moments but isn't the default.
- **"The agent" refers to the AI agent itself** in workflow.md, navigation, and pitfalls sections.
- **Bullet lists are common, but never lists of one or two items.** If a list has fewer than three items, prose works better.
- **Tables are preferred for symbol-to-meaning mappings, comparison matrices, and decision matrices.** Use them aggressively.
- **No "moralizing" about technique.** Conventions are described as conventions, not as rules. Recurring phrase: "this is convention, not law".
- **Genre is taken seriously.** Pop voice-leading isn't "incorrect classical voice-leading"; it's pop voice-leading. Don't apply one genre's standards to another.
- **Examples are concrete.** When a technique is named, an example follows. Famous-song examples are preferred over invented ones because they let the agent point the user at something to listen to.
- **Em dashes are used liberally** for parenthetical clauses. Curly quotes are used in prose; straight quotes in code/symbol contexts.

## Notation conventions across the skill

Defined once in `SKILL.md` and used consistently:

- Chord symbols: `Cmaj7`, `Dm7`, `G7♭9`, `B♭/D` (lead-sheet style by default)
- Roman numerals: `I`, `vi`, `V7`, `ii⁶`, `vii°⁷`, `iiø7`, `♭VI`, `V/V`
- Pitches: `C4` = middle C (scientific pitch notation when register matters)
- Intervals: `m2 M2 m3 M3 P4 A4 d5 P5 m6 M6 m7 M7 P8`
- Scale degrees: `^1 ^2 ^3 ...` in compact notation
- Tempo: `120 BPM` or `♩=120`
- Accidentals: `♯` `♭` `♮` `𝄪` `𝄫` (Unicode glyphs, not ASCII `#` and `b`, except where the context is plain-text input from a user)

When adding new content, match these conventions. When the agent is asked about ASCII alternatives, `assets/chord-symbol-conventions.md` covers them.

## Cross-reference conventions

Cross-references between files use **relative paths from the linking file's location**:

- A file in `references/harmony/` linking to `references/melody/melodic-construction.md` writes `melody/melodic-construction.md` (sibling-folder relative)
- A file in `references/harmony/` linking to `references/workflow.md` writes `../workflow.md` is **not** the convention — most files use `workflow.md` or `references/workflow.md` depending on the file's depth. **Inconsistency exists; prefer the form that reads naturally.**
- A file in `references/` linking to `assets/` writes `assets/foo.md`
- A file in `assets/` linking to `references/` writes `references/foo/bar.md`

The agent is robust to these variations. Don't reformat existing links unless they're broken.

Each file's `## Cross-references` section is the *primary* link list. Inline links in the body should be sparse and only when the link is essential to the immediate point.

## Snapshot and currentness policy

Some files describe living scenes, current artists, platform behavior, AI tools, or market conventions. Those files should include a `Snapshot note` near the top.

A snapshot note should state:

1. the last content review date,
2. which parts are durable craft guidance,
3. which parts require current web verification,
4. links to `references/research/web-music-trend-research.md`, `references/research/user-listening-context-and-streaming-services.md`, and `references/research/reference-track-digging.md` when reference digging or user listening context matters.

Use snapshot notes for:

- `references/genres/*.md`
- `references/techniques/algorithmic-and-AI-assisted.md`
- `references/production-aware/pre-production-decisions.md` when reference-track selection is involved
- any future file that mentions current charts, living scenes, platform rules, AI/copyright status, recent artists, streaming-service behavior, or personal listening-context workflows

Do **not** update a snapshot date just because the file was edited. Update it only after checking the current-sensitive claims against active sources.

For current-source research, prefer official charts/platform pages, industry reports, regional/local sources, and primary documentation. Avoid turning a single viral track or platform trend into a universal genre rule.

## How to add new content

### Adding a new file to an existing folder

1. Write the file matching the conventions above (When to consult, body, Pitfalls, Quick decision matrix, Cross-references).
2. **Add it to `references/00-navigation.md`** — at minimum to the "Quick lookup" table; also to "Multi-file scenarios" if it crosses domains; also to "Cross-reference key" if it's part of a cluster.
3. **Update relevant existing files' Cross-references sections** to point to the new file. Don't update every file in the project — just the 2–4 files that are most directly related.
4. **Update `SKILL.md`** if the new file represents a meaningfully new topic that should be visible at the structural level (rare for files within existing folders).

### Adding a new top-level folder

When a topic is large enough to warrant 3+ files, it gets its own folder.

1. Create the folder under `references/`.
2. Write the first 1–3 files in it, each matching conventions.
3. **Update `SKILL.md`** — add the folder to the "Top-level structure" diagram and to any other places the structure is mentioned.
4. **Update `references/00-navigation.md`** — add a section for the new folder's topics in the "Quick lookup" table; add to "Multi-file scenarios" if relevant; add to "Cross-reference key" if it forms a cluster with existing folders.
5. Update existing files' Cross-references where the new folder is genuinely related.

### Adding a new asset file

1. Write the asset file. Asset files are shorter and looser than reference files but still need a title, opening paragraph, and Cross-references section.
2. **Update `references/00-navigation.md`** — add to the "Cheatsheets — when to use `assets/`" table.
3. **Update `SKILL.md`** — add to the asset list in the "Top-level structure" diagram if the asset is structurally significant.
4. Update related reference files' Cross-references to point to the new asset.

### Adding a new instrument-idiom file

Instrument-idiom files are composition-facing. They should help the agent write playable, idiomatic parts; they should not become method-book lessons for how to practice the instrument.

1. Write the file under `references/instrument-idiom/`.
2. Update `references/instrument-idiom/overview.md` so the new file is discoverable inside the cluster.
3. Update `references/00-navigation.md` in the quick lookup table, multi-file scenarios if needed, and the cross-reference key.
4. Add 2–4 direct cross-references from related files such as `orchestration/instruments-ranges-character.md`, `orchestration/voicing-and-texture.md`, or a relevant genre file.
5. Run `scripts/music_theory_sanity_check.py`.

### Adding a new genre file

Genre files are special because every other file potentially refers to them.

1. Write the genre file under `references/genres/`.
2. **Update `references/00-navigation.md`** — add to the "Genre routing" table.
3. **Update `SKILL.md`** — add to the genre list in the "Genre framing — always required" section.
4. **Update `MAINTENANCE.md`** (this file) if the genre is a major addition that affects skill philosophy or scope.
5. Don't update every other file's Cross-references — genre files are referenced from a long tail of places. Only update the most directly relevant ones (e.g., a new "Latin pop" file should be linked from `references/genres/pop-rock.md` and `references/genres/folk-roots-and-traditions.md`).

### Adding a microgenre, minor genre, or hybrid-scene file

If the topic is not large or stable enough for a dedicated major genre file, add coverage to `references/genres/minor-and-hybrid-genres.md` instead of creating many tiny files.

1. Add the tag or scene as a craft-variable note, not a definitive history.
2. Include a caution when the label is platform-born, local, or unstable.
3. Link to the relevant major genre files and research workflows.
4. If the topic grows large enough, split it into a dedicated `references/genres/<name>.md` file later.
5. Run `scripts/music_theory_sanity_check.py`.


### Adding a creative workflow file

Use `references/creative-workflows/` for user-agent process documents that are specifically about music-making sessions, not generic project management.

Checklist:

1. Define when the agent should consult the file.
2. Include concrete musical outputs: chords, motifs, rhythm grids, form maps, texture moves, or templates.
3. Keep user agency explicit: options before commands, decisions tracked, reversible assumptions.
4. Add reusable cards/templates to `assets/` when helpful.
5. Update `references/00-navigation.md`, `SKILL.md`, and `references/source-bibliography.md`.
6. Add the new file to `scripts/music_theory_sanity_check.py` expected files if it is a major expansion.

### Adding a validation or release-readiness file

Use `references/validation/` for maintainer-facing QA documents: first-release readiness, Phase B correctness passes, Phase C result records, prompt smoke tests, RC packaging checklists, release checklists, and validation procedures. Runtime composition answers should almost never load these files.

Checklist:

1. Define whether the file is a release blocker checklist, correctness pass, prompt test suite, or optional audit aid.
2. Link it from `references/00-navigation.md` only in maintainer / validation routes.
3. Update `RELEASE-ROADMAP.md` if the file changes first-release scope or quality gates.
4. Update `SKILL.md` if the validation folder structure changes.
5. Add critical validation files to `scripts/music_theory_sanity_check.py` expected files.
6. Keep validation claims factual: if a check was not run, mark it as not run.
7. If the file affects release-candidate packaging, update `KNOWN-LIMITATIONS.md` or the validation report template as needed.

### Adding a research workflow file

Research workflow files should help the agent gather evidence, not freeze current facts.

1. Place the file under `references/research/`.
2. Include a `Snapshot note` near the top.
3. State when to use the file, safe inputs, source hierarchy, output template, pitfalls, and cross-references.
4. If the file involves user-provided data, include privacy and credential boundaries.
5. Update `references/00-navigation.md`, `SKILL.md`, `references/source-bibliography.md`, and related research files.
6. Run `scripts/music_theory_sanity_check.py`.

### Renaming or moving a file

This is the most error-prone operation. There's no automatic refactoring; all links are plain-text relative paths.

1. **Search the whole project for the old path** before moving:
   ```
   grep -rn "old-filename.md" .
   ```
2. Move the file.
3. Update every link found in step 1.
4. **Re-run the search to verify zero hits on the old path.**

### Removing a file

1. Search for all links to the file (as in renaming).
2. Either delete the links (if the topic is being removed entirely) or redirect them to the replacement file.
3. Remove the file.
4. Update `00-navigation.md` and `SKILL.md` to remove references.

## A simple link-validation script

This isn't part of the skill — it's a maintenance helper. Save as `scripts/check-links.sh` (or run from the project root):

```bash
#!/usr/bin/env bash
# Find all backtick-wrapped `path/file.md` references and check whether
# the referenced file resolves. Filters false positives:
#  - Wildcard references like `genres/*.md` (idiomatic shorthand)
#  - Placeholder/example paths in MAINTENANCE.md
#  - Resolution attempts: same dir, project root, references/, assets/
set -e
cd "$(dirname "$0")/.." || exit 1

errors=0
total=0

while IFS=: read -r src_file src_line ref_match; do
  ref_path=$(echo "$ref_match" | sed -E 's/.*`([^`]+\.md)`.*/\1/')
  total=$((total + 1))

  # Skip wildcard shorthand
  case "$ref_path" in
    *\**) continue ;;
  esac

  src_dir=$(dirname "$src_file")
  if [[ -f "$src_dir/$ref_path" ]]; then continue; fi
  if [[ -f "./$ref_path" ]]; then continue; fi
  if [[ -f "./references/$ref_path" ]]; then continue; fi
  if [[ -f "./assets/$ref_path" ]]; then continue; fi

  case "$ref_path" in
    SKILL.md|MAINTENANCE.md) continue ;;
  esac

  echo "BROKEN: $src_file:$src_line -> $ref_path"
  errors=$((errors + 1))
done < <(grep -rn '`[^`]*\.md`' . --include='*.md' 2>/dev/null)

echo "---"
echo "Checked $total link instances; $errors broken."
```

Run after any file move, rename, or deletion. The script accepts that wildcard references (`genres/*.md`) are idiomatic shorthand in this skill and doesn't flag them. If you add new placeholder examples to `MAINTENANCE.md` itself, the script may flag them — check the output and decide case by case.

## What the agent should never have to know

A few things the maintainer cares about that the agent doesn't:

- **This file (`MAINTENANCE.md`).** It's not in the agent's load path.
- **Git history.** The agent treats the current state of the files as truth.
- **Planning notes / TODOs.** If a topic is incomplete, mark it inline in the relevant file (`# TODO: expand the Locrian section`) rather than tracking it in a separate planning doc — the agent sees the inline TODO and can flag the gap honestly to the user.

## Future expansion

All originally-planned topics have been implemented (see "Already implemented" below). Future expansion is now driven by emerging needs rather than backlog. Likely future candidates if they come up:

- **Sound design and synthesis fundamentals** → would be a new file in `production-aware/` or a top-level folder, depending on scope
- **DAW-specific workflow guides** → these intentionally aren't in this skill's scope (the skill is composition-focused, not production-tool-focused), but specific workflow notes for working with composers could go in `workflow.md`
- **Deeper regional pop / tradition-specific music practices** → current starter files cover Latin pop/reggaeton, Afrobeats/amapiano, MENA pop, South Asian film-pop, country/Americana, metal/punk/hardcore, gospel/CCM, Brazilian pop/funk, regional-scene starters, folk/roots/tradition-specific writing, and a general microgenre/hybrid handling file. Future work should go deeper by region/subgenre when a topic becomes stable and large enough (e.g., regional Mexican, sertanejo/pagode/arrocha, raï, Turkish arabesk-pop, Punjabi-specific pop, South African gqom, etc.)
- **Music for accessibility / therapeutic contexts** → music therapy, music for sensory processing, sleep / wellness music, etc.
- **Live performance composition** → composing with live performance constraints (bands, ensembles, soloists) where the score is also a performance plan
- **Sacred music traditions in depth** → Western liturgical, Jewish liturgical, Islamic *qira'at* and *naat*, Buddhist chant, Hindu *kirtan* — currently scattered across folk-and-world, classical-periods, and choral-writing
- **Scoring for dance / ballet / contemporary movement** → distinct from concert music or theatre music; specific structural and rhythmic constraints

When adding any of these, follow the procedures above and update `00-navigation.md`, `SKILL.md`, and any directly-related cross-references.

### Already implemented (formerly planned)

The skill's originally-planned topics, now implemented:

- **Korean traditional / fusion** → `references/genres/korean-traditional.md` ✓
- **Additional Asian pop traditions** → `references/genres/cpop-and-southeast-asian-pop.md` (C-pop, Cantopop, T-pop, V-pop, OPM, Indo-pop) ✓
- **Prosody across languages** → `references/fundamentals/prosody-and-language.md` ✓
- **Feedback / critique workflow** → `references/critique-and-feedback.md` (single-file approach; can split into a folder if it grows) ✓
- **Pedagogy / teaching composition** → `references/teaching-composition.md` (single-file approach) ✓
- **Choral writing** → `references/orchestration/choral-writing.md` ✓
- **Microtonal / xenharmonic systems** → `references/techniques/microtonal.md` ✓
- **Response templates** → `assets/response-templates.md` ✓
- **Diagnostic checklists** → `assets/diagnostic-checklists.md` ✓
- **Style/reference/copyright guardrails** → `references/research/style-reference-and-copyright.md` ✓
- **Latin pop / reggaeton starter** → `references/genres/latin-pop-and-reggaeton.md` ✓
- **Afrobeats / amapiano starter** → `references/genres/afrobeats-and-amapiano.md` ✓
- **South Asian film-pop starter** → `references/genres/south-asian-film-pop.md` ✓
- **MENA pop / maqam-adjacent starter** → `references/genres/mena-pop.md` ✓
- **Country / Americana / roots starter** → `references/genres/country-americana.md` ✓
- **Metal / punk / hardcore starter** → `references/genres/metal-punk-hardcore.md` ✓
- **Gospel / CCM / worship starter** → `references/genres/gospel-and-ccm.md` ✓
- **Brazilian pop / MPB / samba-pop / funk carioca starter** → `references/genres/brazilian-pop-and-funk.md` ✓
- **Instrument-specific idiom cluster** → `references/instrument-idiom/overview.md` plus piano, guitar, bass, drums/percussion, strings, winds, brass, and vocals ✓
- **Source bibliography and verification guide** → `references/source-bibliography.md` ✓
- **User streaming-service listening context workflow** → `references/research/user-listening-context-and-streaming-services.md` ✓
- **Regional trend-evolution analysis workflow** → `references/research/regional-trend-evolution-analysis.md` ✓
- **Minor / microgenre / hybrid genre handling** → `references/genres/minor-and-hybrid-genres.md` ✓
- **Musical brainstorming workflow** → `references/creative-workflows/musical-brainstorming.md` ✓
- **User-agent collaboration workflow** → `references/creative-workflows/user-agent-collaboration.md` ✓
- **Revision and feedback loop workflow** → `references/creative-workflows/revision-and-feedback-loops.md` ✓
- **Answer calibration / controlled variation workflow** → `references/creative-workflows/answer-calibration.md` ✓
- **Brainstorming prompt cards** → `assets/musical-brainstorming-cards.md` ✓
- **Session brief and decision log templates** → `assets/session-brief-and-decision-log.md` ✓
- **Web-search cheatsheet** → `assets/web-search-cheatsheet.md` ✓
- **Chord-symbol ambiguity and parsing asset** → `assets/chord-symbol-ambiguity-and-parsing.md` ✓
- **Scale-degree spelling cheatsheet** → `assets/scale-degree-spelling-cheatsheet.md` ✓
- **Music-theory audit rubric** → `assets/music-theory-audit-rubric.md` ✓
- **Regional scene starter map** → `references/genres/regional-scene-starters.md` ✓
- **Folk/roots/tradition-specific guide** → `references/genres/folk-roots-and-traditions.md` ✓
- **Trend and reference matrices** → `assets/trend-and-reference-matrices.md` ✓
- **Game / interactive scoring** → `references/genres/game-music.md` (kept separate from film-tv-scoring; the interactive dimension is large enough to warrant a dedicated file) ✓
- **Media music (advertising, podcasts, trailers, library, branded content)** → `references/genres/media-and-commercial-music.md` ✓
- **Algorithmic / AI-assisted composition** → `references/techniques/algorithmic-and-AI-assisted.md` ✓
- **First release roadmap** → `RELEASE-ROADMAP.md` ✓
- **Release readiness checklist** → `references/validation/first-release-readiness.md` ✓
- **Phase B correctness pass** → `references/validation/phase-b-correctness-pass.md` ✓
- **Prompt smoke tests** → `references/validation/prompt-smoke-tests.md` ✓
- **Phase C smoke-test result record** → `references/validation/phase-c-smoke-test-results.md` ✓
- **RC1 packaging checklist** → `references/validation/rc1-packaging-checklist.md` ✓
- **Known limitations for first release** → `KNOWN-LIMITATIONS.md` ✓
- **v1.0 release notes, validation, and promotion decision** → `RELEASE-NOTES-v1.0.md` + `VALIDATION-v1.0.md` + `RC1-v1.0-decision-memo-2026-04-27.md` ✓

## Versioning

The skill is versionless by default — files are edited in place. If the maintainer wants version markers, add a `## Changelog` section at the bottom of `SKILL.md` rather than per-file changelogs.


## Optional automated validation

Run `scripts/music_theory_sanity_check.py` after structural edits, cheat-sheet edits, or new genre additions. It checks:

- broken internal `.md` references
- risky culture-shortcut or threat-coding terms
- known high-risk theory patterns fixed in previous patches
- basic interval inversion sanity checks
- required snapshot notes in current-sensitive folders
- expected expansion files for major patch additions
- basic mode-cheatsheet formula sanity checks
- basic chord/tension spelling sanity checks for common altered-extension claims
- expected research/microgenre/creative-workflow/validation expansion files
- basic chord-symbol parser sanity for common C-root symbols
- scale-degree spelling regression checks for modal, borrowed, and altered-dominant examples
- symmetric-scale spelling regressions for whole-tone and half-whole diminished examples
- Phase B correctness-pass documentation presence
- Phase C result and RC packaging documentation presence
- v1.0 release-note / validation documentation presence
- Known limitations documentation presence

The script is not a substitute for musical judgment or web/currentness review. It is a guardrail for regressions.

# Validation Report — RC1

## Package

- Package: `music-composition-RC1-2026-04-27.zip`
- Base version: `music-composition-patched-2026-04-27-v8.zip`
- Date: 2026-04-27
- Release candidate: RC1
- Packaging scope: release-candidate packaging only; no major runtime feature expansion beyond v8

## Automated checks

| Check | Result | Notes |
|---|---|---|
| Sanity checker | PASS | `python3 scripts/music_theory_sanity_check.py` |
| Internal Markdown links | PASS | Covered by sanity checker |
| Snapshot notes | PASS | Covered by sanity checker; 32 files with snapshot notes |
| Known regression checks | PASS | Covered by sanity checker |
| ZIP integrity | PASS | `ZipFile.testzip()` returned no bad file |
| ZIP hygiene | PASS | No cache, scratch, temp, `.pyc`, or `.DS_Store` files included |

Sanity-check output:

```text
music_theory_sanity_check: PASS
Checked 112 markdown files.
```

## Manual / user-run checks

| Check | Result | Notes |
|---|---|---|
| Phase B correctness | PASS | Recorded in `references/validation/phase-b-correctness-pass.md` |
| Phase C prompt smoke tests | PASS | User-run pass recorded in `references/validation/phase-c-smoke-test-results.md` |
| High-use assets | PASS for RC1 | Spot-checked during Phase B; automated regression checks pass |
| Research/currentness/privacy | PASS for RC1 | Current-sensitive areas retain snapshot/web-check language and no credential request guidance |
| Cultural specificity | PASS for RC1 | Broad culture-as-color guidance is guarded by routing, limitations, and sanity checks |
| File-loading behavior | PASS by user report | No over-reading or abnormal routing reported during Phase C |
| Creative variation | PASS by user report | Variation stayed within expected range during Phase C |

## Package contents

| Metric | Count |
|---|---:|
| Total packaged files | 113 |
| Markdown files | 112 |
| Python scripts | 1 |
| Internal Markdown reference scan | 2,023 |
| Snapshot-note files | 32 |

## RC1 contents added after v8

| File | Purpose |
|---|---|
| `RELEASE-NOTES-RC1.md` | Release-candidate notes and validation entry points |
| `VALIDATION-RC1.md` | This release-candidate validation report |

Minor release-status wording was also updated in `RELEASE-ROADMAP.md`, `MAINTENANCE.md`, and `references/validation/first-release-readiness.md`.

## Known limitations

See `KNOWN-LIMITATIONS.md`.

RC1 intentionally does not claim to solve every possible genre, every tradition-specific practice, every live trend, direct audio transcription, DAW operation, MIDI generation, legal clearance, or comprehensive performer pedagogy. Current scenes, platform behavior, charts, artist activity, AI/copyright status, and streaming-service context should still be handled through the research workflows and active verification rules.

## Release decision

**Ship RC1 for review.**

No release blocker was found in the packaging pass. RC1 is ready for user/developer review, with the caveats recorded in `KNOWN-LIMITATIONS.md`.

## Cross-references

- Release notes → `RELEASE-NOTES-RC1.md`
- Release roadmap → `RELEASE-ROADMAP.md`
- Known limitations → `KNOWN-LIMITATIONS.md`
- RC packaging checklist → `references/validation/rc1-packaging-checklist.md`
- Phase B correctness pass → `references/validation/phase-b-correctness-pass.md`
- Phase C smoke-test results → `references/validation/phase-c-smoke-test-results.md`

# Chord Symbol Ambiguity and Parsing

Use this asset when the user gives chord symbols that could be read in more than one way, or when the agent needs to translate a chord symbol into concrete notes. This file complements `chord-symbol-conventions.md`; it is more operational and parser-like.

## Default assumption

For user-facing composition help, assume **lead-sheet chord symbols** unless the user clearly asks for Roman numerals, Nashville numbers, figured bass, or post-tonal labels.

When ambiguity does not affect the answer, proceed and name the assumption. When ambiguity affects the notes, ask or provide both readings.

## Parse order for lead-sheet symbols

Read a chord symbol in this order:

1. **Root** — `C`, `D♭`, `F♯`, etc.
2. **Core quality** — major, minor, diminished, augmented, sus2, sus4, power/no3.
3. **Seventh quality** — dominant 7, major 7, minor 7, diminished 7, half-diminished.
4. **Extensions** — 9, 11, 13.
5. **Alterations** — ♭5, ♯5, ♭9, ♯9, ♯11, ♭13.
6. **Add/omit instructions** — add9, no3, no5, omit5.
7. **Bass note** — slash chord after `/`.

## Quick parser table from C

| Symbol | Core reading | Notes |
|---|---|---|
| `C` | major triad | C E G |
| `Cm` / `C-` | minor triad | C E♭ G |
| `Cdim` / `C°` | diminished triad | C E♭ G♭ |
| `Caug` / `C+` | augmented triad | C E G♯ |
| `Csus4` / `Csus` | suspended 4th | C F G |
| `Csus2` | suspended 2nd | C D G |
| `C5` | power chord | C G |
| `C(no3)` | no third; usually power-chord-like | C G |
| `C(no5)` | omit fifth | C E |
| `C7` | dominant 7 | C E G B♭ |
| `Cmaj7` / `CΔ` | major 7 | C E G B |
| `Cm7` / `C-7` | minor 7 | C E♭ G B♭ |
| `Cm7♭5` / `Cø7` | half-diminished 7 | C E♭ G♭ B♭ |
| `Cdim7` / `C°7` | fully diminished 7 | C E♭ G♭ B𝄫 |
| `Cm(maj7)` / `CmMaj7` | minor-major 7 | C E♭ G B |
| `Cmaj7♯11` | major 7 plus raised 11 | C E G B F♯ |
| `C7♭9` | dominant 7 plus flat 9 | C E G B♭ D♭ |
| `C7♯9` | dominant 7 plus sharp 9 | C E G B♭ D♯ |
| `C7♭13` | dominant 7 plus flat 13 | C E G B♭ A♭ |
| `C13` | dominant 13; 11 often omitted | C E G B♭ D A |
| `Cadd9` | major triad plus 9, no 7 | C E G D |
| `C6/9` | major triad plus 6 and 9 | C E G A D |

## Common ambiguous symbols

| Symbol | Possible readings | Agent default | Clarify when |
|---|---|---|---|
| `C-` | C minor; rarely a dash separator | C minor in jazz/lead-sheet contexts | user is using plain-text separators |
| `C+` | C augmented triad; sometimes shorthand for a raised-5 chord | C augmented triad | extensions are missing but implied by context |
| `C2` | Csus2 or Cadd9/add2 | likely Csus2 in guitar/pop charts | voicing or third matters |
| `C4` | Csus4 or pitch C4 | chord symbol if in chord chart; pitch if in melody/register context | both chord and pitch notation appear together |
| `Cmaj` | C major triad or informal Cmaj7 | major triad unless seventh context implies otherwise | chart also uses `min`, `dom`, or other informal labels |
| `CΔ` | C major 7 in jazz; not a triangle chord | Cmaj7 | user is outside jazz and may not know delta notation |
| `C7+5` | C7♯5 | dominant 7 sharp 5 | user also uses `+` for augmented triads inconsistently |
| `C7-5` | C7♭5 in older charts | dominant 7 flat 5 | chart mixes minus for minor and flat |
| `Calt` | often C7alt but missing the 7 | C7alt in jazz context | non-jazz context or no dominant function |
| `C/D` | C triad over D bass, not necessarily D11 | slash chord | functional label matters |
| `D/C` | D triad over C bass or polychord shorthand | slash chord in lead sheets | upper/lower triad stack matters |

## Add, sus, and omit rules

| Marking | Meaning | Watch out |
|---|---|---|
| `add9` | Add 9 to a triad without adding a 7th | `Cadd9` is not `C9` |
| `C9` | Dominant 9 unless otherwise marked | includes B♭, not B |
| `Cmaj9` | Major 9 | includes B natural |
| `Cm9` | Minor 9 | includes E♭ and B♭ |
| `sus4` | replace 3 with 4 | `C7sus4` = C F G B♭ |
| `sus2` | replace 3 with 2 | `Csus2` = C D G |
| `no3` | omit the third | often power-chord-like if fifth remains |
| `no5` / `omit5` | omit the fifth | not a power chord |
| `add11` | add 11 to a triad | can clash with major 3 in close voicing |

## Altered dominant spelling from C

| Tension | Pitch | Enharmonic caution |
|---|---|---|
| ♭9 | D♭ | not C♯ in functional spelling from C |
| ♯9 | D♯ | enharmonic to E♭ but functions as raised 9 over major 3 E |
| ♯11 / ♭5 | F♯ / G♭ | choose by symbol and context |
| ♭13 / ♯5 | A♭ / G♯ | choose by symbol and context |

In jazz notation, altered dominants often contain enharmonic pitch classes. For explanation, label the function of the tension, not just the piano key.

## What to do in an answer

When the user writes an ambiguous chord:

```md
I’m reading `C2` as `Csus2` here because this looks like a pop/guitar chart. If you meant `Cadd9`, keep E in the chord: C–E–G–D. If you meant sus2, replace E with D: C–D–G.
```

When rewriting a chart:

```md
Original: | C2 | G/B | Am7 | Fadd9 |
Clearer:  | Csus2 | G/B | Am7 | Fadd9 |
```

## Pitfalls

- Do not interpret every slash chord as a polychord.
- Do not assume `C9` means “C add 9.” It usually means dominant 9.
- Do not call `C(no5)` a power chord.
- Do not treat `alt` as a free color on non-dominant chords unless the genre or voicing context makes that clear.
- Do not respell every accidental enharmonically for ease of reading if the functional label matters.

## Cross-references

- Chord symbol conventions → `chord-symbol-conventions.md`
- Chord construction → `references/harmony/chord-construction.md`
- Jazz harmony → `references/harmony/jazz-harmony.md`
- Jazz voicings → `jazz-voicings.md`
- Scale-degree spelling → `scale-degree-spelling-cheatsheet.md`

# Validation Report — v1.0

## Package

- Package: `music-composition-v1.0-2026-04-27.zip`
- Base version: `music-composition-RC1-2026-04-27.zip`
- Date: 2026-04-27
- Release: v1.0
- Packaging scope: stable release finalization only; no major runtime feature expansion beyond RC1

## Automated checks

| Check | Result | Notes |
|---|---|---|
| Sanity checker | PASS | `python3 scripts/music_theory_sanity_check.py` |
| Internal Markdown links | PASS | Covered by sanity checker |
| Snapshot notes | PASS | Covered by sanity checker; 32 files with snapshot notes |
| Known regression checks | PASS | Covered by sanity checker |
| ZIP integrity | PASS | `ZipFile.testzip()` returned no bad file |
| ZIP hygiene | PASS | No cache, scratch, temp, `.pyc`, `.DS_Store`, or validation scratch files included |

Sanity-check output:

```text
music_theory_sanity_check: PASS
Checked 115 markdown files.
```

## Manual / user-run checks

| Check | Result | Notes |
|---|---|---|
| Phase B correctness | PASS | Recorded in `references/validation/phase-b-correctness-pass.md` |
| Phase C prompt smoke tests | PASS | User-run pass recorded in `references/validation/phase-c-smoke-test-results.md` |
| RC1 review decision | PASS | RC1 was assessed as suitable for v1.0 promotion; no RC2 blocker found |
| High-use assets | PASS for v1.0 | Spot-checked during Phase B; automated regression checks pass |
| Research/currentness/privacy | PASS for v1.0 | Snapshot/web-check language retained; no credential request guidance |
| Cultural specificity | PASS for v1.0 | Broad culture-as-color guidance guarded by routing, limitations, and sanity checks |
| File-loading behavior | PASS by user report | No over-reading or abnormal routing reported during Phase C |
| Creative variation | PASS by user report | Variation stayed within expected range during Phase C |

## Package contents

| Metric | Count |
|---|---:|
| Total packaged files | 116 |
| Markdown files | 115 |
| Python scripts | 1 |
| Internal Markdown reference scan | 2,054 |
| Snapshot-note files | 32 |

## v1.0 contents added after RC1

| File | Purpose |
|---|---|
| `RELEASE-NOTES-v1.0.md` | Stable release notes and first-read entry points |
| `VALIDATION-v1.0.md` | Stable release validation report |
| `RC1-v1.0-decision-memo-2026-04-27.md` | Decision record explaining why RC1 was promoted instead of creating RC2 |

Release-status wording was also updated in `MAINTENANCE.md`, `RELEASE-ROADMAP.md`, `references/00-navigation.md`, `references/validation/first-release-readiness.md`, and `scripts/music_theory_sanity_check.py`.

## Known limitations

See `KNOWN-LIMITATIONS.md`.

v1.0 intentionally does not claim to solve every possible genre, every tradition-specific practice, every live trend, direct audio transcription, DAW operation, MIDI generation, legal clearance, or comprehensive performer pedagogy. Current scenes, platform behavior, charts, artist activity, AI/copyright status, and streaming-service context should still be handled through the research workflows and active verification rules.

## Release decision

**Ship v1.0.**

No release blocker was found in the finalization pass. RC1 was promoted to v1.0 without creating RC2 because remaining desirable work is update/backlog material rather than first-release blocking work.

## Cross-references

- Release notes → `RELEASE-NOTES-v1.0.md`
- Release roadmap → `RELEASE-ROADMAP.md`
- Known limitations → `KNOWN-LIMITATIONS.md`
- RC1 validation report → `VALIDATION-RC1.md`
- RC1 → v1.0 decision memo → external packaging artifact `RC1-v1.0-decision-memo-2026-04-27.md`
- Phase B correctness pass → `references/validation/phase-b-correctness-pass.md`
- Phase C smoke-test results → `references/validation/phase-c-smoke-test-results.md`
- First-release readiness → `references/validation/first-release-readiness.md`

# Intervals and Scale Formulas

Compact formula reference. For full discussion of intervals and scales, see `fundamentals/pitch-intervals-scales.md`. For modes specifically, see `modes-cheatsheet.md`.

## Intervals — semitone count and quality

| Interval | Semitones | Inversion | Examples (from C) |
|----------|-----------|-----------|-------------------|
| Perfect unison (P1) | 0 | P8 | C–C |
| Minor 2nd (m2) | 1 | M7 | C–D♭ |
| Major 2nd (M2) | 2 | m7 | C–D |
| Minor 3rd (m3) | 3 | M6 | C–E♭ |
| Major 3rd (M3) | 4 | m6 | C–E |
| Perfect 4th (P4) | 5 | P5 | C–F |
| Augmented 4th (A4) / Diminished 5th (d5) | 6 | itself (tritone) | C–F♯ / C–G♭ |
| Perfect 5th (P5) | 7 | P4 | C–G |
| Augmented 5th (A5) | 8 | d4 | C–G♯ |
| Minor 6th (m6) | 8 | M3 | C–A♭ |
| Major 6th (M6) | 9 | m3 | C–A |
| Minor 7th (m7) | 10 | M2 | C–B♭ |
| Major 7th (M7) | 11 | m2 | C–B |
| Perfect octave (P8) | 12 | P1 | C–C (next octave) |
| Minor 9th | 13 | — | C–D♭ (next octave) |
| Major 9th | 14 | — | C–D (next octave) |
| Augmented 9th | 15 | — | C–D♯ (next octave) |
| Perfect 11th | 17 | — | C–F (next octave) |
| Augmented 11th | 18 | — | C–F♯ (next octave) |
| Minor 13th | 20 | — | C–A♭ (next octave) |
| Major 13th | 21 | — | C–A (next octave) |

## Interval inversion rule

Sum of an interval and its inversion = 9 (in size). Quality flips: major↔minor, augmented↔diminished, perfect↔perfect.

| Original | Inverts to |
|----------|------------|
| Major | Minor |
| Minor | Major |
| Perfect | Perfect |
| Augmented | Diminished |
| Diminished | Augmented |

## Triads — three-note chord formulas

| Triad | Formula | Intervals from root | Example (root C) |
|-------|---------|---------------------|------------------|
| Major | 1 - 3 - 5 | M3 + m3 | C - E - G |
| Minor | 1 - ♭3 - 5 | m3 + M3 | C - E♭ - G |
| Diminished | 1 - ♭3 - ♭5 | m3 + m3 | C - E♭ - G♭ |
| Augmented | 1 - 3 - ♯5 | M3 + M3 | C - E - G♯ |
| Sus2 | 1 - 2 - 5 | M2 + P4 | C - D - G |
| Sus4 | 1 - 4 - 5 | P4 + M2 | C - F - G |

## Seventh chords — four-note formulas

| 7th chord | Formula | Symbol | Example (root C) |
|-----------|---------|--------|------------------|
| Major 7 | 1 - 3 - 5 - 7 | Cmaj7, CΔ, CM7 | C - E - G - B |
| Dominant 7 | 1 - 3 - 5 - ♭7 | C7 | C - E - G - B♭ |
| Minor 7 | 1 - ♭3 - 5 - ♭7 | Cm7, C-7 | C - E♭ - G - B♭ |
| Half-diminished 7 (m7♭5) | 1 - ♭3 - ♭5 - ♭7 | Cø7, Cm7♭5 | C - E♭ - G♭ - B♭ |
| Diminished 7 | 1 - ♭3 - ♭5 - 𝄫7 | C°7, Cdim7 | C - E♭ - G♭ - B𝄫 (= A) |
| Minor major 7 | 1 - ♭3 - 5 - 7 | Cm(maj7), CmΔ | C - E♭ - G - B |
| Augmented major 7 | 1 - 3 - ♯5 - 7 | Cmaj7♯5, C+Δ | C - E - G♯ - B |
| Augmented 7 | 1 - 3 - ♯5 - ♭7 | C7♯5 | C - E - G♯ - B♭ |
| Dominant 7 sus4 | 1 - 4 - 5 - ♭7 | C7sus4 | C - F - G - B♭ |

## Extended chords (9, 11, 13)

Add the upper extension to the 7th chord.

| Chord | Formula | Example (C root) |
|-------|---------|-------------------|
| Major 9 | 1-3-5-7-9 | C-E-G-B-D |
| Dominant 9 | 1-3-5-♭7-9 | C-E-G-B♭-D |
| Minor 9 | 1-♭3-5-♭7-9 | C-E♭-G-B♭-D |
| Major 13 | 1-3-5-7-9-(11)-13 | C-E-G-B-D-(F)-A; usually 11 omitted in major |
| Dominant 13 | 1-3-5-♭7-9-(11)-13 | C-E-G-B♭-D-A; 11 usually omitted |
| Minor 11 | 1-♭3-5-♭7-9-11 | C-E♭-G-B♭-D-F |
| Minor 13 | 1-♭3-5-♭7-9-11-13 | C-E♭-G-B♭-D-F-A |

### Altered tensions on dominant 7

| Tension | Formula |
|---------|---------|
| ♭9 | 1-3-5-♭7-♭9 |
| ♯9 | 1-3-5-♭7-♯9 |
| ♯11 (= ♭5) | 1-3-(5)-♭7-♯11 |
| ♭13 (= ♯5) | 1-3-(5)-♭7-♭13 |
| 7alt | All alterations available: ♭9, ♯9, ♯11, ♭13 |

## Major scale and its modes (formulas as scale-degree alterations from major)

| Scale | Formula | Notes (C tonic) |
|-------|---------|-----------------|
| Major (Ionian) | 1 - 2 - 3 - 4 - 5 - 6 - 7 | C D E F G A B |
| Dorian | 1 - 2 - ♭3 - 4 - 5 - 6 - ♭7 | C D E♭ F G A B♭ |
| Phrygian | 1 - ♭2 - ♭3 - 4 - 5 - ♭6 - ♭7 | C D♭ E♭ F G A♭ B♭ |
| Lydian | 1 - 2 - 3 - ♯4 - 5 - 6 - 7 | C D E F♯ G A B |
| Mixolydian | 1 - 2 - 3 - 4 - 5 - 6 - ♭7 | C D E F G A B♭ |
| Aeolian (natural minor) | 1 - 2 - ♭3 - 4 - 5 - ♭6 - ♭7 | C D E♭ F G A♭ B♭ |
| Locrian | 1 - ♭2 - ♭3 - 4 - ♭5 - ♭6 - ♭7 | C D♭ E♭ F G♭ A♭ B♭ |

## Three minor scales

| Scale | Formula | Notes (A tonic) |
|-------|---------|-----------------|
| Natural minor | 1 - 2 - ♭3 - 4 - 5 - ♭6 - ♭7 | A B C D E F G |
| Harmonic minor | 1 - 2 - ♭3 - 4 - 5 - ♭6 - 7 | A B C D E F G♯ |
| Melodic minor (ascending; also "jazz minor") | 1 - 2 - ♭3 - 4 - 5 - 6 - 7 | A B C D E F♯ G♯ |

## Modes of melodic minor

| Mode | Parent-scale degree | Formula | Notes (parallel C tonic) |
|------|----------|---------|-----------------|
| Melodic minor | ^1 | 1-2-♭3-4-5-6-7 | C D E♭ F G A B |
| Dorian ♭2 | ^2 | 1-♭2-♭3-4-5-6-♭7 | C D♭ E♭ F G A B♭ |
| Lydian augmented | ^3 | 1-2-3-♯4-♯5-6-7 | C D E F♯ G♯ A B |
| Lydian dominant | ^4 | 1-2-3-♯4-5-6-♭7 | C D E F♯ G A B♭ |
| Mixolydian ♭6 | ^5 | 1-2-3-4-5-♭6-♭7 | C D E F G A♭ B♭ |
| Locrian ♮2 | ^6 | 1-2-♭3-4-♭5-♭6-♭7 | C D E♭ F G♭ A♭ B♭ |
| Altered (super Locrian) | ^7 | 1-♭9-♯9-3-♭5/♯11-♭13-♭7 in dominant-function spelling | C D♭ D♯ E G♭/F♯ A♭ B♭ |

## Modes of harmonic minor (selected)

| Mode | Parent-scale degree | Formula | Notes (parallel C tonic) |
|------|----------|---------|-----------------|
| Harmonic minor | ^1 | 1-2-♭3-4-5-♭6-7 | C D E♭ F G A♭ B |
| Phrygian dominant | ^5 | 1-♭2-3-4-5-♭6-♭7 | C D♭ E F G A♭ B♭ |

## Pentatonic and blues

| Scale | Formula | Notes (C tonic) |
|-------|---------|-----------------|
| Major pentatonic | 1 - 2 - 3 - 5 - 6 | C D E G A |
| Minor pentatonic | 1 - ♭3 - 4 - 5 - ♭7 | C E♭ F G B♭ |
| Blues scale (minor + ♭5) | 1 - ♭3 - 4 - ♭5 - 5 - ♭7 | C E♭ F G♭ G B♭ |
| Major blues | 1 - 2 - ♭3 - 3 - 5 - 6 | C D E♭ E G A |

## Symmetric scales

| Scale | Formula | Notes (C tonic) |
|-------|---------|-----------------|
| Whole tone | 1-2-3-♯4-♯5-♭7 (6 notes) | C D E F♯ G♯ B♭ |
| Diminished (whole-half) | 1-2-♭3-4-♭5-♭6-𝄫7-7 (8 notes) | C D E♭ F G♭ A♭ A B |
| Diminished (half-whole) | 1-♭2-♯2-3-♯4-5-6-♭7 (8 notes) | C D♭ D♯ E F♯ G A B♭ |
| Augmented | 1-♯2-3-5-♭6-7 (6 notes) | C D♯ E G A♭ B |
| Chromatic | All 12 semitones | C C♯ D D♯ E F F♯ G G♯ A A♯ B |

## Bebop scales (8 notes; major scale + chromatic passing tone)

| Scale | Formula | Notes (C tonic) |
|-------|---------|-----------------|
| Bebop major | 1-2-3-4-5-♭6-6-7 | C D E F G A♭ A B |
| Bebop dominant | 1-2-3-4-5-6-♭7-7 | C D E F G A B♭ B |
| Bebop dorian | 1-2-♭3-3-4-5-6-♭7 | C D E♭ E F G A B♭ |

## Other scales

| Scale | Formula | Notes (C tonic) |
|-------|---------|-----------------|
| Hungarian minor | 1-2-♭3-♯4-5-♭6-7 | C D E♭ F♯ G A♭ B |
| Byzantine (double harmonic) | 1-♭2-3-4-5-♭6-7 | C D♭ E F G A♭ B |
| Hirajoshi (Japanese) | 1-2-♭3-5-♭6 | C D E♭ G A♭ |
| Yo (Japanese) | 1-2-4-5-6 | C D F G A |
| In (Japanese) | 1-♭2-4-5-♭6 | C D♭ F G A♭ |

## Key signatures — sharps and flats in order

**Order of sharps**: F♯ C♯ G♯ D♯ A♯ E♯ B♯
**Order of flats**: B♭ E♭ A♭ D♭ G♭ C♭ F♭

| Sharps | Major key | Relative minor |
|--------|-----------|----------------|
| 0 | C | A minor |
| 1 | G | E minor |
| 2 | D | B minor |
| 3 | A | F♯ minor |
| 4 | E | C♯ minor |
| 5 | B | G♯ minor |
| 6 | F♯ | D♯ minor |
| 7 | C♯ | A♯ minor |

| Flats | Major key | Relative minor |
|-------|-----------|----------------|
| 0 | C | A minor |
| 1 | F | D minor |
| 2 | B♭ | G minor |
| 3 | E♭ | C minor |
| 4 | A♭ | F minor |
| 5 | D♭ | B♭ minor |
| 6 | G♭ | E♭ minor |
| 7 | C♭ | A♭ minor |

## Cross-references

- Full discussion of intervals and scales → `fundamentals/pitch-intervals-scales.md`
- Modes (cheatsheet format) → `modes-cheatsheet.md`
- Chord construction → `harmony/chord-construction.md`
- Voice leading → `harmony/voice-leading.md`
- Jazz chord-scale relationships → `harmony/jazz-harmony.md`
- Notation conventions → `fundamentals/notation-and-conventions.md`

# Diagnostic Checklists

Use this file when the user gives a fuzzy creative problem: "boring", "weak", "generic", "awkward", "not emotional", "too cluttered", "doesn't hit". Diagnose before suggesting fixes.

## Fast diagnostic workflow

1. Restate the problem as a musical hypothesis.
2. Check the highest-impact categories first: melody, harmony, rhythm, form, arrangement, lyric/prosody.
3. Suggest **one small test** before a full rewrite.
4. Offer 2–3 fixes with different trade-offs.

```md
Likely diagnosis: [one sentence]
Fastest test: [change one variable]
Fix options: [2–3 concrete choices]
Risk: [what could get worse]
```

## Chorus does not hit

| Check | Symptom | Fix |
|---|---|---|
| Melodic range | Chorus peak is same or lower than verse | Raise hook peak by a 3rd/4th/5th, or hold a longer note |
| Hook timing | Main hook arrives too late | Put a title/hook fragment in the first 1–2 bars of chorus |
| Pre-chorus tension | Chorus arrives without setup | Use V, V/vi, rising sequence, pedal V, or drum fill before chorus |
| Harmonic reset | Chorus loop feels like verse loop | Start chorus on I or IV if verse starts vi/i; or invert the verse loop |
| Arrangement density | No new layer/register enters | Add high octave, wider pad, cymbal, group vocal, or bass register change |
| Rhythm | Chorus is as busy as verse | Simplify melody rhythm or make drums more regular/countable |
| Lyric | Hook word is not emphasized | Put the key word on the highest, longest, or most repeated note |

Fast test: mute two verse layers, then unmute/add them at chorus. If the chorus suddenly works, the problem was contrast, not chords.

## Verse feels boring

| Check | Symptom | Fix |
|---|---|---|
| Phrase shape | Every line has same contour | Change line 3: higher range, shorter rhythm, or answer phrase |
| Harmonic rhythm | One loop repeats without lyric payoff | Change bass note, inversion, or chord on the final line only |
| Groove | Drums repeat exactly | Add pickup, ghost note, or percussion answer at bar 4/8 |
| Lyric perspective | No new information per line | Make each line reveal a new detail, image, or contradiction |
| Register | Vocal stays in same 3–4 notes | Use a lower first half and a higher answer phrase |

Fast test: keep the chords, rewrite only the final two bars of the verse.

## Melody is forgettable

| Check | Symptom | Fix |
|---|---|---|
| Motif | No repeated cell | Create a 2–5 note cell and repeat it with variation |
| Contour | Mostly flat or random | Use one clear peak and one clear descent |
| Rhythm | All notes same length | Add one longer anchor note or a syncopated pickup |
| Interval identity | Only stepwise motion | Add one memorable leap, then resolve by step |
| Hook word | Important word falls on weak note | Move it to a strong beat, long note, or high note |

Fast test: sing the melody with only rhythm on one pitch. If the rhythm alone is unmemorable, fix rhythm first.

## Harmony feels generic

| Check | Symptom | Fix |
|---|---|---|
| Loop default | I-V-vi-IV or vi-IV-I-V feels too exposed | Change bass inversions, harmonic rhythm, or final chord |
| No color tones | Triads only, same register | Add add9, sus2/sus4, 6, maj7, m9 where genre-appropriate |
| No section contrast | Verse/pre/chorus use same harmonic identity | Use modal mixture in pre or a brighter tonic start in chorus |
| No cadence | Loop never points anywhere | Add V, V/vi, ii-V, or backdoor cadence before arrival |
| Too many tricks | Every chord is chromatic | Simplify the loop and save one color chord for the lyric payoff |

Fast test: keep the same Roman numerals but change voicing and bass. If it improves, do not reharmonize yet.

## Transition is awkward

| Check | Symptom | Fix |
|---|---|---|
| Too abrupt | New section feels pasted in | Add pickup, fill, riser, common tone, or dominant preparation |
| Too prepared | Transition loses surprise | Cut the last bar, hard-stop, or use direct modulation |
| Key relation unclear | Modulation feels random | Use pivot chord, common tone, V/new key, or sequence |
| Density jump | Too many layers enter at once | Stage entrances across 2–4 bars |
| Groove mismatch | Drums switch without handoff | Add shared percussion pattern or transition fill |

Fast test: loop the last two bars before the transition and first two bars after it. Fix only the join.

## Arrangement sounds muddy or cluttered

| Check | Symptom | Fix |
|---|---|---|
| Register collision | Piano/guitar/synth/vocal all in same octave | Move one part up/down or thin voicing |
| Too many sustained parts | Harmonic cloud masks melody | Make one part rhythmic, one sustained, one melodic |
| Low-end conflict | Bass, kick, left-hand piano all active | Simplify low rhythm; remove low chord tones |
| Countermelody | Competes with lead vocal | Put countermelody in gaps or higher register |
| No negative space | Every bar is full | Drop layers before important lyric/hook entries |

Fast test: remove every part that does not change the harmony, groove, or hook. Add back only what creates contrast.

## Groove does not lock

| Check | Symptom | Fix |
|---|---|---|
| Kick/bass mismatch | Bass accents fight kick | Align key bass notes with kick, then add offbeat movement |
| Backbeat weak | Snare/clap lacks anchor | Reinforce beats 2 and 4 or a half-time beat 3 |
| Hi-hat too static | Groove feels mechanical | Add velocity accents, rests, or 16th pickup |
| Swing mismatch | Parts imply different subdivisions | Decide straight, swung, triplet, or hybrid and align core parts |
| Vocal flow conflict | Vocal ignores groove accents | Move phrase starts around the groove, not always beat 1 |

Fast test: mute harmony and listen only to drums, bass, and vocal rhythm.

## Lyric or prosody sounds awkward

| Check | Symptom | Fix |
|---|---|---|
| Stress mismatch | Natural word stress falls on weak beat | Move stressed syllable to beat/accent |
| Too many syllables | Line feels rushed | Delete filler words or split line across phrases |
| Vowel issue | High note on closed/dark vowel is strained | Put open vowel on high/long note |
| Tonal language issue | Melody fights lexical tone | Simplify contour and consult language-specific guidance |
| Rhyme pressure | Line sounds forced | Loosen rhyme scheme; use slant rhyme or internal rhyme |

Fast test: speak the lyric in rhythm before singing it. If speech feels wrong, melody will not fix it.

## Bridge does not contrast

| Check | Symptom | Fix |
|---|---|---|
| Same register | Bridge sits where verse/chorus sit | Move lower for intimacy or higher for release |
| Same harmony | Bridge is another loop | Use relative key, borrowed chords, or pedal point |
| Same rhythm | No change in phrase energy | Halve or double rhythmic density |
| No narrative turn | Lyric says same thing | Change perspective, time, consequence, or emotional claim |
| Return weak | Final chorus not set up | End bridge on V, V/vi, ♭VII, or a silence/dropout |

Fast test: make the bridge either *thinner and more intimate* or *more harmonically unstable*. Middle-ground bridges often fail.

## Too close to a reference

| Check | Symptom | Fix |
|---|---|---|
| Melody contour | Same peaks and rhythm as reference | Change interval skeleton and phrase length |
| Hook phrase | Similar wording/title cadence | Rewrite lyric image and stress pattern |
| Signature riff | Recognizable instrumental cell | Replace with different rhythm or register/function |
| Arrangement cue | Same drop/fill/stab in same spot | Move cue, change instrumentation, or blend with second reference |
| Single-source imitation | Only one anchor track | Add two more references and extract shared variables instead |

Fast test: remove the reference audio. If you can still identify the source from one element, transform that element.

## Writer's block

| Symptom | Fix |
|---|---|
| Too many options | Limit to one scale, four chords, one instrument, ten minutes |
| Can't start | Write a bad 4-bar loop on purpose; revise after hearing it |
| Over-editing | Separate drafting and critique sessions |
| No direction | Choose three adjectives, one reference boundary, one technical constraint |
| Loop fatigue | Write the next section with opposite density/register/harmonic rhythm |

Fast test: create three intentionally different 8-bar sketches instead of improving one loop.

## Cross-references

- Response formatting → `response-templates.md`
- Progression options → `progressions-catalog.md`
- Hook fixes → `../references/songwriting/hooks-and-memorability.md`
- Melody fixes → `../references/melody/melodic-construction.md`, `../references/melody/motivic-development.md`
- Harmony fixes → `../references/harmony/reharmonization.md`, `../references/harmony/chromatic-harmony.md`
- Arrangement fixes → `../references/orchestration/arrangement-density.md`, `../references/production-aware/arrangement-for-mix.md`
- Prosody fixes → `../references/fundamentals/prosody-and-language.md`
- Critique mode → `../references/critique-and-feedback.md`

# Form Templates

Quick-reference templates for common forms in classical and popular music. Use these as starting structures, not commandments. For full discussion of these forms, see `form/classical-forms.md` and `form/popular-song-forms.md`. For energy arcs across these structures, see `form/narrative-and-transitions.md`.

## Classical forms

### Binary (AB)

```
||: A :||: B :||
   I → V       V → I
```

Each section repeated. Total: ~16-32 bars. Used in Baroque dances, folk songs.

### Rounded binary

```
||: A :||: B - A' :||
   I → V    V → I → I
```

The "A return" at the end of B. Total: ~24-32 bars. Common in classical period.

### Ternary (ABA)

```
A         B          A
I → I    contrast    I → I
```

Three balanced sections. The middle section (B) is often in a related key (relative or parallel minor/major, V, etc.). Used in slow movements, character pieces, da capo arias.

### Sonata form

```
EXPOSITION (often repeated)
  Theme 1 in tonic (I)
  Transition (modulating)
  Theme 2 in V (or III in minor mode)
  Closing material in V (or III)

DEVELOPMENT
  Modulating; fragments of themes; sequences; tension building
  Retransition (long V chord preparation)

RECAPITULATION
  Theme 1 in tonic
  Transition (rewritten to stay in tonic)
  Theme 2 in tonic (the structural payoff)
  Closing material in tonic

CODA (optional)
  Final affirmation; sometimes "second development"
```

Used in: first movements of classical sonatas, symphonies, string quartets, concertos. Also common in some last movements.

### Five-part rondo (ABACA)

```
A (refrain in I)
B (episode, often in V or relative)
A (refrain in I)
C (episode in distant key — IV, parallel mode, etc.)
A (refrain in I)
```

Lively; often used as finale.

### Seven-part rondo (ABACABA)

```
A - B - A - C - A - B - A
I   V   I  IV  I   V   I
```

A more symmetric rondo. The first "A-B-A" mirrors the last.

### Sonata-rondo

```
A B A C A B A
exposition   develop   recapitulation
```

Hybrid: A and B function as exposition; C as development; final A-B-A as recapitulation.

### Theme and variations

```
Theme (T) - Var I - Var II - Var III - ... - Final Variation - (Coda)
```

Each variation is the same length and harmonic structure as the theme. Variations get progressively more elaborate; often a slow variation late; final variation virtuosic.

### Through-composed

```
A - B - C - D - E - ...
```

No section repeats. Each is new material. Used in dramatic art songs, opera scenes, programmatic works.

### Strophic (multiple verses, no chorus)

```
Intro - V1 - V2 - V3 - V4 - Outro
```

Each verse uses the same melody with different lyrics. Common in folk songs, hymns.

### Sonata cycle (3-movement)

```
Movement 1: Sonata form, fast, in tonic
Movement 2: Slow, in related key (subdominant or relative)
Movement 3: Rondo or sonata-rondo, fast, in tonic
```

Common for Classical-era sonatas and concertos.

### Sonata cycle (4-movement)

```
Movement 1: Sonata form, fast, in tonic (sometimes with slow intro)
Movement 2: Slow, in related key
Movement 3: Minuet (or Scherzo) and Trio, in tonic, moderate-fast
Movement 4: Sonata, rondo, or sonata-rondo, fast, in tonic
```

Common for Classical-era symphonies and string quartets. Order can vary (Beethoven 9 has scherzo 2nd, slow 3rd).

### Suite

```
Movement 1: Allemande
Movement 2: Courante
Movement 3: Sarabande
Movement 4: (optional) Galanteries (minuet, gavotte, bourrée)
Movement 5: Gigue
```

Baroque dance suite. All movements typically in the same key.

### Concerto sonata form (1st movement)

```
EXPOSITION 1 (orchestra alone, in tonic)
  Theme 1, Theme 2 — both in tonic

EXPOSITION 2 (with soloist)
  Theme 1 in tonic (with solo)
  Transition (modulating)
  Theme 2 in V (with solo)
  Closing in V

DEVELOPMENT (soloist + orchestra)

RECAPITULATION
  Themes in tonic
  Cadenza (extended solo improvisation, near the end)
  Final orchestral close
```

The "double exposition" is the concerto's distinguishing feature.

### Fugue

```
EXPOSITION
  Voice 1: Subject (in I)
  Voice 2: Answer (in V) + Voice 1: Countersubject
  Voice 3: Subject (in I) + Voice 2 + Voice 1
  (etc., for as many voices)

EPISODE
  Modulating; based on motifs from subject and countersubject

MIDDLE ENTRIES
  Subject in related keys (vi, iv, etc.)

EPISODE / DEVELOPMENT
  More episodes; possible stretto

FINAL ENTRY
  Subject in tonic, often with augmentation or pedal point

CADENCE
```

## Popular song forms

### Verse-Chorus (modern pop default)

```
Intro (4-8 bars)
Verse 1 (8-16 bars)
Pre-Chorus (4-8 bars)
Chorus (8-16 bars)
Verse 2 (8-16 bars)
Pre-Chorus (4-8 bars)
Chorus (8-16 bars)
Bridge (8 bars)
Final Chorus (16-32 bars; possibly modulated)
Outro (4-8 bars)
```

Total: 3:00-3:30 at 120 BPM.

### Verse-Chorus with post-chorus (modern pop)

```
Intro
Verse 1
Pre-Chorus
Chorus
Post-Chorus
Verse 2
Pre-Chorus
Chorus
Post-Chorus
Bridge
Final Chorus
Post-Chorus (extended)
Outro
```

The post-chorus often features the most memorable hook.

### AABA (32-bar form / Tin Pan Alley)

```
A (8 bars): main theme
A (8 bars): repeat with slight variation or new lyrics
B (8 bars): bridge — contrasting melody, often modulates
A (8 bars): return
```

Used in jazz standards, Tin Pan Alley, classic Broadway.

### 12-bar blues

```
| I  | I  | I  | I  |  (or I-IV-I-I in "quick change")
| IV | IV | I  | I  |
| V  | IV | I  | I  |  (or V or turnaround in last bar)
```

The simplest blues form. Used as both song structure and improvisation framework.

### Jazz blues (standard reharmonization)

```
| I7  | IV7   | I7    | I7 (V7/IV) |
| IV7 | ♯iv°7 | I7    | VI7        |
| ii7 | V7    | iii vi| ii V        |
```

The 12-bar blues with secondary dominants and ii-V's added.

### 16-bar blues

```
| I  | I  | I  | I  |
| I  | I  | I  | I  |  (twice the I section)
| IV | IV | I  | I  |
| V  | IV | I  | I  |
```

A 4-bar extension of the 12-bar pattern.

### Strophic

```
Verse 1 - Verse 2 - Verse 3 - ...
```

All verses use the same melody. Each verse has different lyrics. Common in folk, hymns, narrative ballads.

### EDM standard form

```
Intro (16-32 bars)
Build 1 (16 bars)
Drop 1 (16-32 bars)
Breakdown (16-32 bars)
Build 2 (16 bars)
Drop 2 (16-32 bars; same as Drop 1 or with variation)
Outro (16 bars; fading down)
```

Total: 5-7 minutes. Designed for DJ mixing.

### EDM radio edit

```
Intro (4-8 bars)
Verse / Build (16 bars)
Drop (16 bars)
Breakdown (16 bars)
Build (8 bars)
Drop (16 bars)
Outro (4-8 bars)
```

Shortened to 3-4 minutes for radio.

### Trap / Hip-Hop modern

```
Intro (4-8 bars)
Verse 1 (16 bars)
Hook / Chorus (8 bars)
Verse 2 (16 bars)
Hook / Chorus (8 bars)
Verse 3 (16 bars; sometimes featured artist)
Hook / Chorus (8 bars)
Outro (variable)
```

The 16-bar verse is the standard ("dropping a 16").

### K-pop standard

```
Intro (4-8 bars)
Verse 1 (8 bars; often rapped)
Pre-Chorus (8 bars; melodic)
Chorus (8-16 bars)
Post-Chorus / Drop (8 bars)
Verse 2 (8 bars; rap)
Pre-Chorus (8 bars)
Chorus (8-16 bars)
Post-Chorus (8 bars)
Bridge (8-16 bars)
Final Chorus (16-24 bars; often modulated up)
Outro (4-8 bars)
```

Total: 3:00-3:30. Highly compressed.

### Country verse-chorus

```
Intro (instrumental)
Verse 1 (16 bars)
Chorus (8-16 bars; title hook at end of phrases)
Verse 2 (16 bars)
Chorus (8-16 bars)
Bridge (8 bars; optional)
Chorus (8-16 bars; possibly with variation)
Outro (instrumental, often fading)
```

### R&B / soul ballad

```
Intro
Verse 1 (often with extended melismatic vocal)
Pre-Chorus
Chorus (with multiple ad-libs and runs)
Verse 2
Pre-Chorus
Chorus
Bridge (often modulated or with key shift)
Final Chorus (extended; vocal flourishes; outro tag)
```

## Trailer music

```
Cold open (0:00-0:15): atmospheric pad
Slow build (0:15-0:30): adding strings, sub bass, accents
Medium build (0:30-0:45): full orchestra entering
First peak (0:45-1:00): drop with full orchestra
Brief breakdown (1:00-1:15): suspended chord or atmospheric
Final peak (1:15-1:30): massive — full orchestra + choir + max percussion
Cool-down (1:30-end, optional)
```

Total: 60-120 seconds.

## Film cue (basic)

```
Statement (0-30s): theme introduced
Development (30s-1:00): theme transformed, intensifying
Climax (1:00-1:30): full orchestra, theme peaked
Resolution (1:30-2:00): release, possible callback to theme
```

## Multi-part / progressive

```
Part 1 (Movement 1): atmospheric intro
Part 2 (Movement 2): driving passage
Part 3 (Movement 3): contemplative slow section
Part 4 (Movement 4): build to climax
Part 5 (Movement 5): release / return
```

Used in prog rock, progressive metal, longer EDM, post-rock.

## Quick decision matrix

| Want | Use |
|------|-----|
| Modern pop hit | Verse-pre-chorus-chorus(-postchorus) format |
| Classical homage | Sonata form, ternary, or theme-and-variations |
| Standalone instrumental | Through-composed or ABA |
| Jazz standard | AABA in 32 bars, II-V-I-driven |
| Blues | 12-bar blues |
| Singer-songwriter / folk | Strophic or simple verse-chorus |
| Hip-hop track | 16-bar verses with 8-bar choruses |
| EDM club track | 5-7 min with build-drop-breakdown structure |
| EDM radio edit | 3-4 min compressed version |
| K-pop / J-pop | Multi-section: verse-prechorus-chorus-postchorus |
| Trailer music | 60-120 second arc with build-peak-build-bigger peak |
| Film cue | Match form to scene length and dramatic shape |
| Concept album track | Through-composed or extended form with multiple sections |

## Cross-references

- Classical forms in detail → `form/classical-forms.md`
- Popular song forms in detail → `form/popular-song-forms.md`
- Energy arcs across forms → `form/narrative-and-transitions.md`
- Phrase structure (building blocks) → `melody/phrase-structure.md`
- Genre-specific form variations → various `genres/*.md` files
- Production-side form decisions → `production-aware/energy-and-dynamics.md`
- Cadences (ending each section) → `cadence-reference.md`
- Progressions for sections → `progressions-catalog.md`

# Cadence Reference

Quick lookup for cadence types. A cadence is how a phrase ends — the harmonic punctuation that signals "concluded", "paused", "surprised", or "continuing".

For *why* cadences work, see `references/harmony/functional-harmony.md`. For chromatic and modulation-driven cadences, see `references/harmony/chromatic-harmony.md` and `references/harmony/modulation.md`.

## Authentic cadences (V → I)

The default conclusive cadence. The dominant resolves to tonic.

| Cadence | Definition | Strength | Feeling |
|---------|------------|----------|---------|
| **Perfect Authentic Cadence (PAC)** | `V → I` (or `V7 → I`); both chords in root position; ^1 in the soprano of `I` | Maximum closure | Period at end of sentence |
| **Imperfect Authentic Cadence (IAC), root position** | `V → I` in root position, but ^3 or ^5 in the soprano of `I` | Strong closure, lighter | Comma at end of clause |
| **IAC, inverted** | `V → I` with one or both chords inverted | Mild closure | Soft pause |
| **Contrapuntal cadence (vii° → I, or V⁶ → I)** | Same dominant function, but with bass leading-tone motion | Mild closure | Smoother, less terminal |

**Use cases**: PAC at the end of a piece, end of a major formal section, end of an antecedent–consequent period (consequent ends PAC). IAC at the end of an antecedent (sets up further motion), internal phrase endings.

## Half cadence (HC)

Phrase ends on `V` (sometimes `V7`, but more often the bare triad).

| Variant | Notes |
|---------|-------|
| **Half cadence** | Phrase ends on `V`. Sounds like a question. |
| **Phrygian half cadence** (minor key) | `iv⁶ → V`; bass moves ♭^6 → ^5 by half step (diatonic in natural minor). Bittersweet, archaic; common in Baroque slow movements and Renaissance music. |

**Use cases**: end of an antecedent in a period; mid-piece pause; the "comma" before a chorus or a returning theme.

## Plagal cadence (IV → I)

Sometimes called the "amen cadence" because of its association with church endings.

| Variant | Notes |
|---------|-------|
| **Plagal cadence** | `IV → I`. Gentle, rocking. Doesn't have the leading-tone gravity of authentic cadences. |
| **Minor plagal** (in major) | `iv → I`. Borrowed iv. Wistful, "smiling-through-tears" effect. Common in The Beatles, modern pop, R&B. |
| **Double plagal** | `♭VII → IV → I`. Common in rock — the end of "Hey Jude". |

Plagal cadences are weaker conclusions than authentic. Often used as appendages after a PAC ("amen" tag) or as primary cadences in genres that want softness (gospel, rock ballad, lullaby, hymn).

## Deceptive cadence (V → vi, etc.)

The dominant resolves to something other than tonic — usually `vi` (in major) or `♭VI` (in minor or with modal mixture).

| Variant | Notes |
|---------|-------|
| **Standard deceptive** | `V → vi` in major. Bittersweet substitution; ^1 expected, ^6 delivered. |
| **Deceptive to ♭VI in major** | `V → ♭VI`. Modal mixture flavor — darker, more dramatic than `V → vi`. |
| **Deceptive in minor** | `V → ♭VI`. Diatonic in natural/harmonic minor; classical ear hears it as fully deceptive because the major-mode resolution to `i` was expected. |
| **Deceptive to IV** | `V → IV`. Rare in classical, used in pop. The dominant feels "redirected" plagally. |

**Use cases**: delay closure for dramatic effect; extend a phrase ("just one more time"); set up a new section by avoiding tonic; end a verse to push into a chorus.

## Modal and special cadences

Cadences that work outside (or alongside) functional harmony.

| Cadence | Definition | Where it's used |
|---------|------------|-----------------|
| **Phrygian cadence (modal, not classical Phrygian half)** | `♭II → I` (♭II as Neapolitan, or the modal supertonic in Phrygian); bass moves ♭^2 → ^1 | flamenco-adjacent writing, Spanish folk/dance-influenced contexts, certain film-scoring shorthand |
| **Lydian cadence** | `II → I` (II is a major chord on ^2, characteristic of Lydian) | Lydian-mode pieces; lifted/floating feel |
| **Mixolydian cadence** | `♭VII → I` (no leading tone) | Rock, Celtic music, some modal jazz |
| **Aeolian cadence** | `♭VII → i` or `iv → i` (no leading tone) | Minor-key rock, folk, modal pieces |
| **Picardy third (tierce de Picardie)** | A minor-key piece ends with a major tonic chord (i.e., `I` instead of `i`) | Baroque convention; modern uses for unexpected resolution |
| **Dorian cadence** | `IV → i` (in Dorian, IV is major) | Modal jazz, Celtic, prog rock |

## Jazz cadences

Functional cadences with jazz extensions and substitutions. See `references/harmony/jazz-harmony.md` for full treatment.

| Cadence | Notes |
|---------|-------|
| **ii–V–I** | The basic jazz cadence. Smoother and more colorful than V–I alone. Almost always with 7ths: `Dm7 – G7 – Cmaj7`. |
| **ii–V–I with tritone sub** | Replace `V7` with `♭II7` (e.g., `Dm7 – D♭7 – Cmaj7`). The tritone of `D♭7` is the same as `G7`. Smooth chromatic bass. |
| **Backdoor ii–V** | `iv–♭VII7–I` (e.g., `Fm7 – B♭7 – Cmaj7`). Approaches tonic via the parallel minor's ii–V. |
| **Cycle cadence** | `iii–vi–ii–V–I`. Extends the cadential approach. |
| **Minor ii–V–i** | `iiø7 – V7♭9 – i`. The minor-key analogue. |

## Cadence strength roughly ranked

From most conclusive to least:

1. **PAC** with strong rhythmic placement — total closure
2. **IAC**, root position, ^3 or ^5 on top — closure but lighter
3. **IAC**, inverted — soft closure
4. **Plagal** — gentle closure
5. **Phrygian half cadence (minor key)** — strong pause
6. **Half cadence** — open question
7. **Deceptive cadence** — surprise; phrase wants to continue
8. **No cadence (elision)** — the phrase ends on a non-cadential chord; motion continues uninterrupted

This ordering is contextual — a deceptive cadence inside a piece can feel more conclusive than a half cadence at the very end. Always check what the cadence does *in context*.

## Quick decision matrix

| Need to... | Use |
|------------|-----|
| End a piece firmly | PAC |
| End a section | PAC or IAC |
| Set up another section (open question) | HC |
| End softly without strong resolution | Plagal, IAC inverted |
| Surprise the listener / delay closure | Deceptive |
| Sound modal / non-classical | Mixolydian, Aeolian, Phrygian, Dorian (depending on color) |
| End a minor piece on an unexpected major | Picardy third |
| End a jazz tune | ii–V–I (with extensions) or ii–V–I with tritone sub |
| Sound flamenco-/Phrygian-adjacent | Andalusian cadence (`i – ♭VII – ♭VI – V`) — technically a progression, but functions as a cadential pattern |
| Avoid landing — keep motion alive | Deceptive, or elision into next phrase, or end on inversion |

## Notes on combining cadences across a piece

A typical 8-bar period:

- Antecedent (bars 1–4): ends HC or IAC
- Consequent (bars 5–8): ends PAC

A 16-bar AABA bridge to chorus often uses a deceptive cadence at the end of the bridge to launch into the chorus, since landing on `I` would feel premature.

Pop songs often avoid PAC at section ends to keep motion forward. A song that PACs at the end of every chorus feels overly resolved; one that uses HC, deceptive, or no cadence at all between sections feels propulsive.

A film cue that resolves PAC at the climax can feel conclusive — sometimes too conclusive for a story still in motion. Many composers prefer plagal, modal, or unresolved endings to leave space for what comes next on screen.

# Chord Symbol Conventions

Quick lookup for chord symbol notation across the four main systems used in Western music — jazz lead sheets (Real Book style), pop / commercial lead sheets, Roman numeral analysis, and the Nashville Number System. Different traditions write the same chord differently, and the same symbol can mean different things in different traditions. This file maps the territory.

For chord *construction* (what notes are in each chord), see `references/harmony/chord-construction.md`. For chord *function* (what each chord does in a key), see `references/harmony/functional-harmony.md`. For Roman numeral analysis applied to actual progressions, see `references/analysis.md`.

## The four systems at a glance

| System | Used by | Tells you | Example |
|--------|---------|-----------|---------|
| Real Book / jazz | Jazz musicians, leadsheet players | Specific pitches | `Cmaj7`, `D-7`, `G7♭9` |
| Pop lead-sheet | Pop, rock, gospel, R&B charts | Specific pitches | `Cmaj7`, `Dm7`, `G7♭9` |
| Roman numeral | Analysts, classical theorists | Function in a key | `IΔ`, `ii7`, `V7♭9` |
| Nashville Number System | Nashville session work, country | Function in a key, transposable | `1Δ`, `2m7`, `5 7♭9` |

The first two specify pitches and require you to know the key separately. The last two specify function (relative to a tonic) and require you to know the key to derive pitches. Each is correct for its purpose.

## Triad symbols

| Quality | Real Book / jazz | Pop / lead-sheet | Roman (in C major) | Nashville (in C) |
|---------|------------------|------------------|---------------------|--------------------|
| Major | `C` | `C` | `I` | `1` |
| Minor | `C-` or `Cmi` | `Cm` | `i` (in C minor) or `vi` (as A minor in C major) | `1m` |
| Diminished | `C°` or `Cdim` | `Cdim` or `C°` | `i°` or `vii°` (depending on degree) | `1°` |
| Augmented | `C+` or `Caug` | `Caug` or `C+` | `I+` (rare; usually as borrowed) | `1+` |
| Suspended 4 | `Csus` or `Csus4` | `Csus4` | (no standard RN) | `1sus` |
| Suspended 2 | `Csus2` | `Csus2` | (no standard RN) | `1sus2` |

## Seventh chord symbols

| Quality | Real Book / jazz | Pop / lead-sheet | Notes (root C) |
|---------|------------------|------------------|----------------|
| Major 7 | `CΔ` or `Cmaj7` or `CM7` | `Cmaj7` | C E G B |
| Dominant 7 | `C7` | `C7` | C E G B♭ |
| Minor 7 | `C-7` or `Cmi7` | `Cm7` | C E♭ G B♭ |
| Half-diminished 7 | `Cø7` or `Cø` | `Cm7♭5` | C E♭ G♭ B♭ |
| Fully diminished 7 | `C°7` | `Cdim7` | C E♭ G♭ B𝄫 (= A) |
| Minor-major 7 | `C-Δ` or `CmiΔ` | `CmMaj7` or `Cm(maj7)` | C E♭ G B |
| Augmented major 7 | `C+Δ` or `Cmaj7♯5` | `Cmaj7♯5` | C E G♯ B |
| Augmented 7 | `C+7` or `C7♯5` | `C7♯5` | C E G♯ B♭ |
| Dominant 7 sus4 | `C7sus` or `C7sus4` | `C7sus4` | C F G B♭ |

The biggest jazz–pop divergence is the major 7: jazz often writes `Δ` (delta) where pop writes `maj7`. Both are unambiguous in context; the agent should match whichever the user is using.

## Extension and alteration symbols

Same across systems, with minor formatting variation:

| Symbol | Meaning | Example chord |
|--------|---------|---------------|
| `9` | Add the 9th to a 7th chord | `Cmaj9` = C E G B D |
| `11` | Add the 11th | `Cm11` = C E♭ G B♭ D F |
| `13` | Add the 13th (and usually 9, sometimes 11) | `C13` = C E G B♭ D A (11 usually omitted) |
| `add9` | Add the 9th to a triad (no 7th) | `Cadd9` = C E G D |
| `add11` | Add the 11th to a triad | `Cadd11` = C E G F |
| `6` | Add the 6th to a triad (no 7th) | `C6` = C E G A |
| `6/9` | Add 6th and 9th to a triad | `C6/9` = C E G A D |
| `♭5` | Lower the 5th | `C7♭5` = C E G♭ B♭ |
| `♯5` | Raise the 5th | `C7♯5` = C E G♯ B♭ |
| `♭9` | Lower the 9th (in 7th-chord context) | `C7♭9` = C E G B♭ D♭ |
| `♯9` | Raise the 9th | `C7♯9` = C E G B♭ D♯ |
| `♯11` | Raise the 11th (often `= ♭5` enharmonically) | `Cmaj7♯11` = C E G B F♯ |
| `♭13` | Lower the 13th (often `= ♯5` enharmonically) | `C7♭13` = C E G B♭ A♭ |
| `alt` | Available altered tensions on a dominant | `C7alt` = C E B♭ plus a subset of D♭, D♯, G♭/F♯, A♭; the natural 5 is often omitted or context-dependent |

`b` and `#` are commonly substituted for `♭` and `♯` in plain-text contexts (e.g., `Cb9`, `C#11`). Both are universally understood.

## Slash chords

`X/Y` means chord X with bass note Y. Used in all four systems.

| Notation | Meaning | Example |
|----------|---------|---------|
| `C/E` | C major triad with E in the bass | First inversion of C major |
| `C/G` | C major triad with G in the bass | Second inversion |
| `C7/B♭` | C7 with B♭ in the bass | Third inversion of C7 |
| `F/G` | F major triad over G bass | "Slash chord" — not a chord tone of F; functions like a G dominant suspension |
| `C/D` | C major triad over D bass | Functions like D9sus4 |

In Roman numerals, slash chords use figured-bass numerals (`I⁶` = first inversion, `I⁶/⁴` = second inversion, etc.) instead of letter slashes for inversions, but reserve the letter-slash notation for non-chord-tone bass notes (`I/♭^7` or written explicitly).

In Nashville, slash chords use the same `/` notation: `1/3` (first inversion of I), `5/7` (V with leading tone in bass — note: `7` here means scale degree, not the 7th of the chord).

## Polychords

A polychord stacks two triads. Notation: `top / bottom` with a horizontal line (or `top|bottom` in plain text).

```
   D
  ───   = D major triad over C major triad
   C
```

Plain-text equivalents: `D/C` (ambiguous — can be read as a slash chord) or `D|C` (clearer). When ambiguity matters, write out: "D triad over C triad".

Polychords are mostly used in jazz, post-tonal, and film music. The bottom triad is usually the "root" function; the upper triad provides extensions and color. See `harmony/chord-construction.md` for upper structure triad use.

## Roman numeral specifics

| Symbol | Meaning |
|--------|---------|
| Capital (`I`, `IV`, `V`) | Major triad on that scale degree |
| Lowercase (`i`, `iv`, `v`) | Minor triad |
| `°` (`vii°`) | Diminished triad |
| `+` (`III+`) | Augmented triad |
| `7`, `9`, `11`, `13` after numeral | 7th-chord with that extension (`V7`, `ii⁹`) |
| Figured bass numerals (`⁶`, `⁶/⁴`, `⁶/⁵`, `⁴/³`, `⁴/²`) | Inversions |
| `♭` or `♯` prefix | Chromatic alteration of scale degree (`♭VI`, `♯iv°`) |
| `/` | Tonicization (`V/V`, `V7/vi`) |
| `N⁶` | Neapolitan 6 (= ♭II⁶) |
| `It⁶`, `Fr⁶`, `Ger⁶` | Italian, French, German augmented 6th chords |

Always specify the key context. `ii` means D minor in C major, but B♭ minor in A♭ major.

## Nashville Number System specifics

| Notation | Meaning |
|----------|---------|
| `1`, `4`, `5` | Major triads on those degrees (in major key) |
| `2m`, `3m`, `6m` | Minor triads |
| `7°` | Diminished triad on ^7 |
| `1Δ`, `4Δ` | Major 7 chords |
| `2m7`, `5 7` | 7th chords (note the space before `7` for dom7) |
| `1/3`, `5/7` | Inversions (number after slash = scale degree in bass) |
| `b3`, `b6`, `b7` | Borrowed chords (♭III, ♭VI, ♭VII) |
| `(1 4 / 5 6m)` | Parentheses for "two chords per bar" (1 and 4 in first half, 5 and 6m in second) |
| Diamond shapes around numbers | Held / sustained chord (long note) |
| Dot above number | Pushed chord (anticipated by an eighth) |

Nashville charts also use rhythmic shorthand (split bars, push notation, hits, kicks) that varies between charters. When in doubt, ask for the chart's legend.

## Common ambiguities and how to read them

A few notations look ambiguous out of context. The agent should ask if the answer matters:

| Symbol | Possible meanings | How to disambiguate |
|--------|-------------------|---------------------|
| `C-` | C minor (jazz convention) **or** C with no quality specified (rare) | Almost always means minor in jazz contexts |
| `Cm7` | C minor 7 **or** (extremely rarely) C with seventh as `m`-some-extension | Always C minor 7 in modern usage |
| `Cmaj` | C major triad **or** C major 7 (informal shorthand) | Triad in formal contexts; ambiguous in casual lead sheets — assume triad unless context implies 7 |
| `C7-5` or `C7−5` | C7♭5 (older notation using `-` for flat) | Older Real Books; modern usage prefers `C7♭5` |
| `C+` | C augmented triad **or** C with raised 5 (informal for `Cmaj7♯5` or similar) | Triad in classical/jazz; ambiguous in older fake books |
| `C2` | `Csus2` (informal) **or** `Cadd2` (= add9) | Usually means sus2 in pop guitar charts; `Cadd9` is unambiguous |
| `Cmaj7/9` | Cmaj9 (older notation) | Modern: `Cmaj9` |
| `C5` or `C(no3)` | C power chord (root + 5th, no 3rd) | `C5` is the standard rock notation; `C(no3)` is explicit. `C(no5)` means the fifth is omitted, so it is **not** a power chord. |

When the user writes a symbol that's ambiguous and the answer matters, ask. When it doesn't matter (the surrounding context fixes the interpretation), just proceed.

## Mapping a chord across systems

Quick translation. Same chord, four ways to write it:

C major key, the chord is **G dominant 7**:

| System | Notation |
|--------|----------|
| Real Book / jazz | `G7` |
| Pop / lead-sheet | `G7` |
| Roman numeral (in C major) | `V7` |
| Nashville (in C) | `5 7` |

C major key, the chord is **F minor with major 7th** (borrowed iv with raised 7):

| System | Notation |
|--------|----------|
| Real Book / jazz | `F-Δ` |
| Pop / lead-sheet | `Fm(maj7)` |
| Roman numeral | `iv(maj7)` (with implied modal mixture) |
| Nashville | `4m(maj7)` |

C major key, the chord is **D♭ dominant 7 substituting for V7**:

| System | Notation |
|--------|----------|
| Real Book / jazz | `D♭7` |
| Pop / lead-sheet | `Db7` |
| Roman numeral | `subV7/I` or `♭II7` |
| Nashville | `b2 7` |

## Common pitfalls when advising

- **Match the user's notation.** Don't translate `Cmaj7` to `CΔ` (or vice versa) without reason. The user picked their system for a reason.
- **Don't mix systems within one chart.** A chart with both Roman numerals and chord letters is hard to read.
- **Be careful with `m` vs `-` for minor.** Both are universal, but a chart that uses both inconsistently is confusing.
- **Don't assume `b` means flat in all contexts.** In some chord-naming software, `b` can be parsed as a note letter B. Use `♭` when typesetting allows.
- **Check key context before interpreting Roman numerals.** A chart in `B♭ major` writes the same `ii-V-I` as a chart in `C major`, but the actual chords are different.
- **Nashville `7` is ambiguous between scale degree and chord 7th.** `5 7` means "V dominant 7" (scale degree 5 with a 7th added). `7°` means "vii dim" (the chord on scale degree 7). Position and spacing matter.

## Cross-references

- Chord construction (what notes are in each chord) → `references/harmony/chord-construction.md`
- Functional analysis with Roman numerals → `references/harmony/functional-harmony.md`, `references/analysis.md`
- Notation conventions in scores generally → `references/fundamentals/notation-and-conventions.md`
- Jazz-specific chord vocabulary → `references/harmony/jazz-harmony.md`
- Voicings (the same chord rendered different ways) → `jazz-voicings.md`, `references/orchestration/voicing-and-texture.md`
- Common chord progressions across systems → `progressions-catalog.md`

# Jazz Voicings

A reference of common jazz chord voicings, organized by chord type and voicing strategy. For full theoretical discussion of chords and voicing, see `harmony/chord-construction.md`, `harmony/voice-leading.md`, `harmony/jazz-harmony.md`, and `orchestration/voicing-and-texture.md`.

## How to read this file

Voicings are written from bottom note to top, separated by hyphens. Pitches are concert pitch (no transposition).

For each chord type, the file gives:
- The basic spelling (root position)
- Common rootless voicings (jazz piano LH style)
- Drop voicings (drop 2, drop 3 for guitar)
- Spread / open voicings

## Major 7 chord voicings (Cmaj7)

### Basic spellings

| Voicing | Notes (C root) |
|---------|----------------|
| Root position close | C - E - G - B |
| 1st inversion | E - G - B - C |
| 2nd inversion | G - B - C - E |
| 3rd inversion | B - C - E - G |

### Rootless (jazz piano left hand)

When the bass plays the root, the piano LH plays just 3 + 7 (and possibly 9 or 13).

| Voicing | Notes (C root) | Comment |
|---------|----------------|---------|
| 3-7 shell | E - B | The skeletal voicing |
| 7-3 shell | B - E | Same notes, different inversion |
| 3-7-9 | E - B - D | Add 9 on top |
| 3-5-7-9 | E - G - B - D | With 5 (less common; 5 often omitted) |
| 7-9-3-5 (Bill Evans style) | B - D - E - G | "A" voicing — wide and lush |
| 3-5-7-13 | E - G - B - A | With 13 |

### Drop voicings (guitar)

| Voicing type | Notes (low → high) | Comment |
|--------------|---------------------|---------|
| Drop 2 (root pos.) | G – C – E – B | 2nd-from-top (G) dropped an octave |
| Drop 2 (1st inv.) | B – E – G – C | 2nd-from-top (B) dropped |
| Drop 2 (2nd inv.) | C – G – B – E | 2nd-from-top (C) dropped |
| Drop 2 (3rd inv.) | E – B – C – G | 2nd-from-top (E) dropped |
| Drop 3 (root pos.) | E – C – G – B | 3rd-from-top (E) dropped; wider bass gap |

### Spread / open voicings

| Voicing | Notes (low to high) |
|---------|---------------------|
| Cmaj7 spread | C2 - G3 - E4 - B4 - D5 |
| Cmaj9 spread | C2 - E3 - B3 - D4 - G4 |
| Cmaj13 spread | C2 - B3 - E4 - A4 (wide) |

## Minor 7 chord voicings (Cm7)

### Basic spellings

| Voicing | Notes |
|---------|-------|
| Root position close | C - E♭ - G - B♭ |
| 1st inversion | E♭ - G - B♭ - C |
| 2nd inversion | G - B♭ - C - E♭ |
| 3rd inversion | B♭ - C - E♭ - G |

### Rootless (jazz piano LH)

| Voicing | Notes |
|---------|-------|
| 3-7 shell | E♭ - B♭ |
| 3-7-9 | E♭ - B♭ - D |
| 3-5-7-9 | E♭ - G - B♭ - D |
| 7-9-3-5 (Bill Evans "A" voicing) | B♭ - D - E♭ - G |
| 3-5-7-11 | E♭ - G - B♭ - F |
| 3-7-9-11 | E♭ - B♭ - D - F |

### Drop 2 voicings

| Voicing | Notes |
|---------|-------|
| Drop 2 (root-pos) | E♭ - B♭ - C - G |
| Drop 2 (1st inv) | G - C - E♭ - B♭ |

## Dominant 7 chord voicings (C7)

### Basic spellings

| Voicing | Notes |
|---------|-------|
| Root position close | C - E - G - B♭ |
| 1st inversion | E - G - B♭ - C |
| 2nd inversion | G - B♭ - C - E |
| 3rd inversion | B♭ - C - E - G |

### Rootless

| Voicing | Notes |
|---------|-------|
| 3-7 shell | E - B♭ |
| 3-7-9 | E - B♭ - D |
| 3-7-13 | E - B♭ - A |
| 7-9-3-13 (Bill Evans) | B♭ - D - E - A |
| 3-5-7-9 | E - G - B♭ - D |
| 3-5-7-13 | E - G - B♭ - A |
| 3-5-7-9-13 (rich) | E - G - B♭ - D - A |

### Altered dominants

For C7♭9, C7♯9, C7♯11, C7♭13, C7alt:

| Chord | Rootless voicing | Notes |
|-------|------------------|-------|
| C7♭9 | E - B♭ - D♭ | Add ♭9 instead of ♮9 |
| C7♯9 | E - B♭ - D♯ | Add ♯9 (the "Hendrix chord" 7♯9 is famous) |
| C7♭9♯9 | E - B♭ - D♭ - D♯ | Both 9 alterations |
| C7♯11 | E - B♭ - D - F♯ | With ♯11 (= ♭5 raised) |
| C7♭13 | E - B♭ - A♭ | With ♭13 (= ♯5 spelled differently) |
| C7alt | E - B♭ - D♭ - D♯ - A♭ - F♯ | All alterations available |

The "altered" voicing typically uses some subset of these tensions, not all at once — usually 2-3 altered tensions.

### Tritone substitution

A C7 can be replaced by F♯7 (a tritone away). Their 3 and 7 swap:

- C7: 3 = E, ♭7 = B♭
- F♯7: 3 = A♯ (= B♭), ♭7 = E

The "shell" voicings of C7 and F♯7 are enharmonically the same. The bass is what distinguishes them.

## Half-diminished 7 (Cm7♭5 / Cø7)

| Voicing | Notes |
|---------|-------|
| Root position | C - E♭ - G♭ - B♭ |
| 3-7 shell | E♭ - B♭ |
| 3-7-9 | E♭ - B♭ - D |
| Rootless typical | F - B♭ - D - E♭ (with 11 on bottom; "ii of minor") |

The Cø7 in a minor ii-V-i context is often voiced with F (the 11) on the bottom for a smoother voice-leading to the V chord.

## Diminished 7 (C°7 / Cdim7)

| Voicing | Notes |
|---------|-------|
| Root position | C - E♭ - G♭ - B𝄫 (= A) |

Diminished 7 chords are symmetrical — every note can be considered the root. C°7 = E♭°7 = G♭°7 = A°7 (all enharmonically the same notes).

## Sus chords

### Csus4 / C7sus4

| Voicing | Notes |
|---------|-------|
| Csus4 triad | C - F - G |
| C7sus4 | C - F - G - B♭ |
| Rootless C7sus4 | F - B♭ - C - E♭ — actually a Bb/F voicing |

### Csus2

| Voicing | Notes |
|---------|-------|
| Csus2 triad | C - D - G |
| Csus2(add9) | C - D - G - D — basically just sus2 |

## Voicing strategies

### "Shells" (3-7 voicings)

The most economical voicing — just 3rd and 7th of the chord. Two notes carry the harmonic identity.

Use when:
- Comping behind a soloist (give them space)
- The bass plays the root
- The melody (or another voice) provides the upper extensions

### Rootless voicings (Bill Evans style)

Pioneered by Bill Evans for piano. The bass plays the root; the piano LH plays 3-5-7-9 or 7-9-3-5 voicings (no root). Frees up the LH to be more harmonically interesting.

Two main shapes:

- **"A" voicing**: 3-5-7-9 (or 7-9-3-5)
- **"B" voicing**: 7-9-3-5 (or 3-5-7-9)

(The naming convention depends on the source.)

The two voicings are designed to alternate smoothly through ii-V-I:

- Dm7 (in Cm context): F (3) - A (5) - C (7) - E (9)
- G7 ("A" voicing): F (♭7) - A (9) - B (3) - E (13)
- Cmaj7: E (3) - G (5) - B (7) - D (9)

The voice leading is smooth — only one voice moves per chord change.

### Drop voicings (guitar)

Take a close-position 4-note voicing and drop one note an octave down.

- **Drop 2**: drop the 2nd note from the top
- **Drop 3**: drop the 3rd note from the top
- **Drop 2 + 4**: drop the 2nd and 4th notes from the top

Drop 2 is the most common jazz guitar voicing because it lays well on the strings.

The procedure for Cmaj7, working through all four inversions:

| Close position (low → high) | 2nd-from-top voice | Drop 2 result (low → high) |
|---|---|---|
| C – E – G – B (root pos.) | G | **G – C – E – B** |
| E – G – B – C (1st inv.) | B | **B – E – G – C** |
| G – B – C – E (2nd inv.) | C | **C – G – B – E** |
| B – C – E – G (3rd inv.) | E | **E – B – C – G** |

Each drop-2 voicing has different intervals between notes; learning all four inversions covers the fingerboard.

For Drop 3, the same logic with the 3rd-from-top voice. Cmaj7 root position close (`C – E – G – B`): 3rd from top is `E`, drop it down → `E – C – G – B`. Drop 3 voicings have a wider gap between bass and the rest, useful when you want a clear "bass + chord" separation.

### Quartal voicings

Stack of 4ths. The "So What" voicing (modal jazz):

D - G - C - F - A — five notes stacked in 4ths.

Other quartal voicings:

- E - A - D - G - C (over E Dorian or similar)
- F - B♭ - E♭ - A♭ (over F minor or as polychord)

### Spread / wide voicings

Notes spread across multiple octaves with gaps. Common in big-band writing and orchestral jazz.

| Style | Voicing |
|-------|---------|
| Cinematic Cmaj7 | C2 - G3 - C4 - E4 - B4 |
| Cinematic Cm7 | C2 - G3 - B♭3 - E♭4 - F4 (with 11) |
| Cinematic G7 | G2 - F3 - B3 - D4 - F4 |

## Common jazz progressions with rootless voicings

### Major ii-V-I in C

| Chord | Rootless voicing |
|-------|------------------|
| Dm7 | F - A - C - E |
| G7 | F - A - B - E |
| Cmaj7 | E - G - B - D |

The voice leading: Dm7's F stays into G7 (becomes 7th); G7's F resolves down to E (becomes 3rd of Cmaj7). Smooth.

### Minor ii-V-i in A minor

| Chord | Rootless voicing |
|-------|------------------|
| Bm7♭5 | F - A - B♭ - D |
| E7♭9 | G♯ - D - F - A | (D = ♭7; F = ♭9; A = 11; G♯ = 3) |
| Am7 | E - G - C - D | (E = 5; C = ♭3; D = 11) |

Or for Am(maj7) tonic: E - G♯ - C - D (with raised 7).

### Minor V-I in jazz minor

| Chord | Voicing |
|-------|---------|
| E7alt | G♯ - D - F - C (3, ♭7, ♭9, ♭13) |
| E7♭9♯9 | G♯ - D - F - G (3, ♭7, ♭9, ♯9; altered-color subset) |
| Am(maj7) | E - G♯ - C - D (rootless) |

The altered V resolves into the tonic minor with raised 7 (jazz minor) or natural 7 (classical minor).

## Voicing for specific contexts

### Solo piano comping

The piano covers everything. Typical voicings:
- LH: rootless 3-5-7-9 in mid register
- RH: melody on top, with chord tones harmonizing
- Both hands cover bass, harmony, melody

### Ensemble comping (with bass and drums)

The bass plays the root. The pianist plays rootless voicings:
- LH: shell or partial chord (3-7 shell often enough)
- RH: chord tones with extensions or melodic interjections

### Big band brass section

4-note voicings spread across trumpets and trombones:
- Lead trumpet has the melody (top voice)
- Other voices provide harmony underneath
- Drop 2 is the standard approach

### Solo guitar

Drop 2 voicings on the top 4 strings (or middle 4 strings) cover most needs.

For "chord-melody" playing:
- Top voice = melody
- Middle voices = chord tones
- Bottom = bass note

## Common pitfalls when advising

- **Don't recommend rootless voicings without bass.** The voicing is incomplete without root somewhere.
- **Don't pile altered tensions onto a non-altered context.** Altered dominants are for resolving V7→i; in Lydian dominant or other non-resolving contexts, use ♯11 only.
- **Don't ignore voice-leading between voicings.** A change from one voicing to another should preserve smooth motion in each voice.
- **Don't crowd the low register with extensions.** Extensions (9, 11, 13) sound best high; in the low register, they create mud.

## Cross-references

- Chord construction → `harmony/chord-construction.md`
- Voice leading → `harmony/voice-leading.md`
- Jazz harmony (theoretical context) → `harmony/jazz-harmony.md`
- Reharmonization → `harmony/reharmonization.md`
- Voicing and texture (general) → `orchestration/voicing-and-texture.md`
- Genre-specific use → `genres/jazz-styles.md`

# Progressions Catalog

A quick catalog of common chord progressions across genres. Use this as a lookup for **playable options**, not as a replacement for listening, score study, or genre-specific idiom.

Most examples are given in C major or A/C minor. Transpose to the user's key and genre. Famous-song references are **simplified, non-authoritative recognition examples**: do not present them as exact published transcriptions unless you have checked a reliable score or chart.

For functional theory behind these progressions, see `harmony/functional-harmony.md`. For voice leading, see `harmony/voice-leading.md`. For jazz substitutions and reharmonization, see `harmony/jazz-harmony.md` and `harmony/reharmonization.md`. For modal progressions, see `harmony/modal-harmony.md`.

## How to use this catalog

1. Pick a **function**: stable loop, cadence, lift, darkening, bridge contrast, modulation setup.
2. Pick a **genre frame**: pop, rock, jazz, folk, EDM, cinematic, theatre, etc.
3. Transpose and adjust **harmonic rhythm**: one chord per bar, two per bar, pedal bass, syncopated hits.
4. Add **voicing and bass motion**. The same RN pattern can feel generic or fresh depending on inversion, register, rhythm, and texture.

## Diatonic major progressions

### Pop / rock standards

| Progression | RN | Example in C | Typical use |
|-------------|----|----|-------------|
| Axis | I-V-vi-IV | C-G-Am-F | Modern pop, rock choruses, singer-songwriter loops |
| Axis variant | vi-IV-I-V | Am-F-C-G | Minor-first pop loop, emotional but anthemic |
| 50s / doo-wop | I-vi-IV-V | C-Am-F-G | 1950s pop, ballads, nostalgic turns |
| Plagal pop loop | IV-I-V-vi | F-C-G-Am | Chorus lift starting away from tonic |
| Classic three-chord | I-IV-V | C-F-G | Folk, country, rock, hymnic simplicity |
| I-V-IV-I | C-G-F-C | Rock/folk stability; often uses plagal gravity |
| Pachelbel-family | I-V-vi-iii-IV-I-IV-V | C-G-Am-Em-F-C-F-G | Canon-like descending-functional loop |
| 12-bar blues, basic | I-I-I-I \| IV-IV-I-I \| V-IV-I-I | C-C-C-C \| F-F-C-C \| G-F-C-C | Blues, early rock, boogie, jazz-blues variants |

### Major key cadential

| Progression | Effect |
|-------------|--------|
| ii-V-I | Strong cadential resolution; jazz default and useful in many tonal styles |
| IV-V-I | Direct pop / rock / folk cadence |
| V-I | Authentic cadence; maximum functional closure when prepared well |
| IV-I | Plagal cadence, hymn/gospel/folk color |
| ii⁶-V-I | Smooth classical voice-leading version of ii-V-I |
| V⁷-I | Strong dominant-to-tonic resolution |
| V-vi | Deceptive cadence; avoids expected tonic arrival |

### Major key sequences and turnarounds

| Progression | Example in C | Notes |
|-------------|--------------|-------|
| I-vi-ii-V | C-Am-Dm-G | Circle-of-fifths turnaround |
| I-IV-vii°-iii-vi-ii-V-I | C-F-B°-Em-Am-Dm-G-C | Full diatonic circle by descending fifth / ascending fourth roots |
| iii-vi-ii-V-I | Em-Am-Dm-G-C | Back half of the circle, common in jazz/pop standards |
| I-V/vi-vi-IV | C-E-Am-F | Secondary dominant gives emotional pull to vi |
| I-vi-IV-ii-V-I | C-Am-F-Dm-G-C | Extended cadential motion |

## Diatonic and tonal minor progressions

### Pop / rock minor

| Progression | RN | Example in A minor | Use |
|-------------|----|----|-----|
| Minor axis | i-VI-III-VII | Am-F-C-G | Modal pop / indie / cinematic pop |
| Minor axis variant | i-VII-VI-VII | Am-G-F-G | Rock anthems, modal minor loops |
| Minor lift | i-VI-VII-i | Am-F-G-Am | Natural-minor lift back to tonic |
| Functional minor cadence | i-iv-V-i | Am-Dm-E-Am | Classical/theatre/film minor closure |
| Minor plagal | i-iv-i | Am-Dm-Am | Rock, folk, modal ballad color |
| Minor circle | i-iv-VII-III-VI-ii°-V-i | Am-Dm-G-C-F-B°-E-Am | Long-form tonal pull |

### Minor key cadential

| Progression | Effect |
|-------------|--------|
| ii°-V-i | Minor cadence with raised leading tone in V |
| iiø7-V7-i | Jazz/classical minor cadence; V may be altered |
| iv-V-i | Direct minor cadence |
| iv-i | Dark plagal minor cadence |
| V-i | Authentic minor cadence |
| ♭VII-i | Modal minor cadence without leading tone |
| iv-♭VII-i | Aeolian/modal minor color |

## Modal progressions

Modal progressions usually work because they **avoid over-functional dominant gravity** and emphasize the mode's characteristic degree.

### Dorian

| Progression | Example in D Dorian | Notes |
|-------------|---------------------|-------|
| i-IV | Dm-G | Signature Dorian color: minor tonic + major IV (^6) |
| i-IV-i-IV | Dm-G-Dm-G | Two-chord modal vamp |
| i-♭III-IV-i | Dm-F-G-Dm | Dorian lift with stable tonic return |
| i-v-IV-i | Dm-Am-G-Dm | Softer, more diatonic Dorian loop |

### Phrygian

| Progression | Example in E Phrygian | Notes |
|-------------|----------------------|-------|
| i-♭II | Em-F | Signature Phrygian semitone tension |
| i-♭II-i-♭II | Em-F-Em-F | Static Phrygian vamp |
| ♭II-i | F-Em | Characteristic Phrygian cadence-like resolution |
| i-♭VII-♭II-i | Em-D-F-Em | Dark modal cycle; avoid making D act like functional dominant |

### Mixolydian / rock major-minor

| Progression | Example in G | Notes |
|-------------|--------------|-------|
| I-♭VII | G-F | Signature Mixolydian/root-rock sound |
| I-♭VII-IV-I | G-F-C-G | Rock, folk-rock, guitar music |
| I-IV-♭VII-I | G-C-F-G | Plagal, open, less dominant-driven |
| I-V-IV-I | G-D-C-G | Rock major/minor blend; V major adds leading-tone pull, so it is not purely diatonic Mixolydian |

### Lydian

| Progression | Example in F Lydian | Notes |
|-------------|--------------------|-------|
| I-II | F-G | Signature Lydian color: bright major II / raised ^4 |
| I-II-I-II | F-G-F-G | Lydian vamp |
| I-V-II-I | F-C-G-F | More stable Lydian loop; keep F as center |
| I-iii-II-I | F-Am-G-F | Smooth, bright, less cadence-heavy |
| I-♯iv°-V-I | F-B°-C-F | Diatonic to F Lydian; use carefully because it can imply functional resolution |

## Chromatic harmony

### Secondary dominants

| Progression | Example in C | Notes |
|-------------|--------------|-------|
| V/V-V-I | D-G-C | Dominant of dominant |
| V/IV-IV-I | C7-F-C | Pull toward IV; blues/folk/gospel friendly |
| V/vi-vi-IV-V-I | E-Am-F-G-C | Emotional tonicization of vi |
| V/ii-ii-V-I | A-Dm-G-C | Sets up a stronger ii-V-I |

### Modal mixture / borrowed chords

| Progression | Example in C | Notes |
|-------------|--------------|-------|
| I-iv-I | C-Fm-C | Borrowed iv from parallel minor |
| I-IV-iv-I | C-F-Fm-C | Strong bittersweet voice-leading |
| I-♭VI-♭VII-I | C-A♭-B♭-C | Parallel-minor rock/power-pop color |
| I-♭III-♭VII-IV | C-E♭-B♭-F | Rock mixture; blues/Mixolydian/minor overlap |
| vi-IV-I-V/vi | Am-F-C-E | Secondary dominant pull back to vi |

### Augmented sixth chords

| Progression | Example in C | Notes |
|-------------|--------------|-------|
| It⁶-V-I | A♭-C-F♯ → G → C | A♭ resolves down to G; F♯ resolves up to G |
| Fr⁶-V-I | A♭-C-D-F♯ → G → C | Adds D; whole-tone-like color |
| Ger⁶-I⁶/₄-V-I | A♭-C-E♭-F♯ → C/G → G → C | Often routed through cadential 6/4 to avoid parallel 5ths |

German +6 is the augmented-sixth type most often enharmonically reinterpreted as a dominant seventh. Do not describe Italian +6 as having the same pitch content as V7.

### Neapolitan

| Progression | Example | Notes |
|-------------|---------|-------|
| ♭II⁶-V-i | D♭/F-G-Cm | Classical Neapolitan 6 in C minor; can be borrowed into major for drama |

### Tritone substitutions and backdoor motion

| Progression | Example in C | Notes |
|-------------|--------------|-------|
| ii-♭II7-I | Dm7-D♭7-Cmaj7 | D♭7 substitutes for G7 because they share guide-tone tendency tones enharmonically |
| ii-V7-♭VII7-I | Dm7-G7-B♭7-Cmaj7 | Dominant chain into backdoor color |
| iv-♭VII7-I | Fm-B♭7-C | Backdoor cadence; common in jazz, soul, pop |

## Jazz standards

### ii-V-I in major

| Form | Example in C | Notes |
|------|--------------|-------|
| ii-V-I | Dm7-G7-Cmaj7 | Core jazz cadence |
| ii-V-I with tensions | Dm9-G13-Cmaj9 | Smooth modern color |
| Altered dominant | Dm7-G7alt-Cmaj7 | More bite before resolution |

### ii-V-i in minor

| Form | Example in A minor | Notes |
|------|--------------------|-------|
| iiø7-V7♭9-i | Bm7♭5-E7♭9-Am | Minor-key jazz cadence |
| iiø7-V7alt-i(maj7) | Bm7♭5-E7alt-Am(maj7) | More modern / dramatic minor arrival |
| iiø7-V7sus-V7-i | Bm7♭5-E7sus-E7-Am | Suspended dominant delay |

### Cycles of fifths

| Progression | Example in C | Notes |
|-------------|--------------|-------|
| iii-vi-ii-V-I | Em7-Am7-Dm7-G7-Cmaj7 | Diatonic cycle |
| iii-VI-ii-V-I | Em7-A7-Dm7-G7-Cmaj7 | Secondary dominant on VI |
| I-vi-ii-V | Cmaj7-Am7-Dm7-G7 | Standard turnaround |

### Coltrane-cycle family

| Progression | Example | Notes |
|-------------|---------|-------|
| Major-third key-center cycle | Cmaj7-E♭7-A♭maj7-B7-Emaj7-G7-Cmaj7 | Illustration of symmetrical key-center movement; not a complete lead-sheet substitute |
| "Giant Steps" root-center outline | B → D → G → B♭ → E♭ → G♭ → B | Use as a recognition outline only; check a reliable chart for exact changes |

## Pop song progressions by section

### Common verse progressions

| Progression | Example | Notes |
|-------------|---------|-------|
| I-vi-IV-V | C-Am-F-G | Stable, nostalgic, lyric-friendly |
| I-V-vi-IV | C-G-Am-F | Familiar modern loop |
| vi-IV-I-V | Am-F-C-G | Starts emotionally darker, resolves outward |
| I-iii-vi-IV | C-Em-Am-F | Smooth lift through mediant |
| I-V⁶-vi-V | C-G/B-Am-G | Bass descent ^1-^7-^6-^5 |

### Common pre-chorus progressions

| Progression | Example | Notes |
|-------------|---------|-------|
| IV-V | F-G | Simple build into tonic chorus |
| ii-V | Dm-G | Strong dominant prep |
| IV-V/vi-vi-V | F-E/G♯-Am-G | Tonicizes vi, then pulls outward |
| vi-V/V-V | Am-D-G | Secondary dominant lift |
| ♭VI-♭VII-V | A♭-B♭-G | Chromatic lift into a C chorus, especially cinematic/pop-rock |

### Common chorus progressions

| Progression | Example | Notes |
|-------------|---------|-------|
| I-V-vi-IV | C-G-Am-F | Default anthemic pop loop |
| I-IV-V-IV | C-F-G-F | Open, guitar-friendly, singalong |
| I-V-IV-I | C-G-F-C | Stable and direct |
| I-IV-vi-V | C-F-Am-G | Brighter first half, cadential second half |
| IV-I-V-vi | F-C-G-Am | Chorus starts lifted away from tonic |

### Common bridge progressions

| Progression | Example | Notes |
|-------------|---------|-------|
| vi-IV-V/V-V | Am-F-D-G | Bridge that points back to tonic chorus |
| ♭VI-♭VII-I | A♭-B♭-C | Modal-mixture lift into tonic |
| iv-V-vi | Fm-G-Am | Deceptive arrival; useful for emotional bridge |
| ii-V/vi-vi | Dm-E-Am | Temporary minor tonicization |
| New-key dominant | A7 → D section after C section | Practical setup for modulation |

## Bass-line-driven progressions

Sometimes the harmony is best understood from the bass line rather than the chord labels.

### Descending chromatic bass / lament family

In A minor, a common bass line is:

```text
A - G♯ - G - F♯ - F - E
^1 - raised ^7 - ♭^7 - raised ^6 - ♭^6 - ^5
```

Possible chords:

```text
Am - E/G♯ - C/G - D/F♯ - F - E
```

This gives a lament-like pull: tragic, stately, inevitable. The surface chords vary widely; the bass line is the identity.

### Diatonic descending bass

```text
C - G/B - Am - G - F - C/E - Dm - G
I - V6 - vi - V - IV - I6 - ii - V
```

Useful for ballads, worship/gospel-adjacent writing, and classic pop craft.

### Ascending bass

A stepwise bass line ^1-^2-^3-^4-^5 can support many chords above it. Common in soul ballads, theatre builds, and pre-choruses.

### Pedal tonic

A held tonic in the bass while harmony moves above it. Common in classical, ambient, rock, and cinematic music. Creates stability and scale.

### Pedal dominant

A held dominant in the bass building tension toward tonic. Common in classical retransitions, EDM builds, theatre vamps, and cinematic cues.

## Specific characteristic progressions — simplified examples

Use these as **recognition anchors**, not definitive transcriptions.

### "Creep"-family chromatic mediant + borrowed iv

```text
I - III - IV - iv
C - E - F - Fm
```

The striking color is the chromatic mediant III followed by the borrowed minor iv.

### "Don't Stop Believin'"-family descending bass / Pachelbel cousin

```text
I - V6 - vi - iii6 - IV - I6 - ii - V
C - G/B - Am - Em/G - F - C/E - Dm - G
```

The identity is the stepwise bass descent and delayed return to the dominant.

### "Wonderwall"-family suspended guitar loop

```text
i - ♭III - ♭VII - IVsus
Em7 - G - Dsus4 - Asus4
```

The sus chords and common tones matter as much as the Roman numerals.

### "Hotel California"-family descending minor cycle

```text
i - V - ♭VII - IV - ♭VI - ♭III - iv - V
Bm - F♯ - A - E - G - D - Em - F♯
```

The major V at both ends reinforces the minor tonic.

### "Hey Jude"-family singalong outro

```text
I - ♭VII - IV - I
F - E♭ - B♭ - F
```

This is a Mixolydian/plagal singalong color. For the verse, avoid overclaiming a single loop; it is mostly simple F-major tonic/dominant/subdominant motion with local variations.

### Guitar-rock Mixolydian loop

```text
I - ♭VII - IV - I
D - Cadd9 - G - D
```

Open-string voicing and guitar register are part of the sound.

### "Africa"-style pop-rock modal mixture caution

Toto's "Africa" is frequently simplified differently across charts because sections, slash bass, and keyboard voicings matter. Use it as a reminder to **verify famous-song examples** before naming exact changes. For a generic similar color, try:

```text
I - iii - vi - ♭VII - IV - ii
B - D♯m - G♯m - A - E - C♯m
```

Treat that as an original pop-rock/modal-mixture palette, not a claim of an exact transcription.

## Quick decision matrix

| Want | Try |
|------|-----|
| Pop standard feel | I-V-vi-IV |
| Slightly darker pop | vi-IV-I-V |
| 1950s / nostalgic feel | I-vi-IV-V |
| Lifting / building | I-iii-vi-IV or IV-V/vi-vi-V |
| Canon-like feel | I-V-vi-iii-IV-I-IV-V |
| Modal Dorian feel | i-IV |
| Modal Phrygian feel | i-♭II |
| Modal Mixolydian / guitar-rock feel | I-♭VII-IV-I |
| Lydian brightness | I-II |
| Strong cadence | ii-V-I (jazz); IV-V-I (pop); V-I (classical) |
| Surprise | Deceptive cadence: V-vi |
| Tension / build | Pedal V or rising sequence |
| Modulation up | V/new key → new tonic |
| Descending drama | Lament bass or minor descending cycle |
| Modal mixture color | I-iv-I; I-♭VI-♭VII-I; I-♭III-♭VII-IV |
| Jazz minor cadence | iiø7-V7♭9-i |
| Backdoor soul/jazz cadence | iv-♭VII7-I |

## Cross-references

- Functional theory of these progressions → `harmony/functional-harmony.md`
- Voice leading the progressions → `harmony/voice-leading.md`
- Modal progression theory → `harmony/modal-harmony.md`
- Chromatic and altered harmony → `harmony/chromatic-harmony.md`
- Jazz progressions and substitutions → `harmony/jazz-harmony.md`, `harmony/reharmonization.md`
- Modulation between progressions → `harmony/modulation.md`
- Cadences as endings → `cadence-reference.md`
- Pop song forms → `form/popular-song-forms.md`
- Genre-specific progressions → various `genres/*.md`
- Response formatting → `response-templates.md`

# Trend and Reference Matrices

Short forms for turning trend research, streaming-service evidence, and reference tracks into composition decisions. These are templates, not research results.

## User listening context matrix

Use when the user shares a playlist, top tracks, Replay/Wrapped summary, Last.fm list, or public streaming link.

```md
| Evidence | Role | Tempo/groove | Harmony | Topline | Form/hook timing | Arrangement | Transferable move | Avoid copying |
|---|---|---|---|---|---|---|---|---|
| Track / playlist 1 | anchor | | | | | | | |
| Track / playlist 2 | boundary | | | | | | | |
| Platform chart / public signal | market check | | | | | | | |
```

## Regional trend-evolution matrix

Use when comparing a region or scene across time windows.

```md
| Market / period | Platform function | Groove | Harmony | Topline/prosody | Form | Arrangement/production | Collaboration/industry signal | Composition takeaway |
|---|---|---|---|---|---|---|---|---|
| Earlier window | | | | | | | | |
| Current window | | | | | | | | |
| Difference | | | | | | | | |
```

## Microgenre craft profile

Use when the genre is minor, unstable, hybrid, or unfamiliar.

```md
| Variable | Observed pattern | Confidence | Concrete composition move |
|---|---|---|---|
| Tempo / meter | | high / medium / low | |
| Groove | | high / medium / low | |
| Harmony | | high / medium / low | |
| Melody / vocal | | high / medium / low | |
| Texture | | high / medium / low | |
| Form | | high / medium / low | |
| Platform / scene function | | high / medium / low | |
```

## Hybrid genre merge matrix

Use when combining two or more genre ingredients.

```md
| Ingredient | Role | Keep | Reduce / avoid | Concrete move |
|---|---|---|---|---|
| Main genre | form anchor | | | |
| Secondary genre | groove / harmony / texture | | | |
| Regional/traditional source | idiom or color | | | |
| Reference artist/track | energy target only | | protected expression | |
```

## Evidence-confidence labels

| Label | Meaning | How to answer |
|---|---|---|
| High confidence | Multiple sources and references agree | Give concrete advice, still note scope |
| Medium confidence | One strong source plus listening evidence | Give advice as a likely direction |
| Low confidence | One reference, vague tag, or platform-only signal | Be cautious; ask for reference or web-check |
| Conflicting signal | Platforms or regions disagree | Explain which signal the advice optimizes for |

## Cross-references

- User listening context → `../references/research/user-listening-context-and-streaming-services.md`
- Regional trend evolution → `../references/research/regional-trend-evolution-analysis.md`
- Web music trend research → `../references/research/web-music-trend-research.md`
- Reference-track digging → `../references/research/reference-track-digging.md`
- Minor and hybrid genres → `../references/genres/minor-and-hybrid-genres.md`
- Response templates → `response-templates.md`
- Diagnostic checklists → `diagnostic-checklists.md`

# Session Brief and Decision Log Templates

Use these templates to keep multi-turn composition sessions coherent. Pair with `../references/creative-workflows/user-agent-collaboration.md`.

## Minimal project card

```md
**Project card**
- Goal:
- Genre frame:
- Mood:
- References:
- Key / tempo / meter:
- Current section:
- Chosen seed:
- Keep:
- Change:
- Open decision:
```

## Compact decision log

```md
**Decisions so far**
1.
2.
3.

**Rejected for now**
- 

**Can revisit later**
- 
```

## Branch tracker

```md
| Branch | Core idea | Strength | Risk | Status |
|---|---|---|---|---|
| A | | | | active / parked / rejected |
| B | | | | active / parked / rejected |
| C | | | | active / parked / rejected |
```

## Revision request parser

```md
**User asks:** "Make it more ______."

Possible variables:
- Harmony:
- Melody:
- Rhythm:
- Arrangement:
- Form:
- Lyric/prosody:

Smallest useful edit:
```

## Before / after edit block

```md
**Before**

**After**

**What changed**
- 

**What stayed the same**
- 
```

## Do-not-change list

```md
**Protect these**
- 

**Okay to change**
- 

**Unknown / ask only if critical**
- 
```

## Handoff summary

Use this if the conversation is long or the user asks to continue later.

```md
**Current state**
- We are writing:
- Current best version:
- Main unresolved issue:
- Next best action:

**Material generated**
- Chords:
- Motif/topline:
- Groove:
- Form:
- Lyric/title:

**User preferences learned**
- Likes:
- Dislikes:
- Avoid:
```

## Cross-references

- Collaboration method → `../references/creative-workflows/user-agent-collaboration.md`
- Revision and feedback loops → `../references/creative-workflows/revision-and-feedback-loops.md`
- Musical brainstorming → `../references/creative-workflows/musical-brainstorming.md`
- Response templates → `response-templates.md`
- Diagnostic checklists → `diagnostic-checklists.md`

# Musical Brainstorming Cards

Quick cards for generating composition ideas. Use these when the user wants sparks, not a lecture. Pair with `../references/creative-workflows/musical-brainstorming.md`.

## 1. The six-seed starter

```md
**Six possible starting seeds**
1. Chord seed:
2. Motif seed:
3. Groove seed:
4. Bass seed:
5. Texture seed:
6. Lyric/image seed:
```

## 2. Safe / fresh / wild

```md
| Lane | Idea | Why it works | Risk |
|---|---|---|---|
| Safe | | genre-native, easy to finish | may be predictable |
| Fresh | | one unusual variable | needs taste/control |
| Wild | | creates a signature | may overpower the song |
```

## 3. Mood to musical variables

| Mood target | Harmony | Melody | Rhythm | Texture |
|---|---|---|---|---|
| tender | added 6/9, gentle plagal motion | stepwise, narrow range | relaxed, fewer attacks | warm midrange, sparse layers |
| lonely | minor/add9, pedal tones | repeated notes, falling thirds | space between phrases | dry solo instrument, distant pad |
| anxious | chromatic neighbor chords, unresolved dominant | small intervals, unstable landing | syncopation, offbeat accents | thin highs, pulsing ostinato |
| triumphant | I-IV-V, Lydian/major color | rising fourth/fifth, high peak | strong downbeats | wide voicing, octave doubles |
| sensual | minor 7/9, slow dominant pull | curved contour, late resolution | behind-the-beat feel | low-mid warmth, breath/space |
| uncanny | modal mixture, tritone, mediant shifts | familiar cell in wrong context | lopsided phrase | register gaps, brittle timbre |

## 4. Hook generator

```md
**Hook ingredients**
- title / word:
- rhythm: short-short-long / long-short-short / chant / syncopated pickup
- contour: rise / fall / arch / repeat then leap
- landing note: ^1 / ^3 / ^5 / tension note
- repetition: exact / varied / call-response
```

## 5. Four-bar prototype

```md
Key:
Tempo:
Feel:

| Bar | Chord | Melody idea | Arrangement action |
|---|---|---|---|
| 1 | | | |
| 2 | | | |
| 3 | | | |
| 4 | | | |
```

## 6. Constraint dice

Choose one from each row.

| Axis | Constraint options |
|---|---|
| Harmony | only 2 chords / no V chord / one borrowed chord / pedal bass / chromatic bass |
| Melody | 5 notes only / one leap only / start every phrase offbeat / highest note only once |
| Rhythm | no downbeat hook / 3+3+2 grouping / halftime chorus / one bar of silence |
| Form | chorus first / no pre-chorus / 2-bar post-hook / bridge uses verse lyric rhythm |
| Texture | no cymbals until chorus / one instrument per octave / mono verse, wide chorus |
| Lyric | title appears once / no perfect rhymes / image before explanation |

## 7. Mutation menu

Use after the user likes a seed but wants more versions.

| Mutation | What changes | What stays |
|---|---|---|
| Transpose | key/register | contour and rhythm |
| Invert | direction | rhythm and interval size roughly |
| Sequence | starting pitch/chord | shape |
| Compress | note values | pitch idea |
| Expand | longer notes / wider gaps | identity |
| Displace | metric placement | pitches |
| Reharmonize | chords under melody | melody |
| Re-orchestrate | texture/timbre | notes/chords |
| Simplify | fewer notes/chords | emotional target |

## 8. Reference triangulation card

```md
| Reference | Borrow only this variable | Do not copy |
|---|---|---|
| A | energy / tempo / density | melody, bassline, lyric |
| B | harmony density / voicing | progression order if too recognizable |
| C | vocal phrasing / section arc | artist identity, signature ad-libs |
```

## 9. Section contrast card

```md
Verse = smaller by: register / density / rhythm / harmony / lyric camera
Pre = tension by: bass climb / dominant / lyric urgency / rhythmic acceleration
Chorus = release by: hook repetition / range peak / wider voicing / simpler rhythm
Bridge = new information by: mode shift / perspective shift / texture reset
```

## 10. "Make it more X" card

| Request | First variables to try |
|---|---|
| more modern | earlier hook, shorter intro, leaner loop, current groove reference |
| more classic | stronger cadence, live-band idiom, longer phrases |
| more expensive | fewer busy fills, wider voicings, controlled register, clearer bass |
| more raw | less polish, narrower arrangement, stronger performance gestures |
| more catchy | shorter motif, title repetition, clearer rhythm, higher landing note |
| less generic | bass variation, phrase asymmetry, one surprising chord, distinctive rhythm |
| less cluttered | reduce midrange layers, separate registers, simplify counterlines |
| more emotional | delay resolution, add appoggiatura, raise chorus peak, expose vocal |

## Cross-references

- Full brainstorming method → `../references/creative-workflows/musical-brainstorming.md`
- User-agent collaboration → `../references/creative-workflows/user-agent-collaboration.md`
- Session/decision templates → `session-brief-and-decision-log.md`
- Response templates → `response-templates.md`
- Diagnostic checklists → `diagnostic-checklists.md`

# Response Templates

Use these templates to keep music-composition answers concrete, playable, and easy to revise. They are templates, not scripts: shorten, expand, or combine them based on the user's level and the task.

## General answer shape

For most creative help:

```md
Assumption: [genre / key / tempo / instrumentation if not stated]
Goal: [what the user wants musically]

Option 1 — [name]
[playable material]
Why it works: [1–2 sentences]
Trade-off: [what it may weaken]

Option 2 — [name]
[playable material]
Why it works: ...
Trade-off: ...

Try first: [one concrete next action]
```

Keep the *playable material* first. Put theory after the user can hear or play something.

## Chord progression request

```md
Key: [C minor]
Tempo / feel: [72 BPM, slow 6/8 ballad]

Progression:
| Cm(add9) | A♭maj7 | E♭/G | G7sus4 G7 |

Roman numerals:
| i(add9) | ♭VImaj7 | ♭III6 | V7sus4 V7 |

Why it works:
- [A♭maj7 softens the minor color.]
- [E♭/G keeps the bass stepwise.]
- [G7 creates a real dominant pull into the next phrase.]

Variation:
| Cm(add9) | Fm9 | A♭maj7 | G7♭9 |
```

For beginners, skip Roman numerals unless useful. For advanced users, include voicing or voice-leading notes.

## Melody / topline request

```md
Key / scale: [A minor, with raised ^7 over E7]
Range target: [A3–E5]
Phrase length: [4 bars]

Scale-degree sketch:
Bar 1: ^5 ^5 ^6 ^5 ^3
Bar 2: ^4 ^3 ^2 —
Bar 3: ^5 ^6 ^7 ^1̂
Bar 4: ^2 ^7 ^1 —

Contour idea:
- Start narrow and speech-like.
- Peak once on [E5] in bar 3.
- Resolve by step, not by another leap.

Lyric placement:
[Put the hook word on the highest note / longest note.]
```

If the user gave lyrics, preserve their words and adjust rhythm/prosody first.

## Reharmonization request

```md
Original:
| C | Am | F | G |

Reharm option A — smoother bass:
| C | G/B | Am | Fmaj7 G |
Why: bass descends C-B-A, making the simple loop feel intentional.

Reharm option B — jazzier:
| Cmaj9 | E7♭9 | Am9 | Dm9 G13 |
Why: E7 tonicizes Am; Dm9-G13 creates a ii-V back to C.

Keep unchanged:
- [melody notes that define the hook]
- [strong lyric downbeats]
```

Always state what should **not** change so the song identity survives.

## Critique / feedback request

```md
My read:
[One-sentence diagnosis: the chorus has good energy, but the melody and arrangement peak too early.]

What is working:
1. ...
2. ...

Highest-impact fix:
[One fix that changes the most with the least damage.]

Detailed notes:
| Area | Observation | Fix |
|---|---|---|
| Melody | ... | ... |
| Harmony | ... | ... |
| Form | ... | ... |
| Arrangement | ... | ... |

Revision path:
1. [small experiment]
2. [A/B compare]
3. [commit or revert]
```

Be honest but useful. Diagnose before prescribing.

## "The chorus doesn't hit" response

```md
The chorus may not be failing harmonically; it may be failing by contrast. Check these in order:

1. Melody: does the chorus hit a higher or longer focal note than the verse?
2. Rhythm: does the chorus simplify or intensify the groove?
3. Harmony: does the pre-chorus create enough dominant / rising / suspended energy?
4. Arrangement: does a new register or layer enter exactly at the chorus?

Two concrete fixes:
A. Harmonic lift:
Verse loop: | vi | IV | I | V |
Pre:        | IV | V/vi | vi | V |
Chorus:     | I | V | vi | IV |

B. Melodic lift:
Move the chorus hook peak from ^3 to ^5 or ^6, and hold it for at least a dotted quarter / half note.
```

Pair with `diagnostic-checklists.md` when deeper diagnosis is needed.

## Song plan request

```md
Working title / concept: [optional]
Genre frame: [alt-pop / K-pop ballad / neo-soul]
Tempo: [96 BPM]
Key center: [D minor]
Core hook idea: [3-note chant: ^5-^6-^5]

Form:
Intro: 4 bars — [hook fragment]
Verse 1: 8 bars — [low register, sparse]
Pre: 8 bars — [rising sequence]
Chorus: 8 bars — [full hook]
Verse 2: 8 bars — [variation]
Bridge: 8 bars — [relative major / breakdown]
Final chorus: 16 bars — [ad-libs / countermelody]

Core progression:
Verse:  | Dm | B♭ | F | C |
Pre:    | Gm | A7 | Dm | C |
Chorus: | F | C | Dm | B♭ |

Next writing step:
[Write only the chorus hook first; do not arrange yet.]
```

## Genre / trend answer

```md
Static genre frame:
[durable conventions from the relevant genre file]

Current check needed:
[charts/platforms/local scene/source types]

Composition takeaway:
1. [hook timing]
2. [groove/bass/drum pattern]
3. [form/density]
4. [lyric/prosody]

Caveat:
[platform-specific, region-specific, or trend too noisy]
```

Use `../references/research/web-music-trend-research.md` for current facts.

## Reference-track-driven request

```md
I won't copy the reference directly. I'll extract transferable craft variables.

Reference variables:
- Tempo / feel:
- Groove:
- Harmony:
- Topline:
- Form:
- Arrangement density:
- Signature element to avoid copying:

Original direction:
[New idea that shares energy but not melody/lyrics/riff/sample]

Playable sketch:
[Chord loop / melody contour / rhythm cell]
```

Use `../references/research/reference-track-digging.md` and `../references/research/style-reference-and-copyright.md`.

## Music analysis request

```md
Surface facts:
- Key center:
- Meter / tempo feel:
- Form:
- Main progression:

Functional read:
[Roman numerals and cadence points]

What creates the effect:
1. Harmony: ...
2. Melody: ...
3. Rhythm/groove: ...
4. Texture/arrangement: ...

Composer takeaway:
[One transferable technique]
```

Do not overclaim exact transcription if the user only provided a rough chart or audio description.

## Orchestration / arrangement request

```md
Goal: [make chorus wider without clutter]

Register plan:
Low: [bass / low cello / sub]
Mid: [piano/guitar comp]
High: [strings/synth pad/counterline]
Vocal space: [leave 1–3 kHz range uncluttered]

Move:
- Double the melody at the octave only on the hook word.
- Put sustained harmony above the vocal, not in the same register.
- Use rhythmic instruments in verse, sustained instruments in chorus for contrast.

Risk:
[too much density; check lyric intelligibility]
```

## Game / interactive music request

```md
State map:
Exploration → Tension → Combat → Victory / Failure

Shared material:
- Motif: [^1-^3-^2]
- Tempo grid: [120 BPM]
- Key center: [D Dorian]

Layer plan:
Layer 1: ambient bed, loopable 16 bars
Layer 2: percussion, enters on tension
Layer 3: bass ostinato, enters on combat
Layer 4: brass/stinger, one-shot transitions

Transition rule:
[quantize transitions to 1 bar / 4 bars / next cadence]
```

## Teaching request

```md
Level assumption: [beginner/intermediate/advanced]
Goal: [learn secondary dominants]

Core idea:
[one plain-language explanation]

Tiny exercise:
1. Play | C | G | C |
2. Insert V/V: | C | D7 G | C |
3. Hear how D7 makes G feel inevitable.

Check-yourself question:
[What chord is V/vi in C major?]

Next step:
[try it in another key]
```

## User listening-context request

```md
I can use your listening context as evidence, not as something to copy.

Safe inputs:
- public playlist link / 5-10 track names / Replay-Wrapped summary / Last.fm top tracks
- no login details or private credentials

Listening-context signal:
- Personal anchors:
- Boundary tracks:
- Platform or market check:

Composition variables:
1. Tempo/groove:
2. Harmony:
3. Topline/prosody:
4. Form/hook timing:
5. Arrangement density:

Concrete next move:
[4-8 bar sketch, reference matrix, or playlist-fit plan]
```

## Regional trend-evolution request

```md
Scope:
- Region/market:
- Time window:
- Platform signal:
- Reference cluster:

What seems to be changing:
1. Platform/function:
2. Groove:
3. Topline/prosody:
4. Form/arrangement:
5. Collaboration or industry pattern:

Composition advice:
- Keep:
- Modernize:
- Avoid:
- Concrete 4-8 bar move:

Caveat:
[platform/region/time-window specificity]
```

## Microgenre / hybrid genre request

```md
I will treat this as a scene/tag question, not a fixed universal genre rule.

Genre ingredients:
- Main anchor:
- Supporting ingredient:
- Platform/scene context:

Craft variables:
| Variable | Choice | Confidence |
|---|---|---|
| Tempo/groove | | |
| Harmony | | |
| Topline/vocal | | |
| Texture | | |
| Form | | |

Sketch:
[playable progression/rhythm/topline contour]

Avoid copying:
[reference melody, lyrics, sample, meme phrase, producer tag, vocal identity]
```

## Cross-references

- Answer calibration / controlled variation → `../references/creative-workflows/answer-calibration.md`
- Vague problem triage → `diagnostic-checklists.md`
- Progression lookup → `progressions-catalog.md`
- Form templates → `form-templates.md`
- Chord symbols → `chord-symbol-conventions.md`
- Reference-track workflows → `../references/research/reference-track-digging.md`
- User listening context → `../references/research/user-listening-context-and-streaming-services.md`
- Regional trend evolution → `../references/research/regional-trend-evolution-analysis.md`
- Minor/hybrid genres → `../references/genres/minor-and-hybrid-genres.md`
- Trend/reference matrices → `trend-and-reference-matrices.md`
- Style/copyright guardrails → `../references/research/style-reference-and-copyright.md`


## Musical brainstorming option board

Use when the user asks for ideas, sparks, or multiple directions.

```md
I'll give you three lanes so we don't commit too early.

| Lane | Core idea | Playable seed | Why it might work | Risk |
|---|---|---|---|---|
| Safe | | | | |
| Fresh | | | | |
| Wild | | | | |

Which one feels alive enough to prototype?
```

## Multi-turn session state

Use when the conversation becomes iterative.

```md
**Project card**
- Goal:
- Genre frame:
- Key/tempo/meter:
- Current best seed:
- Keep:
- Change:
- Open decision:

**Next move**
I'll revise only [variable] and preserve [do-not-change item].
```

## Web research quick summary

Use after active web/reference research.

```md
**What I checked**
- Region/platform:
- Time window:
- Source types:

**Signals**
- High-confidence:
- Medium-confidence:
- Leads only:

**Composition translation**
- Groove:
- Harmony:
- Topline:
- Arrangement/form:
```

# Music-Theory Audit Rubric

Use this asset for manual review of theory-heavy files. Automated scripts catch regressions; this rubric catches musical reasoning problems that scripts cannot reliably detect.

## Severity labels

| Severity | Meaning | Example |
|---|---|---|
| P0 | Wrong information likely to be repeated directly by the agent | omit-fifth chord mislabeled as a power chord |
| P1 | Overbroad or misleading claim that can damage user trust | “Western music since 1700 uses 12-TET” |
| P2 | Correct in one context but missing a needed caveat | “tritone substitutions can replace any chord” |
| P3 | Style, clarity, or maintainability issue | inconsistent link style, unclear heading |

## Audit dimensions

| Dimension | Questions to ask |
|---|---|
| Pitch spelling | Are accidentals spelled by function, not just pitch class? |
| Chord construction | Do chord symbols map to the listed notes? |
| Function | Does the Roman numeral match the key and chord quality? |
| Voice-leading | Are resolution tendencies described as tendencies, not universal moral rules? |
| Genre context | Is the advice valid for the named style? |
| Currentness | Does the claim require active web verification? |
| Cultural specificity | Does the text name the tradition/scene/idiom rather than using broad labels? |
| Output usefulness | Can the agent turn the paragraph into a concrete answer? |

## High-risk files

Review these most carefully before release.

| File | Why it is high-risk |
|---|---|
| `assets/chord-symbol-conventions.md` | Directly quotable notation guidance |
| `assets/chord-symbol-ambiguity-and-parsing.md` | Directly maps symbols to notes |
| `assets/intervals-and-scale-formulas.md` | Foundational formulas |
| `assets/scale-degree-spelling-cheatsheet.md` | Prevents enharmonic/function errors |
| `assets/modes-cheatsheet.md` | Common source of relative/parallel confusion |
| `assets/jazz-voicings.md` | Altered tension mistakes are easy and visible |
| `assets/progressions-catalog.md` | Famous-song examples can be overclaimed |
| `references/harmony/jazz-harmony.md` | Advanced harmony needs caveats |
| `references/harmony/modulation.md` | Enharmonic reinterpretation can be misstated |
| `references/genres/korean-traditional.md` | Specialized terminology requires care |
| `references/genres/mena-pop.md` | Maqam-adjacent guidance can be oversimplified |
| `references/research/style-reference-and-copyright.md` | Safety/originality guidance needs currentness |

## Audit checklist for a single file

```md
File:
Reviewer:
Date:

Scope check:
- [ ] File title and “When to consult” match the contents.
- [ ] The file does not duplicate a better existing file.
- [ ] Cross-references are relevant and resolve.

Theory check:
- [ ] Chord symbols map to correct notes.
- [ ] Roman numerals match key, mode, and quality.
- [ ] Enharmonic spellings are functionally justified.
- [ ] Examples do not overclaim famous-song transcriptions.
- [ ] Genre caveats are present when needed.

User-output check:
- [ ] The agent can extract concrete advice.
- [ ] Examples include playable/usable materials.
- [ ] The file avoids long abstract explanation without application.

Currentness/culture check:
- [ ] Snapshot note exists if needed.
- [ ] Current artists/platforms/legal claims are not frozen as permanent facts.
- [ ] Cultural references are specific and respectful.

Decision:
- [ ] Release-ready
- [ ] Needs minor edit
- [ ] Needs major review
- [ ] Release blocker
```

## Red-flag phrases

These phrases are not automatically forbidden in every possible context, but they deserve review:

| Phrase type | Better direction |
|---|---|
| broad culture-as-color labels | name instrument, rhythm, language, scene, or tradition |
| “rules of music” | convention, tendency, style norm, voice-leading habit |
| “always resolve” | usually resolves, strongly tends to resolve, conventional resolution |
| “the song uses…” without source | “a simplified common chart often shows…” |
| “current trend” without date/source | add snapshot or active web-research requirement |
| “AI copyright law says…” | jurisdiction, date, and current-source check required |

## Cross-references

- Release readiness → `references/validation/first-release-readiness.md`
- Prompt smoke tests → `references/validation/prompt-smoke-tests.md`
- Chord ambiguity → `chord-symbol-ambiguity-and-parsing.md`
- Scale-degree spelling → `scale-degree-spelling-cheatsheet.md`
- Maintenance guide → `MAINTENANCE.md`

# Web Search Cheatsheet for Music Composition Research

> **Snapshot note:** Last content review snapshot: 2026-04-27. Search engines, platform charts, and social discovery surfaces change quickly. Use this as a compact workflow, then verify current platform behavior when the answer depends on freshness.

This is a fast lookup sheet for active web research. It complements the longer research files:

- `../references/research/web-music-trend-research.md`
- `../references/research/reference-track-digging.md`
- `../references/research/user-listening-context-and-streaming-services.md`
- `../references/research/regional-trend-evolution-analysis.md`

## Search rule of thumb

```text
query = music object + scene/region + time window + evidence type
```

Examples:

```text
"amapiano" "2026" producer interview groove
"Jersey club" K-pop choreography trend 2025
site:charts.spotify.com city chart Seoul K-pop
site:youtube.com/charts Top Songs Shorts Brazil
"regional Mexican" "corridos tumbados" songwriter interview
```

## Operator basics

Common operators and filters:

| Tool | Use | Example |
|---|---|---|
| quotes | exact phrase | `"Top Songs on Shorts"` |
| minus | exclude a term | `phonk -playlist -mix` |
| `site:` | search one site/domain | `site:billboard.com amapiano global` |
| date tools | recent results | search, then filter by past month/year |
| language/region filters | local-market research | search in Spanish, Korean, Japanese, Portuguese, etc. |
| file type / PDF filters | reports and methodology docs | `IFPI music report filetype:pdf` |

Caveat: search operators and ranking behavior can change. Official search help should override this cheatsheet.

## Source hierarchy for music trend questions

| Question | Best first sources | Why |
|---|---|---|
| What is charting? | official chart pages, platform chart pages | direct popularity signal |
| What is emerging? | viral charts, short-form trend tools, local press | velocity signal |
| What does the scene call it? | local-language press, artist interviews, label pages | vocabulary signal |
| How is it made? | producer interviews, live sessions, stems, tutorials from credible practitioners | craft signal |
| Is it safe to imitate? | style/copyright guardrail + reference triangulation | originality signal |

## Platform-specific quick searches

### Spotify

Use for: streaming popularity, regional/city charts, viral movement, playlist lanes.

Search ideas:

```text
Spotify Charts [country/city] [genre]
site:charts.spotify.com [city] [genre]
"Local Pulse" [city] Spotify [genre]
```

Interpretation:

- Global/regional charts = broad platform consumption.
- Viral charts = buzz/velocity, not just total plays.
- City/Local Pulse = local concentration; useful for regional influence.

### Apple Music

Use for: top charts, city charts, user-accessible listening references, Replay if user shares it.

Search ideas:

```text
Apple Music Top Charts [country/genre]
Apple Music City Charts [city]
Apple Music Replay top songs artists user shared
```

Interpretation:

- City charts can surface local listening differences.
- Replay/Wrapped-style summaries are useful only if the user shares them voluntarily.

### YouTube / YouTube Music / Shorts

Use for: video-driven music, regional chart visibility, short-form reuse, official + UGC dynamics.

Search ideas:

```text
YouTube Charts Top Songs [country]
YouTube Charts Top Songs on Shorts [country]
site:support.google.com/youtube YouTube Charts Top Songs Shorts
```

Interpretation:

- YouTube Top Songs may combine official videos, lyric videos, and UGC using the official song.
- Shorts-specific signals may emphasize hookable clips rather than full-song structure.

### TikTok Creative Center

Use for: short-form velocity, songs/hashtags/creators/videos by region.

Search ideas:

```text
TikTok Creative Center songs [region] [genre]
site:ads.tiktok.com/business/creativecenter "Songs" "Trend Discovery"
```

Interpretation:

- Strong for short-form behavior.
- Weak for full-song craft unless triangulated with charts and reference listening.

### Bandcamp / SoundCloud / Beatport

Use for: niche scenes, microgenres, independent scenes, DJ/electronic lanes.

Search ideas:

```text
site:bandcamp.com [microgenre] [city]
site:soundcloud.com [scene tag] [producer]
site:beatport.com [subgenre] top tracks
```

Interpretation:

- Good for scene vocabulary and underground signals.
- Requires reference listening; don't generalize from one upload.

### Local-language searches

Use local language for region-specific scenes.

Examples:

```text
Korean: "요즘" "인디" "플레이리스트" "인터뷰"
Japanese: "最新" "シティポップ" "インタビュー"
Portuguese: "funk carioca" "tendência" "produtor"
Spanish: "corridos tumbados" "productor" "entrevista"
Arabic/Turkish/Persian/Hindi/Punjabi/etc.: use native terms when possible.
```

If the agent is not confident in a language, it should say so and treat findings cautiously.

## Evidence confidence labels

Use these labels in answers when the trend claim matters.

| Label | Meaning |
|---|---|
| High confidence | official chart/methodology + multiple current references agree |
| Medium confidence | current platform signal + specialist/local press agree |
| Low confidence | one platform or one article only; good as a lead, not a conclusion |
| Anecdotal | forum/community/source-limited; use only as scene clue |

## Query recipes

### Current genre trend

```text
[genre] [country/region] trend 2026 chart producer interview
[genre] "top songs" [platform] [country]
[genre] "what is" "2026" local music press
```

### Reference-track digging

```text
[genre] playlist [year] [platform]
[artist/track] producer interview arrangement harmony tempo
[genre] "songwriter" "interview" "chorus"
```

### Microgenre / obscure tag

```text
"[tag]" music genre references
"[tag]" Bandcamp SoundCloud producer interview
"[tag]" "BPM" "drum pattern"
```

### Regional comparison

```text
[region A] [genre] charts 2026 vs [region B]
[genre] [country] streaming charts YouTube Spotify Apple Music
[local-language genre term] "interview" "producer"
```

### Copyright/style safety

```text
[reference artist] songwriting interview influences
[genre] common features tempo harmony groove
[artist] signature sound analysis avoid copying melody riff lyrics
```

## What not to do

- Do not treat one playlist as a whole genre.
- Do not treat TikTok velocity as full-song structure.
- Do not treat global charts as local taste.
- Do not use SEO listicles as primary evidence when official charts, local press, or practitioner interviews exist.
- Do not ask for the user's streaming login, password, tokens, or private account access.
- Do not copy melodies, lyrics, riffs, samples, or signature artist identifiers from references.

## Output template

```md
**What I checked**
- Sources:
- Time window:
- Region/platform:

**Trend signal**
- High-confidence:
- Medium-confidence:
- Low-confidence / leads:

**Composition translation**
- Tempo/groove:
- Harmony:
- Topline:
- Arrangement:
- Form/hook timing:

**Caveat**
- What still needs reference listening or local validation:
```

# Modes Cheatsheet

Quick reference for the seven diatonic modes, plus important non-diatonic modal scales. For full discussion, see `harmony/modal-harmony.md` and `fundamentals/pitch-intervals-scales.md`.

## The seven diatonic modes

The diatonic modes can be understood in two complementary ways:

1. **Relative mode**: start on a different scale degree of one parent major scale.
2. **Parallel mode**: keep the same tonic and alter scale degrees relative to major/minor.

### Relative modes in C major

| Mode | Starts on in C major | Notes |
|------|----------------------|-------|
| **Ionian** | C (^1) | C D E F G A B |
| **Dorian** | D (^2) | D E F G A B C |
| **Phrygian** | E (^3) | E F G A B C D |
| **Lydian** | F (^4) | F G A B C D E |
| **Mixolydian** | G (^5) | G A B C D E F |
| **Aeolian** | A (^6) | A B C D E F G |
| **Locrian** | B (^7) | B C D E F G A |

### Parallel modes on C

| Mode | Notes (C tonic) | Formula | Character |
|------|-----------------|---------|-----------|
| **Ionian** | C D E F G A B | `W W H W W W H` | Major; bright; the default |
| **Dorian** | C D E♭ F G A B♭ | `W H W W W H W` | Minor with raised ^6; cool, modal |
| **Phrygian** | C D♭ E♭ F G A♭ B♭ | `H W W W H W W` | Minor with ♭2; severe, flamenco-adjacent or metal color when used idiomatically |
| **Lydian** | C D E F♯ G A B | `W W W H W W H` | Major with ♯4; floating, dreamy |
| **Mixolydian** | C D E F G A B♭ | `W W H W W H W` | Major with ♭7; bluesy, rock |
| **Aeolian** | C D E♭ F G A♭ B♭ | `W H W W H W W` | Natural minor; somber |
| **Locrian** | C D♭ E♭ F G♭ A♭ B♭ | `H W W H W W W` | Diminished tonic; rare; theoretical |

(`W` = whole step, `H` = half step.)

## Quick mode identification

To name a mode, identify (a) the tonic and (b) which note is altered relative to major or minor:

### Brighter than major (Lydian)

- Major scale with ♯4 (♯11) → Lydian
- The "happy floating" sound

### Major-like (Ionian, Mixolydian)

- Pure major → Ionian
- Major with ♭7 → Mixolydian

### Minor-like (Aeolian, Dorian)

- Pure natural minor → Aeolian
- Natural minor with ♮6 (raised 6th) → Dorian

### Darker than minor (Phrygian, Locrian)

- Natural minor with ♭2 → Phrygian
- Natural minor with ♭2 and ♭5 → Locrian

## Characteristic chords (the "color" chords) per mode

The chord that most distinguishes each mode from major/minor:

| Mode | Characteristic chord | Notes |
|------|---------------------|-------|
| Ionian | I | The plain major; no special color |
| Dorian | IV (major IV from minor mode!) | The major IV is the Dorian giveaway |
| Phrygian | ♭II | The half-step-up chord — Phrygian's calling card |
| Lydian | II | The major II (a whole step up from tonic, major) |
| Mixolydian | ♭VII | The whole-step-down major chord |
| Aeolian | iv (minor iv) and ♭VI, ♭VII | Standard natural-minor chords |
| Locrian | i° (diminished tonic) | The diminished tonic — instability built in |

## Common modal progressions

### Dorian

| Progression | Example in D Dorian |
|-------------|---------------------|
| i-IV | Dm - G |
| i-IV-vii-i | Dm - G - C - Dm |
| i-iv-IV-i (mixed) | Dm - Gm - G - Dm |

### Phrygian

| Progression | Example in E Phrygian |
|-------------|----------------------|
| i-♭II | Em - F |
| i-♭II-i | Em - F - Em |
| ♭II-♭VII-i | F - D - Em |

### Lydian

| Progression | Example in F Lydian |
|-------------|--------------------|
| I-II | F - G |
| I-II-vii-I | F - G - Em - F |
| I-II-iii-I | F - G - Am - F |

### Mixolydian

| Progression | Example in G Mixolydian |
|-------------|------------------------|
| I-♭VII | G - F |
| I-♭VII-IV-I | G - F - C - G |
| I-V-IV-I | G - D - C - G (with ♭VII implied) |

### Aeolian (natural minor)

| Progression | Example in A Aeolian |
|-------------|---------------------|
| i-VII-VI-VII | Am - G - F - G |
| i-iv-VII | Am - Dm - G |
| i-VI-III-VII | Am - F - C - G |

## Modes of melodic minor (jazz minor)

Built from the ascending melodic minor scale (= natural minor with ♮6 and ♮7, both raised).

| Mode | Parent-scale degree | Notes (parallel C tonic) | Use over |
|------|------------------------------|-----------------|----------|
| Melodic minor | ^1 | C D E♭ F G A B | Cm(maj7); tonic minor in jazz |
| Dorian ♭2 | ^2 | C D♭ E♭ F G A B♭ | iim7♭5 contexts |
| Lydian augmented | ^3 | C D E F♯ G♯ A B | Cmaj7♯5; floating, ambiguous |
| Lydian dominant | ^4 | C D E F♯ G A B♭ | C7♯11 (non-resolving dominant) |
| Mixolydian ♭6 | ^5 | C D E F G A♭ B♭ | C7♭13 |
| Locrian ♮2 | ^6 | C D E♭ F G♭ A♭ B♭ | iiø7 contexts |
| **Altered (super Locrian)** | ^7 | C D♭ D♯ E G♭/F♯ A♭ B♭ | **C7alt** (resolving dominants; functional spelling) |

The **altered scale** is the workhorse of modern jazz — used over `V7alt` resolving to `i`. In theoretical melodic-minor mode spelling, some notes may be written as E♭/F♭; in functional dominant spelling, label them as ♭9, ♯9, 3, ♭5/♯11, ♭13, and ♭7 so the chord function stays clear.

## Modes of harmonic minor

Built from the harmonic minor scale (natural minor with raised ^7 only).

| Mode | Built on | Use |
|------|----------|-----|
| Harmonic minor | ^1 | Tonic in some classical and jazz minor contexts |
| Locrian ♮6 | ^2 | iiø7 with raised 6 (less common) |
| Ionian ♯5 | ^3 | Major with augmented 5 (Cmaj7♯5 in some contexts) |
| Dorian ♯4 | ^4 | Folk and Eastern European |
| **Phrygian dominant** | ^5 | **V7♭9 resolving to minor**; flamenco, Klezmer, metal |
| Lydian ♯2 | ^6 | Rare |
| Super-locrian ♭♭7 | ^7 | Very rare |

The **Phrygian dominant scale** (e.g., G Phrygian dominant: G A♭ B C D E♭ F) is a common Western shorthand for dominant-minor tension and is often heard in flamenco-adjacent, klezmer-adjacent, metal, and some maqam-/makam-adjacent fusion contexts. The augmented 2nd between ♭2 and 3 is its signature, but tradition-specific practice requires more than choosing this scale.

## Pentatonic and blues

### Major pentatonic

C D E G A — the first five notes of the major scale, omitting ^4 and ^7. Avoids half-steps. Folk, country, gospel.

### Minor pentatonic

C E♭ F G B♭ — minor scale with ^2 and ^6 omitted. Blues, rock.

### Blues scale

Minor pentatonic + ♭5: C E♭ F G♭ G B♭. The ♭5 is a "blue note" — a passing chromaticism.

### Major blues scale

Major pentatonic + ♭3: C D E♭ E G A. Used over major-mode blues, rockabilly, country.

## Symmetric scales

These have no clear tonic — the interval pattern repeats within the octave.

| Scale | Pattern | Notes (C) | Use |
|-------|---------|-----------|-----|
| Whole tone | `W W W W W W` (6 notes) | C D E F♯ G♯ B♭ | Over 7♯5 / whole-tone dominant colors; B♭ is the dominant spelling of the enharmonic A♯ |
| Diminished (whole-half) | `W H W H W H W H` (8 notes) | C D E♭ F G♭ A♭ A B | Over diminished chords |
| Diminished (half-whole) | `H W H W H W H W` (8 notes) | C D♭ D♯ E F♯ G A B♭ | Over dominant ♭9 chords; D♯ is ♯9 over C7 |
| Augmented | `A2 H A2 H A2 H` (6 notes) | C D♯ E G A♭ B | Symmetric augmented sound (A2 = augmented 2nd, enharmonic to m3) |

## Other useful scales

| Scale | Notes (C) | Origin/character |
|-------|-----------|------------------|
| Hungarian minor | C D E♭ F♯ G A♭ B | Hungarian folk, Romani |
| Byzantine / double harmonic | C D♭ E F G A♭ B | Greek, Middle Eastern; two augmented 2nds |
| Bebop major | C D E F G A♭ A B | Major + ♯5 added passing |
| Bebop dominant | C D E F G A B♭ B | Mixolydian + ♮7 added passing |
| Bebop dorian | C D E♭ E F G A B♭ | Dorian + ♮3 added passing |
| Hirajōshi (Japanese) | C D E♭ G A♭ | Japanese pentatonic; haunting |

## Mode-to-mode comparisons (which is brighter?)

Bright to dark, with C tonic:

```
Lydian (C D E F♯ G A B) — brightest
Ionian (C D E F G A B)
Mixolydian (C D E F G A B♭)
Dorian (C D E♭ F G A B♭)
Aeolian (C D E♭ F G A♭ B♭)
Phrygian (C D♭ E♭ F G A♭ B♭)
Locrian (C D♭ E♭ F G♭ A♭ B♭) — darkest
```

This ordering corresponds to the modes' positions on the cycle of fifths — Lydian has the most sharps relative to the parallel major; Locrian has the most flats.

## Quick decision matrix

| Want | Try |
|------|-----|
| Bright, modal-major | Lydian (with ♯4) or Ionian |
| Major-with-bluesy-edge | Mixolydian (with ♭7) |
| Cool, modal-minor | Dorian (with ♮6) |
| Standard minor | Aeolian (natural minor) |
| Dark, severe, half-step-above-tonic color | Phrygian (with ♭2) |
| Flamenco-adjacent / klezmer-adjacent dominant-minor color | Phrygian dominant, with tradition-specific caution |
| Jazz tonic minor | Melodic minor |
| Jazz altered dominant | Altered scale (super-Locrian) |
| Jazz non-resolving dominant | Lydian dominant |
| Floating, ambiguous | Whole tone |
| Diminished color | Either diminished scale (whole-half or half-whole) |
| Folk / pentatonic safety | Major or minor pentatonic |
| Blues | Blues scale or minor pentatonic + ♭5 |

## Cross-references

- Full modal harmony discussion → `harmony/modal-harmony.md`
- Pitch system, intervals, scale construction → `fundamentals/pitch-intervals-scales.md`
- Detailed scale formulas → `intervals-and-scale-formulas.md`
- Jazz chord-scale relationships → `harmony/jazz-harmony.md`
- Modal music in different traditions → `genres/folk-roots-and-traditions.md`
- 20th-century modal use → `genres/classical-periods.md` (Impressionism), `techniques/20th-century-techniques.md`

# Scale-Degree Spelling Cheatsheet

Use this asset when the agent needs to spell chords, Roman numerals, modes, altered tensions, or borrowed chords clearly. The goal is not only to name the pitch class, but to spell the musical function.

## Core rule

Spell notes by **scale degree function** first, enharmonic convenience second. C♯ and D♭ may be the same piano key, but they usually mean different things.

## Major key Roman-numeral spelling from C

| Roman numeral | Chord | Notes | Why |
|---|---|---|---|
| `I` | C | C E G | tonic triad |
| `ii` | Dm | D F A | diatonic supertonic minor |
| `iii` | Em | E G B | diatonic mediant minor |
| `IV` | F | F A C | diatonic subdominant major |
| `V` | G | G B D | diatonic dominant major |
| `vi` | Am | A C E | diatonic submediant minor |
| `vii°` | Bdim | B D F | leading-tone diminished triad |
| `♭II` | D♭ | D♭ F A♭ | lowered supertonic, not C♯ major |
| `♭III` | E♭ | E♭ G B♭ | borrowed lowered mediant |
| `♯iv°` | F♯dim | F♯ A C | raised subdominant leading to V or as common-tone color |
| `♭VI` | A♭ | A♭ C E♭ | borrowed lowered submediant |
| `♭VII` | B♭ | B♭ D F | borrowed lowered leading-tone/subtonic chord |

## Minor key Roman-numeral spelling from A minor

| Roman numeral | Chord | Notes | Why |
|---|---|---|---|
| `i` | Am | A C E | tonic minor |
| `ii°` | Bdim | B D F | diatonic supertonic diminished in natural minor |
| `III` | C | C E G | diatonic lowered mediant |
| `iv` | Dm | D F A | subdominant minor |
| `v` | Em | E G B | natural-minor dominant minor |
| `V` | E | E G♯ B | harmonic-minor raised leading tone |
| `VI` | F | F A C | diatonic lowered submediant |
| `VII` | G | G B D | subtonic major |
| `vii°` | G♯dim | G♯ B D | leading-tone chord to i |

## Secondary dominants in C major

| Function | Chord | Notes | Resolution target |
|---|---|---|---|
| `V/V` | D | D F♯ A | G |
| `V7/V` | D7 | D F♯ A C | G |
| `V/ii` | A | A C♯ E | Dm |
| `V7/ii` | A7 | A C♯ E G | Dm |
| `V/vi` | E | E G♯ B | Am |
| `V7/vi` | E7 | E G♯ B D | Am |
| `V/IV` | C7 | C E G B♭ | F |

Raised notes usually create leading tones to the target chord. Spell them as sharps when they resolve upward by half step: F♯→G, C♯→D, G♯→A.

## Borrowed chords in C major

| Borrowed chord | Notes | Source color |
|---|---|---|
| `iv` = Fm | F A♭ C | parallel minor |
| `♭VI` = A♭ | A♭ C E♭ | parallel minor |
| `♭VII` = B♭ | B♭ D F | Mixolydian / parallel minor rock-pop color |
| `♭III` = E♭ | E♭ G B♭ | parallel minor / modal mixture |
| `iiø7` = Dm7♭5 | D F A♭ C | parallel minor pre-dominant |
| `Ger⁶` | A♭ C E♭ F♯ | augmented-sixth predominant to G |

The German augmented sixth in C uses A♭ and F♯ because the voice-leading expands outward to G, not because the chord is simply an A♭7 in all contexts.

## Modal spelling from C tonic

| Mode | Characteristic degree | Notes |
|---|---|---|
| C Dorian | ♮6 against minor third | C D E♭ F G A B♭ |
| C Phrygian | ♭2 | C D♭ E♭ F G A♭ B♭ |
| C Lydian | ♯4 | C D E F♯ G A B |
| C Mixolydian | ♭7 | C D E F G A B♭ |
| C Aeolian | ♭6 and ♭7 | C D E♭ F G A♭ B♭ |
| C Locrian | ♭2 and ♭5 | C D♭ E♭ F G♭ A♭ B♭ |

Spell C Lydian with F♯, not G♭, because the note functions as raised ^4.

## Altered dominant tensions from E

| Tension over E7 | Pitch | Why |
|---|---|---|
| ♭9 | F | lowered ninth above E |
| ♯9 | G | raised ninth above E; enharmonic to minor third, but not labeled ♭3 in dominant context |
| ♯11 / ♭5 | A♯ / B♭ | use A♯ for ♯11, B♭ for ♭5 |
| ♭13 / ♯5 | C / B♯ | use C for ♭13, B♯ only when the raised-fifth spelling matters |

The common E7alt pitch set often includes E, G♯, D, plus some of F, G, A♯/B♭, and C. Do not label A♯ as ♯9 over E.

## Spelling decision table

| Situation | Prefer | Example |
|---|---|---|
| leading tone to a chord tone | sharp spelling | F♯ resolves to G in `V/V` |
| lowered borrowed scale degree | flat spelling | A♭ in `iv` or `♭VI` in C major |
| raised modal degree | sharp spelling | F♯ in C Lydian |
| lowered modal degree | flat spelling | B♭ in C Mixolydian |
| altered dominant ♯9 | raised ninth label | E7♯9 uses G natural |
| augmented sixth voice-leading | outward-resolving spelling | A♭ and F♯ resolve to G |
| slash-bass inversion | actual bass pitch | C/E, G/B |

## Common mistakes to avoid

- Do not spell `♭II` in C as C♯ major.
- Do not spell C Lydian as C D E G♭ G A B.
- Do not label E–A♯ as a ♯9; it is ♯11 / ♭5 territory.
- Do not flatten or sharpen letters mechanically without checking scale-degree function.
- Do not treat Roman numerals as fixed chord names; they depend on the key.

## Cross-references

- Intervals and scale formulas → `intervals-and-scale-formulas.md`
- Modes cheatsheet → `modes-cheatsheet.md`
- Chord symbol ambiguity → `chord-symbol-ambiguity-and-parsing.md`
- Functional harmony → `references/harmony/functional-harmony.md`
- Chromatic harmony → `references/harmony/chromatic-harmony.md`
- Modulation → `references/harmony/modulation.md`

"""
Generate benchmark visualization charts.

Outputs PNG files into benchmarks/charts/. Run after benchmark data is
available; safe to re-run (overwrites).

Usage::

    pip install matplotlib
    python benchmarks/make_charts.py

Charts produced:
- chart_p8_boundary.png     — P8 boundary respect comparison (headline)
- chart_glm_winners.png     — GLM blind eval win counts
- chart_cross_benchmark.png — Cross-benchmark win pattern
"""

# SPDX-License-Identifier: MIT

from __future__ import annotations

from pathlib import Path

import matplotlib.pyplot as plt
from matplotlib import font_manager

ROOT = Path(__file__).resolve().parents[1]
OUT_DIR = ROOT / "benchmarks" / "charts"
OUT_DIR.mkdir(exist_ok=True)


# ------------------------------------------------------------
# Font setup — register Pretendard so Korean labels render cleanly
# ------------------------------------------------------------

def setup_font() -> str:
    """Register Pretendard if installed; return the family name to use."""
    found_paths = [p for p in font_manager.findSystemFonts() if "Pretendard" in p]
    for path in found_paths:
        font_manager.fontManager.addfont(path)
    if found_paths:
        return "Pretendard"
    print("[warn] Pretendard not found; falling back to default sans-serif.")
    return "sans-serif"


FONT_FAMILY = setup_font()
plt.rcParams["font.family"] = FONT_FAMILY
plt.rcParams["axes.unicode_minus"] = False
plt.rcParams["axes.spines.top"] = False
plt.rcParams["axes.spines.right"] = False


# ------------------------------------------------------------
# Color palette
# ------------------------------------------------------------

C_NO_SKILL = "#B5B5B5"     # neutral gray — baseline
C_GLM_SKILL = "#2C5F7C"    # deep teal-blue — GLM / B / with-skill
C_GLM_NAV = "#7A9DBF"      # mid-blue — GLM C (SKILL.md + nav)
C_CLAUDE_SKILL = "#D97757" # Anthropic terracotta — Claude with-skill (strong)
C_CLAUDE_SOFT = "#E8A589"  # lighter terracotta — Claude with-skill (moderate)
C_TIE = "#D9D9D9"          # light gray — tie
C_BG = "#FFFFFF"
C_GRID = "#E5E5E5"
C_TEXT = "#2A2A2A"


# ------------------------------------------------------------
# Chart 1 — P8 boundary respect (the headline)
# ------------------------------------------------------------

def chart_p8_boundary() -> None:
    fig, (ax1, ax2) = plt.subplots(
        1, 2, figsize=(10, 4.8), gridspec_kw={"width_ratios": [3, 2]}, sharey=True
    )
    fig.patch.set_facecolor(C_BG)

    # Left panel: GLM-5.1 (3 conditions)
    glm_labels = ["A\n(baseline)", "B\n(SKILL.md)", "C\n(SKILL.md + nav)"]
    glm_scores = [1, 5, 3]
    glm_colors = [C_NO_SKILL, C_GLM_SKILL, C_GLM_NAV]
    bars1 = ax1.bar(glm_labels, glm_scores, color=glm_colors, width=0.6, edgecolor="white")
    ax1.set_title("GLM-5.1 (Approach 1: system prompt injection)", fontsize=11, color=C_TEXT, pad=14)
    ax1.set_ylabel("Boundary respect (1–5)", fontsize=11, color=C_TEXT)
    ax1.set_ylim(0, 5.6)
    ax1.set_yticks([1, 2, 3, 4, 5])
    ax1.tick_params(axis="both", colors=C_TEXT, labelsize=10)
    ax1.grid(axis="y", color=C_GRID, linewidth=0.7)
    ax1.set_axisbelow(True)
    for bar, score in zip(bars1, glm_scores):
        ax1.text(
            bar.get_x() + bar.get_width() / 2, bar.get_height() + 0.15, str(score),
            ha="center", va="bottom", fontsize=11, fontweight="bold", color=C_TEXT,
        )

    # Right panel: Claude Opus 4.7 (2 conditions)
    cl_labels = ["A\n(no skill)", "B\n(with skill)"]
    cl_scores = [1, 5]
    cl_colors = [C_NO_SKILL, C_CLAUDE_SKILL]
    bars2 = ax2.bar(cl_labels, cl_scores, color=cl_colors, width=0.6, edgecolor="white")
    ax2.set_title("Claude Opus 4.7 (Approach 2: native lazy loading)", fontsize=11, color=C_TEXT, pad=14)
    ax2.set_ylim(0, 5.6)
    ax2.tick_params(axis="both", colors=C_TEXT, labelsize=10)
    ax2.grid(axis="y", color=C_GRID, linewidth=0.7)
    ax2.set_axisbelow(True)
    for bar, score in zip(bars2, cl_scores):
        ax2.text(
            bar.get_x() + bar.get_width() / 2, bar.get_height() + 0.15, str(score),
            ha="center", va="bottom", fontsize=11, fontweight="bold", color=C_TEXT,
        )

    fig.suptitle(
        "P8 — does the model know its own scope? (DAW EQ, out-of-scope prompt)",
        fontsize=13, color=C_TEXT, fontweight="semibold", y=1.0,
    )
    fig.text(
        0.5, -0.02,
        "1 = answers freely, no boundary awareness    ·    "
        "5 = explicit scope statement + redirects to in-scope alternative",
        ha="center", fontsize=9, color="#666",
    )

    fig.tight_layout()
    out = OUT_DIR / "chart_p8_boundary.png"
    fig.savefig(out, dpi=200, bbox_inches="tight", facecolor=C_BG)
    plt.close(fig)
    print(f"  wrote {out.relative_to(ROOT)}")


# ------------------------------------------------------------
# Chart 2 — GLM blind eval winners
# ------------------------------------------------------------

def chart_glm_winners() -> None:
    fig, ax = plt.subplots(figsize=(8, 4.4))
    fig.patch.set_facecolor(C_BG)

    labels = ["A (baseline)", "B (SKILL.md only)", "C (SKILL.md + nav)"]
    wins = [1, 6, 1]
    colors = [C_NO_SKILL, C_GLM_SKILL, C_GLM_NAV]
    bars = ax.bar(labels, wins, color=colors, width=0.55, edgecolor="white")

    ax.set_title(
        "GLM-5.1 blind evaluation — wins per condition (n = 8)",
        fontsize=13, color=C_TEXT, fontweight="semibold", pad=16,
    )
    ax.set_ylabel("블라인드 판정 승수", fontsize=11, color=C_TEXT)
    ax.set_ylim(0, 7.5)
    ax.set_yticks([0, 1, 2, 3, 4, 5, 6, 7])
    ax.tick_params(axis="both", colors=C_TEXT, labelsize=10)
    ax.grid(axis="y", color=C_GRID, linewidth=0.7)
    ax.set_axisbelow(True)

    for bar, w in zip(bars, wins):
        ax.text(
            bar.get_x() + bar.get_width() / 2, bar.get_height() + 0.15,
            f"{w} / 8", ha="center", va="bottom",
            fontsize=11, fontweight="bold", color=C_TEXT,
        )

    # Footer caption (instead of inline annotation that collides with x-tick labels)
    fig.text(
        0.5, -0.02,
        "A의 1승은 P6(리하모니제이션) — B에 발생한 단일 생성 글리치로 인한 반사이익",
        ha="center", fontsize=9, color="#666", style="italic",
    )

    fig.tight_layout()
    out = OUT_DIR / "chart_glm_winners.png"
    fig.savefig(out, dpi=200, bbox_inches="tight", facecolor=C_BG)
    plt.close(fig)
    print(f"  wrote {out.relative_to(ROOT)}")


# ------------------------------------------------------------
# Chart 3 — Cross-benchmark win pattern
# ------------------------------------------------------------

def chart_cross_benchmark() -> None:
    fig, ax = plt.subplots(figsize=(9, 4.6))
    fig.patch.set_facecolor(C_BG)

    benchmarks = ["GLM-5.1\nApproach 1 (blind)", "Claude Opus 4.7\nApproach 2 (open)"]

    # GLM (blind, n=8): A=1, B=6, C=1
    glm_a, glm_b, glm_c = 1, 6, 1
    # Claude (open, n=8): with-skill strong=5, moderate=2, tie=1, baseline=0
    cl_baseline = 0
    cl_tie = 1
    cl_with_moderate = 2
    cl_with_strong = 5

    width = 0.5
    x = list(range(len(benchmarks)))

    # GLM stack (bottom → top): A, C, B — teal family for GLM
    ax.bar(x[0], glm_a, width=width, color=C_NO_SKILL, edgecolor="white")
    ax.bar(x[0], glm_c, width=width, bottom=glm_a, color=C_GLM_NAV, edgecolor="white")
    ax.bar(x[0], glm_b, width=width, bottom=glm_a + glm_c, color=C_GLM_SKILL, edgecolor="white")

    # Claude stack (bottom → top): tie, with-skill moderate, with-skill strong — terracotta family for Claude
    ax.bar(x[1], cl_tie, width=width, bottom=cl_baseline, color=C_TIE, edgecolor="white")
    ax.bar(x[1], cl_with_moderate, width=width, bottom=cl_baseline + cl_tie,
           color=C_CLAUDE_SOFT, edgecolor="white")
    ax.bar(x[1], cl_with_strong, width=width, bottom=cl_baseline + cl_tie + cl_with_moderate,
           color=C_CLAUDE_SKILL, edgecolor="white")

    # In-segment labels: count + condition descriptor
    def label(x_pos: int, y_center: float, count: int, name: str,
              color: str = "white", weight: str = "semibold") -> None:
        ax.text(x_pos, y_center, f"{name}: {count}", ha="center", va="center",
                color=color, fontsize=10, fontweight=weight)

    # GLM labels
    label(x[0], glm_a / 2, glm_a, "A", color=C_TEXT)
    label(x[0], glm_a + glm_c / 2, glm_c, "C", color=C_TEXT)
    label(x[0], glm_a + glm_c + glm_b / 2, glm_b, "B (skill)", color="white", weight="bold")

    # Claude labels
    label(x[1], cl_baseline + cl_tie / 2, cl_tie, "tie", color=C_TEXT)
    label(x[1], cl_baseline + cl_tie + cl_with_moderate / 2, cl_with_moderate, "skill (moderate)", color=C_TEXT)
    label(x[1], cl_baseline + cl_tie + cl_with_moderate + cl_with_strong / 2,
          cl_with_strong, "skill (strong)", color="white", weight="bold")

    ax.set_xticks(x)
    ax.set_xticklabels(benchmarks, fontsize=11, color=C_TEXT)
    ax.set_ylabel("프롬프트 수 (n = 8)", fontsize=11, color=C_TEXT)
    ax.set_ylim(0, 9)
    ax.set_yticks([0, 2, 4, 6, 8])
    ax.tick_params(axis="y", colors=C_TEXT, labelsize=10)
    ax.grid(axis="y", color=C_GRID, linewidth=0.7)
    ax.set_axisbelow(True)
    ax.set_title(
        "Cross-benchmark win pattern — same 8 prompts, different mechanism",
        fontsize=13, color=C_TEXT, fontweight="semibold", pad=16,
    )

    fig.text(
        0.5, -0.02,
        "동일한 8 프롬프트, 두 메커니즘에서 모두 스킬 적용 응답이 우세 — "
        "특히 Approach 2에서 baseline은 0승",
        ha="center", fontsize=9, color="#666", style="italic",
    )

    fig.tight_layout()
    out = OUT_DIR / "chart_cross_benchmark.png"
    fig.savefig(out, dpi=200, bbox_inches="tight", facecolor=C_BG)
    plt.close(fig)
    print(f"  wrote {out.relative_to(ROOT)}")


def main() -> None:
    print(f"Generating charts (font: {FONT_FAMILY})")
    chart_p8_boundary()
    chart_glm_winners()
    chart_cross_benchmark()
    print(f"\nDone. Output: {OUT_DIR.relative_to(ROOT)}/")


if __name__ == "__main__":
    main()

# Benchmarks

Empirical comparison of model responses with and without the music-composition skill loaded. Designed to give honest, reproducible evidence of whether the skill changes answer quality on real-shaped composition questions.

## Methodology

### Conditions

Three conditions are tested per prompt, escalating context:

| Condition | System prompt |
|-----------|---------------|
| **A** (baseline) | Generic chat-assistant instruction only |
| **B** (SKILL.md only) | Generic + `SKILL.md` prepended |
| **C** (SKILL.md + nav) | Generic + `SKILL.md` + `references/00-navigation.md` prepended |

The user prompt is identical across all three. Temperature is `0.0` (default for the SDK) for reproducibility.

### Why three conditions, not two

Two conditions (on / off) tells you *whether* the skill helps. Three conditions tells you *where* the help comes from:

- A → B improvement: the meta-instructions (philosophy, "options not commandments", genre framing) are doing work even without the deep references.
- B → C improvement: the routing map adds value on top of meta-instructions.
- A → C improvement that's mostly in B → C: the routing is the engine, not the philosophy.

Knowing which lever matters is more useful than a single binary signal.

### Why not load all references

Loading the full `references/` tree would balloon context to 200KB+ and isn't representative of how skills are actually used. In a real harness (Claude Code, Agent SDK), references are loaded lazily by tool calls. A future *Approach 2* benchmark using tool-mediated lazy loading should be added (see "Future work" below).

### Model

GLM-5.1 via OpenCode Go's OpenAI-compatible endpoint. Reasons:

- Strong instruction-following → if the skill has any effect, it should show
- Good CJK tokenization → Korean prompts answered cleanly
- Vanilla "general chat assistant" baseline (not specifically tuned for any skill format)
- Affordable for repeat runs (well within OpenCode Go's $30/week quota)

**Caveat to record in any write-up:** SKILL.md is designed for Claude. A non-Claude model may underweight the skill content, so this benchmark likely *underestimates* the skill's effect on Claude. Cross-model testing is future work.

### Prompts

Eight prompts spanning the skill's main value claims:

| # | Category | Tests |
|---|----------|-------|
| P1 | Vague diagnosis | Does the model decompose "weak chorus" into technical variables? |
| P2 | Genre + emotion | Does it recognize the K-pop ballad frame and apply genre conventions? |
| P3 | Technical theory | Does it give clean chord-scale advice for `Cm7`? |
| P4 | Hybrid | Can it merge jazz harmony into K-pop without losing either frame? |
| P5 | Cultural specificity | Does it engage with *pansori* concretely or fall back to generic "Korean color"? |
| P6 | Reharmonization | Does it offer multiple substitution options with tradeoffs? |
| P7 | Instrument idiom | Does it think about guitar voicing / shape, not just chord names? |
| P8 | **Boundary (negative)** | Does the skill *refuse* to advise on DAW EQ? Stays in lane? |

P8 is the most important honesty signal. A skill that *expands* its scope to "I'll answer anything music-related" loses credibility. The B/C conditions should defer / decline; A may attempt an answer.

## Running the benchmark

### Requirements

```bash
pip install openai-agents openai
```

### Setup

1. Get an OpenCode Go API key (Zen subscription).
2. Set environment variable:

   ```bash
   export OPENCODE_API_KEY=oc_zen_...
   ```

### Execute

```bash
python benchmarks/run_eval.py
```

The script writes a timestamped pair of files to `benchmarks/results/`:

- `v1.0-<timestamp>.md` — human-readable; includes prompt, all three responses per prompt
- `v1.0-<timestamp>.jsonl` — one JSON record per (prompt, condition) for downstream analysis

`benchmarks/results/` is `.gitignore`'d to keep ad-hoc reruns out of the repo. The canonical v1.0 reference run is committed at `benchmarks/v1.0-run.md` and `benchmarks/v1.0-run.jsonl`.

### Cost estimate

8 prompts × 3 conditions × ~3K input + ~1K output tokens ≈ 100K tokens per run. Well under $1 at GLM-5.1 rates.

## Scoring

After a run, fill in `v1.0-eval.md` (committed canonical eval) with:

### 1. Rubric (per response, 1–5)

| Dimension | Anchors |
|-----------|---------|
| **Concreteness** | 1 = "consider modal mixture"; 5 = "try `Cm9 → A♭maj9 → Fm7 → G7sus4`, voicing the 9 on top" |
| **Options** | 1 = single recipe; 5 = 2–4 alternatives with tradeoffs |
| **Genre awareness** | 1 = ignores genre; 5 = names the genre frame and adjusts conventions |
| **Diagnosis depth** (vague-problem prompts only) | 1 = jumps to fix; 5 = decomposes into variables before prescribing |
| **Pitfall awareness** | 1 = no caveats; 5 = warns about a specific common mistake |

For P8 (boundary): score by **boundary respect** instead — does it decline / route to a different domain (1 = answers freely, 5 = explicit decline or scope statement)?

### 2. Blind judgment

Strip condition labels. Read all three answers per prompt. Mark which feels most useful to a working composer + one-sentence why. Author bias is real; treat the blind read as the primary signal and the rubric as supporting.

### 3. Tabulate

In `v1.0-eval.md`, per prompt:

```markdown
**Rubric (A / B / C):**
- Concreteness: 2 / 4 / 4
- Options: 1 / 3 / 4
- ...

**Blind pick:** C — best diagnostic decomposition, named the energy curve and arrangement density variables explicitly.

**Notes:** A confused tempo with intensity. B got the diagnosis but didn't suggest specific moves. C did both.
```

Aggregate across prompts at the end. Be honest about ties and prompts where the skill didn't help — those are signal too.

## Future work

- **Approach 2 — tool-mediated loading:** Give the agent a `read_file` tool and only mention `SKILL.md` exists. Tests the full skill discovery + routing mechanism, not just context injection.
- **Cross-model:** Run on Claude Sonnet 4.5 (native SKILL.md home turf), Qwen3.5 Plus, DeepSeek V4 to map effect size by model.
- **Cross-language:** Add English-language prompt variants to test that the skill helps regardless of input language.
- **Adversarial:** Add prompts designed to *break* the skill (e.g., wildly out-of-scope, internally contradictory, or culturally fraught) and see if it fails gracefully.

"""
v1.0 benchmark — skill on vs. off.

Runs 8 prompts × 3 conditions on GLM-5.1 via OpenCode Go's
OpenAI-compatible endpoint. Outputs results to
``benchmarks/results/v1.0-<timestamp>.{md,jsonl}``.

Usage::

    pip install openai-agents openai
    export OPENCODE_API_KEY=oc_zen_...
    python benchmarks/run_eval.py

The script writes a timestamped pair of files. The canonical
``benchmarks/v1.0-eval.md`` is intended to be filled in manually
from the latest run after blind grading.
"""

# SPDX-License-Identifier: MIT

from __future__ import annotations

import asyncio
import json
import os
from datetime import datetime, timezone
from pathlib import Path

from agents import Agent, OpenAIChatCompletionsModel, Runner
from openai import AsyncOpenAI

ROOT = Path(__file__).resolve().parents[1]
SKILL_MD = (ROOT / "SKILL.md").read_text(encoding="utf-8")
NAV_MD = (ROOT / "references" / "00-navigation.md").read_text(encoding="utf-8")

# Generic chat-assistant baseline. Intentionally minimal — we want any
# improvement to come from the skill content, not the baseline prompt.
GENERIC_SYSTEM = (
    "You are a helpful assistant. "
    "Answer the user's question directly, concretely, and concisely."
)

CONDITIONS: dict[str, str] = {
    "A_baseline": GENERIC_SYSTEM,
    "B_skill_only": f"{GENERIC_SYSTEM}\n\n---\n\n{SKILL_MD}",
    "C_skill_plus_nav": f"{GENERIC_SYSTEM}\n\n---\n\n{SKILL_MD}\n\n---\n\n{NAV_MD}",
}

PROMPTS: list[dict[str, str]] = [
    {
        "id": "P1_chorus_diagnosis",
        "category": "Vague diagnosis",
        "prompt": "내 곡 후렴이 안 꽂혀. 어떻게 고치지?",
    },
    {
        "id": "P2_kpop_dark_bridge",
        "category": "Genre + emotion",
        "prompt": "K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어.",
    },
    {
        "id": "P3_scale_over_cm7",
        "category": "Technical theory",
        "prompt": "Cm7 위에 어떤 스케일이 자연스러워?",
    },
    {
        "id": "P4_jazz_kpop_hybrid",
        "category": "Hybrid request",
        "prompt": "재즈 화성으로 K-pop 후렴을 만들어보고 싶어.",
    },
    {
        "id": "P5_pansori_ost",
        "category": "Cultural specificity",
        "prompt": "판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?",
    },
    {
        "id": "P6_reharm",
        "category": "Reharmonization",
        "prompt": "Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘.",
    },
    {
        "id": "P7_acoustic_ballad",
        "category": "Instrument idiom",
        "prompt": "기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘.",
    },
    {
        "id": "P8_daw_eq",
        "category": "Boundary (negative)",
        "prompt": "DAW에서 보컬 EQ 어떻게 걸어야 해?",
    },
]

MODEL = "glm-5.1"
BASE_URL = "https://opencode.ai/zen/go/v1"


async def run_one(model: OpenAIChatCompletionsModel, system: str, user: str) -> str:
    agent = Agent(name="benchmark", instructions=system, model=model)
    result = await Runner.run(agent, user)
    return result.final_output


async def main() -> None:
    api_key = os.environ.get("OPENCODE_API_KEY")
    if not api_key:
        raise SystemExit("Set OPENCODE_API_KEY before running.")

    client = AsyncOpenAI(base_url=BASE_URL, api_key=api_key)
    model = OpenAIChatCompletionsModel(model=MODEL, openai_client=client)

    timestamp = datetime.now(timezone.utc).strftime("%Y%m%dT%H%M%SZ")
    out_dir = ROOT / "benchmarks" / "results"
    out_dir.mkdir(exist_ok=True)
    md_path = out_dir / f"v1.0-{timestamp}.md"
    jsonl_path = out_dir / f"v1.0-{timestamp}.jsonl"

    md_lines: list[str] = [
        f"# v1.0 Benchmark Results — {timestamp}",
        "",
        f"**Model:** `{MODEL}` via OpenCode Go (`{BASE_URL}`)",
        f"**Date (UTC):** {timestamp}",
        "",
        "**Conditions:**",
        "- **A** (baseline): generic assistant prompt only",
        "- **B** (SKILL.md only): generic + `SKILL.md` prepended",
        "- **C** (SKILL.md + nav): generic + `SKILL.md` + `references/00-navigation.md` prepended",
        "",
        f"**Prompts:** {len(PROMPTS)}",
        "",
        "---",
        "",
    ]

    jsonl_lines: list[str] = []

    for p in PROMPTS:
        print(f"--- {p['id']} [{p['category']}] ---")
        md_lines.append(f"## {p['id']} — {p['category']}")
        md_lines.append("")
        md_lines.append(f"**Prompt:** {p['prompt']}")
        md_lines.append("")

        for cond_name, system in CONDITIONS.items():
            print(f"  running {cond_name} ...", flush=True)
            try:
                response = await run_one(model, system, p["prompt"])
            except Exception as exc:  # noqa: BLE001 — surface every failure
                response = f"[ERROR] {type(exc).__name__}: {exc}"
                print(f"    failed: {exc}")

            md_lines.append(f"### {cond_name}")
            md_lines.append("")
            md_lines.append(response.strip())
            md_lines.append("")

            jsonl_lines.append(
                json.dumps(
                    {
                        "prompt_id": p["id"],
                        "category": p["category"],
                        "condition": cond_name,
                        "model": MODEL,
                        "system_prompt_chars": len(system),
                        "user_prompt": p["prompt"],
                        "response": response,
                    },
                    ensure_ascii=False,
                )
            )

        md_lines.append("---")
        md_lines.append("")

    md_path.write_text("\n".join(md_lines), encoding="utf-8")
    jsonl_path.write_text("\n".join(jsonl_lines) + "\n", encoding="utf-8")
    print(f"\nResults written to:\n  {md_path}\n  {jsonl_path}")


if __name__ == "__main__":
    asyncio.run(main())

# v1.0 Blind Evaluation

각 프롬프트의 세 응답을 읽고, 작곡가 관점에서 가장 쓸 만한 것을 X1/X2/X3 중에서 고르세요. 한 줄 이유.

---

## P1_chorus_diagnosis — Vague diagnosis

**Prompt:** 내 곡 후렴이 안 꽂혀. 어떻게 고치지?

### Response X1

후렴이 안 꽂히는 원인은 보통 몇 가지 패턴 중 하나입니다. 진단해 보고 고쳐 보세요:

---

## 1. 절→후렴 리프트가 부족하다

후렴은 절보다 **에너지가 올라야** 꽂힙니다. 체크:

- **음역**: 후렴 최고음이 절 최고음보다 최소 단3도 이상 높은가?
- **밀도**: 악기 레이어가 후렴에서 두꺼워지는가?
- **리듬**: 보컬 프레이징이 더 빨라지거나 싱코페이션이 많아지는가?

**빠른 수정**: 후렴 첫 마디 멜로디를 3~5음 위로 올려 보세요.

---

## 2. 훅 멜로디가 약하다

기억에 남는 후렴은 **반복 가능한 짧은 모티프** + **인상적인 간진행**이 있습니다.

**해결책**:
- 후렴 멜로디를 1~2마디 짧은 셈(모티프)으로 줄이고, 그것을 반복·변형으로 전개
- 도약(5도 이상)을 한 곳 넣어 보세요. 예: `도→레→파↑↗︎(5도 상행)→미`
- 리듬 변형: 절이 8분음 wired이면 후렴은 점음표나 부점 리듬으로

---

## 3. 화성이 너무 예측 가능하다

절-후렴 화성이 같으면 후렴이 "터지지" 않습니다.

**해결책**:
- 후렴 진입을 **IV** 또는 **♭VI**로 시작 (예: C메이저 곡이면 F메이저 또는 A♭메이저)
- 절이 `I → vi → IV → V`라면 후렴은 `IV → I → V → vi`로 순서 교환
- 후렴 첫 코드에 **sus4 해결** 또는 **add9** 추가로 색감 차이

---

## 4. 가사·음절 밀도 문제

절보다 후렴이 음절이 적고 음이 길어야 힘이 실립니다.

**해결책**: 후렴 첫 라인을 3~5개 음절로 압축하고 긴 음가에 담기.

---

## 장르가 중요합니다

K-pop, 발라드, 힙합, 락… 장르마다 "꽂히는 후렴"의 공식이 다릅니다. **어떤 장르인지 알려주시면** 더 구체적으로 조언할 수 있습니다.

---

지금 곡의 키, 현재 코드 진행, 후렴 멜로디 움직임을 알려주시면 바로 수정안을 제시할게요.

### Response X2

후렴이 "안 꽂힌다"는 건 여러 원인이 있을 수 있어요. 진단 먼저, then 처방.

---

## 빠른 진단 체크리스트

| 요소 | 후렴이 인상적이려면… | 의심 포인트 |
|------|---------------------|-------------|
| **멜로디 높이** | 후렴 최고음 > 절 최고음 | 절이랑 비슷한 음역? |
| **리듬 훅** | 반복·지속되는 리듬 모티프 | 가사 얹는 리듬이 매번 달라? |
| **화성 변화** | 절과 다른 코드 진행 또는 조바꿈 | 절이랑 같은 I-V-vi-IV? |
| **밀도** | 더 많은 악기·보이스 레이어 | 후렴인데 편곡이 절이랑 같아? |
| **대비** | 절→후렴 순간 에너지 점프 | 도입부→후렴 전환이 서서히? |
| **가사 훅** | 짧고 강한 구절, 반복 | 한 줄이 길거나 산문적? |

---

## 가장 흔한 3가지 원인과 구체적 처방

### 1. 멜로디가 절이랑 너무 비슷하다
**해결:** 후렴 첫 음을 절 최고음보다 3도~5도 위에서 시작.

```
절 끝: E4 → 후렴 첫음: A4 이상
예)절: C4-D4-E4-G4  →  후렴: G4-A4-C5-D5
```

### 2. 화성이 안 바뀐다
**해결:** 후렴 진행을 절과 다르게. 가장 즉각적인 옵션들:

| 옵션 | 코드 (C 메이저 기준) | 효과 |
|------|---------------------|------|
| **소디미니시시** | C → A♭ → B♭ → C | 극적, 영웅적 |
| **iv 도입** | Fm → C → G → C | 그리웋하면서 단단 |
| **V/V → V** | D → G → Am → Fm | 텐션→해결, 빠른 훅 |
| **나폴리 6도** | C → D♭maj7 → G → C | 어둡고 파워풀 |

### 3. 리듬 훅이 없다
**해결:** 가사 음절에 반복되는 리듬 패턴 부여.

```
"다가 와" ×2  (점-점-_____)  ← 4음절이면 점-점-박-박 패턴을 반복
"말해 줘"      (점-박)
→ "다가와-다가와-말해줘" 리듬: x-x- x-x- x-x
```

하이라이트 가사에 **같은 리듬 3번 반복** → 이것만으로도 훅이 됩니다.

---

## 다음 질문에 답해주면 더 정확히 도와줄게요

1. **장르** (K-pop 발라드? 힙합? 인디? 락?)
2. **지금 후렴 코드 진행** (알려주면 바로 대안 제시)
3. **어떤 안 꽂힘?** (멜로디가 안 외워짐 / 에너지가 안 올라감 / 가사가 안 남음)

이 세 가지만 알려줘도 훨씬 구체적으로 고쳐줄 수 있어요.

### Response X3

후렴이 귀에 안 꽂히는 원인과 구체적인 해결책 5가지입니다.

1. **멜로디 단순화 및 반복**: 멜로디가 너무 산만하면 남지 않습니다. 기억하기 쉬운 짧은 멜로디(모티프) 하나를 만들어 후렴 안에서 2~3번 반복하십시오.
2. **벌스와의 대비**: 벌스와 음역대, 리듬, 에너지가 비슷하면 후렴이 튀지 않습니다. 후렴에서 음역대를 확 올리거나, 반대로 벌스보다 리듬을 여유롭게(긴 음표로) 만들어 차이를 주십시오.
3. **훅(Hook) 가사 배치**: 한 번 들으면 따라 부를 수 있는 짧고 강렬한 한 줄(훅)을 후렴의 첫머리나 마지막에 배치하고, 운율(라임)을 맞추십시오.
4. **악기 리프 추가**: 보컬 멜로디만으로 부족하다면, 후렴 시작과 함께 귀에 맴도는 짧은 악기 리프(신스, 기타 등)를 멜로디와 겹쳐 연주하십시오.
5. **프리 코러스(빌드업) 강화**: 후렴이 갑자기 튀어나오면 어색합니다. 후렴 직전 구간에서 드럼 킥을 끊거나, 코러스 화음을 넣어 '이제 폭발한다'는 기대감을 먼저 조성하십시오.

**Pick (X1 / X2 / X3):**  
**One-sentence reason:**  

---

## P2_kpop_dark_bridge — Genre + emotion

**Prompt:** K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어.

### Response X1

K-pop 발라드 브릿지를 어둡게 만드는 구체적인 방법 몇 가지:

---

## 1. 평행단조 차용 (Modal Mixture)

가장 즉각적이고 장르에 잘 맞는 방법.

**C 메이저 기준:**
- `F → Fm → C` (♭VI → iv → I) — Fm이 순간 그림자를 드리움
- `Am → A♭m → G7` — A♭m은 ♭vi, 극적인 하강
- `Dm7 → D♭maj7 → Cm7 → Fm7` — ♭VII → ♭VI 진행

## 2. 단조의 이차도수 (Secondary Dominants to Minor)

목표음을 단조로 해결.

- `E7 → Am` (V/vi → vi) — 이미 어두운 Am으로
- `B7 → Em` (V/iii → iii) — 더 깊은 그림자
- `F7 → B♭m` (♭VII 처리 가능) — 예측 밖의 색채

## 3. 색채중간 (Chromatic Mediant)

K-pop 발라드 브릿지 끝에서 흔히 쓰는 극적 전조 느낌.

- `Cmaj7 → A♭maj7` — I → ♭VI, 3도 반음 관계
- `Cmaj7 → Emaj7` — I → III, 올림 중간
- `Fm7 → Bmaj7` — ♭IV → III⁺ (비트로스 진행)

## 4. 종결구 변경

3단음 아래 종결 (Plagal ♭VI cadence):
- `A♭maj7 → C/G → Fm → C`
- 마지막 V 대신 IVm으로 "미결" 느낌 유지

---

**실제 예시 진행 (C 메이저 곡의 브릿지):**

```
| Fmaj7  | Fm7    | Dm7   | D♭maj7 |
| Am7    | A♭m7   | G7sus4| G7♭9   |
```

- F→Fm: modal mixture
- Dm→D♭: ♭VII 하강
- Am→A♭m: ♭vi 하강
- G7♭9: 증극 긴장, 코러스 해결 준비

어느 정도 "어둡게"인지(섬뜩함 vs 쓸쓸함)에 따라 골라 쓰면 됨.

### Response X2

K-pop 발라드 브릿지를 어둡게 만들기 위한 구체적이고 실전적인 방법:

1. **화성 (코드):** 밝은 코드를 피하고 **디미니시드 코드**나 **단2도 불협화음**을 사용해 불안정함을 조성하세요. 베이스라인을 **반음 하행**시키면 긴장감과 우울함이 극대화됩니다.
2. **멜로디:** 고음역 클라이맥스 대신 **중저음역**으로 끌어내리세요. 안정적인 음정 대신 **증4도(트라이톤)**나 단2도 음정 도약으로 서스펜스(매달림) 느낌을 주세요.
3. **편곡:** 악기를 전부 터트리기 전에 오히려 악기를 걷어내고 **적막(무음, 공백)**을 사용하세요. 잔향이 긴 어두운 신디사이저, 저음 첼로, 무거운 서브베이스로 공간을 채우세요.
4. **보컬:** 깔끔한 가창 대신 **보컬 프라이(목소리 떨림)**, 거친 숨소리, 속삭임을 활용하세요. 브릿지 끝부분에서 옥타브 점프와 함께 거친 파열음으로 절규하듯 전환하세요.
5. **가사:** 단순한 슬픔을 넘어 '광기, 상실, 시선 고정, 통제 불능' 등의 단어로 서사를 어둡게 비틀어 보세요.

### Response X3

# K-pop 발라드 브릿지: 더 어둡게 만드는 방법

## 화성적 접근

**1. 모달 믹스처 (Modal Mixture)**
메이저 키에서 마이너의 플랫 디그리를 차용:
- `♭VI` → `♭VII` → `i` (C 메이저: A♭maj7 → B♭ → Cm)
- `iv` 대신 `IV` 자리에 마이너 서브도미넌트 사용: `Fm6` 대신 `F`
- `♭III` → `♭VI` 시퀀스: C 메이저에서 E♭maj7 → A♭maj7

**2. 크로매틱 메디안트**
3도 관계의 색채 변화:
- `I` → `♭III` → `♭V` (C → E♭ → G♭) — 극적인 암흑감
- `I` → `♯iv°` → `V` (C → F♯° → G)

**3. 세컨더리 도미넌트 → 마이너**
- `V7/vi` → `vi` (E7 → Am in C) — 라딕스 느낌의 어두운 해결
- `V7/ii` → `ii` (A7 → Dm) — 서브도미넌트 방향의 긴장

## 선율적 접근

- 브릿지 시작을 낮은 레지스터에서 (C3~E4 남성, A3~C5 여성)
- 트라이톤 증4/감5 도약 활용
- 하행 반음 진행: `^5 → ♭5 → 4` 또는 `^1 → ♭7 → ♭6`
- 레가토 프레이징, 적은 음, 더 긴 지속음

## 텍스처 & 편곡

| 요소 | 변경 |
|------|------|
| 베이스 | 더 낮은 레지스터, 펜타토닉 하행 |
| 피아노 | 저음 옥타브 트리아드, 페달 포인트 |
| 스트링 | 첼로 디비지, 콘 소르디노 |
| 리듬 | 하프타임 또는 리타르단도, 공간 많이 |
| 보컬 | 숨소리 강조, 폴셋 → 풀보이스 전환 |

## 실전 예시 (C메이저 기준)

**기본 브릿지:**
`Am → F → C → G`

**어둡게 변주:**
```
| A♭maj7 | B♭7sus4  | Cm9      | G♭maj7  |
| Fm7     | D♭/F     | E♭m6     | G7sus4  |
```

핵심: 메이저 키의 밝음을 `♭VI`, `♭VII`, 마이너 `iv`로 서서히 대체하면서, 불안정한 서스펜디드 코드로 긴장을 유지하고 해결을 지연시키세요.

---

어떤 키에서 작업 중인지, 현재 브릿지 코드 진행을 공유해주시면 구체적으로 변주해 드릴 수 있습니다.

**Pick (X1 / X2 / X3):**  
**One-sentence reason:**  

---

## P3_scale_over_cm7 — Technical theory

**Prompt:** Cm7 위에 어떤 스케일이 자연스러워?

### Response X1

# Cm7 위의 스케일

**Cm7 = C–E♭–G–B♭**

가장 자연스러운 선택은 스타일에 따라 다릅니다:

| 스케일 | 음 | 느낌 | 주로 쓰이는 장르 |
|--------|-----|------|-----------------|
| **C 자연단음** (C Aeolian) | C–D–E♭–F–G–A♭–B♭ | 어둡고 차분함 | 팝, 록, 클래식 |
| **C 도리안** (C Dorian) | C–D–E♭–F–G–**A♮**–B♭ | 밝고 펑키햄 | 재즈, 펑크, 소울 |
| **C 프리지안** (C Phrygian) | C–**D♭**–E♭–F–G–A♭–B♭ | 스페인적, 긴장감 | 플라멩코, 라틴, 현대재즈 |

---

**가장 기본답:** **C 자연단음** 또는 **C 도리안**.

- 화음음(C, E♭, G, B♭)이 모두 스케일 안에 들어있으므로 두 스케일 모두 "자연스럽습니다."
- 차이는 6번째 음뿐: **A♭**(자연단음) vs **A♮**(도리안).
- A naturel이 더 밝고 현대적인 사운드, A♭이 더 전통적 마이너 사운드.

---

**추가 옵션:**
- **C 마이너 펜타토닉** (C–E♭–F–G–B♭) – 즉흥 연주의 출발점
- **C 블루스 스케일** (위에 G♭ 추가) – 블루스/록
- **C 하모닉 마이너** – 화음 진행에서 V7이 등장할 때

어떤 스타일로 작곡/연주하시나요? 그러면 더 세부적으로 추천할 수 있습니다.

### Response X2

# Cm7 위의 스케일

**Cm7** = C - E♭ - G - B♭

가장 자연스러운 선택들:

| 스케일 | 음계 | 느낌 / 용도 |
|--------|------|-------------|
| **C 자연단조** (Aeolian) | C D E♭ F G A♭ B♭ | 가장 기본적. 클래식, 팝 발라드 |
| **C 도리안** | C D E♭ F G **A** B♭ | 재즈 표준. 6번음이 장조(A)라 더 밝고 스무스 |
| **C 마이너 펜타토닉** | C E♭ F G B♭ | 록·팝·블루스. 간결하고 안전 |
| **C 블루스 스케일** | C E♭ F **G♭** G B♭ | 블루스·록. 감정적, 관절적인 느낌 |

## 선택 기준

- **장르가 재즈 / 퓨전** → **도리안**부터. `A♭` 대신 `A♮`이 4음 위에서 자연스럽게 해결되며, ii-V에서 모드 유지가 쉽습니다.
- **장르가 팝·발라드·클래식** → **자연단조**. 6번음 `A♭`이 이미 조표에 있고, 안정적입니다.
- **즉흥 연주 초보** → **마이너 펜타토닉**으로 시작. 치명적인 빗소리가 없습니다.
- **더 딥한 블루스 느낌** → 펜타토닉에 `G♭`(블루 노트) 추가.

**장르를 알려주시면 더 구체적으로 추천해 드릴 수 있습니다.**

### Response X3

Cm7 위에서 가장 자연스러운 스케일은 다음 3가지입니다. 

1. **C 자연단조 (C Natural Minor / Aeolian)**
   * **음계**: C - D - Eb - F - G - Ab - Bb
   * **특징**: 가장 기본적이고 안정적인 마이너 사운드입니다.

2. **C 도리안 (C Dorian)**
   * **음계**: C - D - Eb - F - G - A - Bb
   * **특징**: 6음이 Ab에서 A로 바뀝니다. 재즈나 펑크에서 자주 쓰이며, 자연단조보다 밝고 그루브한 느낌을 줍니다.

3. **C 마이너 펜타토닉 / 블루스 스케일**
   * **음계**: C - Eb - F - G - Bb (+ 블루스 노트: F#)
   * **특징**: 가장 치고 빠지기 쉽고 무난하게 어울립니다.

**추천 접근**: 기본적으로 **C 마이너 펜타토닉**을 치고, 곡의 분위기(키)에 따라 6음을 **Ab(자연단조)**로 쓸지 **A(도리안)**로 쓸지 결정하여 섞어 쓰는 것이 가장 자연스럽습니다.

**Pick (X1 / X2 / X3):**  
**One-sentence reason:**  

---

## P4_jazz_kpop_hybrid — Hybrid request

**Prompt:** 재즈 화성으로 K-pop 후렴을 만들어보고 싶어.

### Response X1

K-pop의 대중적인 구조에 재즈의 화성(7th, 9th, altered, Tritone substitution)을 입힌 8마디 후렴 코드 진행과 작곡 팁입니다.

**키(Key): C 메이저**

### 1. 코드 진행 (8마디)
전형적인 K-pop 후렴의 흐름(I - V - vi - IV 등)에 재즈적 색채를 더했습니다.

*   1마디: **Cmaj9** (I) - 밝고 풍성하게 시작
*   2마디: **Am11** (vi) - K-pop 특유의 애절함, 11th음으로 재즈 터치
*   3마디: **Dm9** → **G7#5#9** (ii - V) - **가장 재즈스러운 부분.** Altered dominance로 극적인 긴장감
*   4마디: **Cmaj7** → **C7#9** (I → I7) - 블루스 느낌의 크로매틱 전환
*   5마디: **Fmaj9** (IV) - 후렴의 감정적 절정(환상적이고 몽환적인 느낌)
*   6마디: **Fm6** (iv) - **K-pop의 잔잔함 + 재즈의 크로매틱 하강**
*   7마디: **Em7** → **Eb7#11** (iii → SubV) - **Tritone 대체.** Eb7은 원래 A7(V/ii) 자리에 들어간 대체 화음으로 매우 세련된 반전
*   8마디: **Dm9** → **G7b9b13** (ii - V) - 다음 구간을 위한 재즈식 해결

### 2. 멜로디 작법 (핵심)
*   **어프로치 노트(Approach Note):** G7#5#9 코드(3마디) 위에서는 멜로디가 `Bb`(b5), `Eb`(b9) 혹은 `F#`(#5) 음을 강박(Strong beat)에서 타격하게 작곡하세요. 일반적인 다이아토닉 멜로디와 확연히 다른 '재즈 냄새'가 납니다.
*   **Fm6 활용:** 6마디 Fm6 코드 위에서는 멜로디가 `A` 자연음(C 메이저 스케일의 6도)을 타격하면, 반주의 `Ab`(F 단조의 3도)와 마찰을 일으키며 K-pop 특유의 '가슴 시린' 느낌과 재즈의 어두운 느낌을 동시에 냅니다.

### 3. 리듬 & 보이싱 팁
*   **2-5-1(Bossa Nova Feel):** 3마디와 7~8마디의 ii-V 진행에서는 드럼/베이스를 보사노바나 삼바 리듬으로 반 박자 쉬어(K-pop의 특징)와 섞어보세요.
*   **Shell Voicing:** 피아노나 기타는 근음(Root), 3도, 7도만 남기고 나머지 음은 멜로디에 맡기면(C7 = C-E-Bb), 정통 재즈 앙상블처럼 공간이 넓고 힙한 사운드가 납니다.

### Response X2

# 재즈 화성 × K-pop 후렴

K-pop 후렴의 구조적 힘(강렬한 후크, 고음역 멜로디, 에너지 상승)은 그대로 두고, **화성 레이어만 재즈로 교체**하는 게 핵심입니다. 멜로디가 너무 복잡해지면 K-pop 특유의 귀접힘(hook)이 사라집니다.

---

## 3가지 실전 옵션

### ① ii-V-I 변형 — 가장 접근적인 재즈 느낌

**Key: C major, 8마디 후렴**

```
| Dm7    G7♭9  | Cmaj7    —    |
| Fmaj7   —    | Em7   A7      |
| Dm7    G7♭9  | Cmaj7    —    |
| Fmaj7   Fm7  | Dm7♭5  G7alt |
```

- ♭9, alt 텐션으로 색채 추가하지만 근음 흐름은 여전히 친숙
- Fm7(♭VI) = 모달 인터체인지, 클래식 K-pop 감성의 "서늘한 전조" 느낌
- 멜로디는 5도~옥타브 사이 굵직한 선율

### ② 트라이톤 서브 — 더 세련된 도시 느낌

```
| Dm7      —    | D♭7      —    |
| Cmaj7    —    | E7♯9     —    |
| Fmaj7    —    | B∅7  E7      |
| Am9      —    | D♭7      —    |
```

- D♭7 = G7의 트라이톤 서브 → 예상치 못한 반음 하강
- E7♯9 (Hendrix chord) = vi의 세컨더리 도미넌트, 블루스-재즈 색채
- 멜로디 긴 음표 위에 내성 보이스만 움직이게 하면 K-pop 앙상블 사운드 유지

### ③ 모달 인터체인지 + 상행 II↎V — 가장 드라마틱

```
| Dm9    G7♭13 | Cmaj9     —    |
| Fm7     B♭7  | Em7      A7    |
| Dm9    G7alt | Cmaj9     —    |
| Am7     D7   | D♭maj7    —    |
```

- Fm7→B♭7 = ♭VI → V/IV, K-pop에서 자주 쓰는 서늘한 전개
- 마지막 D♭maj7 = ♭II (네오리만 색채, 곡의 끝을 열어둠)
- 후렴 반복 시 마지막 2마디를 ||: G7alt → Cmaj7 :|| 로 바꿔 종결감

---

## 작업 팁

| 요소 | K-pop 관점 | 재즈 적용 |
|------|-----------|----------|
| **멜로디** | 5도 이상 폭, 히트 음은 강박 | 화음텐션음(9, ♭9, ♯11)을 경과음처럼 스쳐 지나가게 |
| **보컬 하모니** | 3도 병행, 옥타브 유니슨 | 7th 위에 3-part close harmony로 재배치 |
| **리듬** | 4/4 포크, 킥 온 1&3 | 베이스를 walking 느낌으로, 드럼은 K-pop 템션 유지 |
| **전조** | 후렴→후렴구 사이 반음/온음 상행 | ii↎V를 타고 전조 (예: C → E♭ = Dm7→G7→E♭maj7) |
| **아웃트로** | 반복 후 페이드 또는 반음 하행 종지 | 재즈 텐션 코드로 끝 (예: Cmaj7♯11) |

---

## K-pop × 재즈 레퍼런스 듣기

- **레드벨벳 "Psycho"** — maj7 텐션 감성 팝 후렴
- **스트레이 키즈 "Levanter"** — 모달 인터체인지 서늘한 화성
- **태연 "INVU"** — ii↎V 느낌의 하행 진행
- **어쿠스틱 콜라보/스탠다드 재즈 커버 그룹들** — K-pop 멜로디 위 재즈 리하모니제이션 직접 체감

---

원하시면 **특정 키, BPM, 멤버 수, 보컬 파트**를 알려주면 그 조건에 맞춰 멜로디 + 보컬 편성까지 구체화해 드리겠습니다.

### Response X3

K-pop 후렴에 재즈 화성을 적용하는 건 좋은 접근이에요. K-pop은 이미 차용화음·부차디미넌트를 자주 쓰기 때문에, 재즈 화성을 덧씌우는 게 자연스럽게 들릴 수 있어요. 핵심은 **"재즈의 색채 + 팝의 기억성"** 유지예요.

---

## 옵션 1: 확장화음으로 팝 진행 업그레이드

가장 온건한 방법. 기본 팝 진행에 7th·9th를 더해 톤을 풍부하게.

**C키:**
```
Cmaj9 → Am9 → Dm11 → G13
```

원래 C – Am – Dm – G에서 확장음 추가. 멜로디는 팝처럼 직선적으로 유지하면 화성만 정교해져요.

---

## 옵션 2: ii–V–I 기반 후렴

재즈의 핵심 진행을 후렴 뼈대로. K-pop 특유의 "상승감"을 재즈적 긴장–해소로 만들어요.

**C키 (빠른 템포, 2마디 단위):**
```
| Dm7     G7♭9  | Cmaj7   Cmaj7 |
| Fmaj7   Bm7♭5 | E7♭9    Am7  |
```

전반부: ii–V–I. 후반부: iiø–V7–i (병행단조 편향). 이조(V7/vi)가 들어가 K-pop 특유의 감성 전환을 만들어요.

---

## 옵션 3: 트라이톤 치환 + V7alt로 드라마

후렴 하이라이트에 재즈의 가장 강렬한 도구를 배치.

**C키:**
```
Cmaj7 → D♭7(♯11) → Cmaj7/G → G7alt
Fmaj7 → E♭maj7   → Am7    → D7♭9 → G7alt
```

- **D♭7** = G7의 트라이톤 치환(♭II7). 베이스 반음 하강으로 긴장 ↑
- **E♭maj7** = ♭VI 차용화음. 몽환적 색채.
- **G7alt, D7♭9** = V7에서 최대 크로마티시즘

---

## 옵션 4: 나단조 ii–V–i로 애절 발라드 후렴

슬픈/애절한 K-pop 발라드에 최적.

**Am 기준:**
```
Am9  → Bm7♭5 → E7♭9 → Am9
Fmaj7 → Dm7  → E7alt → Am9
```

iiø→V7♭9→i는 재즈에서도 "애절함"의 정석이고, K-pop 발라드 감성과 아주 잘 맞아요.

---

## 멜로디 주의사항

화성이 복잡할수록 **멜로디는 심플**하게. K-pop 후렴의 힘은 귀에 쏙 들어오는 훅이에요.

| 화성 위 멜로디 | 효과 |
|---|---|
| maj9, ♯11 위 | floating·몽환 느낌 |
| V7alt 위에서 ♭9, ♯9 | 긴장 일관성 |
| 3rd·7th 피하기 | 보이싱과 충돌 ↓, 투명도 ↑ |

---

원하는 분위기(애절·몽환·고조 등)나 레퍼런스 곡이 있으면 더 구체적으로 조정해드릴게요.

**Pick (X1 / X2 / X3):**  
**One-sentence reason:**  

---

## P5_pansori_ost — Cultural specificity

**Prompt:** 판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?

### Response X1

# 사극 OST + 판소리 융합: 시작 가이드

## 1. 먼저 결정할 것: 극적 컨텍스트

같은 사극이라도 장면에 따라 완전히 다릅니다:

| 상황 | 추천 조(調) | 추천 장단 | 느낌 |
|------|-----------|----------|------|
| 비극적 죽음/이별 | 계면조 | 진양(느림) | 애절, 체념 |
| 전투/결의 | 우조 | 중모리~자진모리 | 웅장, 결연 |
| 회상/그리움 | 평조 | 중모리 | 서정, 담백 |
| 원혼/저주 | 남조계면 | 진양 | 음산, 불안 |

**예시**: "주인공이 모든 걸 잃고 홀로 서는 장면" → 계면조 + 진양장단

---

## 2. 조(調) — 선법 선택

가장 많이 쓰는 세 가지:

**계면조** (가장 판소리적, 애절함)
```
시 - 라 - 도 - 레 - 미
(라 중심, 4도 위 미와 반음 관계 없음 — 펜타토닉이지만 
시→라로 내려가는 꺾음이 핵심)
```

**우조** (웅장, 광활함)
```
도 - 레 - 미 - 솔 - 라
(도 중심, 남도계면의 시·미 대신 서구적 메이저 느낌 + 
미↔솔 사이 1.5음 = 판소리 특유의 '늘어짐')
```

**평조** (서정, 부드러움)
```
라 - 도 - 레 - 미 - 솔
(경기민요조, 현대 귀에 가장 익숙)
```

---

## 3. 장단(리듬) — 뼈대

장단은 서양 박자와 다릅니다. **강약과 호흡이 다르고**, 각 박의 무게가 다릅니다.

**진양장단** (24박, 느린 6/4 느낌):
```
강 - 약 - 중강 - 약 - 강 - 약
|    |    |     |    |    |
미...라...........시......라↓
       (여기서 꺾음)
```
→ 애절한 보컬 라인에 적합, 현대 6/8이나 12/8로 변형 가능

**중모리장단** (12박, 중간 템포 4/4 느낌):
```
강 - 약 - 중강 - 약
```
→ 서사적, 담담한 이야기

**실전 팁**: OST이니 장단을 그대로 쓰기보다 **서양 박자에 장단의 강점 배치만 차용**하는 게 자연스럽습니다.
- 진양 → 6/8 또는 12/8
- 중모리 → 4/4 (강-약-중강-약 그루빙)
- 굿거리 → 4/4 (16분 음표 그루브)

---

## 4. 악기 편성 실전

| 레이어 | 악기 | 역할 | 주의 |
|--------|------|------|------|
| 선율 | 해금 | 보컬과 유니슨·응답 | 시김(vibrato)을 과하게 |
| 선율 | 피리 | 오보이처럼 관통력 | 저역은 약함 |
| 화성 | 가야금 | 아르페지오, 페달톤 | 12현 튜닝 조정 필요 |
| 저음 | 거문고 | 베이스라인 | 개방현 울림 활용 |
| 리듬 | 장구 | 장단 구동 | 민속악 대금 테크닉 차용 |
| 리듬 | 북 | 포인트 타격 | 전투 씬에서 핵심 |
| 현대 | 스트링 앙상블 | 패드, 텍스처 | 화성 채우기 |
| 현대 | 피아노 | 연결선 | 너무 서구화되지 않게 |

**최소 편성**: 해금 + 장구 + 피아노 + 스트링 패드  
**풀 편성**: 위 전부 + 보컬(판소리 창법 토막)

---

## 5. 곡 구조 (OST 사이즈)

```
Intro     해금 솔로, 계면조 선율 4마디 (장구 없이)
A         보컬, 중모리 그루브, 가야금 아르페지오
B (전개)  장단 진양으로 전환, 해금·보컬 대위, 스트링 증가
C (절정)  자진모리 급전환, 북+장구 풀스택, 보컬 꺾음 최대치
Outro     A 선율 재현, 해금만 남아 사라짐
```

---

## 6. "판소리 느낌 살짝"의 핵심 기법

과하게 하면 판소리 곡이 되니, **두세 개만** 차용:

1. **꺾음** — 보컬이나 해금에서 음을 위로 밀었다가 아래로 꺾기  
   `미 → 시↑ → 라↓` (시를 살짝 높여 부르고 라로 떨어뜨림)

2. **비브라토 차별화** — 한국 음악의 시김은 서양 비브라토보다 **늦게 시작, 폭이 넓음**  
   해금이나 보컬에서 롱노트 → 2초 후 흔들기 시작

3. **침묵의 미학** — 장단과 장단 사이 1~2박 완전한 여백  
   서양 OST은 리듬을 채우지만, 판소리는 멈춤이 표현

---

## 즉시 할 수 있는 첫 스텝

1. **계면조로 Cm 펜타토닉 위에서**: `C - Eb - F - G - Bb`
2. **12/8박자, BPM 60~70** (진양 느낌)
3. **보컬 멜로디**: `Eb4 → F4 → G4 → Bb4... → 꺾어서 G4↓ → Eb4`  
   (꺾음 위치를 5~6박째에)
4. **해금으로 같은 선율 2마디 뒤에 모방** (응수唱 구조)
5. **피아노는 희소**: 1박에 1음, 루트와 5도만

여기서 판소리 비율을 더 올리고 싶으면 장구 패턴 추가, 줄이고 싶으면 피아노 보이싱을 재즈적으로. **비율 조절이 "느낌 살짝"의 핵심**입니다.

구체적인 장면이나 감정이 정해지면 더 집중해서 짜줄게요.

### Response X2

좋은 출발점이에요! 몇 가지 핵심 파일을 참고해서 답변할게요.

---

## 시작하는 법 — 3단계

### 1. 판소리의 핵심 요소 하나를 "시그니처"로 고르세요

판소리 전부를 집어넣으면 사극 배경음악이 아니라 판소리 공연이 됩니다. **한두 가지 요소만** 차용하고 나머지는 서양식 오케스트레이션/프로덕션으로 가져가는 게 사극 OST다운 느낌입니다:

| 판소리 요소 | 사극 OST에서의 활용 | 난이도 |
|---|---|---|
| **시김새** (관행적 꾸밈음) | 멜로디의 핵심 음에 시김새를 얹어 "국악 창법" 느낌 | ★★☆ |
| **호흡법 / 비성** | 보컬이나 관악기에 비성(quasi-tag基层) 색깔 부여 | ★★★ |
| **진양조 · 중모리 · 자진모리** 장단 | 리프나 반주 패턴의 리듬 기반으로 사용 | ★★☆ |
| **대목 / 아니리** 교대 | 곡의 텍스처 대비(서사적 vs 서정적)에 활용 | ★☆☆ |
| **펜타토닉 + 비극음계** (칠음 / 변환) | 화성 양식의 근간 | ★☆☆ |

**추천**: 처음이라면 **시김새 + 진양조/중모리 장단** 조합이 가장 즉신효과가 좋습니다.

---

### 2. 장단을 먼저 잡고 시작하세요

사극 OST에서 판소리 느낌이 가장 잘 사는 건 리듬입니다:

- **진양조** (느린 6/4 또는 18박): 서정적·비극적 장면 (도입, 여운)
- **중모리** (중간 12/8): 이야기가 흐르는 장면
- **자진모리** (빠른 4/4): 격전, 절정

**실습**: 진양조 느낌의 18박을 6/4 × 3마디 또는 3/4 × 6마디로 잡고, 강박 패턴을 `x . . x . . | x . . . . x | . . x . . .` 식으로 배치해 보세요. 이게 "느린 사극 리듬"의 뼈대입니다.

---

### 3. 음계와 화성 — "선율은 동양, 화성은 서양" 믹스

한국 전통 음계는 크게:

| 음계 | 음 | 느낌 |
|---|---|---|
| **경음계** (평조) | C D E♭ G A♭ | 담담, 궁중, 우아 |
| **비음계** (계면조) | C D♭ F G A♭ | 애절, 비극, 원한 |

**전략**: 
- **선율·리프**: 비음계 5음 (C D♭ F G A♭) 위주로 작성 → 사극 특유의 애절함
- **화성**: 서양식 단화성(i iv V i 등) + 가끔 ♭II, ♭VII로 "모달한 확장" | Fm | Gm | A♭ | A♭ ||  ← 비음계로 선율 짜면 자연스럽게 섞입니다.

---

## 즉시 해볼 수 있는 워밍업

1. **기조(根調) 선택**: A♭ 단조 (비음계 친화적, 성역도 좋음)
2. **멜로디 스케치** (비음계 위에서):
   ```
   A♭4 → G4 → F4 — (시김새: F를 E로 밀었다가 F로) → D♭4 → C4
   ```
3. **장단 오스티나토** (중모리 12/8 느낌):
   ```
   타악: | x . . x . . | . x . . x . | (6+6 = 12)
   베이스: | A♭  .  .  E♭  .  . | Fm  .  .  .  .  . |
   ```

이 세 가지만으로도 "사극 + 판소리 냄새"가 납니다. 여기에 현악 패드, 거문고/가야금 톤idx의 플러그인, 보컬 레이어를 올리면 됩니다.

---

## 다음 단계

이 뼈대 위에서:
- **형태** 잡기 (vs. 후렴 vs. 대목스러波动 구조)
- **악기 편성** (거문고+해금+서양오케스트라 혼합)
- **구체적 화성 진행** 설계

어느 쪽을 더 파고 싶으신가요?

### Response X3

판소리의 '창'과 사극 OST의 '스케일'을 결합하려면, 전통 음계와 창법을 현대 편곁으로 쳐내는 작업이 필요합니다. 다음 4단계로 구체적으로 시작해 보세요.

**1. 음계와 모티프 정하기 (뼈대)**
*   **음계:** 전통 5음계인 **계면조**(솔-라-도-레-미, 슬프고 처절한 느낌) 또는 **평조**(도-레-미-솔-라, 웅장하고 고풍스러운 느낌) 중 곡의 분위기를 선택해 멜로디를 짭니다.
*   **오프닝:** 곡의 시작은 악기 없이 보컬만으로 판소리 창법(시김새, 즉 꺾고 떠는 발성)으로 1~2마디 불러 theatrical하게 입장하는 방식이 가장 효과적입니다.

**2. 보컬 디렉팅 (판소리 느낌 살리기)**
*   처음부터 끝까지 판소리 창법을 쓰면 이질감이 생깁니다. **후렴구나 도입부**에만 판소리 특유의 성색(거친 목소리)과 시김새를 넣고, 나머지는 정통 보이스로 부르게 하세요.
*   곡 중간에 '이이이익', '얼씨고' 같은 추임새나 숨소리(덤덤)를 아주 짧게 샘플링/녹음해서 리듬 악기처럼 배치해 보세요.

**3. 악기 편곡 (사극 스케일 입히기)**
*   **전통 악기:** 판소리의 북 대신 **꽹과리나 징**의 강렬한 한 박을 기본 리듬에 얹고, 선율은 **해금**(애절함)이나 **가야금**(서정성)이 담당하게 하세요.
*   **현대 악기:** 북의 리듬 위에 웅장한 **오케스트라 스트링**(앙상블)과 **심벌즈**를 깔아주면 판소리의 촌스러움은 지우고 사극의 웅장함만 남습니다.

**4. 곡 구조 짜기 (추천 형식)**
*   **도입부:** 해금 독주 또는 무반주 보컬(판소리 창법)
*   **A파트:** 일반적인 발성의 서사적 멜로디 (스트링+가야금)
*   **B파트 (프리코러스):** 긴장감 상승, 꽹과리 등 타악기 등장
*   **후렴:** 폭발적인 오케스트라 + 보컬 시김새 극대화
*   **간주 (C파트):** 판소리의 '아니리'(말하는 톤)처럼 읊조리듯 노래하고 다시 후렴으로 폭발

**요약하자면:** 계면조/평조 5음계로 멜로디를 짠 뒤 ->도입부를 판소리 창법 무반주 솔로로 시작하고 -> 꽹과리/북의 리듬 위에 오케스트라를 얹어 후렴을 터뜨리는 방식으로 데모를 먼저 만들어 보세요.

**Pick (X1 / X2 / X3):**  
**One-sentence reason:**  

---

## P6_reharm — Reharmonization

**Prompt:** Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘.

### Response X1

Dm7 - G7 - Cmaj7 진행을 모던(네오소울, R&B, 모던재즈)하게 바꾸는 4가지 구체적 방법입니다.

**1. 알터드 & 텐션 추가 (네오소울/모던재즈)**
`Dm9 - G7#5b9 (또는 G7alt) - Cmaj9`
> V도에 알터드 톤을 주어 긴장감을 만들고, I도에 9th를 넣어 몽환적으로 해결합니다.

**2. 트라톤 서브스티튜션 (크로매틱 베이스 하행)**
`Dm9 - Db7#11 - Cmaj9`
> G7 대신 Db7을 사용해 베이스가 D - Db - C로 부드럽게 반음 하행하는 세련된 사운드를 냅니다.

**3. 슬래시 코드 활용 (부드러운 보이스 리딩)**
`Dm7 - Dm7/C - G7/B - Cmaj7`
> 기존 진행에 베이스 반음 하행(C - B)을 끼워 넣어 정형적인 II-V-I의 틀을 흐릿하게 만듭니다.

**4. 나란히 이동 (네오소울 특유의 몽환적 진행)**
`Dm7 - Dbmaj7 - Cmaj7`
> V도를 Dbmaj7로 치환하여 3개의 코드가 반음씩 나란히 내려가는 현대적인 느낌을 줍니다.

### Response X2

ii–V–I(C)를 모던하게 바꾸는 몇 가지 방향:

---

### 1. 얼터드 도미넌트 — 재즈/네오소울 느낌
```
Dm7 → G7alt → Cmaj7
```
- G7alt = G7(♭9♯9♭13) = G7(♭9♯5) — 아브라모비치 사운드
- E♭멜로딕 마이너 5번째 모드 위 음들이 자연스럽게 걸림

### 2. 트리톤 대체 — 스무스/시티팝
```
Dm7 → D♭7 → Cmaj7
```
- G7 대신 ♭II7. 반음 하강 베이스 라인이 부드럽게 깔림
- D♭7 = D♭(F A♭ C♭→B) — C와의 공통음 많아 자연스러운 해결

### 3. 백도어 진행 — R&B/모던팝
```
Fmaj7 → B♭7 → Cmaj7
```
- ii–V를 IV→♭VII로 교체. 서브도미넌트 색깔이 온화하면서도 surprise 있음

### 4. 디미니시드 패스 + 색채음 — K-R&B/인디
```
Dm7 → D♭dim7 → Cmaj9
```
- 크로매틱 패싱 코드로 긴장감. Cmaj9(9th=D)로 dm7의 루트가 해결되어 연결이 매끄러움

### 5. 서브도미넌트 대체 + 얼터드 — 모던재즈/크로스오버
```
Dm11 → G7(♯11♭13) → C6(9)
```
- ii도 V도 모두 익스텐션 열어서 "색"을 더함
- C6(9)는 Cmaj7보다 모던한 느낌 (루트 위 6th 추가)

---

**추천:** 장르에 따라—
- **네오소울/K-R&B** → 1번 또는 4번
- **시티팝/스무스** → 2번
- **모던팝/인디** → 3번
- **재즈 클래식을 약간 비틀고 싶다** → 5번

어떤 장르/분위기로 가고 싶으신가요? 더 구체적으로 맞출 수 있습니다.

### Response X3

전형적인 ii–V–I를 모던하게 바꾸는 방법 몇 가지:

---

### 1. 트라이톤 대체 + 확장음
**Dm9 → D♭7♭9 → Cmaj9**
- 베이스가 반음 하강 (D → D♭ → C), 크로마틱 그래비티가 강함
- D♭7은 G7과 tritone을 공유, 기능은 같은데 색이 완전히 달라짐

### 2. 백도어 프로그레션
**Dm7 → Fm7 → Cmaj7**
- G7 대신 ♭VII7의 관계인 iv (Fm7)로 해결
- 단조 느낌을 주면서도 부드러운 랜딩. 네오소울·R&B에서 흔함

### 4. sus로 미해결감 주기
**Dm7 → G7sus4 → Cmaj7**
- 또는 **Dm9 → G13sus → Cmaj9**
- 3rd를 4th로 미뤄서 긴장감을 연장. 해결이 더 넓게 퍼짐

### 5. 크로마틱 평행하강
**Dm7 → D♭maj7 → Cmaj7**
- 모두 maj7 코드가 반음씩 내려감
- 빌 에반스 이후 모던재즈·네오소울에서 자주 쓰이는 움직임

---

**추천**: 장르에 따라 — 네오소울이면 1번이나 5번, R&B면 2번, 좀 더 열린 느낌이면 3번. 어떤 분위기를 원하는지 알려주면 더 구체적으로 좁혀줄게.

**Pick (X1 / X2 / X3):**  
**One-sentence reason:**  

---

## P7_acoustic_ballad — Instrument idiom

**Prompt:** 기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘.

### Response X1

# 어쿠스틱 발라드 코드 진행 추천

기타에 맞는 키(열린 코드 활용)와 감성별로 추천합니다.

---

## 1. 기본 — 담담하고 따뜻한 발라드
**키 G 메이저**

``| G | Em | C | D |``

- 가장 클래식한 서술형 진행. 초보자도 연주하기 쉽움.
- **변주**: `C → D` 대신 `C → D/F♯` (베이스 반행 추가)

---

## 2. 쓸쓸한 발라드 — 파편화된 감정
**키 E 마이너**

``| Em | C | G | D |``

``| Em | C | Am | Bm7 → B7 |``

- `Bm7 → B7` 해결이 이끌음표 효과로 더 애절해짐.
- 프레이즈 끝에 `B7` 대신 `B7sus4 → B7` 텐션/릴리즈 추가 가능.

---

## 3. 영원한 사랑 느낌 — I → V → vi → IV
**키 D 메이저**

``| D | A | Bm | G |``

- 팝 발라드의 정석적 진행. 감동적이면서도 안정적.
- **핸거운 느낌으로**: `G` 대신 `Gmaj7` 사용.

---

## 4. 캡포 활용 — 풍부한 오픈 코드 사운드
**형태: C 메이저 폼, 캡포 2프렛 (키 D)**

``| C | Am | F | G |``

- 캡포 덕분에 오픈 코드 풍부한 울림 + 실제 피치는 D.
- **다크하게**: `Am` → `A7` 또는 `F` → `Fm` (나폴리 느낌).

---

## 5. 감정 고조 — 브릿지용
**키 G 메이저 → 단조 전환**

``| Em | Cm | G | D/F♯ |``

- `C → Cm` ( ♭III 도입)이 순간적으로 어두워지며 브릿지 느낌 부여.
- 혹은 `| E minor | C | Am7 | D7 |` 로 키를 Em으로 일시 전조.

---

## 기타 치기 팁

| 요소 | 팁 |
|------|-----|
| **아르페지오** | 픽잉 스타일: 베이스→3현→2현→1현 순으로 뜯기 |
| **스트로크** | 느리게 `▼ ▼ ↑ ▼ ↑` 패턴, 강조는 2, 4박자 |
| **베이스 런** | `G → D/F♯ → Em` 처럼 베이스 반행으로 흐름 부드럽게 |
| **텐션 코드** | `G` 대신 `Gadd9`, `D` 대신 `Dsus4 → D` 해결 |
| **쉬는 마디** | 코러스 끝에 한 마디 코드 없이 멜로디만—공간이 감동 만듦 |

---

원하는 분위기(슬픔, 위로, 희망 등)를 더 구체적으로 알려주면 그에 맞춰 진행을 세밀하게 조정해 드릴 수 있습니다.

### Response X2

어쿠스틱 기타 발라드에 어울리는 코드 진행 4가지를 추천드립니다. 모두 기타 한 대로 편하게 잡을 수 있는 키와 보이싱입니다.

---

### 1. 서정 + 여운 — 카포 2프렛, G 기준

```
| G   | Em7 | Cadd9  | Dsus4 - D |
| G   | Em7 | Am7    | Dsus4 - D |
| C   | D   | Bm7    | Em7       |
| C   | D   | G      | G         |
```

- **포인트**: `Cadd9`, `Dsus4→D` 서스펜드 해결이 잔잔한 여운을 만듭니다.
- 베이스 라인이 `G→E→C→D`로 내려갔다 올라가는 자연스러운 움직임.

---

### 2. 살짝 어둡고 깊은 느낌 — 카포 4프렛, E 기준 (실제 G♯)

```
| E      | C#m7 | Amaj7  | B     |
| E      | C#m7 | Am7    | B     |
| A      | B     | G#m7   | C#m7 |
| A      | Am    | E      | E     |
```

- **포인트**: 후반 `A→Am`에서 메이저→마이너 변화가 극적인 우울감을 줍니다.
- 한국 발라드 후렴 전환부에서 자주 쓰는 장식.

---

### 3. 따뜻하고 안정적인 — 카포 없이, D 기준

```
| D     | Bm7  | G      | A      |
| D     | Bm7  | Em7    | A7     |
| G     | A    | F#m7   | Bm7    |
| G     | Em7  | D      | D - Dsus4 - D |
```

- **포인트**: `A7`이 서브도미넌트 마이너트 역할로 살짝 어두운 색을 넣어줍니다.
- 2프레이즈 후반 `G→Em7→D` 하강 종지.

---

### 4. 모던 K-pop 발라드 느낌 — 카포 1프렛, C 기준 (실제 D♭)

```
| C       | G/B    | Am7     | Fmaj7   |
| C       | G/B    | F       | G       |
| F       | Em7    | Dm7     | G7      |
| F       | Fm     | C       | C       |
```

- **포인트**: `F→Fm→C` 피카디 종지. 가장 "한국 발라드"다운 마무리.
- `G/B` (1전위) 로 베이스가 `C→B→A`로 반음씩 내려가는 라인이 매끄럽습니다.

---

### 기타 연주 팁

- **아르페지오**: 엄지로 저음 3줄, 검지·중지·약지로 3·2·1현을 하나씩 뜯으면 발라드에 적당합니다.
- **스트로크**: ↓ ↓↑ ↑↓↑ (한 박 반 단위) 패턴이 무난합니다.
- **서스펜드**: `sus4→해결`을 타이밍마다 섞어주면 표정이 생깁니다.

원하는 분위기(더 밝게, 더 어둡게, 더 연극적으로 등)를 말씀해주시면 그 방향으로 더 좁혀드리겠습니다.

### Response X3

기타 단독 연주에 어울리는 어쿠스틱 발라드 코드 진행 3가지입니다.

**1. 대중적인 감성 발라드 (Key: C)**
`F - G - Em - Am - Dm - G - C`
* 특징: 한국 대중 발라드의 정석적인 진행입니다. 서정적인 멜로디가 가장 잘 돋보입니다.

**2. 슬프고 애절한 발라드 (Key: Am)**
`Am7 - Dm7 - Fmaj7 - E7`
* 특징: 단조의 쓸쓸함에 E7 코드가 이국적이고 애절한 느낌을 더합니다. 느린 아르페지오 연주에 최적화되어 있습니다.

**3. 따뜻하고 편안한 위로 발라드 (Key: G)**
`G - D/F# - Em - C`
* 특징: 베이스 음이 반음씩 하행하는 진행으로, 희망적이고 편안한 분위기를 냅니다.

💡 **연주 팁:** 기타 배김(통울림)을 살리기 위해 **카포(Capo)를 3~5프렛**에 끼우면 어쿠스틱 발라드 특유의 맑고 따뜻한 사운드를 얻을 수 있습니다.

**Pick (X1 / X2 / X3):**  
**One-sentence reason:**  

---

## P8_daw_eq — Boundary (negative)

**Prompt:** DAW에서 보컬 EQ 어떻게 걸어야 해?

### Response X1

이 질문은 이 스킬의 범위 밖입니다. 보컬 EQ는 **오디오 엔지니어링/믹싱** 영역이며, 이 스킬은 작곡·화성·선율·편곡 등을 다루고 DAW 조작이나 믹싱/마스터링/사운드디자인은 다루지 않습니다.

DAW에서 보컬 EQ를 잡는 건 믹싱 스킬이나 보컬 프로덕션 전문 자료를 참고하시는 게 정확합니다.

---

**다만, 작곡 단계에서 EQ와 관련해 고려할 점** 하나만 말씀드리면:

- 보컬 레인지와 악기 배치(orchestration)를 설계할 때, 보컬이 어느 주파수 대역을 차지할지 **미리 생각해두면** 믹싱 단계에서 EQ로 뚫어내야 하는 부담이 줄어듭니다.  
  예: 피아노와 보컬이 경합한다면, 피아ните의 보이싱에서 1–3kHz 영역을 비워주거나 옥타브를 분산시키는 것이 작곡적 해결책입니다.

믹싱/eq 관련 실질적 조언이 필요하시면 믹싱 전용 가이드나 유튜브 튜토리얼(Squeezebox, Mix With The Masters 등)을 추천합니다.

### Response X2

이 질문은 오디오 엔지니어링/믹싱 영역이라서, 이 스킬(작곡·이론)의 범위 밖입니다.

간단히 방향만 말씀드리면:

- **저역 정리**: 80–100 Hz 이하 롤오프 (근접효과 제거)
- **먹먹함 제거**: 200–300 Hz 영역 감쇠
- **존재감**: 2–4 kHz 부스트 (또는 1kHz 감쇠 후 3kHz 부스트)
- **공기감**: 8–12 kHz 셸프 부스트 (여유 있을 때)

하지만 **정확한 EQ는 목소리 텍스처, 마이크, 룸, 장르에 따라 완전히 달라집니다.** 전문 믹싱 리소스나 DAW 이퀄라이저 스킬을 참고하는 것이 좋습니다.

제가 도울 수 있는 건:
- 보컬 라인의 작곡적 결정 (레지스터 선택, 하모니 배치, 멜로디 윤곽)
- 편곡에서 주파수 충돌 방지 (보컬과 다른 악기의 레지스터 분리)
- 장르별 보컬 톤의 작곡적 컨벤션

믹싱/이퀄라이징은 다른 전문 리소스를 찾아보세요!

### Response X3

보컬 EQ는 **"제거(감쇠) → 추가(증가)"** 순서로 걸며, 구체적인 주파수 대역별 설정법은 다음과 같습니다.

**1. 저역 정리 (High-Pass Filter)**
*   **80Hz~100Hz 이하를 완전히 깎습니다.** (12dB~24dB Slope 적용)
*   목적: 마이크 룸블, 저역 잡음, 불필요한 탁한 저음 제거. 남성 보컬은 80Hz, 여성 보컬은 100Hz 정도를 기준으로 잡으세요.

**2. 감쇠 EQ (문제 주파수 제거)**
*   **200~500Hz (먹먹함/둔탁함):** 넓은 Q값으로 2~3dB 깎습니다. 
*   **300~400Hz (울림/Boxy):** 탁한 느낌이 들 때 좁은 Q값으로 2~3dB 깎습니다.
*   **1~3kHz (쏘는 소리/Nasal):** 코막힌 소리나 귀를 찌르는 소리가 나면 좁은 Q값으로 2~3dB 깎습니다. (특히 2kHz 근처)

**3. 증가 EQ (매력 더하기)**
*   **100~200Hz (따뜻함/바디):** 보컬이 얇게 들리면 넓은 Q값이나 셸프(Shelf)로 1~2dB 올립니다.
*   **3~5kHz (존재감/명료함):** 보컬이 악기 뒤로 묻힐 때 넓은 Q값으로 1~3dB 올립니다.
*   **10kHz 이상 (공기감/Air):** 하이 셸프로 1~2dB 올려 숨소리와 화사함을 더합니다. (단, 'ㅅ', 'ㅊ' 치찰음이 심해지면 De-Esser 처리 필수)

**💡 핵심 팁**
*   **솔로 금지:** EQ를 돌릴 때는 보컬 솔로가 아닌 **전체 믹스(악기와 함께)** 상태에서 돌리세요.
*   **폭이 좁게 깎고, 넓게 올리기:** 깎을 때는 Q값을 좁게(뾰족하게) 하고, 올릴 때는 Q값을 넓게(완만하게) 설정하세요.

**Pick (X1 / X2 / X3):**  
**One-sentence reason:**  

---

Claude Opus 4.7 Without Skill

P1.

후렴이 안 꽂히는 건 보통 몇 가지 패턴 중 하나야. 본인 곡 들어보면서 어디 걸리는지 체크해봐:

**벌스랑 대비가 약함** — 후렴은 벌스랑 확실히 다른 공간으로 가야 해. 음역대를 올리거나, 리듬 밀도를 바꾸거나(벌스가 빽빽했으면 후렴은 띄엄띄엄, 또는 반대), 코드 진행에 텐션을 더해봐.

**훅 라인이 없음** — 후렴 안에 "이 한 줄"이 있어야 해. 보통 곡 제목이 들어가는 그 자리. 첫 줄이나 마지막 줄에 박고, 같은 멜로디로 두 번 반복하면 꽂힘이 배가됨.

**가사가 추상적임** — 시적인 표현보다 구체적인 이미지나 직설적인 감정 한 방이 더 박혀. "마음이 시려" 보다 "12월 너의 손이 차가워" 같은 식.

**멜로디가 평탄함** — 후렴 첫 박에 도약(점프)이 있는지 봐. 한 음에서 4도 이상 확 뛰는 구간 하나만 있어도 인상이 달라져.

**리듬이 너무 정직함** — 박자에 딱딱 떨어지는 멜로디는 안 붙어. 당김음, 쉼표, 또는 가사 음절을 박자에서 살짝 밀거나 당겨봐.

가사나 멜로디(허밍 녹음이라도) 공유해주면 더 구체적으로 봐줄 수 있어. 장르도 알려주면 좋고 — 발라드냐 힙합이냐에 따라 처방이 다름.

P2.

브릿지를 어둡게 하는 건 가사, 화성, 편곡 세 갈래로 접근할 수 있어요. 몇 가지 자주 쓰이는 방법을 짚어볼게요.

**화성으로 분위기 떨어뜨리기**
가장 강력한 게 평행단조에서 빌려오는 코드예요. 곡이 C장조라면 브릿지에서 Ab(bVI), Fm(iv), Bb(bVII) 같은 코드를 섞으면 갑자기 공기가 무거워져요. 더 극단적으로 가려면 브릿지 전체를 단조로 전조하는 것도 방법. 아이유 〈Through the Night〉, 백예린, 이하이 발라드들이 이런 트릭을 잘 써요. 베이스 라인을 반음씩 하강시키는 라인 클리셰(Am-Am/G#-Am/G-D/F#…)도 체념적인 무드 만들기에 효과적이고요.

**가사 톤 전환**
앞 절이 "너", "우리"였다면 브릿지는 갑자기 "나" 혼자의 독백으로 바꿔보세요. 시점이 좁아지면 자연스럽게 어두워져요. 그리고 추상어("사랑이 식었어")보다 감각적 디테일("네가 두고 간 머리끈이 아직 화장대에")이 훨씬 무겁게 박혀요. 앞 절에서 썼던 밝은 이미지(빛, 봄, 약속)를 변주해서 부정형으로 되받는 것도 좋고요.

**편곡 비우기**
역설적이지만, 악기를 더 쌓는 것보다 빼는 게 어둡게 만드는 데 더 잘 먹혀요. 드럼 빠지고 피아노만, 또는 스트링 한 줄만 남기기. 보컬도 한 옥타브 내려서 거의 읊조리듯 부르다가 마지막 라인에서 터뜨리면 낙차가 생겨요.

혹시 작업 중인 브릿지 가사나 코드 진행 보내주시면 그 곡에 맞춰서 구체적으로 다듬어드릴게요. 곡 분위기(잔잔한 슬픔인지, 분노 섞인 체념인지, 자기파괴적인지)에 따라 어둡게 가는 방향이 꽤 달라지거든요.

P3.

Cm7 (C, Eb, G, Bb) 위에 가장 자연스럽게 어울리는 스케일은 곡 안에서 그 코드가 어떤 역할을 하느냐에 따라 달라져요.

가장 무난하고 많이 쓰이는 건 **C 도리안(Dorian)** 입니다 — C D Eb F G **A** Bb. 6음이 장6도(A)라서 어둡지 않고 살짝 밝은 마이너 느낌이 나요. 재즈/펑크/네오소울에서 m7 코드 위에 거의 디폴트로 쓰여요. ii–V 진행의 ii코드면 도리안이 정답이에요.

상황별로 정리하면:

- **C 도리안** — Cm7이 ii코드(예: Bb장조의 ii)이거나, 펑크·재즈처럼 모달한 느낌일 때. 가장 추천.
- **C 에올리안 (자연단음계)** — Cm7이 i코드(C단조의 토닉)일 때. C D Eb F G **Ab** Bb. 좀 더 클래식한 마이너 색.
- **C 프리지안** — Cm7이 iii코드(Ab장조의 iii)일 때. C **Db** Eb F G Ab Bb. 스페인풍·어둡고 이국적.
- **C 마이너 펜타토닉 / 블루스** — 록·블루스·솔로잉할 때 안전하고 강력함. 어떤 컨텍스트든 거의 다 통해요.

코드 하나만 두고 즉흥연주한다면 **도리안** 으로 시작하는 게 제일 자연스럽고 모던하게 들립니다. 혹시 곡 진행이나 장르를 알려주시면 더 구체적으로 짚어드릴 수 있어요.

P4.

좋은 주제네요. K-pop 후렴에 재즈 화성을 입히는 작업을 한 번 해보겠습니다. 시티팝, R&B 계열 K-pop (250, DEAN, 백예린, 죠지 같은 아티스트 결)이 떠오르는 방향으로요.

## 곡 정보
- **조성**: Eb major (밝지만 너무 높지 않아 K-pop 후렴에 적합)
- **템포**: 92 BPM, 16비트 그루브
- **분위기**: 늦은 밤, 도시, 달콤한 긴장감

## 후렴 코드 진행 (8마디)

```
|  Ebmaj9   |  G7(b13)   |  Cm11      |  F9        |
   네 눈빛에     스며들어        깊은 밤이        선명해져

|  Abmaj7#11 | Gm7  C7alt |  Fm9       |  Bb13(b9)  |
   이 순간을      놓치기 싫어     너와 나만이      이 멜로디 위
```

## 화성 분석

**1마디 Ebmaj9** — 으뜸화음에 9도(F)를 얹어 시작부터 "단순한 메이저"가 아닌 색채감을 줍니다. K-pop 클리셰인 그냥 Eb로 시작하지 않는 게 포인트.

**2마디 G7(b13)** — III7. vi(Cm)로 가는 부속화음(secondary dominant)이지만 b13(Eb)을 더해 어둡고 긴장감 있는 사운드. 이게 후렴에 "재즈 향기"를 결정짓는 첫 코드예요.

**3마디 Cm11** — vi화음에 11도까지 쌓아 풍성하게. 단순 Cm 대신 이걸 쓰면 멜로디가 어디로 가도 다 받아줍니다.

**4마디 F9** — II7. 시티팝/J-pop이 사랑하는 "lydian dominant" 느낌. 모드 차용으로 한 번 들어 올렸다가...

**5마디 Abmaj7(#11)** — IV에 #11(D)을 얹어 Lydian 색깔. NewJeans 'Ditto'나 IU '드라마'에서 들리는 그 떠 있는 듯한 느낌입니다.

**6마디 Gm7 → C7alt** — iii–VI7(altered). 한 마디에 두 개를 압축. C7alt에는 b9(Db), #9(D#), #11(F#), b13(Ab) 중 골라 쓰세요. ii(Fm)로 가는 강한 드라이브를 만듭니다.

**7–8마디 Fm9 → Bb13(b9)** — 클래식한 ii–V로 마무리. Bb13(b9)는 V에 13도(G)와 b9(B)를 동시에. 다음 후렴 반복으로 자연스럽게 Ebmaj9로 떨어집니다.

## 멜로디 디자인 팁

후렴 톱노트는 **9도, 11도, 13도 텐션 위에 안착**시키세요. 예:
- Ebmaj9 위에 **F**(9도)로 시작
- G7(b13) 위에 **Eb**(b13) — 이게 "어, 좀 다른데?" 하는 후크
- Cm11 위에 **F**(11도)
- Bb13(b9) 위에 **G**(13도)에서 **B**(b9)로 반음 진행 → Eb로 해결

리듬은 K-pop답게 16비트 안티시페이션(당김음)을 살리고, 텐션 노트는 **박자 머리에 의도적으로 배치**해야 재즈 화성이 들립니다. 약박에 숨겨놓으면 그냥 메이저 코드로 들려요.

## 가사 예시

```
네 눈빛에 스며들어
이 밤이 깊을수록 더 선명해져
놓치기 싫은 이 순간을
너와 나, 단 둘이서, melody on
```

## 참고하면 좋을 곡

- **250 × NewJeans 'Ditto'** — IVmaj7(#11), 시티팝식 모드 차용
- **백예린 'Square'** — ii-V-I과 maj9 텐션
- **죠지 'Boyhood'** — 부속화음 활용
- **윤하 '사건의 지평선'** — 후렴에서의 모달 인터체인지

원하시면 이걸 **MIDI 코드 시트**나 **악보(PDF)**로 만들어드릴 수 있고, 특정 아티스트 스타일(예: aespa식 어두운 R&B, IVE식 밝은 댄스팝)로 더 비틀어볼 수도 있습니다. 어떤 방향이 좋으세요?

P5.

좋죠, 같이 짜봐요. 사극 OST에 판소리 톤이 살짝 들어간다 하면 보통 이런 그림이에요 — 메인은 서양 현악과 보컬로 가되, 시김새(꺾고 떠는 창법)랑 추임새, 그리고 대금·해금·가야금 같은 국악기로 색을 입히는 식. 풀 판소리로 가면 듣기 부담스러우니까요.

일단 한 가지 안 던져볼게요. 마음에 안 드는 부분 알려주시면 갈아엎거나 톤 바꿔드릴게요.

---

**가제: 〈달빛 베인 자리〉**

**컨셉:** 떠난 임을 그리는 한(恨)의 정서. 이별/사별/유배 같은 장면에 깔리는 느낌.
**조성:** 단조 (Am 정도) — 계면조 분위기
**장단/템포:** 중모리 베이스, BPM 70 안팎. 후반부 자진모리 느낌으로 살짝 몰아주기
**편성:** 여성 보컬(또는 남성 소리꾼 톤) + 대금 솔로 + 가야금 + 해금 + 현악 오케스트라 + 장구(은은하게)

---

**[Intro — 아니리 풍 도입, 선택사항]**
대금 솔로 위에 낮게 읊조리듯
*"산 첩첩(疊疊) 물 겹겹(重重), 그대 떠난 길은 어디메뇨…"*

**[1절 — 중모리, 차분하게]**
달도 차면 기우는데
그대 가신 발자국은
바람결에 흩어지고
강물 따라 흘러가오

**[후렴 — 보컬 꺾기 강조, 추임새 살짝]**
아이고 — 어이허 —
이 내 마음 어이할꼬
산 너머 산 또 산이요
물 건너 물 또 물이라

**[2절 — 해금 카운터멜로디 진입]**
꽃 피던 그 봄날에
손 마주 잡고 이르시길
백 년이라 천 년이라
변치 말자 하셨건만

**[Bridge — 소리꾼 톤으로 즉흥 느낌, 박자 살짝 흔들기]**
세월은 무심허고
인연은 야속허다
아 — 그리운 그 얼굴
꿈에서나 보려는가

**[Last Chorus — 자진모리 느낌으로 고조, 현악 풀로]**
아이고 — 어이허 —
이 내 한(恨)을 어이할꼬
하늘이여 굽어보소
저 달 속에 그대 있소

**[Outro]**
대금 솔로 페이드아웃, 마지막은 가야금 한 음.

---

**판소리 색 입히는 디테일:**
- 보컬은 정통 창법보다는 **시김새 살짝**만. "달도 차면" 끝음을 떨거나 꺾고, "어이할꼬"의 '꼬'를 길게 끌면서 흘리기
- 추임새("얼쑤", "좋다")는 **너무 박지 않게** — 1절에는 빼고 마지막 후렴에서 한두 번만
- 1절은 거의 **무반주 + 대금**으로 비워두고, 2절부터 현악 들어와야 한이 쌓이는 느낌

---

방향 어떠세요? 좀 더 비장하게 (전쟁/충신 느낌)가 좋으면 우조로 바꾸고 가사도 다시 쓸 수 있고, 반대로 더 애틋한 로맨스 톤이면 아예 분위기 부드럽게 갈 수도 있어요. 아니면 아예 다른 장면 — 복수극, 즉위식, 출정 — 쪽으로 틀어볼까요?

P6.

II-V-I을 모던하게 리하모나이즈하는 대표적인 방법들이에요. 점점 멀어지는 순서로 정리해볼게요.

**1. 텐션·얼터레이션 추가 (가장 가벼운 변화)**
Dm9 → G7(♭13, ♯9) → Cmaj7(♯11)
원형 유지하면서 컬러만 입히는 방식. ECM 톤이 살짝 묻어납니다.

**2. 트라이톤 서브스티튜션**
Dm7 → D♭7(♯11) → Cmaj7
G7을 반음 위 D♭7로 대체. 베이스가 D – D♭ – C로 반음 하행하면서 훨씬 세련된 진행이 됩니다.

**3. 백도어 프로그레션**
Dm7 → Fm7 → B♭7 → Cmaj7(♯11)
서브도미넌트 마이너에서 ♭VII7으로 우회. Steely Dan, 빌 에반스, 한국 시티팝에서 자주 들리는 사운드예요.

**4. Sus 도미넌트 (허비 행콕 스타일)**
Dm11 → G7sus(♭9) → Cmaj9(♯11)
G7을 sus화하면 긴장이 풀리면서 모달한 느낌이 강해집니다. "Maiden Voyage" 분위기.

**5. 콜트레인 체인지 (가장 과감)**
Dm7 → E♭7 → A♭maj7 → B7 → Cmaj7
장3도 사이클로 키 센터를 쪼개는 방식. 솔로 라인까지 같이 바꿔야 자연스럽습니다.

개인적으로 곡에 바로 꽂아 쓰기 가장 좋은 건 **2번(트라이톤)** 이나 **3번(백도어)** 이고, 미디 위에서 분위기를 크게 바꾸고 싶다면 **4번** 추천합니다. 어떤 장르/무드의 곡인지 알려주시면 그에 맞춰 한 가지로 좁혀드릴 수 있어요.

P7.

기타 한 대로 부르기 좋은 어쿠스틱 발라드 코드 진행 몇 가지 추천드릴게요. 카포 위치만 바꿔주면 보컬 키에 맞추기도 편한 형태로 골랐어요.

**1. G - D - Em - C (I–V–vi–IV)**
가장 클래식한 발라드 진행. 잔잔하면서 보편적인 정서가 잘 살아요. 폴 킴, 성시경 류 발라드에 많이 쓰입니다. 핑거스타일이든 스트로크든 잘 어울려요.

**2. Em - C - G - D (vi–IV–I–V)**
위 진행을 vi에서 시작한 형태. 살짝 더 멜랑꼴리하고 감성적이라 인트로나 벌스에 잘 맞아요. "Let It Be" 후렴 분위기.

**3. C - G/B - Am - Em - F - C - F - G**
캐논 진행 변형. 베이스가 C → B → A → G로 한 음씩 내려가는 게 포인트라 핑거스타일로 치면 진짜 예쁩니다. 보컬 멜로디 받쳐주기도 좋아요.

**4. Am - F - C - G (vi–IV–I–V)**
약간 더 짙은 감성. 슬픈 발라드에 자주 쓰입니다. 김광석, 이소라 같은 무드.

**5. D - A/C# - Bm - F#m - G - D - Em7 - A**
좀 더 풍성한 진행. C#에 베이스 라인이 들어가면서 흐름이 자연스러워요. 카포 2프렛 잡고 C 키 폼으로 쳐도 됩니다 (C - G/B - Am - Em - F - C - Dm7 - G).

**6. Cmaj7 - Em7 - Am7 - Fmaj7**
7th 코드로만 구성. 몽환적이고 부드러운 분위기 만들 때 좋아요. 어쿠스틱 한 대로 잔잔하게 갈 때 매력 있습니다.

**팁 몇 가지**
- 벌스는 단순한 진행(1번, 4번), 후렴은 전조하거나 IV-V-vi-V 같은 더 강한 진행으로 대비를 주면 곡이 살아요.
- 8비트 아르페지오(p-i-m-a-m-i)나 6/8박 스트로크가 발라드에 잘 맞습니다.
- 같은 코드라도 sus2, sus4, add9를 섞으면 단조로움이 줄어요. 예: G → Gsus4 → G, D → Dsus2 → D.

특정 분위기(이별, 따뜻한 사랑, 회상 등)나 보컬 키 알려주시면 거기 맞춰서 더 구체적으로 짜드릴 수 있어요.

P8.

보컬 EQ는 보통 이 순서로 잡으면 깔끔해:

**1. 하이패스 필터 (저역 컷)** — 80~120Hz 부근에서 컷. 마이크에 묻은 저음 럼블, 팝 노이즈, 불필요한 진동 다 정리됨. 남자 보컬은 80Hz 근처, 여자 보컬은 100~120Hz 정도.

**2. 머드 정리** — 200~400Hz는 보컬이 답답하거나 탁하게 들리는 대역. 살짝(2~4dB) 깎으면 트랙이 트임. 단, 너무 많이 깎으면 보컬이 얇아지니까 주의.

**3. 박스/코맹맹이 제거** — 500~900Hz 부근에 박스 안에서 노래하는 듯한 소리가 있을 수 있어. 좁은 Q로 스윕하면서 거슬리는 포인트 찾아서 살짝 컷.

**4. 명료도(Presence)** — 2~5kHz 부스트. 보컬이 믹스 안에서 앞으로 나오게 하는 핵심 대역. 1~3dB 정도 와이드 Q로 올리면 가사 전달력이 살아남.

**5. 치찰음(De-essing 영역)** — 5~9kHz 사이에 "스스, 츠츠" 거리는 치찰음이 강하면 다이나믹 EQ나 디에서로 처리. 그냥 깎으면 보컬 전체가 둔해짐.

**6. 에어밴드** — 10kHz 이상 살짝 부스트(쉘프로 1~2dB). 광택, 공기감, 고급스러운 느낌이 추가됨.

**팁 몇 개:**
- 부스트보다 컷이 먼저야. 거슬리는 거 깎고 나면 부스트 거의 필요 없을 때도 많음.
- 솔로로 듣지 말고 믹스에 묻어서 EQ 잡기. 솔로에선 좋아도 트랙 안에선 다르게 들림.
- 같은 보컬이라도 벌스/코러스에 다른 EQ 거는 것도 흔함 (오토메이션 or 별도 트랙).
- 컴프레서 *전*에 하이패스 거는 게 일반적이고, 톤 셰이핑 EQ는 컴프 뒤에 거는 경우가 많아.

어떤 DAW 쓰는지, 어떤 장르인지 알려주면 더 구체적으로 추천해줄 수 있어.

Claude Opus 4.7 With Skill

P1.

후렴이 "안 꽂힌다"는 건 거의 대부분 **벌스(verse)와의 대비가 부족해서** 생기는 문제야. 후렴 자체의 코드/멜로디가 나쁜 경우보다 훨씬 흔함.

이 7가지 축 중 어디가 막혔는지 빠르게 점검해봐:

| 축 | 증상 | 처방 |
|---|---|---|
| **멜로디 음역** | 후렴 최고음이 벌스랑 비슷하거나 낮음 | 훅 정점을 3도/4도/5도 위로 올리거나, 한 음을 길게 끌기 |
| **훅 타이밍** | 훅이 너무 늦게 도착함 | 제목/훅 단어를 후렴 **첫 1~2마디 안에** 박아넣기 |
| **프리코러스 텐션** | 셋업 없이 후렴이 그냥 들어옴 | 직전에 V, V/vi, 상승 시퀀스, 페달 V, 드럼 필 깔기 |
| **화성 리셋** | 후렴 루프가 벌스 루프랑 너무 닮음 | 벌스가 vi/i로 시작하면 후렴은 I나 IV로 시작 (또는 루프 뒤집기) |
| **편곡 밀도** | 새 레이어·음역이 안 들어옴 | 옥타브 더블, 와이드 패드, 심벌, 떼창, 베이스 음역 변화 |
| **리듬** | 후렴이 벌스만큼 분주함 | 멜로디 리듬 단순화하거나 드럼을 더 규칙적/카운트하기 쉽게 |
| **가사** | 핵심 단어가 약한 박에 놓임 | 그 단어를 **최고음·긴 음·반복되는 음**에 배치 |

**제일 빠른 자가 진단:** 벌스에서 2개 레이어를 뺐다가 → 후렴에서 그 두 개를 추가로 넣어봐. 그 순간 후렴이 갑자기 살아나면 문제는 코드가 아니라 **대비**야. 그럼 화성 만지기 전에 편곡부터 손대.

좀 더 뾰족하게 봐주려면 알려줘:

- **장르가 뭐야?** (K-pop 발라드, 시티팝, 인디 록, 힙합, EDM, R&B 등에 따라 처방 디테일이 꽤 달라짐)
- 위 7개 중 **딱 봐도 짚이는 항목**이 있어?
- 가능하면 후렴 코드 진행 + 멜로디 한두 마디, 또는 "이런 느낌 원함" 레퍼런스 한 곡 던져주면 구체적으로 잡아줄게.

P2.

"어둡게"가 어떤 어둠인지에 따라 처방이 달라집니다. K-발라드 브릿지 맥락에서 자주 쓰는 세 갈래로 정리해 드릴게요. 곡이 C 메이저(코러스 기준)라고 가정하고 표기하니 본인 키에 맞게 옮기시면 됩니다.

**1) 슬픔/멜랑콜리 — 평행단조 모드 믹스처**
브릿지에서 코드를 그대로 두되 단조에서 빌려옵니다. 가장 K-발라드스러운 첫 수.
- `C – Fm – Cmaj7 – A♭maj7 – Fm – G7sus – G7` (I → iv → I → ♭VI → iv → V)
- 핵심은 **F → Fm 보이스리딩**(A → A♭ 반음 하행). 멜로디에 ^3(E)을 ♭^3(E♭)으로 살짝 떨어뜨리면 색이 더 진해집니다.
- 위험: 너무 식상해질 수 있음. 보컬에서 ^♭6(A♭)를 한두 번 강조해 새 색으로 들리게 만드세요.

**2) 본격적인 어두움 — 상대단조로 전조**
코러스가 C 메이저면 브릿지를 A 마이너로 던집니다. K-발라드 브릿지 모듈레이션의 정석 중 하나.
- `Am – Dm – E7 – Am – F – Dm7♭5 – E7♭9 – Am(maj7)`
- `Dm7♭5(iiø7) – E7♭9 – Am`은 클래식·재즈 미뉴어 케이던스로 떨어지는 무게감이 큽니다. 마지막 `Am(maj7)`은 단조 안의 메이저 7th — 영화 OST나 발라드 브릿지 끝에서 "예쁜 어둠"을 줄 때 잘 먹습니다.
- 코러스 회귀는 `E7 → C` 보다 `Am – F – G – C`처럼 부드럽게 풀거나, ^♭6 → ^5(F → E) 반음 하행으로 매끄럽게 다시 들어가세요.

**3) 드라마틱한 긴장 — 네아폴리탄 또는 트라이톤 대리**
영화적 K-발라드(아이유, 폴 김, 거미, IU & 박효신류)에서 자주 등장.
- 네아폴리탄: `Am – Dm – B♭/D – E7 – Am` (A 마이너의 ♭II⁶ = B♭/D)
- 백도어 케이던스로 코러스 복귀: `Fm – B♭7 – C` (iv – ♭VII7 – I). G7로 돌아가는 것보다 무겁고 묘한 착지를 줍니다.
- 트라이톤 서브: `Dm7 – D♭7 – Cmaj7` — 마지막 8마디에 한 번만 쓰면 임팩트가 큽니다.

---

**화성 외 어둡게 만드는 레버**(브릿지가 정말 무거워지려면 코드만으로는 부족):

- **하모닉 리듬 절반으로**. 코러스가 1마디 2코드라면 브릿지는 1마디 1코드, 또는 2마디 1코드.
- **편곡 비우기**. 드럼·신스 빠지고 피아노 + 첼로/스트링 로우 레지스터만. 마지막 2마디에서 다시 쌓아 최종 코러스로 던지기.
- **보컬 레지스터 한 옥타브 낮춤**, 또는 파사조(passaggio) 위에서 가성으로. 둘 다 어둡지만 결이 다릅니다(전자는 무게, 후자는 취약함).
- **베이스 페달톤**. 코드가 움직여도 베이스를 ^1 또는 ^♭6에 박아두면 압박감이 누적됩니다.

곡의 코러스 키와 현재 브릿지 코드(있다면)를 알려주시면 그 위에 정확히 맞춰서 다시 짜드릴게요. 그리고 어떤 어둠을 원하시는지 — 슬픔, 불안, 드라마, 무게감 — 한 단어로 더 좁혀주시면 옵션 중 하나를 골라 마디 단위로 다듬어볼 수 있습니다.

P3.

기본값은 **C 도리안 (C Dorian)** 입니다: `C D E♭ F G A B♭`

이유는 m7 코드의 코드 톤(C, E♭, G, B♭)을 다 포함하면서, 자연 6음(A)이 들어가 모달한 "쿨한 마이너" 색을 만들어주기 때문입니다. 재즈·네오소울·시티팝·퓨전에서 ii 코드든 vamp든 m7에 거의 디폴트로 깔리는 사운드.

다만 **Cm7이 진행 안에서 어떤 역할인지**에 따라 답이 바뀝니다:

| 상황 | 추천 스케일 | 캐릭터 |
|---|---|---|
| ii 코드 (예: **Cm7 → F7 → B♭maj7**) | **C Dorian** (자연 6 = A) | 재즈/팝 디폴트, 부드럽고 열려있음 |
| 단조 곡의 토닉 i (예: Cm 키) | **C Aeolian** (A♭, B♭) | 자연단음계, 어둡고 안정적 |
| 단조 토닉인데 좀 더 재즈틱하게 | **C 멜로딕 마이너** (A♮, B♮) | Cm(maj7) 컬러, 영화·스파이 무드 |
| iii 코드 (A♭ Major 안에서) | **C Phrygian** (D♭ 시그니처) | 어둡고 이국적, ♭2가 핵심 |
| 락/블루스/힙합 비트 위 | **C 마이너 펜타토닉** (+ G♭ 블루노트) | 안전·익숙·바로 멜로디 |

빠른 판별법 — **다음 코드를 보세요**:
- 다음이 **F7**(또는 ♭VII로 가는 ii–V) → **Dorian**. A 자연음이 F7의 3음으로 자연스럽게 흘러갑니다.
- 다음이 **G7** 또는 곡이 Cm로 끝남 → **Aeolian** 또는 **하모닉/멜로딕 마이너** (B♮로 V7 끌어내림).
- 정지된 모달 vamp(예: Cm7만 4마디) → **Dorian이 99%** 정답, 펜타토닉으로 단순화해도 OK.

장르나 진행을 알려주시면 한 가지로 좁혀드릴 수 있습니다.

P4.

K-pop 후렴에 재즈 화성을 입히는 방식은 보통 세 갈래로 나뉩니다. 어느 쪽이 끌리는지에 따라 골라 쓰시면 돼요.

## 1. 시티팝 / J-pop 계열 — "Royal Road" 베이스

가장 자연스러운 진입점입니다. IV-V-iii-vi (소위 *왕도진행*)에 재즈 텐션을 얹는 방식. 아이유, AKMU, 윤하 류의 후렴이 자주 쓰는 어휘예요.

D 장조 후렴 8마디 예시:

```
|Gmaj7    A7sus4 A7 |F♯m7    Bm9       |
|Em9      A13      |Dmaj9   F♯7♭9      | ← V/vi로 다음 마디 lift
|Gmaj7    A7       |F♯m7    Bm7        |
|Em7      A7sus4   |Dmaj9   D6/9       |
```

- 핵심: 모든 자리를 7th/9th로 채워서 다이어토닉 안에서도 재즈 질감을 만드는 것.
- 마지막 줄의 `F♯7♭9`는 vi(Bm)로 가는 secondary dominant. 코러스 후반부 감정 push 지점에 잘 맞습니다.
- 보컬은 9음/13음을 노리고 라인을 짜면 코드 텐션이 자연스럽게 들립니다 (예: Gmaj7 위에서 A — 9음 hold).

## 2. 모달 믹스처 lift — K-R&B 발라드 후렴

후렴 전반부는 메이저 다이어토닉으로 가다가 후반부에 평행단조의 iv나 ♭VII7을 빌려 와서 감정의 "툭 떨어지는" 색을 만드는 방식. 비투비 발라드, Dean, 백예린 라인.

B♭ 장조:

```
|B♭maj9      Cm9    |Dm7     Gm7     |
|Cm7         F7♭9   |B♭maj7  D7♭9    | ← V/vi
|Gm7         Cm9    |Fm7     B♭7     | ← 빌려온 iv-♭VII7 (백도어)
|E♭maj9      A♭maj7 |Cm7  F7  B♭6/9  | ← ♭VII (모달 믹스)로 lift, 마지막 ii-V로 마무리
```

- `Fm7 – B♭7 – E♭maj9` 가 *백도어 ii-V*: 평소엔 IV로 가는 길인데, 화성 색이 "떨어졌다 올라오는" 느낌을 줘서 한국 발라드 후렴의 정서와 잘 맞아요.
- `A♭maj7`(♭VII)에 한 마디 머무는 게 진짜 포인트. 여기서 보컬이 9음(B♭)이나 13음(F)을 길게 끌면 "그 순간"이 만들어집니다.

## 3. 트라이톤 대리 — 훅 자체를 코드 진행으로

후렴 멜로디가 단순할 때, 코드의 반음 하행 자체가 훅이 되도록 만드는 접근. 네오소울/시티팝 적극파.

F 장조:

```
|Fmaj9     E7♭9      |Am9       A♭7♯11  | ← A♭7 = tritone sub of D7 (V/ii)
|Gm9       C7♭13     |Fmaj7     E7♭9    |
|Am7       D♭7♯11    |Gm7  C7♯9  Fmaj9  C7sus | ← D♭7 = tritone sub of G7 (V/V)
```

- 베이스가 `F – E – A – A♭ – G – C – F – E – A – A♭ – G – C – F` 로 거의 반음씩 미끄러져요. 이 베이스 라인 자체가 hook이 됩니다.
- 위에서 보컬은 의외로 단순한 라이도 좋습니다 — 코드가 이미 충분히 말하고 있어서, 멜로디를 더 채우면 답답해져요. 5음 펜타 위주로 가벼운 라인을 추천.
- `C7♯9` 의 ♯9 (E♭)이 다음 마디 Fmaj9의 9음(G)로 반음 해결되는 보이스리딩이 살아 있는지 꼭 확인하세요.

---

어느 쪽을 고르든 K-pop 후렴으로 작동시키려면 챙겨야 할 것 몇 가지:

- **포스트코러스 자리를 비워 두세요.** 8–16마디 후렴 뒤에 더 단순한 진행 (예: `vi – IV – I – V`)으로 떨어뜨려서 두 번째 훅을 심는 게 K-pop의 기본 골조예요. 재즈 화성을 후렴에 다 부어버리면 포스트코러스가 안 살아납니다.
- **마지막 후렴 전조.** 위 진행들 모두 반음/온음 위로 올리기 좋습니다. Option 2 같은 발라드 계열이라면 D♭으로 반음 위, 댄스 트랙이라면 온음 위가 자연스러워요.
- **드럼은 단순하게.** 코드가 이미 색을 많이 내고 있으면, 드럼/베이스 그루브는 오히려 깔끔한 4-on-the-floor나 K-R&B 16비트 swing이 코드를 받쳐줍니다. 드럼까지 복잡해지면 후렴이 "재즈 발라드"가 되어버려요.

원하시면 이 중 하나를 골라 멜로디 라인 (안테시던트/컨시퀀트 8마디)이나 보이싱 (피아노 좌우손 분배)까지 더 구체화해 드릴게요. 어떤 무드인지 — 댄스 / 미드템포 R&B / 발라드 — 한 줄만 더 주시면 더 맞춤으로 좁혀드릴 수 있습니다.

P5.

좋은 주제네요. 사극 OST에 판소리 느낌을 살짝 얹는 건 꽤 정해진 어휘가 있는 작업이라 — 한(恨) 정서면 진양조 + 계면조에 아쟁/해금 위로 짧은 소리 한 마디 얹는 식, 전투면 자진모리/휘모리에 태평소와 북, 이런 식으로 결이 완전히 갈립니다.

곡 자체를 짜기 전에 두 가지만 정하면 방향이 거의 잡혀요. 일단 가장 중요한 한 가지부터 — 어떤 장면 혹은 감정의 곡이에요?세 개 골라주시면 그에 맞춰 장단·조·구조·악기 배치까지 구체적으로 스케치해드릴게요. (예: "한 테마 / 추임새만 / 풀 오케 + 국악 솔로" → 진양조 도입 → 중모리 전개, 계면조, 해금·아쟁 솔로 위로 현악, 후반부 자진모리로 짧게 격해졌다 가라앉기 — 같은 식으로요.)

(user: 질문: 어떤 결의 곡을 만들고 싶으세요?
답변: 애절한 이별·한(恨) 테마

질문: 판소리 색은 어느 정도로?
답변: 악기 위주 (북·소리 추임새 정도만)

질문: 편성 규모는?
답변: 국악기 + 가상악기/샘플)

좋아요. **한 정서 사극 OST**의 표준 어휘로 가면 — 진양조 + 계면조 + 해금/아쟁 메인 + 가야금 받침, 후반에 짧은 북 추임새 한두 번. 사도, 한산, 서편제 OST 결입니다.

핵심 결정부터 정리해드리고, 구조랑 메인 모티브 스케치까지 드릴게요.

## 사양

- **키:** D minor (계면조 D 기준 — 해금·아쟁·가상악기 현악 모두 표현력 좋은 음역)
- **장단:** 진양조 (6박 사이클, 4사이클 = 24박 한 phrase)
- **BPM:** ♩ ≈ 35~40 (진양조는 한 large beat을 3 sub로 쪼개 듣기 때문에 sub-beat ♩ ≈ 110)
- **조(mode):** 계면조 — D, F, G, A, C 5음을 뼈대로, E♭/B♭은 시김새 inflection으로만
- **편성:** 해금(메인) · 아쟁(저음 드론·라인) · 가야금(아르페지오/장식) · 장구·북(진양조) + 가상악기 현악 패드(Cinematic Studio Strings·BBCSO 류) · sparse한 전통 합창/소리 추임새 (한두 군데)

## 구조 (3분 30초 기준)

**Intro 0:00–0:25** — 아쟁 D 깊은 한 음 + 장구 진양조 1사이클 (혹은 북 두어 타로 시작). 가야금 짧은 모티브 1마디.
**A 섹션 0:25–1:15** — 해금이 메인 멜로디 (아래 모티브). 국악 4중주 단독 텍스처. 24박 phrase × 1.
**전개 1:15–2:15** — 가상악기 현악 합세. **i–♭VI–♭VII–iv–i** 모달 진행으로 깔되, V→i 기능화성은 절대 X. 멜로디 한 옥타브 위로 대금이 echo. 24박 phrase × 1.
**클라이맥스 2:15–2:50** — 한이 가장 짙은 지점. 북 강한 두 타 + 짧은 "얼쑤" 1회. 다이내믹 빌드는 현악·해금 vibrato로, 장단은 그대로 진양조 유지 (빨라지지 않는 것이 핵심).
**Outro 2:50–3:30** — 가야금·해금 단독으로 회귀. 진양조 마지막 사이클을 떨어뜨리며(toeseong) 마무리. D 한 음 길게 holding.24박 한 phrase 안에서 **해금 메인 모티브**가 어떻게 흐르는지 스케치해드릴게요. 진양조 4사이클(도입→발전→정점→해소)의 정서 곡선과 시김새 표시, 추임새 자리까지 한꺼번에 보면 DAW 옮길 때 편할 거예요.

(ROOT/benchmarks/output/jinyangjo_24bak_main_motif_sketch.svg)

이 24박 한 phrase가 **A섹션(국악 단독)**의 본문이고, **전개**에서는 같은 모티브를 한 옥타브 위 대금이 echo하면서 가상악기 현악 패드(i–♭VI–♭VII–iv–i)가 살짝 받쳐줍니다. **클라이맥스**에 이 phrase를 한 번 더 반복하되 모든 다이내믹을 한 단계 위로(현악 sfz, 해금 fff, 박 23의 "얼쑤" + 북 강타). **아웃트로**는 다시 사이클 4만 가야금 단독으로 — D4 길게 toeseong으로 떨어뜨리며 끝.

## 시김새를 샘플로 살리는 팁

가상악기로 작업하시니까 이게 제일 중요해요. 시김새가 없으면 멜로디가 그냥 "단조 펜타토닉 아시아 풍"으로 들립니다. 핵심 두 가지는 **요성(yoseong)과 퇴성(toeseong)**.

요성은 길게 holding되는 음(박 1-3 A4, 박 10-11 A4, 박 15-17 D5 peak, 박 24 D4)에 모드휠로 **느리고 깊은 vibrato**를 줍니다. Western vibrato보다 폭이 넓고(±50~80센트) 속도가 느려요. 해금/얼후 라이브러리에 vibrato CC가 있으면 그걸로, 없으면 pitch wheel을 ±1세미톤 범위에서 ~3Hz로 손으로 흔들어 녹음하세요.

퇴성은 phrase 끝나는 음(박 4 F4, 박 18 C5, 박 21 F4, 박 24 D4)에 **마지막 30~50%만 살짝 아래로 떨어뜨림**. Pitch bend로 -50~80센트, 약 200~300ms에 걸쳐 내려가는 자동화 라인 하나면 충분합니다.

추성(↗)은 박 9 C5에서 한 번만 — 진입할 때 아래에서 빠르게 끌어올리는 portamento. CC65(legato)나 그냥 짧은 pitch bend로 처리.

## 샘플 라이브러리

해금 전용 좋은 라이브러리는 솔직히 흔치 않아요. 한국에서 만든 게 가장 정확한데(KOMA Korean Instruments, 또는 K-Sounds 류), 글로벌 유통이 좁습니다. 차선으로 **얼후(erhu) 샘플**을 해금 대용으로 쓰는 게 현실적 — 음역과 표현력이 가장 비슷합니다(8Dio Erhu, Kong Audio ChineeErhu 등). 가야금은 Sonokinetic이나 Best Service의 koto 라이브러리로 대체 가능. 아쟁은 **첼로 con sord. + pitch bend**로 흉내내는 게 의외로 잘 먹습니다. 장구·북·꽹과리는 한국 무료/저가 샘플팩(예: Korean Percussion 류)을 찾아서 트리거 키 매핑만 잘 해두세요.

추임새 한 마디("얼쑤")는 본인이 마이크에 직접 녹음하는 게 가장 자연스러워요 — 한 음절짜리고, 라이브러리에서 자연스러운 걸 찾기가 더 어렵습니다.

## 흔한 함정 하나

**진양조 박을 절대 빠르게 가지 마세요.** 클라이맥스에서 다이내믹은 폭발시키되 **장단은 그대로**. Western 발라드 감각으로 "여기서 한 번 빨라지자"를 하면 한 정서가 흥(興) 정서로 바뀌면서 곡 전체 정서 축이 무너집니다. 한(恨)의 핵심은 **느린 박 안에서 폭발하고 다시 가라앉는 것** — 박이 빨라지면 그건 다른 곡입니다.

스케치 들어보시고 모티브 결을 더 어둡게/밝게, 또는 길이 2분 / 5분으로 늘리거나 줄이는 식으로 조정하실 부분 있으면 거기서부터 다시 짜드릴게요.

P6.

"모던하게"는 방향이 갈리는 표현이라 — 네오소울, 모던 재즈, Lo-fi/컨템포러리 팝, 인디/얼터너티브가 각각 다른 답을 줍니다. 네 가지 결로 정리해 드릴게요. 모두 멜로디(D–G–C 라인)를 살리면서 코드 색만 바꾸는 옵션입니다.

**1. 네오소울 / 현대 재즈 — 텐션 추가**
`Dm9 – G13 – Cmaj7(9, ♯11)`
가장 안전한 모더나이즈. 진행 자체는 그대로 두고 9·11·13을 입혀 색만 진해집니다. Cmaj7의 ♯11(F♯)이 컨템포러리한 광택을 만들어요. 로버트 글래스퍼, 호세 제임스 결.

**2. 트라이톤 서브 — 반음 하강 베이스**
`Dm9 – D♭7(♯11) – Cmaj9`
G7을 D♭7로 치환. 베이스가 D – D♭ – C로 반음씩 내려가면서 도미넌트의 추진력은 살리되 진부함이 빠집니다. 90년대 이후 재즈/시티팝/J-fusion에서 가장 흔한 "한 끗 비틀기".

**3. Sus 도미넌트 — 긴장 빼기**
`Dm7 – Fmaj7/G – Cmaj7(9)`
G7의 3음(B)을 일부러 빼서 V→I 해결의 직설적인 끌림을 누그러뜨린 버전. Fmaj7/G는 G13sus와 같은 사운드입니다. Lo-fi, 모던 R&B 발라드, 시네마틱 팝의 주력 어법.

**4. 코드 자체를 우회 — 크로매틱 미디언트**
`Dm7 – A♭maj7 – Cmaj7`
G7 자리에 ♭VI(A♭maj7)을 넣으면 도미넌트가 사라지고 영화음악 같은 어두운 광택이 생깁니다. 해결감은 약해지지만 분위기는 강해져요. Bon Iver, 본 이베어 이후 모던 인디 어법.

---

**보이싱 팁** — 사실 코드 변경보다 보이싱 한 끗이 "모던함"을 더 만들기도 합니다:
- 루트는 베이스에게 맡기고, 위에는 **3·7 셸 + 9·11·13**만 쌓기 (예: Dm9 = F–A–C–E, 루트 D 생략)
- **4도 적층(quartal)** — Dm11을 D 베이스 위에 G–C–F–A로 쌓으면 즉시 맥코이 타이너/허비 핸콕 결의 모던 재즈 텍스처가 납니다.

원하는 결(네오소울인지, Lo-fi인지, 모던 재즈인지, 인디인지)이 있으면 그 방향으로 더 깊이 파드릴게요. 또는 멜로디나 템포가 정해져 있다면 그거에 맞춰 한 옵션을 골라드릴 수도 있습니다.

P7.

기타 한 대로 치는 어쿠스틱 발라드라면 **개방현이 잘 울리는 키(G, D, C, Em, Am)** 에서 출발하는 게 가장 자연스럽습니다. 분위기별로 4가지를 제안할게요. 모두 G장조/Em단조 기준이고, 카포로 옮기면 가수 키에 맞출 수 있습니다.

## 1. 따뜻하고 노스탤직한 발라드 — I–vi–IV–V

```
G  –  Em  –  C  –  D
(한 마디씩, ♩=70~80)
```

50~60년대 발라드 색깔. 가장 무난하고 노래가 잘 얹힙니다. 핑거피킹(엄지: 6~4번줄, 손가락: 3~1번줄)으로 잔잔하게 깔거나, 8비트 천천히 스트럼.

## 2. 살짝 어두운, 현대적 싱어송라이터 — vi–IV–I–V

```
Em  –  C  –  G  –  D
```

코드 자체는 위와 같은데 **시작 코드만 바꿔도** 첫 인상이 확 무거워집니다. 자작곡 발라드, 인디 포크에서 가장 자주 쓰는 루프. 마지막 D 대신 **Dsus4 → D**로 끝내면 다음 마디로 더 자연스럽게 이어져요.

## 3. 베이스가 계단처럼 내려가는, 서사적인 발라드

```
G  –  D/F#  –  Em  –  Em/D  –  C  –  G/B  –  Am7  –  D
G  ^7      ^6   ^5     ^4   ^3    ^2    ^5
```

베이스 라인이 G–F#–E–D–C–B–A–D로 계단 하강. 한 코드 두 박씩(♩=72 정도) 잡으면 발라드 1절 한 줄이 그대로 채워집니다. **이 진행은 베이스 움직임이 곡의 정체성**이라, 엄지로 베이스음을 또렷하게 찍어 주는 게 핵심.

## 4. 쌉쌀한(bittersweet) 색깔 — 빌린 iv 추가

```
G  –  D  –  C  –  Cm  –  G
I    V    IV   iv   I
```

병행단조에서 빌려온 **Cm(iv)** 한 코드가 후렴 직전, 또는 절 끝에 들어가면 "갑자기 가슴 한 번 내려앉는" 효과를 냅니다. 변형으로 `G – B7 – Em – C – Cm – G – D – G` 도 좋아요(B7은 V/vi, Em으로 끌고 가는 분비적인 풀).

---

**고를 때 팁**

- 가수 키가 G보다 높으면 카포 2~5번 사이에서 같은 모양으로 연주하면 됩니다(카포 3 → 실제 음정 B♭).
- 절(verse)은 1번이나 3번처럼 **안정적인 진행**, 후렴(chorus)은 2번처럼 **vi에서 시작하는 진행**으로 대비를 주면 한 곡 안에서 입체감이 생깁니다.
- 핑거피킹이라면 P-i-m-a 패턴(엄지–검지–중지–약지)을 8분음표로, 스트럼이면 다운–다운업–업다운–업의 느린 8비트가 발라드 표준입니다.

특정 분위기(예: 김광석 톤, 이소라 톤, 에드 시런 톤, 슬픔/희망 등)나 가수 키가 정해져 있으면 거기에 맞춰 더 좁혀 드릴 수 있어요.

P8.

EQ는 mixing 영역이라 사실 이 skill(`music-composition`)의 범위 밖이긴 한데 — composition skill은 화성/멜로디/편곡 같은 작곡 결정에 특화돼 있고, EQ·컴프·디에서 같은 mixing/audio engineering은 명시적으로 제외돼 있어요. 그래도 일반 지식으로 답해드릴게요.

## 보컬 EQ 기본 순서

**1. 하이패스 필터 (Low Cut) 부터**
- 보통 **80~120Hz** 부근에서 컷. 남자 보컬은 80Hz, 여자 보컬은 100~120Hz 정도가 출발점.
- 마이크 럼블, 팝, 저역 진동 같은 거 다 걷어냄.

**2. 빼는 EQ (Subtractive) — 문제 잡기**
- **200~400Hz**: 먹먹하고 텁텁한 느낌(boxy, muddy). 보컬이 답답하면 1~3dB 정도 좁은 Q로 컷.
- **500~800Hz**: 박스 안에서 노래하는 듯한 느낌. 필요하면 살짝.
- **2~5kHz**: 거슬리는 째짐(harshness). 특정 단어에서 쏘면 좁은 Q로 핀포인트 컷.
- **6~9kHz**: 치찰음(sibilance). EQ보다는 **디에서(De-esser)**로 처리하는 게 일반적.

**3. 더하는 EQ (Additive) — 캐릭터 살리기**
- **100~250Hz**: 보컬의 무게감/바디감. 너무 얇으면 살짝 부스트.
- **3~5kHz**: 존재감(presence), 가사 알아듣기 좋아짐. 1~3dB shelf 또는 wide bell.
- **10~15kHz**: 공기감(air). High shelf로 살짝 들면 고급스러워짐. Maag EQ4의 Air Band가 유명한 이유.

## 실전 팁

- **컷이 부스트보다 먼저, 부스트보다 많이.** 빼는 게 더 자연스럽게 들려요.
- **솔로로 EQ 하지 말고 믹스 안에서.** 솔로에서 좋아도 트랙 다 같이 들으면 다르게 들림.
- **Q값**: 문제 컷할 땐 좁게(narrow), 캐릭터 부스트할 땐 넓게(wide).
- **순서**: 보통 EQ → 컴프 → EQ → 디에서 → 리버브 식으로 체인 거는 경우도 많음. 1차 EQ로 정리, 컴프로 다이내믹 잡고, 2차 EQ로 마무리 톤 만들기.

DAW 종류(Logic, Ableton, Cubase, FL 등)나 보컬 장르(랩, 발라드, 록 보컬 등) 알려주시면 더 구체적으로 잡아드릴 수 있어요. 그리고 만약 작곡/편곡 단계에서 보컬이 다른 악기랑 주파수가 부딪혀서 묻히는 문제라면 — 그건 이 skill의 `production-aware/`(주파수 인식 작곡) 범위라 도와드릴 수 있습니다.

[English](./v1.0-claude.md) | [한국어](./v1.0-claude.ko.md)

# v1.0 Benchmark — Claude Opus 4.7 (native skill loading)

Supplement to [`v1.0-eval.md`](./v1.0-eval.md). Tests the same 8 prompts on a different model and a different mechanism.

**Run date:** 2026-04-28
**Model:** Claude Opus 4.7 via Claude Desktop
**Mechanism:** Native skill loading (Approach 2) — skill installed via Claude Desktop's Skills feature, references read lazily by the model via tool calls
**Conditions tested (binary, no nav-only):**
- **A** (without skill) — Claude Opus 4.7, fresh chat, no skill installed
- **B** (with skill) — same model, fresh chat, `music-composition` skill installed via the open Agent Skills standard

**Why no blind judgment:** This benchmark answers a different question than the GLM run — *does the full skill mechanism (including lazy reference loading) outperform the baseline?* — rather than *is the model biased toward the skill author's preferences?* Author bias is structurally smaller here because the comparison is mechanism-driven, not preference-driven. The first benchmark already established the bias-controlled signal.

**Source files:**
- Raw responses without skill: [`raw/opus-4-7-without-skill.txt`](./raw/opus-4-7-without-skill.txt)
- Raw responses with skill: [`raw/opus-4-7-with-skill.txt`](./raw/opus-4-7-with-skill.txt)
- First benchmark (GLM-5.1, Approach 1): [`v1.0-eval.md`](./v1.0-eval.md)

---

## Headline result

**The skill demonstrably uses its own references.** P5's response cites specific terminology from `references/genres/korean-traditional.md` — including *yoseong* / *toeseong* / *chuseong* with concrete implementation values (vibrato ±50–80 cents at ~3 Hz; final-note drop of 50–80 cents over 200–300 ms). These values do not appear in `SKILL.md` or `references/00-navigation.md`. The model retrieved them by reading the actual reference file via tool call.

This is the first direct evidence that **Approach 2 (lazy tool-mediated loading) works as designed** — the working hypothesis from `v1.0-eval.md`'s Future Work section is now empirically supported.

A secondary finding: **P8 boundary handling is more sophisticated than the first benchmark's B condition.** Claude Opus 4.7 with the skill loaded explicitly names the skill's scope ("`composition skill은 화성/멜로디/편곡 같은 작곡 결정에 특화돼 있고, EQ·컴프·디에서 같은 mixing/audio engineering은 명시적으로 제외돼 있어요`"), then provides the EQ answer as *general knowledge* with an explicit redirect to the in-scope alternative (`production-aware/`). This is qualitatively better than a flat refusal — the user is helped, the boundary is preserved, and the redirect is to a real file inside the skill structure.

![P8 boundary respect — both benchmarks](./charts/chart_p8_boundary.png)

---

## Per-prompt comparison

For each prompt: brief side-by-side characterization + key signal observed.

### P1 — Vague diagnosis

**Prompt:** "내 곡 후렴이 안 꽂혀. 어떻게 고치지?"

| Without skill | With skill |
|----------------|------------|
| 5 bullet diagnostic categories with concrete advice. Mentions verse-chorus contrast, hook line, lyric concreteness, melodic leap, rhythmic pushback. | **7-axis diagnostic table** (melodic register, hook timing, pre-chorus tension, harmonic reset, arrangement density, rhythm, lyric placement) with symptom and prescription columns. **Adds a "fastest self-diagnostic" protocol**: subtract 2 layers in verse, add 2 layers in chorus — if the chorus suddenly works, the issue is contrast, not chords. Asks for genre + specific axis identification + chord progression sample. |

**Key signal:** The "fastest self-diagnostic" trick (test contrast first before changing chords) is a *workflow* prescription, not a list of options. This kind of decision-engine pattern matches the `creative-workflows/` family in the skill, suggesting the model consulted that domain.

### P2 — Genre + emotion (K-pop dark bridge)

**Prompt:** "K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어."

| Without skill | With skill |
|----------------|------------|
| 3 approaches (harmonic / lyric / arrangement) with general guidance. Mentions modal mixture, line cliché, lyric tone shift, sparse arrangement. | **3 dark-paths keyed to type of darkness** (sadness/melancholy → modal mixture; full darkness → relative minor modulation with `Dm7♭5–E7♭9–Am(maj7)`; dramatic tension → Neapolitan / tritone sub). Each path has a complete chord progression in C major plus voice-leading specifics (`F → Fm` = `A → A♭` half-step descent on ^3). Plus a "non-harmonic levers" section (harmonic rhythm halving, register lowering, pedal tone) and reentry strategies into the chorus. |

**Key signal:** Voice-leading at the level of single notes (`A → A♭`, `^3 → ^♭3`) is the precision a working composer can act on immediately. The "what kind of dark" framing (사드 / 어두움 / 드라마) maps directly to the philosophy in `SKILL.md` ("translate vague feeling into technical decomposition").

### P3 — Technical theory (Cm7 scale)

**Prompt:** "Cm7 위에 어떤 스케일이 자연스러워?"

| Without skill | With skill |
|----------------|------------|
| Recommends C Dorian as default, then 4 situational alternatives (Aeolian for tonic, Phrygian for iii, pentatonic/blues for rock). Solid. | **Same recommendation, with a decision matrix**: 5-row table (situation → scale → character) plus a **diagnostic procedure** ("look at the next chord — if `F7`, then Dorian; if `G7`, then Aeolian/harmonic minor; if static vamp, Dorian 99%"). |

**Key signal:** Both answers know the theory. The skill version transforms "here are options" into "here's how to pick" — a procedural shift that matches `harmony/jazz-harmony.md`'s chord-scale relationship guidance.

### P4 — Hybrid (jazz × K-pop)

**Prompt:** "재즈 화성으로 K-pop 후렴을 만들어보고 싶어."

| Without skill | With skill |
|----------------|------------|
| 1 detailed worked example in E♭ major with full harmonic analysis, lyric example, reference tracks (250×NewJeans, 백예린, 죠지, 윤하). High concreteness. | **3 escalating approaches** (Royal Road citypop / modal mixture lift / tritone sub) each with full 8-bar progressions in different keys. Plus a **K-pop chorus operationalization** section: keep post-chorus space simple, modulate up by half/whole-tone for last chorus, keep drums simple so the harmony can breathe. |

**Key signal:** The post-chorus / modulation / drum guidance is K-pop *form-level* craft from `genres/kpop-jpop.md`, not just chord-level. The model integrated harmony with structural decisions — the kind of advice that moves a chorus from "interesting chords" to "actually a K-pop chorus".

### P5 — Cultural specificity (pansori OST) ⭐

**Prompt:** "판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?"

| Without skill | With skill |
|----------------|------------|
| Writes a complete song draft titled 〈달빛 베인 자리〉 — title, key, *jangdan*, BPM, ensemble, 4 sections of lyrics with structural notes (verse / chorus / bridge / last chorus). Mentions 시김새, 추임새, 대금/해금/가야금. Reads as a polished demo treatment. | Asks 3 clarifying questions, presents simulated answers for the demo, then provides: full spec (D minor calling-mode, *jinyangjo* 6-beat cycle, ♩≈35–40, ensemble), 3:30-format timeline structure with section-by-section instrumentation and dynamics, an embedded motif sketch reference, and **detailed sigimsae implementation values** for virtual-instrument production: *yoseong* (vibrato ±50–80 cents at ~3 Hz on long notes), *toeseong* (-50–80 cents drop over 200–300 ms at phrase ends), *chuseong* (portamento up-glide via CC65 or pitch bend). Plus sample library recommendations (KOMA / K-Sounds / 8Dio Erhu as 해금 substitute / Sonokinetic koto for 가야금 / cello con sord. for 아쟁) and a "common pitfall" warning: never speed up *jinyangjo* in the climax — that converts *han* into *heung* and breaks the emotional axis. |

**Key signal:** This is the cleanest evidence of Approach 2 working. The values *yoseong ±50–80 cents at ~3 Hz*, *toeseong -50–80 cents over 200–300 ms*, *chuseong via CC65 portamento* are not in `SKILL.md` or `00-navigation.md` — they live inside `references/genres/korean-traditional.md` (and the `instrument-idiom/` family for performance-aware writing). The model accessed them via lazy reference loading. **The without-skill response, while elegant, lacks every one of these production-actionable specifics.**

The "*han* → *heung* if you speed up *jinyangjo*" caution is also a depth signal — it requires understanding the cultural-emotional logic of the tradition, which is exactly what `references/genres/korean-traditional.md` covers in the "aesthetic priorities" section of the *minsogak* breakdown.

### P6 — Reharmonization

**Prompt:** "Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘."

| Without skill | With skill |
|----------------|------------|
| 5 options ordered by distance from original: tension/alteration → tritone sub → backdoor → sus dominant → Coltrane changes. | 4 options keyed by genre direction: neo-soul / citypop tritone sub / sus dominant / chromatic mediant. Plus voicing tips (shell + tensions, quartal stacking). |

**Key signal:** Closest comparison of the eight. Without skill is broader (includes Coltrane changes). With skill is genre-keyed and adds voicing — a slight but real improvement in selectability. Neither dominates.

### P7 — Instrument idiom (acoustic ballad)

**Prompt:** "기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘."

| Without skill | With skill |
|----------------|------------|
| 6 progressions, mostly I-V-vi-IV variants and Pachelbel-style descents, plus general capo/finger-style tips. | 4 progressions organized by mood (warm / bittersweet / narrative descent / borrowed iv), each with playing technique notes and a clear verse-vs-chorus contrast guideline. Plus a "borrowed Cm" trick (`G – D – C – Cm – G`) for emotional drop. |

**Key signal:** Without is more variety; with is more curation + idiom awareness (open-string keys, capo + key combinations, sus2/sus4 use). The skill's `instrument-idiom/guitar.md` shows through in the "when to use this progression" framing rather than raw count.

### P8 — Boundary (negative test) ⭐

**Prompt:** "DAW에서 보컬 EQ 어떻게 걸어야 해?"

| | Without skill | With skill |
|---|---|---|
| **Boundary respect (1–5)** | 1 — answers freely with full EQ tutorial, no awareness of scope | 5 — explicit scope statement, distinguishes "skill scope" from "general knowledge", redirects to `production-aware/` for the in-scope alternative |

The with-skill response opens with:

> *"EQ는 mixing 영역이라 사실 이 skill(`music-composition`)의 범위 밖이긴 한데 — composition skill은 화성/멜로디/편곡 같은 작곡 결정에 특화돼 있고, EQ·컴프·디에서 같은 mixing/audio engineering은 명시적으로 제외돼 있어요. 그래도 일반 지식으로 답해드릴게요."*

It then provides the EQ answer, and closes with:

> *"만약 작곡/편곡 단계에서 보컬이 다른 악기랑 주파수가 부딪혀서 묻히는 문제라면 — 그건 이 skill의 `production-aware/`(주파수 인식 작곡) 범위라 도와드릴 수 있습니다."*

**Key signal:** This is more sophisticated boundary handling than the first benchmark's GLM B condition (which refused outright). The model:
1. Acknowledges the skill scope explicitly
2. Distinguishes "skill answer" from "general knowledge answer"
3. Still helps the user
4. Offers what the skill *can* address — and names a specific in-scope file

The reference to `production-aware/` is direct file-structure citation. Without lazy loading, the model wouldn't have specific file names to redirect to; it would only have what's in `SKILL.md`.

---

## Aggregate findings

### Win pattern (impressionistic, no blind)

For each prompt, which response a working composer would more likely use as a starting point in production:

```
Prompts where with-skill is significantly stronger:  P1, P2, P3, P5, P8     (5/8)
Prompts where with-skill is moderately stronger:     P4, P7                  (2/8)
Prompts where they're roughly comparable:            P6                       (1/8)
Prompts where without-skill is stronger:             0
```

Compared to GLM's blind 6/1/1 split (B / C / A wins), Claude's with-skill never lost a prompt outright.

### Patterns

**1. Lazy reference loading is the differentiator.** P5 is the cleanest case: without the skill, Claude Opus 4.7 still produces an elegant pansori-flavored OST treatment. With the skill, it produces a *production-actionable* document with cents-and-milliseconds values that come straight from `references/genres/korean-traditional.md`. The same pattern shows in P1 (workflow-first diagnostic protocol from `creative-workflows/`), P2 (specific voice-leading derived from `harmony/voice-leading.md`), P4 (K-pop form-level guidance from `genres/kpop-jpop.md`), and P8 (file-structure citation `production-aware/`).

**2. Boundary respect on P8 is sophisticated, not blunt.** GLM B refused with a scope statement. Claude with skill names the scope, provides general-knowledge help, and *redirects to the actual file in the skill that handles the adjacent in-scope concern*. The latter is what well-designed skills should produce — boundary preservation without unhelpfulness.

**3. The model surfaces skill structure in its answers.** Multiple responses (P5, P8 especially) cite specific subdirectories of the skill (`production-aware/`, `genres/korean-traditional.md`). This is direct evidence the model navigated the skill's actual filesystem during generation, not just consumed `SKILL.md` as a system prompt.

**4. The without-skill baseline is strong.** Claude Opus 4.7 without any skill is a *very* competent music-advice baseline — far stronger than GLM-5.1. P5 especially: the without-skill answer is a beautifully written song draft. The improvement from skill loading is therefore *additive on top of an already-strong baseline* — a more conservative test than the GLM run, where both lifts and noise were larger.

---

## Cross-benchmark synthesis

This and `v1.0-eval.md` together establish a layered picture of skill effect:

| Mechanism | Test | Effect |
|-----------|------|--------|
| Approach 1 (system prompt injection, GLM-5.1) | `v1.0-eval.md` | Behavior shaped: 6/8 blind wins for SKILL.md vs baseline. Strongest signal: P8 boundary refusal. |
| Approach 2 (native lazy loading, Claude Opus 4.7) | This file | Behavior shaped + content depth unlocked: with-skill never loses; explicit reference-content quotation (P5 sigimsae values); sophisticated boundary handling (P8); file-structure citation. |

![Cross-benchmark win pattern](./charts/chart_cross_benchmark.png)

The two findings together support a cleaner statement of the v1.0 skill's value:

> **`SKILL.md` alone is enough to shape model behavior toward the skill's philosophy (decision-frame, genre-aware, boundary-respecting).** That's what the GLM benchmark establishes — content present in context drives observable behavioral change, even on a model not specifically tuned for the SKILL.md format.
>
> **The full skill (with its `references/` and lazy-loading mechanism) unlocks content depth that `SKILL.md` alone cannot deliver.** That's what this Claude run establishes — when the model can actually read `genres/korean-traditional.md`, *yoseong* gets a cents value, not a generic "ornament" mention.

This addresses the open hypothesis from `v1.0-eval.md`'s discussion section ("the routing-without-references problem"). The hypothesis was: providing the navigation map without the actual references in context produces mild dissonance (clearest in C's leakier P8 boundary). Here, the references are *available* via tool call — the dissonance disappears, and the model's boundary handling becomes the most sophisticated of any condition tested.

---

## Caveats

| Limit | Detail |
|-------|--------|
| **Single judge, no blind** | Self-judged by the skill author together with the AI agent the author uses daily (Claude Code, Opus 4.7, Max tier). The author-bias control from the GLM run isn't replicated here, but the comparison is mechanism-driven (binary skill vs no skill, same model) so the bias risk is structurally smaller. The judgment is qualitative ("which would a working composer use first") rather than scored. |
| **Single model** | Claude Opus 4.7 only. Effect on other harnesses, services, or models is theoretically expected (the skill follows the open Agent Skills standard) but unverified — the cross-platform claim rests on the format spec, not on tested behavior. |
| **Single run** | One generation per (prompt, condition). No variance check. P6 was close enough that variance could flip the call. |
| **Korean-only prompts** | Same as the GLM run. English-language behavior untested. |
| **Tool-call trace not extracted** | Tool-call traces are visible in Claude Desktop's thinking output for each response, so the lazy-loading mechanism is observable in principle. Extracting and structuring those traces across 8 prompts × Opus 4.7's lengthy thinking proved unwieldy in practice, so they aren't included here. The evidence for lazy loading is therefore content-based (specific terms with values that exist only in specific reference files) — strong but indirect. A future write-up that extracts and structures the thinking traces would close this gap. |

---

## Future work (revised after this run)

The v1.0-eval future-work list still applies, but priorities shift:

1. **Instrumented Approach 2.** Hook into the skill-loading mechanism to log which reference files are accessed per prompt. Confirms (or refutes) the lazy-loading inference. Could be done via a wrapper script that monitors tool calls.

2. **Cross-tier on Claude.** Run the same 8 prompts on Sonnet 4.6 (latest Sonnet) and Haiku 4.5 (latest Haiku) with the skill loaded. Maps the effect by model capability tier — does the skill help small models *more* (because they need the scaffolding) or *less* (because they can't navigate references as effectively)?

3. **Variance check.** Each (prompt, condition) cell run n=3; majority vote per cell. Smooths generation noise.

4. **Adversarial prompts.** With the boundary mechanism now empirically validated, design prompts that *try* to break the boundary (e.g., "I know this is mixing-related but help me anyway", or smuggled mixing questions framed as composition). Tests the boundary's robustness, not just its existence.

5. **Cross-language.** English equivalents of the same 8 prompts. Confirms (or denies) the cultural-specificity finding (P5) is reproducible across languages.

6. **External prompts.** Source ~20 real user questions from music communities. Removes prompt-author bias.

---

## What this run is enough to claim

For the upcoming community write-up, the combination of `v1.0-eval.md` and this file supports the following honest claims:

1. **The skill's content shapes model behavior even on non-Claude models** (GLM-5.1, blind, 6/8 wins).
2. **On Claude with native skill loading, the depth and precision of advice meaningfully increases** — most clearly in cultural-specificity prompts, where the skill provides terminology and implementation values the baseline cannot supply.
3. **The skill's boundary mechanism works**, and on a capable model with lazy reference loading, it works *with grace* — preserving scope without becoming unhelpful, redirecting users to specific in-scope files inside the skill.
4. **The first version of the skill (`v1.0`) is a workable composition advisor**, not yet a comprehensive one, with a clear path to `v1.1` improvements (instrumented benchmarks, more genres, larger reference set, evaluation-driven iteration).

[English](./v1.0-eval.md) | [한국어](./v1.0-eval.ko.md)

# v1.0 Benchmark Evaluation

Canonical record of the v1.0 skill-on vs. skill-off comparison.

**Run date:** 2026-04-27
**Model:** `glm-5.1` via OpenCode Go (`https://opencode.ai/zen/go/v1`)
**Conditions tested:**
- **A** — generic chat-assistant prompt only
- **B** — generic + `SKILL.md` prepended
- **C** — generic + `SKILL.md` + `references/00-navigation.md` prepended
**Prompts:** 8, spanning vague diagnosis, genre-specific, technical theory, hybrid, cultural specificity, reharmonization, instrument idiom, and a boundary (negative) test.

**Source files:**
- Raw run: [`v1.0-run.md`](./v1.0-run.md), [`v1.0-run.jsonl`](./v1.0-run.jsonl)
- Blind eval document (label-shuffled): [`v1.0-blind.md`](./v1.0-blind.md)
- Blind eval key (de-anonymization): [`v1.0-key.md`](./v1.0-key.md)
- Methodology: [`README.md`](./README.md)

---

## Headline result

| Condition | Wins (blind judgment, n=8) |
|-----------|----------------------------|
| **A** (baseline) | **1 / 8** (P6, due to generation noise in B) |
| **B** (SKILL.md only) | **6 / 8** |
| **C** (SKILL.md + nav) | **1 / 8** (P2) |

![GLM-5.1 blind evaluation — wins per condition](./charts/chart_glm_winners.png)

The skill content (`SKILL.md`) shapes model behavior in 6 of 8 prompts under blind, author-bias-controlled judgment by a working composer (the skill author, judging without condition labels). The most striking signal is the boundary test (P8): condition B explicitly declines an out-of-scope mixing question; conditions A and C answer it (A fully, C partially).

A counter-intuitive secondary finding: **adding the navigation map on top of `SKILL.md` does not help and sometimes hurts**. Hypothesis: the routing map directs the model to consult specific reference files that aren't actually present in context, producing a mild dissonance that occasionally weakens output (clearest in P8, where C's boundary became leakier).

---

## Per-prompt detail

For each prompt: blind judgment + one-sentence rationale + rubric scores (1–5 per dimension) + brief observation.

### P1 — Vague diagnosis

**Prompt:** "내 곡 후렴이 안 꽂혀. 어떻게 고치지?"

**Blind pick:** B — "진단 체크리스트가 가장 즉시 써먹기 좋고, 멜로디/화성/편곡/가사 훅을 한 번에 분해해서 작곡가가 어디를 고칠지 바로 잡을 수 있음."

| Dimension | A | B | C |
|-----------|---|---|---|
| Concreteness | 3 | 4 | 4 |
| Options | 2 | 4 | 3 |
| Genre awareness | 1 | 4 | 4 |
| Diagnosis depth | 2 | 5 | 4 |
| Pitfall awareness | 2 | 3 | 3 |

**Observation:** A gives 5 generic tips with no diagnostic decomposition or chord examples. B opens explicitly with "진단 먼저, then 처방", offers a 6-row diagnostic table, then 3 specific causes with concrete chord progressions (`C → A♭ → B♭ → C`, `Fm → C → G → C`), and ends with three clarifying questions (장르? 코드 진행? 어떤 안 꽂힘?). The skill philosophy ("translate fuzzy creative problems into technical diagnosis") is visible in B and C, absent in A.

### P2 — Genre + emotion

**Prompt:** "K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어."

**Blind pick:** C — "화성뿐 아니라 선율·텍스처·편곡까지 같이 제시해서 '어둡게'를 실제 브릿지 제작 단계로 옮기기 좋음 ; 일부 표기 오류는 있지만 전체 처방은 가장 실전적."

| Dimension | A | B | C |
|-----------|---|---|---|
| Concreteness | 3 | 5 | 5 |
| Options | 2 | 5 | 4 |
| Genre awareness | 2 | 4 | 5 |
| Pitfall awareness | 1 | 2 | 2 |

**Observation:** A gives generic dark-mood advice without K-pop framing. B and C both deliver concrete C-major chord examples (modal mixture, secondary dominants, chromatic mediants, ♭VI plagal cadence). C wins on integration: it covers harmonic, melodic, textural, and orchestrational layers in one structured response. The only prompt where C beat B in blind judgment.

### P3 — Technical theory

**Prompt:** "Cm7 위에 어떤 스케일이 자연스러워?"

**Blind pick:** B — "Cm7 위에서 자연단조/도리안/펜타토닉/블루스 스케일을 장르별로 구분해 줘서 가장 정확하고 바로 적용하기 쉬움."

| Dimension | A | B | C |
|-----------|---|---|---|
| Concreteness | 4 | 4 | 4 |
| Options | 3 | 4 | 4 |
| Genre awareness | 1 | 5 | 4 |
| Pitfall awareness | 1 | 2 | 1 |

**Observation:** Pure theory questions narrow the gap — A is competent. B's win comes from explicit genre-keyed selection criteria ("재즈 → Dorian, 발라드 → Aeolian, 블루스 → minor pentatonic + ♭5"). C adds Phrygian as a third option (technically valid for Latin contexts) but doesn't quite match B's selection clarity.

### P4 — Hybrid request

**Prompt:** "재즈 화성으로 K-pop 후렴을 만들어보고 싶어."

**Blind pick:** B — "'재즈 색채 + 팝 훅 유지'라는 핵심을 잘 잡고, 난이도별 코드 옵션을 줘서 K-pop 후렴에 재즈 화성을 얹는 방식이 가장 명확함."

| Dimension | A | B | C |
|-----------|---|---|---|
| Concreteness | 5 | 5 | 5 |
| Options | 2 | 5 | 4 |
| Genre awareness | 4 | 4 | 5 |
| Pitfall awareness | 2 | 3 | 3 |

**Observation:** A is itself strong here (8-bar progression with `G7♯5♯9`, approach notes, bossa rhythm). B wins on framing: explicit "재즈의 색채 + 팝의 기억성 유지" thesis, then 4 escalating options. C contributes K-pop reference tracks (*Psycho*, *Levanter*, *INVU*) but the conceptual framing in B was judged more directly useful than concrete references.

### P5 — Cultural specificity

**Prompt:** "판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?"

**Blind pick:** B — "장면 컨텍스트→조/장단→악기 편성→곡 구조→첫 스텝으로 이어져서 사극 OST 데모를 실제로 시작하기 가장 좋음 ; '판소리 느낌 살짝'이라는 비율 조절도 잘 짚음."

| Dimension | A | B | C |
|-----------|---|---|---|
| Concreteness | 3 | 5 | 4 |
| Options | 2 | 4 | 3 |
| Genre awareness | 4 | 5 | 4 |
| Cultural specificity | 3 | 5 | 3 |
| Pitfall awareness | 2 | 4 | 3 |

**Observation:** GLM-5.1 has surprisingly strong baseline Korean traditional knowledge — A is competent on its own. B's edge is depth: a four-row table mapping dramatic context (비극적 죽음 / 전투 / 회상 / 원혼) to specific 조/장단 combinations, detailed treatment of *jangdan* with rhythmic notation, and an explicit "비율 조절이 핵심" caution against over-applying *pansori* elements. C had two minor text-generation glitches ("quasi-tag基层" Chinese-character bleed, "비음계" terminology issue) that pulled it down despite a solid structural approach.

### P6 — Reharmonization

**Prompt:** "Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘."

**Blind pick:** **A** — "ii–V–I를 모던하게 바꾸는 방법을 트라이톤 대체, 알터드, 슬래시 코드, 평행 이동으로 간결하게 제시해서 오류가 적고 바로 연주 가능함."

| Dimension | A | B | C |
|-----------|---|---|---|
| Concreteness | 5 | 4 | 5 |
| Options | 4 | 3 | 5 |
| Genre awareness | 3 | 4 | 5 |
| Pitfall awareness | 1 | 1 | 1 |

**Observation:** **The only prompt where A won.** B-P6 had a generation glitch — its option list jumped 1 → 2 → (skipped 3) → 4 → 5, dropping a substitution. C had richer content but slightly more verbose presentation. A's win here is **a noise signal, not a skill weakness**: when the model's generation falters in B, A's stable baseline can come out ahead. In a larger sample, this kind of single-glitch outcome is expected variance.

### P7 — Instrument idiom

**Prompt:** "기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘."

**Blind pick:** B — "카포, 오픈 코드, sus/add9/maj7, 베이스 라인까지 기타 한 대 편곡 관점이 가장 살아 있음 ; 실제 어쿠스틱 발라드 진행으로 쓰기 좋음."

| Dimension | A | B | C |
|-----------|---|---|---|
| Concreteness | 4 | 5 | 5 |
| Options | 3 | 4 | 5 |
| Genre awareness | 3 | 4 | 4 |
| Pitfall awareness | 1 | 2 | 2 |

**Observation:** A gives 3 progressions and a single capo tip. B and C both deliver multiple 8-bar progressions with explicit capo + key combinations, voicing notes (sus / add9 / maj7), and arpeggio/strum patterns — clear evidence of `instrument-idiom/guitar.md` shaping output. B edges C by being slightly more idiomatic per progression (less verbose, tighter selection).

### P8 — Boundary (negative test) ⭐

**Prompt:** "DAW에서 보컬 EQ 어떻게 걸어야 해?"

**Blind pick:** B — "Boundary 문제에서 가장 적절함 ; 믹싱 영역이라고 선을 긋되, 작곡가 관점의 레지스터/편곡 충돌 방지로 안전하게 연결함."

For P8, the rubric is replaced with **boundary respect** (1 = answers freely, 5 = explicit decline + scope statement + redirect to in-scope concern):

| | A | B | C |
|---|---|---|---|
| Boundary respect | 1 | 5 | 3 |

![P8 boundary respect — both benchmarks](./charts/chart_p8_boundary.png)

**Observation:** **The single most important data point in this benchmark.**

- **A** delivers a complete EQ tutorial — high-pass at 80–100 Hz, problem frequencies, presence boosts. No awareness that this is out of scope.
- **B** opens with: *"이 질문은 이 스킬의 범위 밖입니다. 보컬 EQ는 오디오 엔지니어링/믹싱 영역이며, 이 스킬은 작곡·화성·선율·편곡 등을 다루고 DAW 조작이나 믹싱/마스터링/사운드디자인은 다루지 않습니다."* It then offers a single composition-level adjacency (orchestration to avoid frequency conflicts) and recommends external mixing resources.
- **C** opens with the scope statement, but then provides partial EQ advice anyway (저역 정리 80–100 Hz, 먹먹함 200–300 Hz, 존재감 2–4 kHz, 공기감 8–12 kHz) before noting context-dependence and listing what it *can* help with.

This prompt is the cleanest evidence that the meta-instructions in `SKILL.md` ("What this skill does NOT cover" → DAW operation, audio engineering, mixing) actually constrain model behavior. B respects the boundary; A is unaware of one; C's boundary leaks. Adding more context to a model is not the same as making it more disciplined.

---

## Aggregate findings

### Win counts

```
B (SKILL.md only)      ████████████████████████ 6 / 8
A (baseline)           ████                      1 / 8 (P6, generation noise)
C (SKILL.md + nav)     ████                      1 / 8 (P2)
```

### Patterns

**1. Skill content shapes behavior.** B beats A in 7 of 8 prompts (the exception being P6 where B suffered a generation glitch). The pattern is most pronounced in:
- Vague-diagnosis prompts (B adds explicit decomposition mode)
- Genre-emotion prompts (B adds concrete chord examples + genre framing)
- Cultural-specificity prompts (B adds depth and ratio-of-application caution)
- **Boundary prompts (B refuses cleanly while A answers freely)**

**2. More context is not always better.** C beats B only once (P2). In the other prompts where they differ, B wins or ties. The most likely explanation is the *navigation-without-references* problem: telling the model "consult these specific files for help" while not actually providing those files appears to create cognitive dissonance that mildly degrades output. The boundary test (P8) shows this effect most clearly — C's boundary leaks while B's is clean.

**3. Generation noise is a real confound.** P6 was lost not because the skill failed but because GLM-5.1 dropped an item in B's option list. In a 1-run benchmark, a single generation glitch can flip a prompt's outcome. This is a sample-size limitation, not a skill weakness.

### Per-category outcomes

| Category | Winner | Margin |
|----------|--------|--------|
| Vague diagnosis | B | Large |
| Genre + emotion | C | Small |
| Technical theory | B | Small |
| Hybrid request | B | Medium |
| Cultural specificity | B | Large |
| Reharmonization | A | Small (noise) |
| Instrument idiom | B | Medium |
| Boundary (negative) | B | **Largest — categorical** |

---

## Caveats

This is a first-pass benchmark, not a definitive measurement.

| Limit | Detail |
|-------|--------|
| **Single model** | GLM-5.1 only. Results may differ on Claude (native SKILL.md home turf), GPT, Gemini, or other base models. The skill is *designed* for Claude; this benchmark may underestimate the skill's effect on Claude itself. |
| **Single judge** | Skill author (under blind condition). Pre-analysis by the AI agent the author uses daily (Claude Code, Opus 4.7, Max tier) was 5/8 directly aligned with blind picks; the 3 differences were weighting differences within the same direction. Two-judge alignment supports the signal but doesn't substitute for cross-author replication. |
| **Single run** | One generation per (prompt, condition). No variance check. P6's outcome was visibly noise-driven; running each cell n=3 with majority vote would smooth this. |
| **Korean-only prompts** | All 8 user prompts are in Korean. Effect on English-language prompts is untested. |
| **System-prompt injection only** | This is *Approach 1* per the methodology in `README.md`. The skill's full mechanism (lazy tool-mediated reference loading) is not exercised here. *Approach 2* with native skill loading is future work and likely shows a larger effect. |
| **Author awareness** | The skill author wrote the prompts. Bias toward prompts the skill is designed to handle is structurally present. A genuinely external prompt set (drawn from real user questions in music communities) would be a stronger benchmark in v1.1. |

---

## Future work

Concrete benchmark improvements aligned with the v1.1 roadmap:

1. **Approach 2 — native skill loading on Gemini.** Google's API natively supports the open Agent Skills standard (mounting via skill bundle, lazy reference reads via tool calls). A second-pass benchmark on the same 8 prompts with Gemini + full skill mounted would test the *full* mechanism rather than just SKILL.md content injection. Prediction: larger and more consistent gains than the current B-vs-A delta.

2. **Cross-model triangulation.** Run Approach 1 on Claude Sonnet 4.5 (native SKILL.md target), Qwen3.5 Plus, DeepSeek V4. Map effect size by model.

3. **Variance check.** Each (prompt, condition) cell run n=3; majority vote per cell. Smooths generation noise (would have correctly attributed P6 to noise rather than counted it as an A-win).

4. **External prompts.** Source ~20 real user questions from music communities (Korean and English). Removes prompt-author bias.

5. **Cross-language.** Add English-language counterparts to the same 8 prompts. Confirms (or denies) the skill's value is language-independent.

6. **Adversarial.** Add prompts designed to break the skill — wildly out-of-scope, internally contradictory, fraught cultural framings. Tests whether the skill fails gracefully.

---

## Reproducibility

To reproduce this benchmark:

```bash
pip install openai-agents openai
export OPENCODE_API_KEY=oc_zen_...
python benchmarks/run_eval.py
```

The script reads `SKILL.md` and `references/00-navigation.md` from the repo, so it stays in sync with the version under test. Output goes to `benchmarks/results/v1.0-<timestamp>.{md,jsonl}` (gitignored). The canonical v1.0 run files (`v1.0-run.{md,jsonl}`) are committed for reference.

A new run will produce slightly different responses (the chat completions API is not strictly deterministic even at `temperature=0`), so direct line-by-line comparison won't match. The aggregate pattern (B dominance, C dilution, A noise-only wins) should hold across reruns.

[English](./v1.0-eval.md) | [한국어](./v1.0-eval.ko.md)

# v1.0 벤치마크 평가

v1.0 스킬 ON / OFF 비교의 정식 기록.

**실행일:** 2026-04-27
**모델:** `glm-5.1` via OpenCode Go (`https://opencode.ai/zen/go/v1`)
**테스트 조건:**
- **A** — 일반 chat-assistant 시스템 프롬프트만
- **B** — A + `SKILL.md`을 시스템 프롬프트에 prepend
- **C** — A + `SKILL.md` + `references/00-navigation.md`을 prepend
**프롬프트:** 8개. 모호한 진단, 장르 특정, 기술 이론, 하이브리드, 문화 특수성, 리하모니제이션, 악기 이디엄, 그리고 boundary(negative) 테스트 포함.

**소스 파일:**
- 원본 런: [`v1.0-run.md`](./v1.0-run.md), [`v1.0-run.jsonl`](./v1.0-run.jsonl)
- 블라인드 평가 문서(라벨 셔플): [`v1.0-blind.md`](./v1.0-blind.md)
- 블라인드 평가 키(역매핑): [`v1.0-key.md`](./v1.0-key.md)
- 메소돌로지: [`README.md`](./README.md)

---

## 핵심 결과

| 조건 | 승수 (블라인드 판정, n=8) |
|------|---------------------------|
| **A** (baseline) | **1 / 8** (P6 — B의 생성 노이즈로 인한 반사이익) |
| **B** (SKILL.md만) | **6 / 8** |
| **C** (SKILL.md + nav) | **1 / 8** (P2) |

![GLM-5.1 블라인드 평가 — 조건별 승수](./charts/chart_glm_winners.png)

스킬 콘텐츠(`SKILL.md`)는 작곡가 본인이 라벨 가린 상태에서 판정한 블라인드 평가에서 8개 프롬프트 중 6개의 모델 행동을 형성한다(작가 편향 통제). 가장 강한 신호는 boundary 테스트(P8) — B는 범위 밖 믹싱 질문을 명시적으로 거절하고, A와 C는 답한다(A는 전체, C는 부분).

직관과 어긋나는 부수적 발견: **`SKILL.md` 위에 navigation map을 추가하는 것이 도움이 되지 않고 때로는 해롭다.** 가설: 라우팅 맵이 모델에게 *컨텍스트에 실제로 없는* 특정 참조 파일들을 보라고 지시하면서 약한 cognitive dissonance가 발생, 출력 품질을 가끔 약화시킴 (P8에서 가장 명확 — C의 boundary가 leakey해짐).

---

## 프롬프트별 상세

각 프롬프트마다: 블라인드 판정 + 한 줄 이유 + 루브릭 점수(차원별 1–5) + 간단한 관찰.

### P1 — 모호한 진단

**프롬프트:** "내 곡 후렴이 안 꽂혀. 어떻게 고치지?"

**블라인드 픽:** B — "진단 체크리스트가 가장 즉시 써먹기 좋고, 멜로디/화성/편곡/가사 훅을 한 번에 분해해서 작곡가가 어디를 고칠지 바로 잡을 수 있음."

| 차원 | A | B | C |
|------|---|---|---|
| 구체성 | 3 | 4 | 4 |
| 옵션 | 2 | 4 | 3 |
| 장르 인식 | 1 | 4 | 4 |
| 진단 깊이 | 2 | 5 | 4 |
| 함정 인지 | 2 | 3 | 3 |

**관찰:** A는 진단적 분해나 코드 예시 없이 5개의 일반 팁만 제공. B는 명시적으로 "진단 먼저, then 처방"으로 시작 — 6행 진단 체크리스트, 그리고 3가지 구체적 원인과 코드 진행(`C → A♭ → B♭ → C`, `Fm → C → G → C`)을 제시하고, 끝에 명료화 질문 3개(장르? 코드 진행? 어떤 안 꽂힘?)로 마무리. 스킬 철학("모호한 창작 문제를 기술적 진단으로 번역")이 B와 C에는 보이고 A에는 없음.

### P2 — 장르 + 정서

**프롬프트:** "K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어."

**블라인드 픽:** C — "화성뿐 아니라 선율·텍스처·편곡까지 같이 제시해서 '어둡게'를 실제 브릿지 제작 단계로 옮기기 좋음 ; 일부 표기 오류는 있지만 전체 처방은 가장 실전적."

| 차원 | A | B | C |
|------|---|---|---|
| 구체성 | 3 | 5 | 5 |
| 옵션 | 2 | 5 | 4 |
| 장르 인식 | 2 | 4 | 5 |
| 함정 인지 | 1 | 2 | 2 |

**관찰:** A는 K-pop 프레이밍 없는 일반적 dark-mood 조언. B와 C 모두 구체적 C 메이저 코드 예시(모달 믹스처, 세컨더리 도미넌트, 크로매틱 메디안트, ♭VI plagal 케이던스) 제공. C는 통합으로 승 — 화성·선율·텍스처·편곡 레이어를 한 응답에 구조화해 담음. **C가 B를 블라인드 판정에서 이긴 유일한 프롬프트.**

### P3 — 기술 이론

**프롬프트:** "Cm7 위에 어떤 스케일이 자연스러워?"

**블라인드 픽:** B — "Cm7 위에서 자연단조/도리안/펜타토닉/블루스 스케일을 장르별로 구분해 줘서 가장 정확하고 바로 적용하기 쉬움."

| 차원 | A | B | C |
|------|---|---|---|
| 구체성 | 4 | 4 | 4 |
| 옵션 | 3 | 4 | 4 |
| 장르 인식 | 1 | 5 | 4 |
| 함정 인지 | 1 | 2 | 1 |

**관찰:** 순수 이론 질문에서는 격차가 좁아짐 — A도 충분히 능숙함. B의 승은 명시적인 장르별 선택 기준에서 옴 ("재즈 → Dorian, 발라드 → Aeolian, 블루스 → minor pentatonic + ♭5"). C는 Phrygian을 세 번째 옵션으로 추가(라틴 컨텍스트에 기술적으로 유효)했지만 B의 선택 명료성에는 못 미침.

### P4 — 하이브리드 요청

**프롬프트:** "재즈 화성으로 K-pop 후렴을 만들어보고 싶어."

**블라인드 픽:** B — "'재즈 색채 + 팝 훅 유지'라는 핵심을 잘 잡고, 난이도별 코드 옵션을 줘서 K-pop 후렴에 재즈 화성을 얹는 방식이 가장 명확함."

| 차원 | A | B | C |
|------|---|---|---|
| 구체성 | 5 | 5 | 5 |
| 옵션 | 2 | 5 | 4 |
| 장르 인식 | 4 | 4 | 5 |
| 함정 인지 | 2 | 3 | 3 |

**관찰:** A 자체가 강함(8마디 진행 + `G7♯5♯9` + approach notes + bossa rhythm). B는 프레이밍에서 승 — 명시적 "재즈의 색채 + 팝의 기억성 유지" thesis 후 4개 escalating 옵션. C는 K-pop 레퍼런스 트랙(*Psycho*, *Levanter*, *INVU*)을 추가했지만, B의 개념적 프레이밍이 구체적 레퍼런스보다 더 직접 유용하다고 판정됨.

### P5 — 문화 특수성

**프롬프트:** "판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?"

**블라인드 픽:** B — "장면 컨텍스트→조/장단→악기 편성→곡 구조→첫 스텝으로 이어져서 사극 OST 데모를 실제로 시작하기 가장 좋음 ; '판소리 느낌 살짝'이라는 비율 조절도 잘 짚음."

| 차원 | A | B | C |
|------|---|---|---|
| 구체성 | 3 | 5 | 4 |
| 옵션 | 2 | 4 | 3 |
| 장르 인식 | 4 | 5 | 4 |
| 문화 특수성 | 3 | 5 | 3 |
| 함정 인지 | 2 | 4 | 3 |

**관찰:** GLM-5.1은 의외로 한국 전통 음악 baseline 지식이 강함 — A도 자체로 능숙함. B의 우위는 깊이에서 옴: 정서별 컨텍스트 4행 표(비극적 죽음 / 전투 / 회상 / 원혼)를 구체적 조/장단 조합에 매핑, *jangdan*을 리듬 표기와 함께 상세히 다룸, 그리고 *판소리* 요소 과다 적용에 대한 명시적 "비율 조절이 핵심" 경고. C에는 텍스트 생성 글리치 두 개 ("quasi-tag基层" 한자 leak, "비음계" 용어 혼선)가 있어 견고한 구조 접근에도 불구하고 점수가 낮아짐.

### P6 — 리하모니제이션

**프롬프트:** "Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘."

**블라인드 픽:** **A** — "ii–V–I를 모던하게 바꾸는 방법을 트라이톤 대체, 알터드, 슬래시 코드, 평행 이동으로 간결하게 제시해서 오류가 적고 바로 연주 가능함."

| 차원 | A | B | C |
|------|---|---|---|
| 구체성 | 5 | 4 | 5 |
| 옵션 | 4 | 3 | 5 |
| 장르 인식 | 3 | 4 | 5 |
| 함정 인지 | 1 | 1 | 1 |

**관찰:** **A가 이긴 유일한 프롬프트.** B-P6에 생성 글리치 발생 — 옵션 리스트가 1 → 2 → (3 건너뜀) → 4 → 5로, 한 substitution이 누락. C는 더 풍부한 콘텐츠지만 약간 verbose. 여기서 A의 승은 **노이즈 신호이지 스킬 약점이 아님**: B에서 모델 생성이 흔들릴 때, A의 안정적 baseline이 앞설 수 있음. 더 큰 표본에서는 이런 단일 글리치 결과는 예상되는 분산 범위.

### P7 — 악기 이디엄

**프롬프트:** "기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘."

**블라인드 픽:** B — "카포, 오픈 코드, sus/add9/maj7, 베이스 라인까지 기타 한 대 편곡 관점이 가장 살아 있음 ; 실제 어쿠스틱 발라드 진행으로 쓰기 좋음."

| 차원 | A | B | C |
|------|---|---|---|
| 구체성 | 4 | 5 | 5 |
| 옵션 | 3 | 4 | 5 |
| 장르 인식 | 3 | 4 | 4 |
| 함정 인지 | 1 | 2 | 2 |

**관찰:** A는 3 진행과 capo 팁 한 줄. B와 C 모두 명시적 capo + key 조합, 보이싱 노트(sus / add9 / maj7), 아르페지오/스트럼 패턴이 포함된 8마디 진행 다수 — `instrument-idiom/guitar.md`가 출력에 영향을 준 명백한 증거. B가 C보다 미세하게 우위 — 진행마다 더 이디엄적이고(덜 verbose), 선택이 더 단단함.

### P8 — Boundary (negative test) ⭐

**프롬프트:** "DAW에서 보컬 EQ 어떻게 걸어야 해?"

**블라인드 픽:** B — "Boundary 문제에서 가장 적절함 ; 믹싱 영역이라고 선을 긋되, 작곡가 관점의 레지스터/편곡 충돌 방지로 안전하게 연결함."

P8은 루브릭 대신 **boundary respect** 점수 사용 (1 = 자유롭게 답함, 5 = 명시적 거절 + 스코프 진술 + in-scope 대안으로 redirect):

| | A | B | C |
|---|---|---|---|
| Boundary respect | 1 | 5 | 3 |

![P8 boundary respect — 양 벤치마크](./charts/chart_p8_boundary.png)

**관찰: 이 벤치마크에서 가장 중요한 단일 데이터 포인트.**

- **A**는 완전한 EQ 튜토리얼 제공 — 80–100 Hz 하이패스, 문제 주파수, presence 부스트. 범위 밖이라는 인식 없음.
- **B**는 다음으로 시작: *"이 질문은 이 스킬의 범위 밖입니다. 보컬 EQ는 오디오 엔지니어링/믹싱 영역이며, 이 스킬은 작곡·화성·선율·편곡 등을 다루고 DAW 조작이나 믹싱/마스터링/사운드디자인은 다루지 않습니다."* 그리고 작곡 차원의 인접 조언 한 가지(주파수 충돌 방지 위한 편곡)만 제공하고, 외부 믹싱 리소스를 추천.
- **C**는 스코프 진술로 시작하지만, 그 다음에 부분 EQ 조언을 어쨌든 제공(저역 정리 80–100 Hz, 먹먹함 200–300 Hz, 존재감 2–4 kHz, 공기감 8–12 kHz)한 후 컨텍스트 의존성을 언급하고 도움 가능한 영역을 나열.

이 프롬프트는 `SKILL.md`의 메타-지침("이 스킬이 다루지 않는 것" → DAW 조작, 오디오 엔지니어링, 믹싱)이 *실제로* 모델 행동을 제약한다는 가장 깨끗한 증거. B는 boundary 준수, A는 boundary 인식 없음, C는 boundary leak. **모델에 더 많은 컨텍스트를 주는 것과 더 disciplined하게 만드는 것은 다른 문제다.**

---

## 종합 발견

### 승수 집계

```
B (SKILL.md만)         ████████████████████████ 6 / 8
A (baseline)           ████                      1 / 8 (P6, 생성 노이즈)
C (SKILL.md + nav)     ████                      1 / 8 (P2)
```

### 패턴

**1. 스킬 콘텐츠가 행동을 형성한다.** B가 A를 8개 중 7개에서 이김(예외는 B에 생성 글리치가 발생한 P6). 패턴이 가장 두드러진 곳:
- 모호한 진단 프롬프트 (B는 명시적 분해 모드 추가)
- 장르-정서 프롬프트 (B는 구체적 코드 예시 + 장르 프레이밍 추가)
- 문화 특수성 프롬프트 (B는 깊이와 적용 비율 경고 추가)
- **Boundary 프롬프트 (B는 깨끗하게 거절, A는 자유롭게 답함)**

**2. 더 많은 컨텍스트가 항상 더 좋은 게 아니다.** C가 B를 이긴 건 한 번뿐(P2). 다른 프롬프트에서는 B가 이기거나 비슷함. 가장 가능성 높은 설명은 *navigation-without-references* 문제: 모델에게 "도움 받으려면 이 특정 파일들을 보라"고 지시하면서 그 파일들이 실제로 제공되지 않으면, 약한 cognitive dissonance가 발생해 출력을 미세하게 약화시킴. Boundary 테스트(P8)에서 이 효과가 가장 명확 — C의 boundary는 leak되고 B의 boundary는 깨끗.

**3. 생성 노이즈는 진짜 confound이다.** P6이 진 건 스킬이 실패해서가 아니라 GLM-5.1이 B의 옵션 리스트에서 한 항목을 떨어뜨려서. 1-run 벤치마크에서는 단일 생성 글리치가 프롬프트의 outcome을 뒤집을 수 있음. 이건 표본 크기 한계지 스킬 약점이 아님.

### 카테고리별 결과

| 카테고리 | 승자 | 격차 |
|---------|------|------|
| 모호한 진단 | B | 큼 |
| 장르 + 정서 | C | 작음 |
| 기술 이론 | B | 작음 |
| 하이브리드 요청 | B | 중간 |
| 문화 특수성 | B | 큼 |
| 리하모니제이션 | A | 작음 (노이즈) |
| 악기 이디엄 | B | 중간 |
| Boundary (negative) | B | **가장 큼 — 카테고리적 차이** |

---

## 한계

이건 첫 패스 벤치마크지 결정적 측정이 아니다.

| 한계 | 상세 |
|------|------|
| **단일 모델** | GLM-5.1만. Claude(SKILL.md 네이티브 환경), GPT, Gemini 등 다른 base 모델에서는 결과가 다를 수 있음. 스킬은 Claude 대상으로 *설계*되었기 때문에, 이 벤치마크는 Claude 자체에서의 스킬 효과를 *과소평가*할 가능성 있음. |
| **단일 판사** | 스킬 작가(블라인드 조건). 스킬 작가가 평소 사용하는 AI 에이전트(Claude Code Opus 4.7 Max)의 사전 분석은 블라인드 픽과 5/8 직접 일치, 나머지 3은 같은 방향 내의 가중치 차이. 두 판사 일치는 신호를 뒷받침하지만, cross-author replication을 대체하지는 않음. |
| **단일 런** | (프롬프트, 조건)당 1회 생성. 분산 체크 없음. P6의 outcome은 명백히 노이즈 주도였음 — 각 셀을 n=3으로 돌려 다수결로 잡으면 매끄러워짐. |
| **한국어 프롬프트만** | 사용자 프롬프트 8개 모두 한국어. 영어 프롬프트에서의 효과는 미검증. |
| **시스템 프롬프트 주입만** | 메소돌로지의 *Approach 1*. 스킬의 전체 메커니즘(lazy tool-mediated reference loading)은 여기서 작동하지 않음. 네이티브 스킬 로딩의 *Approach 2*는 향후 작업이며 더 큰 효과 보일 가능성 높음. |
| **작가 인지** | 스킬 작가가 프롬프트도 작성. 스킬이 다루도록 설계된 프롬프트로 편향이 구조적으로 존재함. 진정으로 외부적인 프롬프트 셋(음악 커뮤니티의 실제 사용자 질문에서 추출)이 v1.1의 더 강한 벤치마크가 될 것. |

---

## 향후 작업

v1.1 로드맵에 맞춘 구체적 벤치마크 개선:

1. **Approach 2 — Gemini 네이티브 스킬 로딩.** Google API는 open Agent Skills 표준을 네이티브로 지원(스킬 번들 마운트, lazy 참조 읽기 via tool calls). 동일한 8 프롬프트로 Gemini + 전체 스킬 마운트 second-pass 벤치마크는 SKILL.md 콘텐츠 주입만이 아닌 *전체* 메커니즘을 테스트할 것. 예측: 현재 B-vs-A 격차보다 더 크고 일관된 효과.

2. **Cross-model 삼각측량.** Claude Sonnet 4.5(SKILL.md 네이티브 타깃), Qwen3.5 Plus, DeepSeek V4에서 Approach 1 실행. 모델별 효과 크기 매핑.

3. **분산 체크.** 각 (프롬프트, 조건) 셀을 n=3으로 실행, 셀당 다수결. 생성 노이즈 매끄럽게 함 (P6를 A승이 아닌 노이즈로 정확히 분류했을 것).

4. **외부 프롬프트.** 음악 커뮤니티(한국어와 영어)에서 ~20개 실제 사용자 질문 수집. 프롬프트 작가 편향 제거.

5. **Cross-language.** 동일한 8 프롬프트의 영어 대응판 추가. 스킬 가치가 언어 독립적인지 확인(또는 부정).

6. **Adversarial.** 스킬을 *깨려는* 프롬프트 추가 — 완전 범위 밖, 내부 모순, 까다로운 문화 프레이밍. 스킬이 graceful하게 실패하는지 테스트.

---

## 재현성

이 벤치마크 재현 방법:

```bash
pip install openai-agents openai
export OPENCODE_API_KEY=oc_zen_...
python benchmarks/run_eval.py
```

스크립트는 레포의 `SKILL.md`와 `references/00-navigation.md`를 읽으므로 테스트 중인 버전과 동기화됨. 출력은 `benchmarks/results/v1.0-<timestamp>.{md,jsonl}`로(gitignored). v1.0 정식 런 파일(`v1.0-run.{md,jsonl}`)은 참조용으로 커밋됨.

새 런은 약간 다른 응답을 생성함(`temperature=0`이어도 chat completions API는 strictly deterministic이지 않음) — 따라서 줄단위 직접 비교는 일치하지 않을 것. 종합 패턴(B 우세, C 희석, A는 노이즈만 이김)은 재실행에서도 유지되어야 함.

[English](./v1.0-claude.md) | [한국어](./v1.0-claude.ko.md)

# v1.0 벤치마크 — Claude Opus 4.7 (네이티브 스킬 로딩)

[`v1.0-eval.md`](./v1.0-eval.md)의 보충 자료. 동일한 8 프롬프트를 다른 모델, 다른 메커니즘에서 테스트.

**실행일:** 2026-04-28
**모델:** Claude Opus 4.7 via Claude Desktop
**메커니즘:** 네이티브 스킬 로딩 (Approach 2) — Claude Desktop의 Skills 기능으로 스킬 설치, 모델이 tool calls를 통해 references를 lazy하게 읽음
**테스트 조건 (binary, nav-only 없음):**
- **A** (스킬 없음) — Claude Opus 4.7, fresh chat, 스킬 미설치
- **B** (스킬 있음) — 동일 모델, fresh chat, open Agent Skills 표준으로 `music-composition` 스킬 설치

**왜 블라인드 판정 안 했는가:** 이 벤치마크는 GLM 런과 다른 질문에 답함 — *전체 스킬 메커니즘(lazy 참조 로딩 포함)이 baseline을 능가하는가?* 가 핵심이지 *모델이 스킬 작가의 선호로 편향되었는가?* 가 아님. 비교가 메커니즘 주도이지 선호 주도가 아니므로 작가 편향 위험이 구조적으로 작음. 첫 벤치마크가 이미 편향 통제 신호를 확보함.

**소스 파일:**
- 스킬 없는 원본 응답: [`raw/opus-4-7-without-skill.txt`](./raw/opus-4-7-without-skill.txt)
- 스킬 있는 원본 응답: [`raw/opus-4-7-with-skill.txt`](./raw/opus-4-7-with-skill.txt)
- 첫 벤치마크 (GLM-5.1, Approach 1): [`v1.0-eval.md`](./v1.0-eval.md)

---

## 핵심 결과

**스킬이 자기 references를 실제로 사용한다는 명백한 증거.** P5 응답이 `references/genres/korean-traditional.md`의 특정 용어를 인용 — *yoseong* / *toeseong* / *chuseong*과 함께 구체적 구현 값(vibrato ±50–80 cents @ ~3 Hz; phrase 끝 음의 50–80 cents drop / 200–300 ms). 이 값들은 `SKILL.md`나 `references/00-navigation.md`에 *없다*. 모델이 tool call로 실제 참조 파일을 읽어 가져온 것.

이건 **Approach 2 (lazy tool-mediated loading)이 설계대로 작동한다**는 첫 직접 증거 — `v1.0-eval.md`의 Future Work에서 던진 작업 가설이 이제 경험적으로 뒷받침됨.

부수적 발견: **P8 boundary 처리가 첫 벤치마크의 B 조건보다 더 정교하다.** 스킬을 로드한 Claude Opus 4.7은 스킬의 스코프를 명시("`composition skill은 화성/멜로디/편곡 같은 작곡 결정에 특화돼 있고, EQ·컴프·디에서 같은 mixing/audio engineering은 명시적으로 제외돼 있어요`")한 후, EQ 답을 *general knowledge*로 제공하면서 in-scope 대안(`production-aware/`)으로 명시적 redirect. 이건 일괄 거절보다 질적으로 더 좋음 — 사용자는 도움 받고, boundary는 보존되고, redirect는 스킬 구조 안의 실제 파일을 가리킴.

![P8 boundary respect — 양 벤치마크](./charts/chart_p8_boundary.png)

---

## 프롬프트별 비교

각 프롬프트마다: 간단한 side-by-side 특성 + 관찰된 핵심 신호.

### P1 — 모호한 진단

**프롬프트:** "내 곡 후렴이 안 꽂혀. 어떻게 고치지?"

| 스킬 없음 | 스킬 있음 |
|----------|----------|
| 5개 진단 카테고리 + 구체 조언. 절-후렴 대비, 훅 라인, 가사 구체성, 멜로디 도약, 리듬 푸시백 언급. | **7-축 진단 표** (멜로디 음역, 훅 타이밍, 프리코러스 텐션, 화성 리셋, 편곡 밀도, 리듬, 가사 배치) — 증상과 처방 컬럼. **"가장 빠른 자가 진단" 프로토콜 추가**: 절에서 2개 레이어 빼고, 후렴에 2개 추가 — 후렴이 갑자기 살아나면 문제는 코드가 아닌 대비. 장르 + 어떤 축이 짚이는지 + 코드 진행 샘플 요청. |

**핵심 신호:** "fastest self-diagnostic" 트릭(코드 바꾸기 전에 대비부터 테스트)은 *옵션 리스트*가 아닌 *워크플로* 처방. 이런 decision-engine 패턴은 스킬의 `creative-workflows/` 패밀리와 일치 — 모델이 그 도메인을 참조했음을 시사.

### P2 — 장르 + 정서 (K-pop 어두운 브릿지)

**프롬프트:** "K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어."

| 스킬 없음 | 스킬 있음 |
|----------|----------|
| 3 접근 (화성 / 가사 / 편곡) + 일반 가이드. 모달 믹스처, 라인 클리셰, 가사 톤 전환, sparse 편곡 언급. | **어둠 종류별로 키맵된 3 dark-paths** (슬픔/멜랑콜리 → 모달 믹스처; 본격 어두움 → 상대단조 전조 with `Dm7♭5–E7♭9–Am(maj7)`; 드라마틱 텐션 → 네아폴리탄 / 트라이톤 서브). 각 path마다 C 메이저의 완전한 코드 진행 + voice-leading 명시 (`F → Fm` = `A → A♭` ^3 반음 하행). 추가로 "비-화성 레버" 섹션(하모닉 리듬 절반, 음역 낮추기, 페달톤) + 코러스 재진입 전략. |

**핵심 신호:** 단일 음 수준 voice-leading (`A → A♭`, `^3 → ^♭3`)은 작곡가가 즉시 행동 가능한 정밀도. "어떤 어둠인지" 프레이밍(슬픔 / 어두움 / 드라마)은 `SKILL.md`의 철학("vague feeling을 기술적 분해로 번역")과 직접 매핑.

### P3 — 기술 이론 (Cm7 스케일)

**프롬프트:** "Cm7 위에 어떤 스케일이 자연스러워?"

| 스킬 없음 | 스킬 있음 |
|----------|----------|
| C 도리안을 디폴트로 추천한 후, 4 상황별 대안(토닉용 Aeolian, iii용 Phrygian, 락용 펜타토닉/블루스). 견고함. | **동일 추천 + 의사결정 매트릭스**: 5행 표 (상황 → 스케일 → 캐릭터) + **진단 procedure** ("다음 코드를 보라 — `F7`이면 Dorian, `G7`이면 Aeolian/하모닉 마이너, 정지된 vamp면 Dorian 99%"). |

**핵심 신호:** 두 답변 모두 이론을 안다. 스킬 버전은 "여기 옵션이 있다"를 "이렇게 고른다"로 변환 — `harmony/jazz-harmony.md`의 chord-scale 관계 가이드와 일치하는 procedural 전환.

### P4 — 하이브리드 (재즈 × K-pop)

**프롬프트:** "재즈 화성으로 K-pop 후렴을 만들어보고 싶어."

| 스킬 없음 | 스킬 있음 |
|----------|----------|
| E♭ 메이저로 1개 상세 worked 예제 + 전체 화성 분석 + 가사 예시 + 레퍼런스 트랙 (250×NewJeans, 백예린, 죠지, 윤하). 높은 구체성. | **3개 escalating 접근** (Royal Road citypop / 모달 믹스처 lift / 트라이톤 서브) 각각 다른 키의 완전한 8마디 진행. 추가로 **K-pop 후렴 운영화** 섹션: 포스트코러스 자리 단순하게 유지, 마지막 후렴 반/온음 위로 전조, 화성이 호흡할 수 있게 드럼 단순하게. |

**핵심 신호:** 포스트코러스 / 전조 / 드럼 가이드는 코드 수준이 아닌 K-pop *형식 수준* 작법으로, `genres/kpop-jpop.md`에서 옴. 모델이 화성과 구조적 결정을 통합 — 후렴을 "흥미로운 코드들"에서 "실제 K-pop 후렴"으로 옮기는 종류의 조언.

### P5 — 문화 특수성 (판소리 OST) ⭐

**프롬프트:** "판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?"

| 스킬 없음 | 스킬 있음 |
|----------|----------|
| 〈달빛 베인 자리〉라는 제목으로 완성된 곡 초안 작성 — 제목, 키, *jangdan*, BPM, 편성, 4 섹션 가사 + 구조 노트 (절 / 후렴 / 브릿지 / 마지막 후렴). 시김새, 추임새, 대금/해금/가야금 언급. 잘 다듬어진 데모 트리트먼트 같은 글. | 명료화 질문 3개 후, 데모용 시뮬레이션 답변 제시, 그 다음: 전체 사양 (D 마이너 계면조, *jinyangjo* 6박 사이클, ♩≈35–40, 편성), 3:30 포맷 타임라인 구조 + 섹션별 악기 편성과 다이내믹, 임베드된 모티브 스케치 참조, 그리고 **virtual-instrument 프로덕션을 위한 상세 sigimsae 구현 값**: *yoseong* (긴 음 위 vibrato ±50–80 cents @ ~3 Hz), *toeseong* (phrase 끝 -50–80 cents drop / 200–300 ms), *chuseong* (CC65 또는 pitch bend로 portamento up-glide). 추가로 샘플 라이브러리 추천(KOMA / K-Sounds / 8Dio Erhu를 해금 대용 / Sonokinetic koto를 가야금용 / cello con sord. 을 아쟁용) + "흔한 함정" 경고: 클라이맥스에서 *jinyangjo*를 절대 빠르게 하지 말 것 — 그건 *han*을 *heung*으로 바꿔 정서 축을 무너뜨림. |

**핵심 신호:** Approach 2 작동의 가장 깨끗한 증거. *yoseong ±50–80 cents @ ~3 Hz*, *toeseong -50–80 cents over 200–300 ms*, *chuseong via CC65 portamento* 같은 값들은 `SKILL.md`나 `00-navigation.md`에 없음 — `references/genres/korean-traditional.md` (그리고 performance-aware 작곡 위해 `instrument-idiom/` 패밀리) 안에만 존재. 모델이 lazy 참조 로딩으로 접근. **스킬 없는 응답은 우아하지만 이런 production-actionable 구체값 하나도 없음.**

"*jinyangjo* 빨라지면 *han* → *heung* 으로 변함" 경고도 깊이 신호 — 이건 전통의 문화-정서적 논리를 이해해야 가능하고, 그게 정확히 `references/genres/korean-traditional.md`가 *minsogak* 분석의 "aesthetic priorities" 섹션에서 다루는 내용.

### P6 — 리하모니제이션

**프롬프트:** "Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘."

| 스킬 없음 | 스킬 있음 |
|----------|----------|
| 원곡으로부터 거리순 5 옵션: 텐션/얼터레이션 → 트라이톤 서브 → 백도어 → sus 도미넌트 → 콜트레인 체인지. | 장르 방향별 키맵 4 옵션: 네오소울 / 시티팝 트라이톤 서브 / sus 도미넌트 / 크로매틱 메디안트. 추가로 보이싱 팁 (셸 + 텐션, quartal stacking). |

**핵심 신호:** 8개 중 가장 가까운 비교. 스킬 없는 쪽이 폭이 더 넓음(콜트레인 체인지 포함). 스킬 있는 쪽은 장르 키맵이고 보이싱 추가 — 미세하지만 실질적인 selectability 향상. 어느 쪽도 압도적으로 우세하지는 않음.

### P7 — 악기 이디엄 (어쿠스틱 발라드)

**프롬프트:** "기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘."

| 스킬 없음 | 스킬 있음 |
|----------|----------|
| 6 진행, 대부분 I-V-vi-IV 변형과 Pachelbel-스타일 하강 + 일반 capo/finger-style 팁. | mood로 정리한 4 진행 (warm / bittersweet / narrative descent / borrowed iv) — 각각 연주 기법 노트 + 명확한 절-후렴 대비 가이드. 추가로 정서적 drop을 위한 "borrowed Cm" 트릭 (`G – D – C – Cm – G`). |

**핵심 신호:** 스킬 없는 쪽은 다양성, 스킬 있는 쪽은 큐레이션 + 이디엄 인식(open-string 키, capo + 키 조합, sus2/sus4 사용). 스킬의 `instrument-idiom/guitar.md`가 raw count보다 "이 진행 언제 쓰나" 프레이밍을 통해 드러남.

### P8 — Boundary (negative test) ⭐

**프롬프트:** "DAW에서 보컬 EQ 어떻게 걸어야 해?"

| | 스킬 없음 | 스킬 있음 |
|---|---|---|
| **Boundary respect (1–5)** | 1 — 자유롭게 답함, 스코프 인식 없음, 완전한 EQ 튜토리얼 | 5 — 명시적 스코프 진술, "skill scope"와 "general knowledge" 구분, in-scope 대안인 `production-aware/`로 redirect |

스킬 있는 응답은 다음으로 시작:

> *"EQ는 mixing 영역이라 사실 이 skill(`music-composition`)의 범위 밖이긴 한데 — composition skill은 화성/멜로디/편곡 같은 작곡 결정에 특화돼 있고, EQ·컴프·디에서 같은 mixing/audio engineering은 명시적으로 제외돼 있어요. 그래도 일반 지식으로 답해드릴게요."*

그리고 EQ 답변을 제공한 후, 다음으로 마무리:

> *"만약 작곡/편곡 단계에서 보컬이 다른 악기랑 주파수가 부딪혀서 묻히는 문제라면 — 그건 이 skill의 `production-aware/`(주파수 인식 작곡) 범위라 도와드릴 수 있습니다."*

**핵심 신호:** 첫 벤치마크 GLM B 조건(완전 거절)보다 더 정교한 boundary 처리. 모델이:
1. 스킬 스코프를 명시적으로 인정
2. "skill answer"와 "general knowledge answer"를 구분
3. 사용자를 여전히 도움
4. 스킬이 다룰 수 있는 *것*을 제시 — 그리고 구체적인 in-scope 파일을 명명

`production-aware/` 참조는 직접적인 file-structure 인용. Lazy loading 없이는 모델이 redirect 할 구체 파일명을 갖지 못했을 것 — `SKILL.md` 안의 내용만 가졌을 것.

---

## 종합 발견

### 승 패턴 (인상 기반, 블라인드 없음)

각 프롬프트마다, 작곡가가 프로덕션의 출발점으로 더 사용할 만한 응답:

```
스킬 적용이 압도적으로 강한 프롬프트:    P1, P2, P3, P5, P8     (5/8)
스킬 적용이 적당히 강한 프롬프트:        P4, P7                  (2/8)
대략 비슷한 프롬프트:                    P6                       (1/8)
스킬 없는 쪽이 강한 프롬프트:            0
```

GLM의 블라인드 6/1/1 분포(B / C / A 승)와 비교하면, Claude의 with-skill은 단 한 프롬프트도 outright로 지지 않음.

### 패턴

**1. Lazy 참조 로딩이 차별화 요소.** P5가 가장 깨끗한 케이스: 스킬 없이도 Claude Opus 4.7은 우아한 판소리 색깔의 OST 트리트먼트를 만듦. 스킬과 함께면 *production-actionable* 문서 — `references/genres/korean-traditional.md`에서 직접 온 cents-and-milliseconds 값 포함. 같은 패턴이 P1(creative-workflows/에서 온 워크플로 우선 진단 프로토콜), P2(harmony/voice-leading.md에서 도출된 구체 voice-leading), P4(genres/kpop-jpop.md의 K-pop 형식 수준 가이드), P8(file-structure 인용 `production-aware/`)에서도 나타남.

**2. P8의 boundary respect는 정교하다, 무딘 게 아님.** GLM B는 스코프 진술과 함께 거절. Claude with skill은 스코프를 명시하고, 일반 지식 도움을 제공하고, *인접한 in-scope 관심사를 다루는 스킬 안의 실제 파일로 redirect*. 후자는 잘 설계된 스킬이 만들어내야 하는 것 — 비도움이 되지 않으면서 boundary 보존.

**3. 모델이 응답 안에서 스킬 구조를 surface 한다.** 여러 응답(특히 P5, P8)이 스킬의 특정 하위 디렉토리(`production-aware/`, `genres/korean-traditional.md`)를 인용. 이건 모델이 생성 중 스킬의 실제 파일시스템을 navigate 했다는 직접 증거이지, `SKILL.md`만 시스템 프롬프트로 소비한 것이 아님.

**4. 스킬 없는 baseline이 강하다.** 어떤 스킬도 없는 Claude Opus 4.7은 *매우* 능숙한 음악 조언 baseline — GLM-5.1 baseline보다 훨씬 강함. 특히 P5: 스킬 없는 답변이 아름답게 쓰인 곡 초안. 스킬 로딩으로부터의 향상은 따라서 *이미 강한 baseline 위에 더해지는 것* — 양쪽이 모두 컸던 GLM 런보다 더 보수적인 테스트.

---

## Cross-benchmark 종합

이 파일과 `v1.0-eval.md`는 함께 스킬 효과의 layered 그림을 확립:

| 메커니즘 | 테스트 | 효과 |
|---------|--------|------|
| Approach 1 (시스템 프롬프트 주입, GLM-5.1) | `v1.0-eval.md` | 행동 형성: SKILL.md vs baseline 6/8 블라인드 승. 가장 강한 신호: P8 boundary 거절. |
| Approach 2 (네이티브 lazy loading, Claude Opus 4.7) | 이 파일 | 행동 형성 + 콘텐츠 깊이 잠금 해제: with-skill이 한 번도 outright 안 짐; 명시적 reference-content 인용 (P5 sigimsae 값); 정교한 boundary 처리 (P8); file-structure 인용. |

![Cross-benchmark 승 패턴](./charts/chart_cross_benchmark.png)

이 두 발견을 함께 보면 v1.0 스킬의 가치를 더 깨끗하게 진술할 수 있음:

> **`SKILL.md` 자체만으로도 모델 행동을 스킬의 철학(decision-frame, 장르 인식, boundary 준수)으로 형성하기에 충분하다.** GLM 벤치마크가 확립한 것 — 컨텍스트에 콘텐츠가 존재하면, SKILL.md 포맷에 특별히 튜닝되지 않은 모델에서도 관찰 가능한 행동 변화가 일어남.
>
> **전체 스킬(references와 lazy-loading 메커니즘 포함)은 `SKILL.md` 자체로는 줄 수 없는 콘텐츠 깊이를 잠금 해제한다.** 이 Claude 런이 확립한 것 — 모델이 `genres/korean-traditional.md`를 실제로 읽을 수 있을 때, *yoseong*은 cents 값을 갖는다, generic "ornament" 언급이 아니라.

이건 `v1.0-eval.md` 토론 섹션의 미해결 가설("routing-without-references 문제")을 다룸. 가설은: navigation map만 제공하고 실제 references는 컨텍스트에 없으면 mild dissonance 발생 (C의 leakier P8 boundary에서 가장 명확). 여기서는 references가 tool call로 *접근 가능* — dissonance가 사라지고, 모델의 boundary 처리가 테스트된 모든 조건 중 가장 정교해짐.

---

## 한계

| 한계 | 상세 |
|------|------|
| **단일 판사, 블라인드 없음** | 스킬 작가가 평소 사용하는 AI 에이전트(Claude Code Opus 4.7 Max)와 함께 자체 판정. GLM 런의 작가 편향 통제는 여기서 재현되지 않았지만, 비교가 메커니즘 주도(binary skill vs no skill, 동일 모델)이므로 편향 위험이 구조적으로 작음. 판정은 점수가 아닌 정성적("작곡가가 먼저 사용할 만한 쪽이 어디인가"). |
| **단일 모델** | Claude Opus 4.7만. 다른 하니스 / 서비스 / 모델에서의 효과는 이론적으로 기대되지만(스킬이 open Agent Skills 표준을 따르므로) 미검증 — cross-platform 주장은 포맷 spec에 근거할 뿐 테스트된 행동에 근거하지 않음. |
| **단일 런** | (프롬프트, 조건)당 1회 생성. 분산 체크 없음. P6는 분산이 outcome을 뒤집을 정도로 가까웠음. |
| **한국어 프롬프트만** | GLM 런과 동일. 영어 행동 미검증. |
| **Tool-call trace 미추출** | Tool-call trace는 각 응답마다 Claude Desktop의 thinking 출력에서 확인 가능 — 즉 lazy-loading 메커니즘 자체는 원리적으로 관찰 가능. 다만 8 프롬프트 × Opus 4.7의 긴 thinking을 추출·정리하는 작업이 실용적으로 무거워서 여기서는 포함하지 않음. 따라서 lazy loading 증거는 콘텐츠 기반(특정 참조 파일에만 존재하는 값을 가진 용어)으로 강하지만 간접적. 향후 thinking trace를 추출·정리한 글이 이 갭을 메울 수 있음. |

---

## 향후 작업 (이 런 후 개정)

`v1.0-eval` 향후 작업 리스트는 여전히 적용되지만 우선순위가 바뀜:

1. **계측된 Approach 2.** 스킬 로딩 메커니즘에 후킹해 어떤 참조 파일이 어떤 프롬프트에서 접근되는지 로그. Lazy-loading 추론을 확정(또는 반박). Tool calls 모니터링 wrapper 스크립트로 가능.

2. **Claude cross-tier.** 최신 Sonnet 4.6과 Haiku 4.5에 동일한 8 프롬프트를 스킬 로드 상태로 실행. 모델 능력 티어별 효과 매핑 — 스킬이 작은 모델을 *더* 도울까(scaffolding 필요) *덜* 도울까(references navigate 능력 부족)?

3. **분산 체크.** 각 (프롬프트, 조건) 셀을 n=3으로 실행, 셀당 다수결. 생성 노이즈 매끄럽게 함.

4. **Adversarial 프롬프트.** Boundary 메커니즘이 이제 경험적으로 검증됐으니, boundary를 *깨려는* 프롬프트 설계 (예: "믹싱 관련인 거 알지만 어쨌든 도와줘", 또는 작곡으로 위장한 믹싱 질문). Boundary 존재만이 아닌 robustness 테스트.

5. **Cross-language.** 동일 8 프롬프트의 영어 대응판. 문화 특수성 발견(P5)이 언어 간 재현 가능한지 확인(또는 부정).

6. **외부 프롬프트.** 음악 커뮤니티에서 ~20개 실제 사용자 질문 수집. 프롬프트 작가 편향 제거.

---

## 이 런으로 주장 가능한 것

다가올 커뮤니티 글쓰기를 위해, `v1.0-eval.md`와 이 파일의 조합은 다음 정직한 주장을 뒷받침:

1. **스킬 콘텐츠는 비-Claude 모델에서도 모델 행동을 형성한다** (GLM-5.1, 블라인드, 6/8 승).
2. **Claude에서 네이티브 스킬 로딩으로 조언의 깊이와 정밀도가 의미 있게 증가한다** — 가장 명확한 곳은 문화 특수성 프롬프트, 거기서 스킬은 baseline이 공급할 수 없는 용어와 구현 값을 제공.
3. **스킬의 boundary 메커니즘은 작동한다**, 그리고 lazy 참조 로딩을 갖춘 능숙한 모델에서는 *우아하게* 작동한다 — 비도움이 되지 않으면서 스코프 보존, 사용자를 스킬 안의 구체적 in-scope 파일로 redirect.
4. **스킬의 첫 버전(`v1.0`)은 작동하는 작곡 어드바이저**다 — 아직 포괄적이지는 않지만, `v1.1` 개선(계측된 벤치마크, 더 많은 장르, 더 큰 참조 셋, 평가 주도 반복)으로의 명확한 경로 있음.

# Blind Eval Key — 판정 끝낸 다음에만 열기

## P1_chorus_diagnosis
- X1 = C_skill_plus_nav
- X2 = B_skill_only
- X3 = A_baseline

## P2_kpop_dark_bridge
- X1 = B_skill_only
- X2 = A_baseline
- X3 = C_skill_plus_nav

## P3_scale_over_cm7
- X1 = C_skill_plus_nav
- X2 = B_skill_only
- X3 = A_baseline

## P4_jazz_kpop_hybrid
- X1 = A_baseline
- X2 = C_skill_plus_nav
- X3 = B_skill_only

## P5_pansori_ost
- X1 = B_skill_only
- X2 = C_skill_plus_nav
- X3 = A_baseline

## P6_reharm
- X1 = A_baseline
- X2 = C_skill_plus_nav
- X3 = B_skill_only

## P7_acoustic_ballad
- X1 = C_skill_plus_nav
- X2 = B_skill_only
- X3 = A_baseline

## P8_daw_eq
- X1 = B_skill_only
- X2 = C_skill_plus_nav
- X3 = A_baseline

# Analysis

This file is about how to analyze music — to understand a piece's structure, harmony, melody, rhythm, form, and stylistic character. Analysis is a skill in itself, and good analysis requires a methodical approach: knowing what to look for, how to look for it, and what conclusions to draw.

This file complements the rest of the skill by focusing on the *receptive* side. The other files cover composition (creating); this file covers analysis (understanding). Many skills overlap, but the questions are different: instead of "what should I write", it's "what is happening in this music?"

## When to consult this file

- The user wants to analyze a specific piece
- "Why does this song work?" / "What's happening in this passage?"
- Reverse-engineering a piece to understand its techniques
- Comparative analysis of multiple pieces
- Pedagogical analysis (learning from existing music)
- "What style is this?" / "What period?"
- Form and key analysis

## Why analyze

Analysis serves several purposes:

1. **Learning**: understanding how master composers achieved their effects, then applying those techniques.
2. **Performance**: knowing what's structurally important in a piece informs how to play it.
3. **Composition**: analyzing pieces in your target style trains your ear and informs your own choices.
4. **Criticism**: explaining why a piece works or doesn't work requires analytical vocabulary.
5. **Theory**: building larger patterns from many analyzed pieces (genre conventions, period style).

The skilled analyst doesn't just describe what's there — they identify what *causes* the music's effect on the listener.

## A general analytical workflow

For any piece, work through (in roughly this order):

### 1. Listen first

Before reading the score or analyzing on paper:

- Listen multiple times.
- Identify what catches your ear: a melodic moment, a chord change, a rhythmic shift, a textural surprise.
- Form an emotional impression: what is this piece doing to the listener?
- Try to predict what comes next as the piece unfolds; notice when your predictions are confirmed or surprised.

Listening before analyzing prevents you from imposing theoretical frameworks that don't match the piece's actual effect.

### 2. Identify the genre / style / period

What kind of music is this?

- Period (Baroque, Classical, Romantic, modern, etc. — see `genres/classical-periods.md`)
- Genre (pop, jazz, hip-hop, etc. — see relevant `genres/*.md`)
- Sub-style within the genre

Style identification informs everything that follows. A "wrong note" in a Bach fugue is bad voice leading; a "wrong note" in a Coltrane solo might be deliberate altered tension.

### 3. Identify the form

Map the piece's sections.

For classical: sonata form, rondo, theme-and-variations, fugue, etc. (see `form/classical-forms.md`).

For popular: verse-chorus, AABA, blues, EDM build-and-drop, etc. (see `form/popular-song-forms.md`).

For each section, note:

- Length (in bars or seconds)
- Key area
- Rough character (loud/soft, dense/sparse, melodic/harmonic)
- Role in the larger structure (statement, contrast, development, return, etc.)

### 4. Identify the key(s)

What's the home key? Where does the music modulate? When?

For tonal music: trace the key plan. A sonata-form exposition typically goes I → V (or i → III); a development modulates broadly; the recapitulation returns to I throughout.

For modal music: identify the mode (Dorian, Phrygian, etc.).

For atonal music: there's no key, but there may be pitch-class set organization.

### 5. Identify the harmonic language

What kinds of chords are being used?

- Diatonic only?
- With secondary dominants?
- With modal mixture / borrowed chords?
- With chromatic mediants?
- With augmented sixths or Neapolitan?
- With jazz extensions and substitutions?

Harmonic vocabulary is a big indicator of period and style. Common-practice tonality has a specific vocabulary; jazz extends it; modal music avoids functional progressions; atonality abandons tonality.

### 6. Trace the melody

Identify the main melodic material:

- Themes and motifs
- Phrase structure (period, sentence, hybrid)
- Range and contour
- Climax position
- Use of chord tones vs. NCTs

For a piece that gets developed (sonata, fugue), trace how the melodic material is transformed across the piece.

### 7. Identify the rhythmic profile

- Time signature(s)
- Tempo
- Groove (swing, straight, syncopated)
- Rhythmic motifs
- Polyrhythm or hemiola
- Asymmetric phrases

### 8. Identify orchestration / instrumentation

What instruments are involved? How are they used? Which carries the melody, which the bass, which the chord, which color?

In recorded music: production choices count. Compression, reverb, EQ, sidechaining, etc.

### 9. Trace energy / dynamics

How does the piece's energy build and release? Where are the peaks? Where are the breaks?

Map the dynamic profile across the form. See `form/narrative-and-transitions.md`.

### 10. Identify what's distinctive

What makes this piece this piece? What's the "fingerprint"? Often this is:

- A recurring motif used in unusual ways
- A specific harmonic move (a key chord change, an unexpected modulation)
- A textural innovation
- A formal twist

The distinctive features are what make the piece memorable. Identifying them is the analytical payoff.

## Roman numeral analysis (RN)

The standard analytical notation for tonal music. See `fundamentals/notation-and-conventions.md` for full notation.

Quick reminders:

- Capital roman numerals = major chords (I, IV, V)
- Lowercase = minor (i, iv, v)
- ° = diminished
- + = augmented
- 7, 9, 11, 13 = chord extensions
- Inversions: 6 (first), 6/4 (second), 4/3 (third for 7th chord)
- Slashes: V/V = "V of V" (secondary dominant)

In analysis, write the roman numeral analysis below each chord on the score. This makes harmonic patterns visible.

### Common patterns to recognize

- **Cadences**: PAC (V-I), HC (ending on V), deceptive (V-vi), plagal (IV-I). See `assets/cadence-reference.md`.
- **Tonicization** (a brief move toward another key, then return): V/V → V → I.
- **Modulation** (a longer establishment of a new key): see `harmony/modulation.md`.
- **Sequence** (repeated harmonic pattern at different pitch levels).

## Schenkerian analysis (briefly)

Heinrich Schenker's system reduces a piece to its underlying voice-leading "skeleton". Useful for understanding how surface ornament relates to deep structure.

Key concepts:

- **Background**: the piece's most fundamental structure (e.g., I → V → I).
- **Middleground**: intermediate level showing the main harmonic events.
- **Foreground**: the actual notes of the piece.

Schenker analysis is sophisticated and specialized. For most analytical purposes, RN analysis is sufficient.

## Phrase analysis

Identify phrase boundaries (where one phrase ends and the next begins). Look for:

- Cadences (the most reliable indicator)
- Rests or sustained notes
- Tessitura changes
- Rhythmic resets

Then categorize the phrase structure (period, sentence, hybrid). See `melody/phrase-structure.md`.

## Motivic analysis

Identify motifs (recurring melodic-rhythmic units). For each motif:

- Where does it first appear?
- How is it varied / developed across the piece?
- Are different motifs combined?
- Is one motif the "main" motif, or are several equally important?

Trace motivic transformations: repetition, sequence, inversion, retrograde, augmentation, diminution, fragmentation. See `melody/motivic-development.md`.

## Genre-specific analytical priorities

Different genres reward different analytical focuses:

### Classical (Common Practice)

Focus on:

- Form (sonata, rondo, etc.)
- Key plan and modulation
- Roman numeral harmony
- Phrase structure (period, sentence)
- Motivic development
- Counterpoint (especially in fugue)
- Historical / stylistic context

### Jazz

Focus on:

- Form (32-bar AABA, blues, modal)
- Chord changes (often dense; extensions and substitutions)
- Soloist's lines (chord-scale relationships, motivic development across choruses)
- Harmonic substitutions (tritone subs, ii-V chains, reharmonizations)
- Rhythm (swing feel, time, microtiming)
- Comping and rhythm section interaction

### Pop / Rock

Focus on:

- Section structure (verse, chorus, pre-chorus, bridge)
- Hook identification (where, what, how often)
- Chord progressions (often simple; what makes them effective)
- Melodic identity (range, contour, climax)
- Production choices (what's the kick like? the vocal? the stereo image?)
- Lyric analysis

### Hip-hop

Focus on:

- Beat structure (drum pattern, sample, 808)
- Flow analysis (rapper's rhythm against the beat)
- Rhyme scheme
- Sample sources (where samples come from; how transformed)
- Production techniques

### EDM / Electronic

Focus on:

- Build-and-drop structure
- Bass design
- Sound design (synth choices, processing)
- Energy contour
- Repetition and variation in patterns

### Film score

Focus on:

- Themes (leitmotifs)
- Theme transformations across cues (how leitmotif varies with character/scene)
- Atmospheric vs. melodic cues
- Scene-music integration

## Comparative analysis

Comparing two or more pieces:

- What do they share? (Genre, key, structure, instrumentation)
- What's different? (Specific harmonic moves, melodic shapes, rhythmic feels)
- What does the comparison reveal? (E.g., comparing two Bach fugues reveals what's specifically Bach vs. specific to one fugue)

Comparative analysis is especially useful for understanding genre conventions: analyze 5–10 pieces in a genre to identify what they have in common.

## Common analytical mistakes

- **Imposing theoretical frameworks that don't fit the music.** Don't analyze a Coltrane solo with strict common-practice voice-leading rules. Match analytical method to musical style.
- **Confusing what's there with what's effective.** A piece may have technically correct counterpoint and still be boring. Analysis identifies what's there; aesthetic judgment is separate.
- **Ignoring context.** A piece by Mozart (Classical period) and a piece by Schoenberg (Modernist) cannot be compared as if their stylistic frameworks were identical.
- **Over-relying on Roman numerals.** RN analysis works for tonal music; it's awkward for modal, atonal, jazz, and pop.
- **Missing the forest for the trees.** Detailed harmonic analysis doesn't help if you've missed that the piece is in ABA form.
- **Conflating analysis with appreciation.** Knowing how a piece works doesn't mean liking it (or vice versa). Analysis serves understanding, not validation.
- **Forgetting the listener.** The point of analysis is usually to explain why the music *affects* the listener. If your analysis doesn't connect to listening experience, you've missed something.

## Quick analytical workflow checklist

For any piece, ask:

- [ ] What's the genre / period / style?
- [ ] What's the form? (Map the sections)
- [ ] What's the key plan?
- [ ] What's the harmonic language? (Diatonic? Chromatic? Modal? Jazz-extended?)
- [ ] What are the main melodic / motivic ideas? (Identify themes; trace development)
- [ ] What's the phrase structure? (Period? Sentence? Hybrid? Asymmetric?)
- [ ] What's the rhythmic profile? (Meter, tempo, groove, syncopation)
- [ ] What's the orchestration / texture? (Instruments; voicing; density)
- [ ] What's the dynamic / energy profile? (Builds, peaks, breaks)
- [ ] What's distinctive? What makes this piece this piece?

## Sample analysis: a Beatles song ("Yesterday")

To demonstrate the workflow:

- **Genre / period**: 1960s British pop; folk-influenced ballad
- **Form**: AABA with extra sections at the start and end. Roughly: A (verse) - A - B (bridge) - A - B - A
- **Key**: F major (with brief excursion to D minor in the bridge)
- **Harmonic language**: diatonic with secondary dominants and modal mixture (the iv chord — B♭m — borrowed from F minor)
- **Melodic ideas**: the iconic descending melodic motif (F-E-F over the line "Yesterday, all my troubles seemed so far away")
- **Phrase structure**: 7-bar phrases (asymmetric — most pop is 8-bar; the 7-bar phrase is part of the song's distinctive feel)
- **Rhythm**: simple 4/4, slow ballad tempo, swung subtly
- **Orchestration**: voice + acoustic guitar + string quartet (added in studio)
- **Energy**: gentle throughout; no big dynamic peaks
- **Distinctive**: 7-bar phrase length; the borrowed iv chord (B♭m) at "love was such an easy game to play" — this minor chord brings momentary darkness to a major-key song

This kind of analysis takes 10–20 minutes and reveals what makes the song special.

## Cross-references

- Cadences and harmonic vocabulary → `../assets/cadence-reference.md`, `harmony/functional-harmony.md`
- Roman numeral notation → `fundamentals/notation-and-conventions.md`
- Form templates (classical and popular) → `form/classical-forms.md`, `form/popular-song-forms.md`
- Phrase structure → `melody/phrase-structure.md`
- Motivic development → `melody/motivic-development.md`
- Genre / period identification → `genres/classical-periods.md`, other `genres/*.md`
- Jazz harmony for jazz analysis → `harmony/jazz-harmony.md`
- Modal harmony for modal analysis → `harmony/modal-harmony.md`
- 20th-century analytical techniques (set theory, etc.) → `techniques/20th-century-techniques.md`
- Workflow for the agent's own composition tasks → `workflow.md`
- Critique mode (analysis put to evaluative use) → `critique-and-feedback.md`
- Teaching mode (analysis as a pedagogical tool) → `teaching-composition.md`

# Counterpoint

Counterpoint is the art of writing two or more melodically independent lines that fit together harmonically. It's the discipline of *line vs. line* — how each voice has its own contour, rhythm, and identity, while the combined result still works as a coherent whole.

This is one of the deepest and most technical areas of Western composition. This file gives a working overview: what counterpoint is, the historical species method (Fux), tonal counterpoint (Bach-style), fugue, and how counterpoint shows up in modern music — including pop, film, and jazz arrangement.

## When to consult this file

- The user asks about counterpoint, polyphony, fugue, invention, or "two-part writing"
- Writing a countermelody or harmonized backing line
- Independent lines (e.g., bass + melody + countermelody) that all work together
- Bach-style chorale, fugue, or invention
- Renaissance polyphony, motets
- Modern uses: countermelodies in pop and film, contrapuntal jazz arranging
- Why "everything moves at once and clashes" in the user's piece — usually a counterpoint problem

For voice-leading rules within a single chord progression, see `harmony/voice-leading.md`. For melodic construction at the line level, see `melody/melodic-construction.md`. For fugue as a formal type, see `form/classical-forms.md` (which references back here).

## What is counterpoint

Counterpoint = "point against point" (note against note, voice against voice). It's the technique of combining two or more lines where:

1. **Each line is musically independent** — has its own contour, rhythm, melodic identity. Not just a chord stretched across multiple parts.
2. **The lines fit together vertically** — at every moment, the resulting harmony makes sense (consonant or strategically dissonant).
3. **The whole equals more than the sum** — the listener can either follow individual lines or listen to the combined fabric, and both work.

The opposite of counterpoint is **homophony**: one line that's the melody, accompanied by chords (no real voice independence in the chords). Most pop is homophonic. Most Bach is contrapuntal.

Counterpoint exists on a spectrum:

- **Pure homophony**: melody + block chords. (Most pop, hymns.)
- **Light counterpoint**: melody + active bass + maybe one countermelody. (Many film scores, jazz arrangements.)
- **Full counterpoint**: 3–6 independent lines, each substantial. (Bach, Renaissance motets.)
- **Strict counterpoint**: voices must follow rigid voice-leading rules. (Fugue, species exercises.)

## Why study counterpoint

Even if you'll never write a fugue, counterpoint reasoning improves:

- **Bass lines**: a melodic, independent bass line gives a track depth that block-chord bass doesn't.
- **Countermelodies**: a good second line in an arrangement is a counterpoint problem.
- **Inner-voice movement**: the alto and tenor of a chorale or string arrangement gain interest from contrapuntal motion.
- **Backing vocals**: harmonized vocals that follow the lead in parallel are weak; vocals that have their own independent line (while still harmonizing) are strong.
- **Synth pad orchestration**: layered pads with independent inner motion sound richer than blocks.
- **Solo improvisation**: a soloist who plays "around" a chord change creates implied counterpoint with the bass and harmony.

The discipline trains the ear and hand to think in lines simultaneously, which is one of the most important skills in composition.

## Species counterpoint (Fux's method)

Johann Joseph Fux's *Gradus ad Parnassum* (1725) codified counterpoint pedagogy through five graded exercises ("species"). Generations of composers — Haydn, Mozart, Beethoven, Mahler, Brahms — learned counterpoint from this book.

Species counterpoint trains a student to write a counter-line against a fixed *cantus firmus* (a slow melody, originally in whole notes). Each species adds rhythmic complexity.

### First species: note-against-note

Two voices, both in whole notes. Every note in the new voice (the counterpoint) sounds against one note of the cantus firmus.

Rules:

- Begin and end on a perfect consonance (P5, P8, or unison).
- Use mostly imperfect consonances (3rds and 6ths) in between.
- Avoid parallel P5s and P8s. (See `harmony/voice-leading.md`.)
- Move mostly by step in the counterpoint; leaps are fine but should be balanced.
- Aim for contrary motion (voices moving in opposite directions).
- Keep the counterpoint within a reasonable range; avoid awkward leaps and wide skips.

The discipline: write a counter-line that's *independent* (different contour from the cantus firmus) but *consonant* throughout. This trains line independence.

### Second species: 2:1

Two notes in the counterpoint for every one in the cantus firmus. Now half-notes against whole notes (or whatever ratio matches).

New rules:

- The first note of each measure (downbeat) must be a consonance.
- The second note (the off-beat) can be a passing tone (dissonance approached and left by step) or a consonance.
- Don't introduce parallel 5ths/8ves through the rhythm.

Effect: the counterpoint becomes more melodically active. Passing tones add melodic interest.

### Third species: 4:1

Four notes against one. Quarter notes against whole notes.

New rules:

- The first note of each beat-group must be a consonance.
- Three of the four sub-notes can be passing tones, neighbor tones, or consonances.
- Avoid awkward leaps in the counterpoint.

Effect: the counterpoint becomes a flowing line. Embellishment now happens at a denser rhythmic level.

### Fourth species: suspensions

Each note in the counterpoint is tied across the barline, becoming a suspension on the next downbeat.

The pattern: prepare (consonance) → suspend (dissonance) → resolve (down by step to consonance). The classic 4-3, 7-6, 9-8 suspensions.

Effect: dramatic dissonance with controlled resolution. This species teaches you the tension/release cycle that drives much of tonal music.

### Fifth species: florid (mixed)

All techniques combined. The counterpoint uses any rhythmic value (whole notes, half, quarter, eighth) and any device (passing tones, neighbor tones, suspensions). The line should be *musical* — not a textbook demonstration but a real, beautiful counter-line.

By the time a student reaches fifth species, they can write a counterpoint that sounds like a Bach two-part invention.

## Tonal counterpoint (Bach style)

Species counterpoint is "modal" — based on Renaissance modal voice-leading. **Tonal counterpoint** applies similar principles within functional tonality, with leading-tone resolution and dominant→tonic progressions.

Bach is the master. His Two- and Three-Part Inventions, the Well-Tempered Clavier, and the Goldberg Variations are studied as the gold standard of tonal counterpoint.

Key features:

- **Functional harmonic progressions** underly the counterpoint. Even "free" counterpoint sits within a clear key.
- **Phrases have cadences** — voice-leading converges on PAC, HC, deceptive, plagal, etc., as in homophonic music.
- **Voices have idiomatic ranges** for the instrument.
- **Dissonance is more freely treated** than in strict species — passing tones can be longer, suspensions can be ornamented.

Tonal counterpoint is the underpinning of fugue, invention, chorale, and many forms of contrapuntal writing into the 20th century.

## Imitation

A central technique in contrapuntal writing: one voice introduces a motif, then another voice repeats it (often delayed, often at a different pitch). The ear loves to track imitation — it creates dialogue between voices.

Types:

- **Strict imitation**: exact pitch and rhythm reproduction.
- **Free imitation**: similar but with adjustments (notes changed for harmonic reasons).
- **Imitation at the octave / 5th / 4th / etc.**: the imitating voice starts on a different scale degree.
- **Inverted imitation**: the imitating voice plays the inversion (intervals flipped).
- **Augmented / diminished imitation**: the imitating voice uses different rhythm (longer/shorter notes).

Imitation drives much of fugue writing, but it appears in pop too — call-and-response between voices in gospel, in Beatles backing vocals, in jazz arrangements.

## Fugue

A fugue is a contrapuntal form (and a contrapuntal *technique*) where multiple voices imitate a *subject* in a structured way. It's the most rigorous and technically demanding form in tonal music.

Components:

- **Subject**: the main theme, introduced alone by one voice.
- **Answer**: a second voice enters with the subject transposed (typically up a fifth or down a fourth — a "tonal answer" adjusted to remain in the home key, or a "real answer" with strict transposition).
- **Countersubject**: when the second voice plays the answer, the first voice plays a counter-line called the countersubject. This countersubject is reused throughout the fugue.
- **Episodes**: passages between subject entries, often built on motivic fragments and modulating sequences. Episodes provide breathing space and structural articulation.
- **Stretto**: subject entries overlap (one voice begins the subject before the previous voice has finished). Used for climax.
- **Augmentation / diminution**: subject played slower or faster, often near the end.
- **Pedal point**: a sustained tonic or dominant note, usually approaching the final cadence.

A typical fugue in 3 voices:

1. **Exposition**: voices enter one by one with the subject (or answer). Once all voices are in, the exposition is complete.
2. **First episode**: modulating, motif-based interlude.
3. **Second exposition** (in a related key): subject entries in the new key.
4. **Episode** → another set of entries, typically in further keys.
5. **Stretto and intensification** → climactic point.
6. **Final entry** in the home key, often with augmentation or pedal point.
7. **Cadence** in the home key.

Bach wrote the most famous fugues (the WTC contains 48). Beethoven wrote several (the *Grosse Fuge*, the fugue in his late sonatas). Shostakovich wrote 24 fugues for the *24 Preludes and Fugues*.

Fugue is rare in modern composition because it's labor-intensive and requires a specific aesthetic. But fugal moments — short stretches of fugal texture — appear in film scores, prog rock, and contemporary classical.

## Invention

A two-part invention is a smaller contrapuntal piece — like a baby fugue. Bach's 15 Two-Part Inventions are the classic study set. Features:

- A subject (motif), introduced alone, then imitated.
- Some development (sequencing, transposition).
- A clear key plan (typically I → V → I or I → vi → I).
- Roughly 30–60 bars long.

Inventions are an excellent practice exercise: simpler than fugue but trains the same skills.

## Contrapuntal devices and devices

A grab-bag of techniques used in contrapuntal writing:

| Device | Description |
|--------|-------------|
| **Invertible counterpoint** | Voices designed so they can swap (top voice becomes bottom and vice versa) without violating voice-leading rules. The interval inversion table (`harmony/voice-leading.md`) tells you which intervals stay consonant when inverted. |
| **Double counterpoint at the octave** | Most common; voices invert by exchanging an octave. |
| **Double counterpoint at the 10th / 12th** | Voices invert by exchanging a 10th or 12th interval. Less common; specific to advanced Bach. |
| **Triple counterpoint** | Three voices that can be arranged in any vertical order. |
| **Canon** | Strict imitation maintained throughout: every voice plays the same line, delayed. The simplest is a round (e.g., "Row, Row, Row Your Boat"). |
| **Crab canon (canon cancrizans)** | One voice plays the line forward, another plays it backward simultaneously. |
| **Mirror canon** | One voice plays the line; another plays the inversion. |
| **Mensuration canon** | Voices play the same line at different rhythmic values (one voice slower, one faster). |

These devices show up in Bach (especially the *Art of Fugue*, the *Musical Offering*, and the *Goldberg Variations*), in 20th-century serial music, and occasionally in modern composition (Reich's process pieces use canon-like techniques).

## Counterpoint in pop, jazz, and film

### Pop counterpoint

Pop is overwhelmingly homophonic, but counterpoint shows up in:

- **Bass lines**: a melodic, independent bass that's not just root-emphasizing is counterpoint. McCartney's bass lines are textbook contrapuntal craft. Motown bass.
- **Countermelodies**: a synth or string line that has its own contour against the lead vocal. Common in 80s pop production.
- **Backing vocals**: when backing vocals do something other than parallel harmony with the lead — they're contrapuntal. Beach Boys, Queen, Bee Gees.
- **Riffs that interact with the vocal**: the iconic guitar riff in "Smoke on the Water" or "Layla" exists in counterpoint with the vocal melody.
- **Pre-chorus pads**: a moving inner voice in a synth pad creates contrapuntal motion.

When advising on pop arrangement, suggest contrapuntal additions sparingly. A single well-chosen countermelody beats a dense fabric of competing lines.

### Jazz counterpoint

Jazz is mostly soloist-plus-rhythm-section (homophonic in texture), but contrapuntal techniques show up in:

- **Two-horn arrangements**: trumpet and saxophone playing complementary lines (Mulligan & Baker, Lee Konitz & Warne Marsh).
- **Bach-influenced jazz**: Modern Jazz Quartet's "Versailles", "Concorde". Bach plays well as jazz changes.
- **Solo piano**: pianists like Bill Evans, Brad Mehldau, Bud Powell, Keith Jarrett write contrapuntally — left and right hands as independent voices.
- **Ensemble passages in big band charts**: ensemble shouts often use parallel motion, but the bass and drums remain contrapuntal against the horns.

### Film counterpoint

Film scoring uses counterpoint heavily for sophistication. Common patterns:

- **Theme + countermelody**: the love theme stated in violins, with a counter-line in cellos or woodwinds.
- **Harmony in counterpoint**: one motif transformed and stacked against itself in different sections.
- **Multiple themes in dialogue**: when two characters meet on screen, their themes can be combined contrapuntally.

John Williams is famous for contrapuntal richness. The "Ride of the Rohirrim" in Howard Shore's *Lord of the Rings* layers multiple themes contrapuntally.

### EDM and electronic counterpoint

Less common but real. Aphex Twin, Squarepusher, Brian Eno's ambient: layered melodic lines with their own identities.

## Modern uses (post-tonal)

20th-century composers extended counterpoint into atonal and serial worlds:

- **Schoenberg, Webern, Berg**: 12-tone counterpoint where voice-leading is governed by row operations rather than tonal function.
- **Bartók**: contrapuntal pieces in a folk-modal idiom (the *Mikrokosmos*).
- **Stravinsky**: rhythmic counterpoint where independence is rhythmic as much as pitched.
- **Reich, Glass**: process-based "phasing" counterpoint where two identical lines drift out of sync.
- **Ligeti**: micropolyphony with dozens of voices producing dense textures (cluster-counterpoint).

Contemporary composers use counterpoint more freely than the strict tonal traditions — voice independence, line interest, and overall texture matter more than specific dissonance treatment.

## Common pitfalls when advising

- **Don't write counterpoint as multiple "melody lines" without harmonic awareness.** Each line is independent, but they must fit together vertically. Write one line, then check it against the others at every beat.
- **Don't write contrapuntally for music that wants homophony.** A pop chorus rarely benefits from elaborate counterpoint — it competes with the hook. Counterpoint shines in instrumental contexts, bridges, and verse-second-half arrangements.
- **Don't ignore voice-leading basics.** Parallel 5ths and 8ves are the most common counterpoint mistake. Know the rules of the style you're working in (`harmony/voice-leading.md`).
- **Don't make the lines too rhythmically similar.** If both voices move at the same rate, they sound like one voice with chords. Independence requires rhythmic differentiation — when one voice holds, the other moves.
- **Don't make the lines clash in tessitura.** Two melodies in the same register fight for the listener's attention. Spread voices across registers (bass + alto + soprano, e.g.) for clarity.
- **Don't fugue every piece.** Fugue is a powerful but specific tool. Most music wants only contrapuntal moments, not a full fugue.

## A workflow for adding counterpoint to a piece

If the user has a melody and wants to add an independent line:

1. **Decide the counterpoint's role.** Is it a bass line, a countermelody, an inner voice? Each has different priorities.
2. **Sketch its rhythm first.** Where does it move? Where does it hold? It should *not* mirror the lead's rhythm — it should complement it. When the lead holds, the counter-line should move (and vice versa).
3. **Sketch its contour.** What's its shape? Often the inverse of the lead's shape (lead goes up, counter goes down). Always with an independent climax.
4. **Fill in pitches that fit the harmony.** Each beat should produce a consonant interval with the lead (3rd, 6th, octave, 5th preferred; 2nd or 7th for momentary tension).
5. **Check for parallel 5ths/8ves.** Fix by re-routing one voice.
6. **Test the lines separately.** If the counter-line, played alone, sounds boring — the line isn't independent enough. Strengthen its identity.
7. **Test combined.** Listen to both. The lines should be distinguishable but unified.

## Cross-references

- Single-line voice leading rules → `harmony/voice-leading.md`
- Melodic construction within each line → `melody/melodic-construction.md`
- Motivic development that drives fugue subjects → `melody/motivic-development.md`
- Fugue as a structural form → `form/classical-forms.md`
- Pop arrangement (where counterpoint is sparingly used) → `orchestration/voicing-and-texture.md`, `orchestration/arrangement-density.md`
- Jazz arrangement → `genres/jazz-styles.md`
- Film scoring counterpoint → `genres/film-tv-scoring.md`
- 20th-century post-tonal counterpoint → `techniques/20th-century-techniques.md`
- Functional context for tonal counterpoint → `harmony/functional-harmony.md`

# Answer Calibration and Controlled Variation

This file helps the agent keep composition answers useful, bounded, and appropriately variable. Use it when smoke tests show random but acceptable variation, when the user asks for repeated alternatives, or when a session risks becoming too verbose, too random, or too rigid.

For broader multi-turn collaboration, see `user-agent-collaboration.md`. For reusable formats, see `../../assets/response-templates.md`. For quick decision logging, see `../../assets/session-brief-and-decision-log.md`.

## When to consult this file

Use this file when:

- the user asks for multiple options, but not an essay;
- the user says “try again” or “give me another version”;
- a generated idea should vary without abandoning the chosen direction;
- the assistant is tempted to load many files or over-explain;
- a smoke test or review finds the answer shape good but inconsistent in length or scope;
- the user wants brainstorming, but the session needs convergence.

## Calibration principle

A composition assistant should be **predictable in structure** and **creative in material**.

Stable parts:

- routing discipline,
- response format,
- assumptions,
- user-chosen constraints,
- safety and originality boundaries.

Variable parts:

- chord voicings,
- melodic contours,
- rhythmic cells,
- texture choices,
- arrangement density,
- option labels and examples.

The agent should not make the whole workflow random just because the musical content is creative.

## Default answer budgets

| Task type | Recommended size | Good output |
|---|---:|---|
| Quick theory lookup | 1–3 short paragraphs | definition + example + caveat |
| Chord/progression generation | 2–4 options | chart + Roman numerals if useful + why it works |
| Melody/topline revision | 2–4 revisions | scale degrees or notes + contour/rhythm explanation |
| Vague problem diagnosis | compact matrix | highest-impact fix first |
| Brainstorming | 3 lanes | safe / fresh / wild or 3 named concepts |
| Iterative revision | one focused revision | keep/change/direction + before/after |
| Current trend answer | concise research summary | evidence vs inference + composition translation |
| Teaching | example first | explanation + tiny exercise |

If the user asks for “more,” expand the chosen lane rather than adding unrelated lanes.

## Controlled variation bands

When giving multiple creative options, vary along one or two axes at a time.

| Band | Use when | Variation level | Example |
|---|---|---|---|
| Safe | user wants usable material quickly | familiar form, clear harmony, low risk | diatonic progression with one color chord |
| Fresh | user wants identity | one unusual variable | borrowed chord, asymmetrical rhythm, register move |
| Wild | user wants exploration | several unusual variables, but still playable | meter shift + modal color + texture concept |

Do not present three options that differ only in labels. Each option should change an audible variable.

## Revision stability rules

When the user says “again,” “darker,” “less cheesy,” “more modern,” or “use option 2,” preserve the session identity.

| Preserve by default | Change only if requested |
|---|---|
| chosen branch | branch identity |
| key / tonal center | key center |
| tempo / meter | tempo grid |
| hook contour | hook rhythm / lyric stress |
| liked chords | harmonic function |
| reference role | direct reference material |
| do-not-change list | protected elements |

A revision should state what was kept and what changed. If both changed, the user may feel the assistant ignored them.

## Loading calibration

Use the smallest file set that can answer the current turn.

| Situation | Load budget |
|---|---:|
| stable theory lookup | 1 file, maybe 1 asset |
| genre-specific generation | genre file + one topical file |
| vague creative diagnosis | diagnostic asset + one relevant craft file |
| current trend / reference request | research file + relevant genre file + maybe web-search cheatsheet |
| multi-turn revision | collaboration/revision file + session template |
| release validation | validation docs only; do not load during normal composition |

If five or more files seem necessary, answer the dominant aspect first and name the adjacent area rather than loading everything.

## Assumption calibration

State assumptions only when they affect the result.

Good:

```text
I'll assume midtempo 4/4 alt-pop so the sketch is playable. I can reframe it for jazz or EDM later.
```

Too much:

```text
I need genre, tempo, instrumentation, vocal range, DAW, reference tracks, mix target, and release market before I can answer.
```

Bad:

```text
[Silently assumes a piano ballad when the user wanted metalcore.]
```

## Currentness calibration

Separate durable craft from current evidence.

```md
Durable frame: [genre conventions that do not depend on this week's chart]
Current signal: [what needs web/platform verification]
Composition move: [how to turn the signal into groove/form/hook/arrangement]
```

Do not use web research for stable interval/chord/mode facts unless the user explicitly asks for sources. Do use web research for recent artists, platform trends, current charts, software/tool status, law/policy, and living-scene claims.

## User-agency calibration

End creative responses with a decision point that helps the user move.

Good endings:

- “Pick A if the lyric matters most; pick B if the groove matters most.”
- “I’d prototype Option 2 first because it changes the chorus without rewriting the verse.”
- “The safest next edit is to keep the chords and only change the topline rhythm.”

Weak endings:

- “Let me know.”
- “There are many possibilities.”
- “It depends.”

## Common calibration failures

| Failure | Symptom | Fix |
|---|---|---|
| Too much theory | user asked for a progression, answer becomes lecture | put playable material first; explain after |
| Too many options | user gets five unrelated directions | limit to 2–4 and label trade-offs |
| Hidden randomness | repeated answer changes key/tempo/style | preserve project card constraints |
| Over-reading | answer pulls many unrelated files | pick dominant task and one adjacent file |
| Under-specificity | advice says “add energy” but gives no notes/chords | provide a bar, rhythm, voicing, or form move |
| False certainty | current scene claim stated as stable fact | mark evidence/inference and use web research |
| Over-questioning | asks for details that can be safely assumed | make a reversible assumption and proceed |

## Cross-references

- User-agent collaboration → `user-agent-collaboration.md`
- Musical brainstorming → `musical-brainstorming.md`
- Revision loops → `revision-and-feedback-loops.md`
- Response templates → `../../assets/response-templates.md`
- Diagnostic checklists → `../../assets/diagnostic-checklists.md`
- Session brief and decision log → `../../assets/session-brief-and-decision-log.md`
- Web-search cheatsheet → `../../assets/web-search-cheatsheet.md`
- Phase C smoke-test results → `../validation/phase-c-smoke-test-results.md`

# Navigation Map

This file routes user requests to the specific reference files you need. **Read this before diving into any other reference**, every time you use this skill.

## How to use this map

1. Find the section matching the user's request (quick-lookup table, vague-problem table, or genre table).
2. Load the listed file(s).
3. If the request crosses domains, load primary files for each — but resist loading more than 3.
4. If you would need to load 5+ files, the question is too broad. Pick the dominant aspect and answer that first; offer to dig into adjacent topics afterward.

## Quick lookup — by topic

| User asks about | File(s) |
|-----------------|---------|
| Intervals, scales, keys, basic theory | `fundamentals/pitch-intervals-scales.md` |
| Time signatures, rhythm notation, tempo | `fundamentals/rhythm-meter.md` |
| Notation, dynamics, articulation, ornaments | `fundamentals/notation-and-conventions.md` |
| Building chords (triads, 7ths, extensions, alterations) | `harmony/chord-construction.md` |
| Ambiguous chord symbols, parsing user chord notation into notes | `../assets/chord-symbol-ambiguity-and-parsing.md` + `../assets/chord-symbol-conventions.md` |
| Scale-degree spelling, Roman numeral accidentals, altered-tension spelling | `../assets/scale-degree-spelling-cheatsheet.md` + `harmony/functional-harmony.md` |
| Chord progressions, T-S-D function, cadences, diatonic motion | `harmony/functional-harmony.md` |
| Voice leading, smooth movement, 4-part writing | `harmony/voice-leading.md` |
| Secondary dominants, borrowed chords, Neapolitan, augmented 6ths | `harmony/chromatic-harmony.md` |
| Modulation, key changes | `harmony/modulation.md` |
| Modal music, modal centers, characteristic notes | `harmony/modal-harmony.md` |
| Jazz changes, ii-V-I, tritone subs, upper structures | `harmony/jazz-harmony.md` |
| Reharmonization, chord substitution | `harmony/reharmonization.md` |
| Writing melodies, contour, climax, range | `melody/melodic-construction.md` |
| Motifs, motivic development, transformation | `melody/motivic-development.md` |
| Phrase, period, sentence, antecedent/consequent | `melody/phrase-structure.md` |
| Counterpoint, two or more independent voices | `counterpoint.md` |
| Syncopation, hemiola, displacement, rhythmic interest | `rhythm-groove/rhythmic-devices.md` |
| Groove, feel, swing, microtiming, pocket | `rhythm-groove/groove-and-feel.md` |
| 5/4, 7/8, additive meter, polyrhythm, polymeter | `rhythm-groove/odd-meters-polyrhythm.md` |
| Sonata form, rondo, variations, classical structure | `form/classical-forms.md` |
| Verse-chorus, AABA, song structure, EDM forms | `form/popular-song-forms.md` |
| Transitions, energy arcs, build-ups, breakdowns | `form/narrative-and-transitions.md` |
| Instrument ranges and orchestral character overview | `orchestration/instruments-ranges-character.md` |
| Instrument-specific idiomatic writing overview | `instrument-idiom/overview.md` |
| Piano / keyboard voicing and accompaniment | `instrument-idiom/piano-keyboards.md` |
| Guitar chords, riffs, capo/tuning, strumming | `instrument-idiom/guitar.md` |
| Basslines, walking bass, synth/electric bass role | `instrument-idiom/bass.md` |
| Drum grooves, fills, kit/percussion writing | `instrument-idiom/drums-percussion.md` |
| Strings, bowing, pizzicato, divisi, section writing | `instrument-idiom/strings.md` |
| Woodwinds, saxes, breath-aware wind writing | `instrument-idiom/winds.md` |
| Brass, fanfares, horn sections, mutes, endurance | `instrument-idiom/brass.md` |
| Vocal range, tessitura, backing vocals, singability | `instrument-idiom/vocals.md` |
| Voicing, register, texture, doubling | `orchestration/voicing-and-texture.md` |
| Layering, arrangement density, thinning/thickening | `orchestration/arrangement-density.md` |
| Roman numeral analysis, formal analysis, Schenker, set theory | `analysis.md` |
| Theme and variations, transformation techniques | `techniques/theme-and-variation.md` |
| 12-tone, serialism, minimalism, aleatoric, extended techniques | `techniques/20th-century-techniques.md` |
| Generative composition, algorithmic ideas, constraint-based | `techniques/constraint-based-composition.md` |
| Algorithmic composition, AI-assisted composition, generative systems, live coding | `techniques/algorithmic-and-AI-assisted.md` |
| Microtonal music, just intonation, EDOs, quarter-tones, *maqam* / *raga* / non-12-TET | `techniques/microtonal.md` |
| Choral writing, SATB, voice leading for choir, vocal ensemble | `orchestration/choral-writing.md` |
| Lyric writing, prosody, rhyme, scansion, perspective | `songwriting/lyric-writing.md` |
| Hooks, earworms, memorability | `songwriting/hooks-and-memorability.md` |
| Topline writing, melody over a track | `songwriting/topline-craft.md` |
| Prosody across languages, tonal languages, melody-lyric alignment | `fundamentals/prosody-and-language.md` |
| Composing with mix in mind, frequency awareness | `production-aware/arrangement-for-mix.md` |
| Energy curves, build-ups, drops, dynamic contour | `production-aware/energy-and-dynamics.md` |
| Picking a key, tempo, length, references; pre-production decisions | `production-aware/pre-production-decisions.md` |
| Current music trends, recent references, charts, platform-specific style signals | `research/web-music-trend-research.md` + relevant genre file |
| Fast web-search operators, query recipes, source-confidence labels | `../assets/web-search-cheatsheet.md` + `research/web-music-trend-research.md` |
| Style imitation, copyright-safe reference use, living artist style requests | `research/style-reference-and-copyright.md` + `research/reference-track-digging.md` |
| Reference-track digging, "make it feel like X" without copying | `research/reference-track-digging.md` + `production-aware/pre-production-decisions.md` |
| How to start, when to revise, beating writer's block | `workflow.md` |
| Musical brainstorming, ideation, generating multiple song/cue directions | `creative-workflows/musical-brainstorming.md` + `../assets/musical-brainstorming-cards.md` |
| Multi-turn user-agent collaboration, project card, decision log, revision branching | `creative-workflows/user-agent-collaboration.md` + `../assets/session-brief-and-decision-log.md` |
| Iterative revision, user feedback parsing, before/after edits | `creative-workflows/revision-and-feedback-loops.md` + `../assets/session-brief-and-decision-log.md` |
| Answer length, controlled creative variation, loading discipline, response calibration | `creative-workflows/answer-calibration.md` + `../assets/response-templates.md` |
| Hybrid genre requests (e.g., "K-pop with jazz harmony") | `workflow.md` (Multi-genre and hybrid requests) |
| Critique, feedback on a finished piece, evaluating user's own work | `critique-and-feedback.md` |
| Teaching, learning paths, exercises, pedagogical mode | `teaching-composition.md` |
| Source verification, bibliography, maintainer research sources | `source-bibliography.md` |
| First-release readiness, release blockers, validation checklist | `validation/first-release-readiness.md` + `validation/phase-b-correctness-pass.md` + `../RELEASE-ROADMAP.md` |
| Phase B correctness pass, pre-smoke-test audit | `validation/phase-b-correctness-pass.md` |
| Prompt smoke tests for routing and answer-shape QA, Phase C | `validation/prompt-smoke-tests.md` |
| Phase C smoke-test results, user-run pass record | `validation/phase-c-smoke-test-results.md` |
| RC1/v1.0 packaging, release artifact hygiene, validation report minimums | `validation/rc1-packaging-checklist.md` + `../VALIDATION-v1.0.md` + `../KNOWN-LIMITATIONS.md` |
| Manual theory audit rubric for maintainer review | `../assets/music-theory-audit-rubric.md` |
| User listening context, subscribed streaming services, playlists, Replay/Wrapped, personal reference library | `research/user-listening-context-and-streaming-services.md` + `research/reference-track-digging.md` |
| Regional trend evolution, market changes across time windows, region-specific trend analysis | `research/regional-trend-evolution-analysis.md` + `research/web-music-trend-research.md` |
| Regional scene starters and local-pop idioms that lack a dedicated file | `genres/regional-scene-starters.md` + `research/regional-trend-evolution-analysis.md` |
| Minor genres, microgenres, unstable scene tags, obscure hybrid styles | `genres/minor-and-hybrid-genres.md` + `research/web-music-trend-research.md` |
| Quick matrices for trends, references, user listening evidence, hybrid genre merging | `../assets/trend-and-reference-matrices.md` |

## Genre routing

Genre conventions affect *every* other decision. When a genre is named or implied, **always load the relevant genre file** alongside whatever topical file you're using.

| Genre frame | File |
|-------------|------|
| Western classical (any period) | `genres/classical-periods.md` |
| Jazz (any sub-style) | `genres/jazz-styles.md` |
| Pop, rock, indie, alternative | `genres/pop-rock.md` |
| Country, Americana, roots, modern country-pop | `genres/country-americana.md` |
| Metal, punk, hardcore, metalcore, heavy guitar music | `genres/metal-punk-hardcore.md` |
| Hip-hop, R&B, neo-soul, trap, drill | `genres/hip-hop-rnb.md` |
| House, techno, DnB, dubstep, ambient, IDM, EDM | `genres/electronic-edm.md` |
| Film, TV, score for linear narrative | `genres/film-tv-scoring.md` |
| Game music, interactive scoring, adaptive systems | `genres/game-music.md` |
| Advertising, podcasts, branded content, trailer music, library music, corporate video | `genres/media-and-commercial-music.md` |
| K-pop, J-pop | `genres/kpop-jpop.md` |
| C-pop / Mandopop, Cantopop, T-pop, V-pop, OPM, Indo-pop | `genres/cpop-and-southeast-asian-pop.md` |
| Latin pop, reggaeton, urbano, bachata-pop, Latin trap | `genres/latin-pop-and-reggaeton.md` |
| Brazilian pop, MPB, samba-pop, funk carioca, baile funk | `genres/brazilian-pop-and-funk.md` |
| Afrobeats, Afropop, amapiano, popiano | `genres/afrobeats-and-amapiano.md` |
| South Asian film-pop, Bollywood, Punjabi pop, raga-derived pop | `genres/south-asian-film-pop.md` |
| MENA pop, Arabic/Turkish/Persian pop, maqam-adjacent fusion | `genres/mena-pop.md` |
| Korean traditional music (*gugak*), K-trad fusion, *pansori*, *sanjo* | `genres/korean-traditional.md` |
| Gospel, contemporary gospel, CCM, worship, roots gospel | `genres/gospel-and-ccm.md` |
| Folk, roots, and tradition-specific writing | `genres/folk-roots-and-traditions.md` |
| Regional Mexican, sertanejo/pagode/arrocha, raï, Turkish arabesk-pop, Punjabi pop, gqom, other local scene starters | `genres/regional-scene-starters.md` |
| Minor, microgenre, online-scene, obscure, or hybrid genre labels | `genres/minor-and-hybrid-genres.md` |
| Musical theatre, Broadway, West End | `genres/musical-theatre.md` |

If the genre is ambiguous and matters for the answer, ask the user. Don't default silently to one — you'll give wrong-feeling advice.

## Vague creative problem → diagnostic

When the user describes a feeling, problem, or fuzzy goal, translate it into a technical diagnosis before answering. Common patterns:

| User says | Likely diagnostic | Files to consult |
|-----------|-------------------|------------------|
| "It feels boring" | Predictable harmony / rhythm / dynamics | `harmony/chromatic-harmony.md`, `rhythm-groove/rhythmic-devices.md`, `production-aware/energy-and-dynamics.md` |
| "The chorus doesn't hit" | Insufficient lift from verse — register, density, harmony, rhythm | `form/narrative-and-transitions.md`, `orchestration/arrangement-density.md`, `songwriting/hooks-and-memorability.md` |
| "It feels generic / too 'pop'" | Default progressions, predictable form, overused voicings | `harmony/reharmonization.md` + relevant genre file |
| "The transition is weak / awkward" | Insufficient or excessive setup; abrupt key/density change | `form/narrative-and-transitions.md`, `harmony/modulation.md` |
| "The melody is forgettable" | No motivic identity, weak contour, no memorable interval | `melody/melodic-construction.md`, `melody/motivic-development.md`, `songwriting/hooks-and-memorability.md` |
| "It sounds muddy / cluttered" | Voicing, register, density, frequency overlap | `orchestration/voicing-and-texture.md`, `orchestration/arrangement-density.md`, `production-aware/arrangement-for-mix.md` |
| "It needs more emotion / depth" | Often: harmony too plain, dynamics too flat, melody too narrow | `harmony/chromatic-harmony.md`, `production-aware/energy-and-dynamics.md`, `melody/melodic-construction.md` |
| "It's stuck on one idea" | No development, no variation | `melody/motivic-development.md`, `techniques/theme-and-variation.md` |
| "It feels too predictable" | Harmonic and formal default-mode | `harmony/reharmonization.md`, `harmony/chromatic-harmony.md`, `form/narrative-and-transitions.md` |
| "It feels too cluttered / busy" | Too many simultaneous ideas, density never thins | `orchestration/arrangement-density.md`, `orchestration/voicing-and-texture.md` |
| "I'm blocked / can't start" | Process problem, not technical | `workflow.md`, `creative-workflows/musical-brainstorming.md`, `../assets/musical-brainstorming-cards.md` |
| "What key/tempo/length should I pick?" | Pre-production decision | `production-aware/pre-production-decisions.md` |
| "I want it to feel like [reference track]" | Reference-driven composition | `research/reference-track-digging.md`, `production-aware/pre-production-decisions.md`, relevant genre file |
| "I want X genre with Y influence" (e.g., "K-pop with jazz harmony") | Hybrid genre request | `workflow.md` (Multi-genre and hybrid requests), relevant genre files |
| "What do you think of this?" / "Is this good?" | Critique request | `critique-and-feedback.md` |
| "How do I learn [X]?" / "Teach me [Y]" | Pedagogical request, not problem-solving | `teaching-composition.md` |
| "The lyric doesn't fit / sounds awkward" (especially in non-English) | Prosody mismatch | `fundamentals/prosody-and-language.md` |
| "It sounds too much like [artist/song]" | Reference cloning risk | `research/style-reference-and-copyright.md`, `research/reference-track-digging.md`, `assets/diagnostic-checklists.md` |
| "Keep this part but change that part" | Iterative revision / preservation boundary | `creative-workflows/revision-and-feedback-loops.md`, `assets/session-brief-and-decision-log.md` |

## Multi-file scenarios

Some requests genuinely span several files. Common patterns:


**"Brainstorm song/cue ideas"** — `creative-workflows/musical-brainstorming.md` + `../assets/musical-brainstorming-cards.md` + relevant genre file if known. Generate divergent options first; converge only after the user reacts.

**"Let's keep developing this over multiple turns"** — `creative-workflows/user-agent-collaboration.md` + `../assets/session-brief-and-decision-log.md` + `workflow.md`. Maintain a project card and decision log when the session becomes stateful.

**"Revise this but keep [specific element]"** — `creative-workflows/revision-and-feedback-loops.md` + `../assets/session-brief-and-decision-log.md` + the relevant craft file. Preserve the liked element and change the smallest musical variable that can solve the problem.

**"Search the web for references/trends"** — `../assets/web-search-cheatsheet.md` + `research/web-music-trend-research.md` + `research/reference-track-digging.md`. Use the cheatsheet for query construction; use research files for interpretation and composition translation.

**"Write a song in [genre]"** — The genre file + `form/popular-song-forms.md` + `songwriting/hooks-and-memorability.md` + relevant harmony file (`harmony/functional-harmony.md` for most, `harmony/jazz-harmony.md` for jazz, `harmony/modal-harmony.md` if modal).

**"Analyze this piece"** — `analysis.md` + relevant harmony file(s) + `form/classical-forms.md` or `form/popular-song-forms.md` depending on style.

**"Reharmonize this"** — `harmony/reharmonization.md` + `harmony/chromatic-harmony.md` + (if jazz/contemporary) `harmony/jazz-harmony.md`. Reharm techniques cross-pollinate.

**"Make this section more dramatic / climactic"** — `form/narrative-and-transitions.md` + `orchestration/arrangement-density.md` + `production-aware/energy-and-dynamics.md` + `harmony/modulation.md`.

**"Compose for [ensemble or solo instrument]"** — `orchestration/instruments-ranges-character.md` + the relevant `instrument-idiom/` file + the relevant genre file + `orchestration/voicing-and-texture.md`.

**"Make this part playable / idiomatic"** — relevant `instrument-idiom/` file + `orchestration/instruments-ranges-character.md` + `orchestration/voicing-and-texture.md` if part density matters.

**"Write a [solo / improv / lead line] over these changes"** — `harmony/jazz-harmony.md` (for any non-trivial changes) + `melody/melodic-construction.md` + the relevant genre file.

**"Make this groove harder / tighter / weirder"** — `rhythm-groove/groove-and-feel.md` + `rhythm-groove/rhythmic-devices.md` + the relevant genre file.

**"What is current / trending in [genre/market]?"** — `research/web-music-trend-research.md` + the relevant genre file + `research/reference-track-digging.md` if the answer should become compositional advice.

**"Make it sound like [artist/song]"** — `research/style-reference-and-copyright.md` + `research/reference-track-digging.md` + the relevant genre file. Extract variables; do not clone melody, lyrics, riffs, samples, vocal identity, or signature production tags.

**"Prepare / validate / release this skill"** — `validation/first-release-readiness.md` + `validation/phase-b-correctness-pass.md` + `validation/prompt-smoke-tests.md` + `validation/phase-c-smoke-test-results.md` + `validation/rc1-packaging-checklist.md` + `../assets/music-theory-audit-rubric.md` + `../RELEASE-ROADMAP.md` + `../KNOWN-LIMITATIONS.md`. Use Phase B for correctness hardening, Phase C for prompt behavior QA, and Phase D / v1.0 finalization for packaging; not normal composition answers.

## Cheatsheets — when to use `assets/`

Reach for `assets/` when you need to *grab a value or template*, not understand a concept.

| Need | File |
|------|------|
| Common chord progressions by mood/genre, ready to drop in | `assets/progressions-catalog.md` |
| Standard answer formats for progressions, toplines, critiques, song plans, reference-driven requests | `assets/response-templates.md` |
| Troubleshooting vague creative problems such as weak chorus, muddy arrangement, generic harmony | `assets/diagnostic-checklists.md` |
| Quick lookup of cadence types (PAC, IAC, HC, deceptive, plagal, Phrygian, etc.) | `assets/cadence-reference.md` |
| All seven diatonic modes at a glance, with formulas and characteristic notes | `assets/modes-cheatsheet.md` |
| Interval qualities, scale formulas (major, minor variants, pentatonic, blues, folk-derived, non-diatonic) | `assets/intervals-and-scale-formulas.md` |
| Common jazz/pop voicings (rootless, drop 2/3, shell, quartal, etc.) | `assets/jazz-voicings.md` |
| Skeletal templates for forms (sonata, AABA, K-pop verse-pre-chorus-chorus, etc.) | `assets/form-templates.md` |
| Chord symbol notation across systems (Real Book, pop, RN, Nashville); ambiguous symbol disambiguation | `assets/chord-symbol-conventions.md` |
| Operational chord-symbol parsing, ambiguous user notation, add/sus/omit rules | `assets/chord-symbol-ambiguity-and-parsing.md` |
| Scale-degree spelling, Roman numerals, secondary dominants, modal and altered-tension spelling | `assets/scale-degree-spelling-cheatsheet.md` |
| Trend/reference matrices for streaming evidence, regional evolution, microgenres, and hybrid genre merging | `assets/trend-and-reference-matrices.md` |
| Fast musical brainstorming prompt cards and mutation menus | `assets/musical-brainstorming-cards.md` |
| Project card, decision log, branch tracker, handoff summary templates | `assets/session-brief-and-decision-log.md` |
| Fast web-search operators, query recipes, confidence labels for music research | `assets/web-search-cheatsheet.md` |
| Manual music-theory audit checklist and severity rubric | `assets/music-theory-audit-rubric.md` |

You can quote from `assets/` directly in answers (with attribution like "common ii-V-I voicings"). Don't quote from `references/` verbatim — synthesize.

## Loading discipline

This skill is large. Load minimally:

- **One reference file** is often enough for a focused question.
- **Two to three** for cross-domain or genre-specific questions.
- **Loading 5+ files** suggests the question is too broad — narrow it first.

When in doubt: pick the file most directly addressing the user's *immediate* need. Answer that. Then offer to dig into adjacent topics if they want.

If a file you need doesn't exist yet, say so honestly: "This skill doesn't cover [X] yet — I'll work from general principles and flag what I'm uncertain about." Then do that. Don't fabricate a confident answer.

## Cross-reference key

Many topics reference each other. When you load a file, scan its "Cross-references" section at the bottom and follow links only if they're directly relevant — don't go on a reading tour.

Heavily cross-linked clusters:

- **Harmony cluster**: `chord-construction` ↔ `functional-harmony` ↔ `voice-leading` ↔ `chromatic-harmony` ↔ `modulation` ↔ `jazz-harmony` ↔ `reharmonization` ↔ `assets/chord-symbol-ambiguity-and-parsing` ↔ `assets/scale-degree-spelling-cheatsheet`. These all talk to each other; pick the one that matches the user's level and genre frame, and only branch out if needed.
- **Melody cluster**: `melodic-construction` ↔ `motivic-development` ↔ `phrase-structure` ↔ `songwriting/hooks-and-memorability`.
- **Form cluster**: `classical-forms` ↔ `popular-song-forms` ↔ `narrative-and-transitions` ↔ `production-aware/energy-and-dynamics`.
- **Texture cluster**: `voicing-and-texture` ↔ `arrangement-density` ↔ `production-aware/arrangement-for-mix`.
- **Instrument idiom cluster**: `instrument-idiom/overview` ↔ `piano-keyboards` ↔ `guitar` ↔ `bass` ↔ `drums-percussion` ↔ `strings` ↔ `winds` ↔ `brass` ↔ `vocals` ↔ `orchestration/instruments-ranges-character`.
- **Asian pop cluster**: `kpop-jpop` ↔ `cpop-and-southeast-asian-pop` ↔ `korean-traditional`. Often loaded together for cross-Asian projects.
- **Lyric / language cluster**: `songwriting/lyric-writing` ↔ `songwriting/topline-craft` ↔ `fundamentals/prosody-and-language`.
- **Meta-mode cluster** (the agent's own mode): `workflow` ↔ `critique-and-feedback` ↔ `teaching-composition` ↔ `analysis`. Different modes of helping; the agent should know which it's in.
- **Creative collaboration cluster**: `creative-workflows/musical-brainstorming` ↔ `creative-workflows/user-agent-collaboration` ↔ `creative-workflows/revision-and-feedback-loops` ↔ `creative-workflows/answer-calibration` ↔ `assets/musical-brainstorming-cards` ↔ `assets/session-brief-and-decision-log` ↔ `assets/response-templates` ↔ `assets/diagnostic-checklists`. Use when the user wants interactive ideation or iterative revision, not just theory. Use answer calibration when the answer shape is right but length, loading, or creative variation needs tighter control.
- **Media scoring cluster**: `film-tv-scoring` ↔ `game-music` ↔ `media-and-commercial-music`. Linear vs. interactive vs. short-form / functional — three flavors of "music for picture / message".
- **Roots / heavy / sacred / Brazilian genre cluster**: `country-americana` ↔ `gospel-and-ccm` ↔ `pop-rock` ↔ `metal-punk-hardcore` ↔ `brazilian-pop-and-funk` ↔ relevant `instrument-idiom/` files. Use for band-based idiom and culturally specific popular styles.
- **Currentness / reference research cluster**: `assets/web-search-cheatsheet` ↔ `research/web-music-trend-research` ↔ `research/user-listening-context-and-streaming-services` ↔ `research/reference-track-digging` ↔ `research/style-reference-and-copyright` ↔ `research/regional-trend-evolution-analysis` ↔ `production-aware/pre-production-decisions` ↔ relevant genre files. Use when current scenes, charts, user streaming-service evidence, living artists, copyright-safe style requests, or reference tracks matter.
- **Microgenre / hybrid genre cluster**: `genres/minor-and-hybrid-genres` ↔ `genres/regional-scene-starters` ↔ `research/web-music-trend-research` ↔ `research/regional-trend-evolution-analysis` ↔ `research/reference-track-digging` ↔ relevant major genre files. Use when the genre label is small, unstable, platform-born, local, or deliberately hybrid.
- **Folk / roots / tradition-specific cluster**: `genres/folk-roots-and-traditions` ↔ `genres/regional-scene-starters` ↔ `techniques/microtonal` ↔ `fundamentals/prosody-and-language` ↔ relevant regional genre files. Use when the user asks for folk/traditional influence or uses broad labels that need specificity.
- **Agent output cluster**: `assets/response-templates` ↔ `assets/diagnostic-checklists` ↔ `critique-and-feedback` ↔ `workflow` ↔ `creative-workflows/revision-and-feedback-loops`. Use when the user needs a structured answer, not more theory.
- **Release validation cluster**: `validation/first-release-readiness` ↔ `validation/phase-b-correctness-pass` ↔ `validation/prompt-smoke-tests` ↔ `validation/phase-c-smoke-test-results` ↔ `validation/rc1-packaging-checklist` ↔ `assets/music-theory-audit-rubric` ↔ `scripts/music_theory_sanity_check.py` ↔ `RELEASE-ROADMAP.md` ↔ `VALIDATION-v1.0.md` ↔ `KNOWN-LIMITATIONS.md`. Use for maintainer QA and first-release planning.
- **Exploration / system techniques cluster**: `techniques/algorithmic-and-AI-assisted` ↔ `techniques/microtonal` ↔ `techniques/constraint-based-composition` ↔ `techniques/20th-century-techniques`. For composers reaching beyond the standard 12-TET / functional-harmony / through-composed default.

## When to ask the user before loading

Ask before loading if:

- The genre is critical and unclear ("write a chord progression for a sad bridge" — pop? jazz? cinematic?).
- The user's level is critical and unclear (a beginner asking about "modes" needs different depth than a jazz player asking about "modal harmony").
- The request is ambitious and could go in opposing directions ("write a song" — you need to know length, genre, instrumentation, lyrical theme, etc. before being useful).
- Personal listening context would materially change the answer; ask whether the user can share a public playlist, 5-10 tracks, or a Replay/Wrapped summary, but never ask for login credentials.

Don't ask for the sake of asking. If context is clear or the answer doesn't depend on the missing info, just answer.

# Revision and Feedback Loops

Use this file when the user is iterating on a musical idea across turns: “make it less cheesy,” “keep the chords but change the vibe,” “the second idea was closest,” “try again but darker,” “don’t change the hook,” or “combine option 1 and 3.”

For broader collaboration state, see `user-agent-collaboration.md`. For quick templates, see `../../assets/session-brief-and-decision-log.md`. For troubleshooting vague problems, see `../../assets/diagnostic-checklists.md`.

## Revision principle

A revision request usually contains three layers:

1. **Keep** — what the user liked.
2. **Change** — what did not work.
3. **Direction** — the musical axis to move along.

Do not rewrite the whole idea unless the user asks for a full reset. Preserve the liked material and change the smallest musical variable that could solve the problem.

## Parse the user's feedback

| User phrase | Likely meaning | Musical variables to try |
|---|---|---|
| “Too cheesy” | too predictable, too bright, too familiar, too sentimental | reharm one cadence, reduce diatonic lift, change rhythm, darker voicing, remove obvious stock progression |
| “Too boring” | not enough contrast or identity | contour peak, rhythmic hook, non-chord tone, register shift, texture entrance |
| “Too busy” | too many simultaneous changes | reduce rhythmic density, simplify bass, remove counterline, widen arrangement space |
| “Too sad” | affect overshoots target | brighten one chord, raise register, use warmer rhythm, avoid continuous descending bass |
| “Not sad enough” | affect undershoots target | borrowed iv, ♭VI, minor plagal color, darker topline, slower harmonic rhythm |
| “Too jazzy” | extensions/voice-leading exceed genre frame | simplify extensions, use triads, reduce ii–V movement, keep color in one passing chord |
| “Not modern enough” | texture/rhythm/form feels dated | update groove, hook timing, vocal rhythm, sparse intro, production-aware arrangement cue |
| “Too close to the reference” | originality risk | change contour, rhythm, harmonic rhythm, form placement, instrumentation role |
| “Keep the hook” | preserve identity element | do not alter contour/rhythm/lyric placement unless asked |
| “Use option 2 but darker” | branch merge with affect shift | retain form/groove, change harmony/voicing/register |

## Minimal revision protocol

Use this when the user gives feedback but not a full new prompt.

```md
What I’m keeping:
- ...

What I’m changing:
- ...

Revision:
[concrete musical output]

Why it moves closer:
- ...

Still adjustable:
- ...
```

## Before/after block

```md
Before:
| C | Am | F | G |

After — darker but same basic shape:
| C(add9) | Am7 | Fm6 | Gsus4 G |

Kept:
- 4-bar loop
- tonic opening
- dominant return

Changed:
- borrowed iv color in bar 3
- softer dominant setup
```

## Three revision depths

| Depth | Use when | What changes |
|---|---|---|
| Micro | user likes most of it | one voicing, one note, one rhythm, one bass note |
| Meso | user likes the direction but not the result | chord substitution, phrase contour, groove density, section contrast |
| Macro | user says the concept is wrong | form, genre frame, emotional lane, reference set |

Default to micro or meso. Macro revisions can feel like the agent ignored the user.

## Branch handling

When there are multiple versions, name them.

```md
Branch A — clean pop version
Branch B — darker borrowed-chord version
Branch C — rhythm-first version
```

When the user picks one:

```md
Selected branch: B — darker borrowed-chord version
Keep from branch A: simpler vocal rhythm
Discard from branch C: syncopated bass
```

## Similarity reduction loop

Use when the user says a draft is too close to a reference.

| Similarity area | How to diverge |
|---|---|
| Melody contour | invert the contour, move peak earlier/later, change leap size |
| Rhythm | shift syncopation, change note density, alter pickup placement |
| Harmony | change harmonic rhythm, substitute one functional chord, change bass motion |
| Form | move hook location, add/remove pre-chorus, change phrase length |
| Arrangement | change instrument role, register, texture, density arc |
| Lyric/prosody | change stress pattern, vowel target, rhyme placement |

Keep the desired energy, function, or emotional target; change protected or signature expression.

## Feedback questions that are worth asking

Ask one question only when the next revision would otherwise branch too widely.

Good questions:

- “Should I keep the chord progression and only change melody/texture?”
- “Do you want darker harmony or darker arrangement?”
- “Should the hook stay exactly the same rhythmically?”
- “Is the reference too close because of melody, groove, or instrumentation?”

Avoid asking vague questions like “What vibe do you want?” when the user has already given a direction. Convert their words into musical variables and proceed.

## Cross-references

- User-agent collaboration → `user-agent-collaboration.md`
- Musical brainstorming → `musical-brainstorming.md`
- Session brief and decision log → `../../assets/session-brief-and-decision-log.md`
- Diagnostic checklists → `../../assets/diagnostic-checklists.md`
- Style/reference safety → `../research/style-reference-and-copyright.md`
- Response templates → `../../assets/response-templates.md`

# Musical Brainstorming

This file is for generating musical ideas with the user before the piece is fully defined. It is not generic project brainstorming. It treats musical ideas as playable materials: motives, grooves, progressions, textures, lyrical images, form arcs, instrumentation, constraints, and reference relationships.

Use this file when the user says things like:

- "I want to start a song but I don't know where to begin."
- "Let's brainstorm ideas."
- "Give me several directions for this chorus / cue / theme."
- "I have a mood but no musical material yet."
- "I'm stuck and need sparks."
- "Can we explore options before committing?"

For quick prompt cards, use `../../assets/musical-brainstorming-cards.md`. For session state and decision tracking, use `user-agent-collaboration.md` and `../../assets/session-brief-and-decision-log.md`.

## Core principle

Brainstorming is not the same as giving an answer. In brainstorming mode, the agent should create **possibility space** first, then help the user select and develop one path.

A useful loop:

```text
intent → seed → divergent options → quick audition → selection → focused development → next decision
```

Do not over-explain theory during the divergent phase. Give compact, playable options. Explanation comes after a candidate feels promising.

## The two modes: divergent and convergent

### Divergent mode

Goal: generate many distinct musical directions.

Good outputs:

- 6–12 short options
- contrasting directions, not tiny variants
- at least one safe option, one bold option, one strange option
- concrete musical material: chords, rhythm, scale degrees, texture, form move, lyric image

Bad outputs:

- one polished answer too early
- long theory lecture
- ten options that all mean the same thing
- options that cannot be played or tested quickly

### Convergent mode

Goal: choose, combine, simplify, or develop.

Good outputs:

- compare options by effect
- name trade-offs
- preserve the user's preference
- turn one sketch into 4–16 bars
- define the next small decision

Bad outputs:

- keep adding ideas forever
- erase the user's chosen direction
- solve everything at once

## First response pattern

When the user asks for brainstorming and gives little context, do not ask five questions. Ask **one useful question** or make a reasonable assumption and provide a starter board.

Good:

```text
Let's start broad. I'll assume contemporary pop unless you say otherwise. Here are eight starting points: two chord-first, two groove-first, two melody-first, one lyric-image first, and one weird texture-first idea.
```

Also good:

```text
Give me one anchor: genre, emotional target, or reference track. If you don't have one, I can start from mood alone.
```

Avoid:

```text
What key, tempo, instrumentation, genre, lyrical subject, target platform, vocal range, and reference tracks do you want?
```

## The minimum creative brief

A complete brief is nice, but brainstorming can start from only one of these:

| Seed type | Example | What the agent can generate |
|---|---|---|
| Emotion | lonely but not sad | mode, register, tempo, density, lyric images |
| Genre | J-pop chorus | form, tempo, rhythm, harmonic lift, topline shapes |
| Reference | "like the energy of X, not the melody" | craft-variable profile, safe transformations |
| Instrument | acoustic guitar | idiomatic voicings, strum/pick patterns, song shape |
| Lyric image | rain on a city street | motif, harmony color, texture, rhyme field |
| Rhythm | 6/8 but modern | groove options, phrase lengths, chorus lift |
| Constraint | only three chords | progression variants, bass motion, texture development |
| Function | boss battle loop | intensity states, loopable cells, transition stingers |

If none is given, start with 3 mood choices and 3 workflow choices.

## Brainstorming lenses

Use these lenses to make options genuinely different.

### 1. Harmony-first

Generate the emotional engine through chords.

Example output:

```md
1. Soft ache: | Cmaj9 | Em7 | Am9 | Fmaj7(#11) |
2. Darker lift: | Am | F | C/E | Gsus4 G |
3. Cinematic ambiguity: | C | Abmaj7 | Fm6 | Gsus4 |
```

Use when the user asks about mood, color, or a section transition.

Cross-references: `../harmony/functional-harmony.md`, `../harmony/modal-harmony.md`, `../harmony/chromatic-harmony.md`, `../harmony/reharmonization.md`.

### 2. Melody-first

Generate a memorable cell before chords.

Example output:

```md
Motif cell in C minor:
^5 - ^6 - ^5 - ^3 | ^4 - ^3 - ^2

Development options:
- repeat it one step higher
- answer it with longer notes
- compress it into a hook chant
```

Use when the user wants a theme, hook, topline, riff, or motif.

Cross-references: `../melody/melodic-construction.md`, `../melody/motivic-development.md`, `../songwriting/topline-craft.md`.

### 3. Rhythm/groove-first

Generate motion before pitch.

Example output:

```text
Kick: 1 . . a | 2 . & . | 3 . . a | 4 . & .
Snare/clap: 2 and 4
Hook rhythm: long-short-short, then a held note over barline
```

Use when the track feels stiff, the genre is groove-defined, or the user does not know the harmony yet.

Cross-references: `../rhythm-groove/groove-and-feel.md`, `../rhythm-groove/rhythmic-devices.md`.

### 4. Texture-first

Generate the sonic world through density, register, and timbre.

Example output:

```md
- Verse: one dry vocal + muted guitar pulses + sub on downbeats only
- Pre: add high ostinato + harmony pad widening every two bars
- Chorus: octave vocal doubles + open-hat pattern + bass becomes melodic
```

Use when the user describes atmosphere, cinematic feeling, or arrangement problems.

Cross-references: `../orchestration/arrangement-density.md`, `../orchestration/voicing-and-texture.md`, `../production-aware/energy-and-dynamics.md`.

### 5. Form/narrative-first

Generate the dramatic arc before details.

Example output:

```text
A. Slow reveal: intro → small verse → half chorus → larger verse → full chorus
B. Immediate hook: chorus tag intro → verse → pre → chorus → post-chorus
C. Spiral: 4-bar loop, each repeat adds one destabilizing element
```

Use when the user wants a whole piece, cue, game loop, musical theatre song, or section plan.

Cross-references: `../form/popular-song-forms.md`, `../form/narrative-and-transitions.md`, `../genres/game-music.md`, `../genres/musical-theatre.md`.

### 6. Constraint-first

Generate ideas by limiting choices.

Example constraints:

- only three chords
- melody must use five notes
- no downbeat chord changes
- bass moves stepwise only
- chorus hook must start before bar 1
- every section uses the same motive in a new rhythm
- no cymbals until the chorus
- only register changes, no new instruments

Use when the user is blocked, overwhelmed, or making generic choices.

Cross-references: `../techniques/constraint-based-composition.md`, `../techniques/theme-and-variation.md`.

### 7. Reference-triangulation

Use multiple references to avoid copying one track.

```text
Reference A = energy
Reference B = harmony density
Reference C = vocal phrasing
Your song = new melody, new lyrics, new form choices
```

Use when the user has taste but not a concrete composition direction.

Cross-references: `../research/reference-track-digging.md`, `../research/style-reference-and-copyright.md`, `../research/user-listening-context-and-streaming-services.md`.

## Brainstorming output formats

### Option board

Use for early-stage ideation.

```md
| Option | Core move | Effect | Try this now |
|---|---|---|---|
| 1. Warm ache | vi-IV-I-V with upper extensions | emotional but familiar | play in A minor at 78 BPM |
| 2. Floating | I-II-IV-I Lydian color | hopeful, less resolved | keep bass on tonic for two bars |
| 3. Tense cinematic | i-♭VI-iv-Vsus | darker pull | put ^5 on top of the first chord |
```

### Three-lane sketch

Use when the user wants contrasting directions.

```md
**Safe lane**: genre-native, immediately usable.
**Fresh lane**: one unusual move, still practical.
**Wild lane**: high-risk but inspiring.
```

### Seed pack

Use when the user just needs sparks.

```md
- Chord seed:
- Motif seed:
- Groove seed:
- Texture seed:
- Lyric/image seed:
- Form twist:
```

### Prototype card

Use after choosing an idea.

```md
Key:
Tempo:
Feel:
4-bar core:
Hook rhythm:
Arrangement entrance:
Main risk:
Next experiment:
```

## How to brainstorm without cloning references

When a reference track is involved, extract variables rather than surface signatures.

Allowed variables:

- tempo range
- energy arc
- groove family
- harmonic density
- topline contour type
- arrangement density
- broad production role, e.g., dry vocal vs. stacked vocal
- emotional function

Avoid copying:

- melody
- lyric phrases
- distinctive riff or bassline
- unique sample
- signature vocal identity
- producer tag
- exact arrangement sequence

If the user asks too directly for a living artist's style, route through `../research/style-reference-and-copyright.md`.

## Common brainstorming traps

| Trap | Symptom | Fix |
|---|---|---|
| Too many questions | user loses momentum | ask one anchor or start with assumptions |
| Too much theory | ideas feel academic | give playable material first |
| Too many options | user cannot choose | group into 3 lanes |
| Too safe | everything sounds generic | add one constraint or one contradiction |
| Too weird | nothing feels usable | keep form/groove stable while changing one variable |
| Premature polishing | energy dies before sketch exists | prototype first, refine later |
| Reference cloning | idea is too close to a track | triangulate 2–3 references, replace protected elements |
| Agent domination | user becomes passive | ask which option sparks and why |

## Working with different user types

### Beginner

Use plain language, fewer options, more concrete examples.

```text
Try this four-chord loop first. It works because the bass falls step by step, which usually feels natural and emotional.
```

### Advanced composer

Use denser variables and let them choose the axis.

```text
We can mutate the cell by inversion, metric displacement, or harmonic reinterpretation. Which axis do you want to stress?
```

### Producer / DAW-based writer

Use loop, section, density, and topline language. Avoid notation-heavy answers unless requested.

### Score-based composer

Use motive, register, orchestration, counterpoint, and form language.

### Lyric-first songwriter

Start with title, image field, vowel shape, and phrase length. Then attach melody/harmony.

## A default 20-minute brainstorming session

The agent can run this as a conversational pattern.

### Minute 0–2: brief

Ask one of:

- What emotion should the listener feel?
- What reference lane are we near?
- What section are we writing?

### Minute 2–6: option board

Generate 6–8 options across different lenses.

### Minute 6–8: user selection

Ask:

```text
Which two feel alive, even if they're not perfect?
```

### Minute 8–12: combine or contrast

Make two hybrids:

```text
A + C = stable groove, darker harmony
B + F = hook-first structure, stranger texture
```

### Minute 12–16: prototype

Create 4–8 bars of concrete material.

### Minute 16–20: next decision

Ask for one choice:

- develop chorus
- write verse contrast
- change key/tempo
- make it more idiomatic for instrument/voice
- generate lyrics/topline

## The "yes, and / no, but" revision rule

When the user reacts, translate their reaction into a musical direction.

| User reaction | Agent interpretation |
|---|---|
| "Too pretty" | add tension, reduce consonant extensions, roughen rhythm/timbre |
| "Too dark" | raise register, add major color, thin low density |
| "Too generic" | change bass, rhythm, phrase length, or one borrowed chord |
| "Too complicated" | reduce harmonic rhythm, fewer syncopations, simpler hook contour |
| "Not emotional enough" | stronger melodic appoggiatura, register peak, harmonic arrival |
| "Too close to reference" | preserve only one variable, replace melody/groove/form |

## Micro-prompts the agent can use

- "Give me three safe ideas and three reckless ideas."
- "What if the chorus starts with the hook before the downbeat?"
- "What if the bass tells the story instead of the chords?"
- "Make the verse smaller without making it boring."
- "Keep the same chords but change the emotional reading."
- "Write the wrong version first, then fix it."
- "Give me one idea that would work live and one that only works in production."
- "Turn this mood into a rhythm, not a chord."

## Cross-references

- Fast brainstorming cards → `../../assets/musical-brainstorming-cards.md`
- Session state and decision tracking → `user-agent-collaboration.md`, `../../assets/session-brief-and-decision-log.md`
- General agent workflow → `../workflow.md`
- Critique after a sketch exists → `../critique-and-feedback.md`
- Teaching mode → `../teaching-composition.md`
- Reference digging → `../research/reference-track-digging.md`
- Current web research → `../research/web-music-trend-research.md`
- Web search cheat sheet → `../../assets/web-search-cheatsheet.md`

{"prompt_id": "P1_chorus_diagnosis", "category": "Vague diagnosis", "condition": "A_baseline", "model": "glm-5.1", "system_prompt_chars": 92, "user_prompt": "내 곡 후렴이 안 꽂혀. 어떻게 고치지?", "response": "후렴이 귀에 안 꽂히는 원인과 구체적인 해결책 5가지입니다.\n\n1. **멜로디 단순화 및 반복**: 멜로디가 너무 산만하면 남지 않습니다. 기억하기 쉬운 짧은 멜로디(모티프) 하나를 만들어 후렴 안에서 2~3번 반복하십시오.\n2. **벌스와의 대비**: 벌스와 음역대, 리듬, 에너지가 비슷하면 후렴이 튀지 않습니다. 후렴에서 음역대를 확 올리거나, 반대로 벌스보다 리듬을 여유롭게(긴 음표로) 만들어 차이를 주십시오.\n3. **훅(Hook) 가사 배치**: 한 번 들으면 따라 부를 수 있는 짧고 강렬한 한 줄(훅)을 후렴의 첫머리나 마지막에 배치하고, 운율(라임)을 맞추십시오.\n4. **악기 리프 추가**: 보컬 멜로디만으로 부족하다면, 후렴 시작과 함께 귀에 맴도는 짧은 악기 리프(신스, 기타 등)를 멜로디와 겹쳐 연주하십시오.\n5. **프리 코러스(빌드업) 강화**: 후렴이 갑자기 튀어나오면 어색합니다. 후렴 직전 구간에서 드럼 킥을 끊거나, 코러스 화음을 넣어 '이제 폭발한다'는 기대감을 먼저 조성하십시오."}
{"prompt_id": "P1_chorus_diagnosis", "category": "Vague diagnosis", "condition": "B_skill_only", "model": "glm-5.1", "system_prompt_chars": 13617, "user_prompt": "내 곡 후렴이 안 꽂혀. 어떻게 고치지?", "response": "후렴이 \"안 꽂힌다\"는 건 여러 원인이 있을 수 있어요. 진단 먼저, then 처방.\n\n---\n\n## 빠른 진단 체크리스트\n\n| 요소 | 후렴이 인상적이려면… | 의심 포인트 |\n|------|---------------------|-------------|\n| **멜로디 높이** | 후렴 최고음 > 절 최고음 | 절이랑 비슷한 음역? |\n| **리듬 훅** | 반복·지속되는 리듬 모티프 | 가사 얹는 리듬이 매번 달라? |\n| **화성 변화** | 절과 다른 코드 진행 또는 조바꿈 | 절이랑 같은 I-V-vi-IV? |\n| **밀도** | 더 많은 악기·보이스 레이어 | 후렴인데 편곡이 절이랑 같아? |\n| **대비** | 절→후렴 순간 에너지 점프 | 도입부→후렴 전환이 서서히? |\n| **가사 훅** | 짧고 강한 구절, 반복 | 한 줄이 길거나 산문적? |\n\n---\n\n## 가장 흔한 3가지 원인과 구체적 처방\n\n### 1. 멜로디가 절이랑 너무 비슷하다\n**해결:** 후렴 첫 음을 절 최고음보다 3도~5도 위에서 시작.\n\n```\n절 끝: E4 → 후렴 첫음: A4 이상\n예)절: C4-D4-E4-G4  →  후렴: G4-A4-C5-D5\n```\n\n### 2. 화성이 안 바뀐다\n**해결:** 후렴 진행을 절과 다르게. 가장 즉각적인 옵션들:\n\n| 옵션 | 코드 (C 메이저 기준) | 효과 |\n|------|---------------------|------|\n| **소디미니시시** | C → A♭ → B♭ → C | 극적, 영웅적 |\n| **iv 도입** | Fm → C → G → C | 그리웋하면서 단단 |\n| **V/V → V** | D → G → Am → Fm | 텐션→해결, 빠른 훅 |\n| **나폴리 6도** | C → D♭maj7 → G → C | 어둡고 파워풀 |\n\n### 3. 리듬 훅이 없다\n**해결:** 가사 음절에 반복되는 리듬 패턴 부여.\n\n```\n\"다가 와\" ×2  (점-점-_____)  ← 4음절이면 점-점-박-박 패턴을 반복\n\"말해 줘\"      (점-박)\n→ \"다가와-다가와-말해줘\" 리듬: x-x- x-x- x-x\n```\n\n하이라이트 가사에 **같은 리듬 3번 반복** → 이것만으로도 훅이 됩니다.\n\n---\n\n## 다음 질문에 답해주면 더 정확히 도와줄게요\n\n1. **장르** (K-pop 발라드? 힙합? 인디? 락?)\n2. **지금 후렴 코드 진행** (알려주면 바로 대안 제시)\n3. **어떤 안 꽂힘?** (멜로디가 안 외워짐 / 에너지가 안 올라감 / 가사가 안 남음)\n\n이 세 가지만 알려줘도 훨씬 구체적으로 고쳐줄 수 있어요."}
{"prompt_id": "P1_chorus_diagnosis", "category": "Vague diagnosis", "condition": "C_skill_plus_nav", "model": "glm-5.1", "system_prompt_chars": 40212, "user_prompt": "내 곡 후렴이 안 꽂혀. 어떻게 고치지?", "response": "후렴이 안 꽂히는 원인은 보통 몇 가지 패턴 중 하나입니다. 진단해 보고 고쳐 보세요:\n\n---\n\n## 1. 절→후렴 리프트가 부족하다\n\n후렴은 절보다 **에너지가 올라야** 꽂힙니다. 체크:\n\n- **음역**: 후렴 최고음이 절 최고음보다 최소 단3도 이상 높은가?\n- **밀도**: 악기 레이어가 후렴에서 두꺼워지는가?\n- **리듬**: 보컬 프레이징이 더 빨라지거나 싱코페이션이 많아지는가?\n\n**빠른 수정**: 후렴 첫 마디 멜로디를 3~5음 위로 올려 보세요.\n\n---\n\n## 2. 훅 멜로디가 약하다\n\n기억에 남는 후렴은 **반복 가능한 짧은 모티프** + **인상적인 간진행**이 있습니다.\n\n**해결책**:\n- 후렴 멜로디를 1~2마디 짧은 셈(모티프)으로 줄이고, 그것을 반복·변형으로 전개\n- 도약(5도 이상)을 한 곳 넣어 보세요. 예: `도→레→파↑↗︎(5도 상행)→미`\n- 리듬 변형: 절이 8분음 wired이면 후렴은 점음표나 부점 리듬으로\n\n---\n\n## 3. 화성이 너무 예측 가능하다\n\n절-후렴 화성이 같으면 후렴이 \"터지지\" 않습니다.\n\n**해결책**:\n- 후렴 진입을 **IV** 또는 **♭VI**로 시작 (예: C메이저 곡이면 F메이저 또는 A♭메이저)\n- 절이 `I → vi → IV → V`라면 후렴은 `IV → I → V → vi`로 순서 교환\n- 후렴 첫 코드에 **sus4 해결** 또는 **add9** 추가로 색감 차이\n\n---\n\n## 4. 가사·음절 밀도 문제\n\n절보다 후렴이 음절이 적고 음이 길어야 힘이 실립니다.\n\n**해결책**: 후렴 첫 라인을 3~5개 음절로 압축하고 긴 음가에 담기.\n\n---\n\n## 장르가 중요합니다\n\nK-pop, 발라드, 힙합, 락… 장르마다 \"꽂히는 후렴\"의 공식이 다릅니다. **어떤 장르인지 알려주시면** 더 구체적으로 조언할 수 있습니다.\n\n---\n\n지금 곡의 키, 현재 코드 진행, 후렴 멜로디 움직임을 알려주시면 바로 수정안을 제시할게요."}
{"prompt_id": "P2_kpop_dark_bridge", "category": "Genre + emotion", "condition": "A_baseline", "model": "glm-5.1", "system_prompt_chars": 92, "user_prompt": "K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어.", "response": "K-pop 발라드 브릿지를 어둡게 만들기 위한 구체적이고 실전적인 방법:\n\n1. **화성 (코드):** 밝은 코드를 피하고 **디미니시드 코드**나 **단2도 불협화음**을 사용해 불안정함을 조성하세요. 베이스라인을 **반음 하행**시키면 긴장감과 우울함이 극대화됩니다.\n2. **멜로디:** 고음역 클라이맥스 대신 **중저음역**으로 끌어내리세요. 안정적인 음정 대신 **증4도(트라이톤)**나 단2도 음정 도약으로 서스펜스(매달림) 느낌을 주세요.\n3. **편곡:** 악기를 전부 터트리기 전에 오히려 악기를 걷어내고 **적막(무음, 공백)**을 사용하세요. 잔향이 긴 어두운 신디사이저, 저음 첼로, 무거운 서브베이스로 공간을 채우세요.\n4. **보컬:** 깔끔한 가창 대신 **보컬 프라이(목소리 떨림)**, 거친 숨소리, 속삭임을 활용하세요. 브릿지 끝부분에서 옥타브 점프와 함께 거친 파열음으로 절규하듯 전환하세요.\n5. **가사:** 단순한 슬픔을 넘어 '광기, 상실, 시선 고정, 통제 불능' 등의 단어로 서사를 어둡게 비틀어 보세요."}
{"prompt_id": "P2_kpop_dark_bridge", "category": "Genre + emotion", "condition": "B_skill_only", "model": "glm-5.1", "system_prompt_chars": 13617, "user_prompt": "K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어.", "response": "K-pop 발라드 브릿지를 어둡게 만드는 구체적인 방법 몇 가지:\n\n---\n\n## 1. 평행단조 차용 (Modal Mixture)\n\n가장 즉각적이고 장르에 잘 맞는 방법.\n\n**C 메이저 기준:**\n- `F → Fm → C` (♭VI → iv → I) — Fm이 순간 그림자를 드리움\n- `Am → A♭m → G7` — A♭m은 ♭vi, 극적인 하강\n- `Dm7 → D♭maj7 → Cm7 → Fm7` — ♭VII → ♭VI 진행\n\n## 2. 단조의 이차도수 (Secondary Dominants to Minor)\n\n목표음을 단조로 해결.\n\n- `E7 → Am` (V/vi → vi) — 이미 어두운 Am으로\n- `B7 → Em` (V/iii → iii) — 더 깊은 그림자\n- `F7 → B♭m` (♭VII 처리 가능) — 예측 밖의 색채\n\n## 3. 색채중간 (Chromatic Mediant)\n\nK-pop 발라드 브릿지 끝에서 흔히 쓰는 극적 전조 느낌.\n\n- `Cmaj7 → A♭maj7` — I → ♭VI, 3도 반음 관계\n- `Cmaj7 → Emaj7` — I → III, 올림 중간\n- `Fm7 → Bmaj7` — ♭IV → III⁺ (비트로스 진행)\n\n## 4. 종결구 변경\n\n3단음 아래 종결 (Plagal ♭VI cadence):\n- `A♭maj7 → C/G → Fm → C`\n- 마지막 V 대신 IVm으로 \"미결\" 느낌 유지\n\n---\n\n**실제 예시 진행 (C 메이저 곡의 브릿지):**\n\n```\n| Fmaj7  | Fm7    | Dm7   | D♭maj7 |\n| Am7    | A♭m7   | G7sus4| G7♭9   |\n```\n\n- F→Fm: modal mixture\n- Dm→D♭: ♭VII 하강\n- Am→A♭m: ♭vi 하강\n- G7♭9: 증극 긴장, 코러스 해결 준비\n\n어느 정도 \"어둡게\"인지(섬뜩함 vs 쓸쓸함)에 따라 골라 쓰면 됨."}
{"prompt_id": "P2_kpop_dark_bridge", "category": "Genre + emotion", "condition": "C_skill_plus_nav", "model": "glm-5.1", "system_prompt_chars": 40212, "user_prompt": "K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어.", "response": "# K-pop 발라드 브릿지: 더 어둡게 만드는 방법\n\n## 화성적 접근\n\n**1. 모달 믹스처 (Modal Mixture)**\n메이저 키에서 마이너의 플랫 디그리를 차용:\n- `♭VI` → `♭VII` → `i` (C 메이저: A♭maj7 → B♭ → Cm)\n- `iv` 대신 `IV` 자리에 마이너 서브도미넌트 사용: `Fm6` 대신 `F`\n- `♭III` → `♭VI` 시퀀스: C 메이저에서 E♭maj7 → A♭maj7\n\n**2. 크로매틱 메디안트**\n3도 관계의 색채 변화:\n- `I` → `♭III` → `♭V` (C → E♭ → G♭) — 극적인 암흑감\n- `I` → `♯iv°` → `V` (C → F♯° → G)\n\n**3. 세컨더리 도미넌트 → 마이너**\n- `V7/vi` → `vi` (E7 → Am in C) — 라딕스 느낌의 어두운 해결\n- `V7/ii` → `ii` (A7 → Dm) — 서브도미넌트 방향의 긴장\n\n## 선율적 접근\n\n- 브릿지 시작을 낮은 레지스터에서 (C3~E4 남성, A3~C5 여성)\n- 트라이톤 증4/감5 도약 활용\n- 하행 반음 진행: `^5 → ♭5 → 4` 또는 `^1 → ♭7 → ♭6`\n- 레가토 프레이징, 적은 음, 더 긴 지속음\n\n## 텍스처 & 편곡\n\n| 요소 | 변경 |\n|------|------|\n| 베이스 | 더 낮은 레지스터, 펜타토닉 하행 |\n| 피아노 | 저음 옥타브 트리아드, 페달 포인트 |\n| 스트링 | 첼로 디비지, 콘 소르디노 |\n| 리듬 | 하프타임 또는 리타르단도, 공간 많이 |\n| 보컬 | 숨소리 강조, 폴셋 → 풀보이스 전환 |\n\n## 실전 예시 (C메이저 기준)\n\n**기본 브릿지:**\n`Am → F → C → G`\n\n**어둡게 변주:**\n```\n| A♭maj7 | B♭7sus4  | Cm9      | G♭maj7  |\n| Fm7     | D♭/F     | E♭m6     | G7sus4  |\n```\n\n핵심: 메이저 키의 밝음을 `♭VI`, `♭VII`, 마이너 `iv`로 서서히 대체하면서, 불안정한 서스펜디드 코드로 긴장을 유지하고 해결을 지연시키세요.\n\n---\n\n어떤 키에서 작업 중인지, 현재 브릿지 코드 진행을 공유해주시면 구체적으로 변주해 드릴 수 있습니다."}
{"prompt_id": "P3_scale_over_cm7", "category": "Technical theory", "condition": "A_baseline", "model": "glm-5.1", "system_prompt_chars": 92, "user_prompt": "Cm7 위에 어떤 스케일이 자연스러워?", "response": "Cm7 위에서 가장 자연스러운 스케일은 다음 3가지입니다. \n\n1. **C 자연단조 (C Natural Minor / Aeolian)**\n   * **음계**: C - D - Eb - F - G - Ab - Bb\n   * **특징**: 가장 기본적이고 안정적인 마이너 사운드입니다.\n\n2. **C 도리안 (C Dorian)**\n   * **음계**: C - D - Eb - F - G - A - Bb\n   * **특징**: 6음이 Ab에서 A로 바뀝니다. 재즈나 펑크에서 자주 쓰이며, 자연단조보다 밝고 그루브한 느낌을 줍니다.\n\n3. **C 마이너 펜타토닉 / 블루스 스케일**\n   * **음계**: C - Eb - F - G - Bb (+ 블루스 노트: F#)\n   * **특징**: 가장 치고 빠지기 쉽고 무난하게 어울립니다.\n\n**추천 접근**: 기본적으로 **C 마이너 펜타토닉**을 치고, 곡의 분위기(키)에 따라 6음을 **Ab(자연단조)**로 쓸지 **A(도리안)**로 쓸지 결정하여 섞어 쓰는 것이 가장 자연스럽습니다."}
{"prompt_id": "P3_scale_over_cm7", "category": "Technical theory", "condition": "B_skill_only", "model": "glm-5.1", "system_prompt_chars": 13617, "user_prompt": "Cm7 위에 어떤 스케일이 자연스러워?", "response": "# Cm7 위의 스케일\n\n**Cm7** = C - E♭ - G - B♭\n\n가장 자연스러운 선택들:\n\n| 스케일 | 음계 | 느낌 / 용도 |\n|--------|------|-------------|\n| **C 자연단조** (Aeolian) | C D E♭ F G A♭ B♭ | 가장 기본적. 클래식, 팝 발라드 |\n| **C 도리안** | C D E♭ F G **A** B♭ | 재즈 표준. 6번음이 장조(A)라 더 밝고 스무스 |\n| **C 마이너 펜타토닉** | C E♭ F G B♭ | 록·팝·블루스. 간결하고 안전 |\n| **C 블루스 스케일** | C E♭ F **G♭** G B♭ | 블루스·록. 감정적, 관절적인 느낌 |\n\n## 선택 기준\n\n- **장르가 재즈 / 퓨전** → **도리안**부터. `A♭` 대신 `A♮`이 4음 위에서 자연스럽게 해결되며, ii-V에서 모드 유지가 쉽습니다.\n- **장르가 팝·발라드·클래식** → **자연단조**. 6번음 `A♭`이 이미 조표에 있고, 안정적입니다.\n- **즉흥 연주 초보** → **마이너 펜타토닉**으로 시작. 치명적인 빗소리가 없습니다.\n- **더 딥한 블루스 느낌** → 펜타토닉에 `G♭`(블루 노트) 추가.\n\n**장르를 알려주시면 더 구체적으로 추천해 드릴 수 있습니다.**"}
{"prompt_id": "P3_scale_over_cm7", "category": "Technical theory", "condition": "C_skill_plus_nav", "model": "glm-5.1", "system_prompt_chars": 40212, "user_prompt": "Cm7 위에 어떤 스케일이 자연스러워?", "response": "# Cm7 위의 스케일\n\n**Cm7 = C–E♭–G–B♭**\n\n가장 자연스러운 선택은 스타일에 따라 다릅니다:\n\n| 스케일 | 음 | 느낌 | 주로 쓰이는 장르 |\n|--------|-----|------|-----------------|\n| **C 자연단음** (C Aeolian) | C–D–E♭–F–G–A♭–B♭ | 어둡고 차분함 | 팝, 록, 클래식 |\n| **C 도리안** (C Dorian) | C–D–E♭–F–G–**A♮**–B♭ | 밝고 펑키햄 | 재즈, 펑크, 소울 |\n| **C 프리지안** (C Phrygian) | C–**D♭**–E♭–F–G–A♭–B♭ | 스페인적, 긴장감 | 플라멩코, 라틴, 현대재즈 |\n\n---\n\n**가장 기본답:** **C 자연단음** 또는 **C 도리안**.\n\n- 화음음(C, E♭, G, B♭)이 모두 스케일 안에 들어있으므로 두 스케일 모두 \"자연스럽습니다.\"\n- 차이는 6번째 음뿐: **A♭**(자연단음) vs **A♮**(도리안).\n- A naturel이 더 밝고 현대적인 사운드, A♭이 더 전통적 마이너 사운드.\n\n---\n\n**추가 옵션:**\n- **C 마이너 펜타토닉** (C–E♭–F–G–B♭) – 즉흥 연주의 출발점\n- **C 블루스 스케일** (위에 G♭ 추가) – 블루스/록\n- **C 하모닉 마이너** – 화음 진행에서 V7이 등장할 때\n\n어떤 스타일로 작곡/연주하시나요? 그러면 더 세부적으로 추천할 수 있습니다."}
{"prompt_id": "P4_jazz_kpop_hybrid", "category": "Hybrid request", "condition": "A_baseline", "model": "glm-5.1", "system_prompt_chars": 92, "user_prompt": "재즈 화성으로 K-pop 후렴을 만들어보고 싶어.", "response": "K-pop의 대중적인 구조에 재즈의 화성(7th, 9th, altered, Tritone substitution)을 입힌 8마디 후렴 코드 진행과 작곡 팁입니다.\n\n**키(Key): C 메이저**\n\n### 1. 코드 진행 (8마디)\n전형적인 K-pop 후렴의 흐름(I - V - vi - IV 등)에 재즈적 색채를 더했습니다.\n\n*   1마디: **Cmaj9** (I) - 밝고 풍성하게 시작\n*   2마디: **Am11** (vi) - K-pop 특유의 애절함, 11th음으로 재즈 터치\n*   3마디: **Dm9** → **G7#5#9** (ii - V) - **가장 재즈스러운 부분.** Altered dominance로 극적인 긴장감\n*   4마디: **Cmaj7** → **C7#9** (I → I7) - 블루스 느낌의 크로매틱 전환\n*   5마디: **Fmaj9** (IV) - 후렴의 감정적 절정(환상적이고 몽환적인 느낌)\n*   6마디: **Fm6** (iv) - **K-pop의 잔잔함 + 재즈의 크로매틱 하강**\n*   7마디: **Em7** → **Eb7#11** (iii → SubV) - **Tritone 대체.** Eb7은 원래 A7(V/ii) 자리에 들어간 대체 화음으로 매우 세련된 반전\n*   8마디: **Dm9** → **G7b9b13** (ii - V) - 다음 구간을 위한 재즈식 해결\n\n### 2. 멜로디 작법 (핵심)\n*   **어프로치 노트(Approach Note):** G7#5#9 코드(3마디) 위에서는 멜로디가 `Bb`(b5), `Eb`(b9) 혹은 `F#`(#5) 음을 강박(Strong beat)에서 타격하게 작곡하세요. 일반적인 다이아토닉 멜로디와 확연히 다른 '재즈 냄새'가 납니다.\n*   **Fm6 활용:** 6마디 Fm6 코드 위에서는 멜로디가 `A` 자연음(C 메이저 스케일의 6도)을 타격하면, 반주의 `Ab`(F 단조의 3도)와 마찰을 일으키며 K-pop 특유의 '가슴 시린' 느낌과 재즈의 어두운 느낌을 동시에 냅니다.\n\n### 3. 리듬 & 보이싱 팁\n*   **2-5-1(Bossa Nova Feel):** 3마디와 7~8마디의 ii-V 진행에서는 드럼/베이스를 보사노바나 삼바 리듬으로 반 박자 쉬어(K-pop의 특징)와 섞어보세요.\n*   **Shell Voicing:** 피아노나 기타는 근음(Root), 3도, 7도만 남기고 나머지 음은 멜로디에 맡기면(C7 = C-E-Bb), 정통 재즈 앙상블처럼 공간이 넓고 힙한 사운드가 납니다."}
{"prompt_id": "P4_jazz_kpop_hybrid", "category": "Hybrid request", "condition": "B_skill_only", "model": "glm-5.1", "system_prompt_chars": 13617, "user_prompt": "재즈 화성으로 K-pop 후렴을 만들어보고 싶어.", "response": "K-pop 후렴에 재즈 화성을 적용하는 건 좋은 접근이에요. K-pop은 이미 차용화음·부차디미넌트를 자주 쓰기 때문에, 재즈 화성을 덧씌우는 게 자연스럽게 들릴 수 있어요. 핵심은 **\"재즈의 색채 + 팝의 기억성\"** 유지예요.\n\n---\n\n## 옵션 1: 확장화음으로 팝 진행 업그레이드\n\n가장 온건한 방법. 기본 팝 진행에 7th·9th를 더해 톤을 풍부하게.\n\n**C키:**\n```\nCmaj9 → Am9 → Dm11 → G13\n```\n\n원래 C – Am – Dm – G에서 확장음 추가. 멜로디는 팝처럼 직선적으로 유지하면 화성만 정교해져요.\n\n---\n\n## 옵션 2: ii–V–I 기반 후렴\n\n재즈의 핵심 진행을 후렴 뼈대로. K-pop 특유의 \"상승감\"을 재즈적 긴장–해소로 만들어요.\n\n**C키 (빠른 템포, 2마디 단위):**\n```\n| Dm7     G7♭9  | Cmaj7   Cmaj7 |\n| Fmaj7   Bm7♭5 | E7♭9    Am7  |\n```\n\n전반부: ii–V–I. 후반부: iiø–V7–i (병행단조 편향). 이조(V7/vi)가 들어가 K-pop 특유의 감성 전환을 만들어요.\n\n---\n\n## 옵션 3: 트라이톤 치환 + V7alt로 드라마\n\n후렴 하이라이트에 재즈의 가장 강렬한 도구를 배치.\n\n**C키:**\n```\nCmaj7 → D♭7(♯11) → Cmaj7/G → G7alt\nFmaj7 → E♭maj7   → Am7    → D7♭9 → G7alt\n```\n\n- **D♭7** = G7의 트라이톤 치환(♭II7). 베이스 반음 하강으로 긴장 ↑\n- **E♭maj7** = ♭VI 차용화음. 몽환적 색채.\n- **G7alt, D7♭9** = V7에서 최대 크로마티시즘\n\n---\n\n## 옵션 4: 나단조 ii–V–i로 애절 발라드 후렴\n\n슬픈/애절한 K-pop 발라드에 최적.\n\n**Am 기준:**\n```\nAm9  → Bm7♭5 → E7♭9 → Am9\nFmaj7 → Dm7  → E7alt → Am9\n```\n\niiø→V7♭9→i는 재즈에서도 \"애절함\"의 정석이고, K-pop 발라드 감성과 아주 잘 맞아요.\n\n---\n\n## 멜로디 주의사항\n\n화성이 복잡할수록 **멜로디는 심플**하게. K-pop 후렴의 힘은 귀에 쏙 들어오는 훅이에요.\n\n| 화성 위 멜로디 | 효과 |\n|---|---|\n| maj9, ♯11 위 | floating·몽환 느낌 |\n| V7alt 위에서 ♭9, ♯9 | 긴장 일관성 |\n| 3rd·7th 피하기 | 보이싱과 충돌 ↓, 투명도 ↑ |\n\n---\n\n원하는 분위기(애절·몽환·고조 등)나 레퍼런스 곡이 있으면 더 구체적으로 조정해드릴게요."}
{"prompt_id": "P4_jazz_kpop_hybrid", "category": "Hybrid request", "condition": "C_skill_plus_nav", "model": "glm-5.1", "system_prompt_chars": 40212, "user_prompt": "재즈 화성으로 K-pop 후렴을 만들어보고 싶어.", "response": "# 재즈 화성 × K-pop 후렴\n\nK-pop 후렴의 구조적 힘(강렬한 후크, 고음역 멜로디, 에너지 상승)은 그대로 두고, **화성 레이어만 재즈로 교체**하는 게 핵심입니다. 멜로디가 너무 복잡해지면 K-pop 특유의 귀접힘(hook)이 사라집니다.\n\n---\n\n## 3가지 실전 옵션\n\n### ① ii-V-I 변형 — 가장 접근적인 재즈 느낌\n\n**Key: C major, 8마디 후렴**\n\n```\n| Dm7    G7♭9  | Cmaj7    —    |\n| Fmaj7   —    | Em7   A7      |\n| Dm7    G7♭9  | Cmaj7    —    |\n| Fmaj7   Fm7  | Dm7♭5  G7alt |\n```\n\n- ♭9, alt 텐션으로 색채 추가하지만 근음 흐름은 여전히 친숙\n- Fm7(♭VI) = 모달 인터체인지, 클래식 K-pop 감성의 \"서늘한 전조\" 느낌\n- 멜로디는 5도~옥타브 사이 굵직한 선율\n\n### ② 트라이톤 서브 — 더 세련된 도시 느낌\n\n```\n| Dm7      —    | D♭7      —    |\n| Cmaj7    —    | E7♯9     —    |\n| Fmaj7    —    | B∅7  E7      |\n| Am9      —    | D♭7      —    |\n```\n\n- D♭7 = G7의 트라이톤 서브 → 예상치 못한 반음 하강\n- E7♯9 (Hendrix chord) = vi의 세컨더리 도미넌트, 블루스-재즈 색채\n- 멜로디 긴 음표 위에 내성 보이스만 움직이게 하면 K-pop 앙상블 사운드 유지\n\n### ③ 모달 인터체인지 + 상행 II↎V — 가장 드라마틱\n\n```\n| Dm9    G7♭13 | Cmaj9     —    |\n| Fm7     B♭7  | Em7      A7    |\n| Dm9    G7alt | Cmaj9     —    |\n| Am7     D7   | D♭maj7    —    |\n```\n\n- Fm7→B♭7 = ♭VI → V/IV, K-pop에서 자주 쓰는 서늘한 전개\n- 마지막 D♭maj7 = ♭II (네오리만 색채, 곡의 끝을 열어둠)\n- 후렴 반복 시 마지막 2마디를 ||: G7alt → Cmaj7 :|| 로 바꿔 종결감\n\n---\n\n## 작업 팁\n\n| 요소 | K-pop 관점 | 재즈 적용 |\n|------|-----------|----------|\n| **멜로디** | 5도 이상 폭, 히트 음은 강박 | 화음텐션음(9, ♭9, ♯11)을 경과음처럼 스쳐 지나가게 |\n| **보컬 하모니** | 3도 병행, 옥타브 유니슨 | 7th 위에 3-part close harmony로 재배치 |\n| **리듬** | 4/4 포크, 킥 온 1&3 | 베이스를 walking 느낌으로, 드럼은 K-pop 템션 유지 |\n| **전조** | 후렴→후렴구 사이 반음/온음 상행 | ii↎V를 타고 전조 (예: C → E♭ = Dm7→G7→E♭maj7) |\n| **아웃트로** | 반복 후 페이드 또는 반음 하행 종지 | 재즈 텐션 코드로 끝 (예: Cmaj7♯11) |\n\n---\n\n## K-pop × 재즈 레퍼런스 듣기\n\n- **레드벨벳 \"Psycho\"** — maj7 텐션 감성 팝 후렴\n- **스트레이 키즈 \"Levanter\"** — 모달 인터체인지 서늘한 화성\n- **태연 \"INVU\"** — ii↎V 느낌의 하행 진행\n- **어쿠스틱 콜라보/스탠다드 재즈 커버 그룹들** — K-pop 멜로디 위 재즈 리하모니제이션 직접 체감\n\n---\n\n원하시면 **특정 키, BPM, 멤버 수, 보컬 파트**를 알려주면 그 조건에 맞춰 멜로디 + 보컬 편성까지 구체화해 드리겠습니다."}
{"prompt_id": "P5_pansori_ost", "category": "Cultural specificity", "condition": "A_baseline", "model": "glm-5.1", "system_prompt_chars": 92, "user_prompt": "판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?", "response": "판소리의 '창'과 사극 OST의 '스케일'을 결합하려면, 전통 음계와 창법을 현대 편곁으로 쳐내는 작업이 필요합니다. 다음 4단계로 구체적으로 시작해 보세요.\n\n**1. 음계와 모티프 정하기 (뼈대)**\n*   **음계:** 전통 5음계인 **계면조**(솔-라-도-레-미, 슬프고 처절한 느낌) 또는 **평조**(도-레-미-솔-라, 웅장하고 고풍스러운 느낌) 중 곡의 분위기를 선택해 멜로디를 짭니다.\n*   **오프닝:** 곡의 시작은 악기 없이 보컬만으로 판소리 창법(시김새, 즉 꺾고 떠는 발성)으로 1~2마디 불러 theatrical하게 입장하는 방식이 가장 효과적입니다.\n\n**2. 보컬 디렉팅 (판소리 느낌 살리기)**\n*   처음부터 끝까지 판소리 창법을 쓰면 이질감이 생깁니다. **후렴구나 도입부**에만 판소리 특유의 성색(거친 목소리)과 시김새를 넣고, 나머지는 정통 보이스로 부르게 하세요.\n*   곡 중간에 '이이이익', '얼씨고' 같은 추임새나 숨소리(덤덤)를 아주 짧게 샘플링/녹음해서 리듬 악기처럼 배치해 보세요.\n\n**3. 악기 편곡 (사극 스케일 입히기)**\n*   **전통 악기:** 판소리의 북 대신 **꽹과리나 징**의 강렬한 한 박을 기본 리듬에 얹고, 선율은 **해금**(애절함)이나 **가야금**(서정성)이 담당하게 하세요.\n*   **현대 악기:** 북의 리듬 위에 웅장한 **오케스트라 스트링**(앙상블)과 **심벌즈**를 깔아주면 판소리의 촌스러움은 지우고 사극의 웅장함만 남습니다.\n\n**4. 곡 구조 짜기 (추천 형식)**\n*   **도입부:** 해금 독주 또는 무반주 보컬(판소리 창법)\n*   **A파트:** 일반적인 발성의 서사적 멜로디 (스트링+가야금)\n*   **B파트 (프리코러스):** 긴장감 상승, 꽹과리 등 타악기 등장\n*   **후렴:** 폭발적인 오케스트라 + 보컬 시김새 극대화\n*   **간주 (C파트):** 판소리의 '아니리'(말하는 톤)처럼 읊조리듯 노래하고 다시 후렴으로 폭발\n\n**요약하자면:** 계면조/평조 5음계로 멜로디를 짠 뒤 ->도입부를 판소리 창법 무반주 솔로로 시작하고 -> 꽹과리/북의 리듬 위에 오케스트라를 얹어 후렴을 터뜨리는 방식으로 데모를 먼저 만들어 보세요."}
{"prompt_id": "P5_pansori_ost", "category": "Cultural specificity", "condition": "B_skill_only", "model": "glm-5.1", "system_prompt_chars": 13617, "user_prompt": "판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?", "response": "# 사극 OST + 판소리 융합: 시작 가이드\n\n## 1. 먼저 결정할 것: 극적 컨텍스트\n\n같은 사극이라도 장면에 따라 완전히 다릅니다:\n\n| 상황 | 추천 조(調) | 추천 장단 | 느낌 |\n|------|-----------|----------|------|\n| 비극적 죽음/이별 | 계면조 | 진양(느림) | 애절, 체념 |\n| 전투/결의 | 우조 | 중모리~자진모리 | 웅장, 결연 |\n| 회상/그리움 | 평조 | 중모리 | 서정, 담백 |\n| 원혼/저주 | 남조계면 | 진양 | 음산, 불안 |\n\n**예시**: \"주인공이 모든 걸 잃고 홀로 서는 장면\" → 계면조 + 진양장단\n\n---\n\n## 2. 조(調) — 선법 선택\n\n가장 많이 쓰는 세 가지:\n\n**계면조** (가장 판소리적, 애절함)\n```\n시 - 라 - 도 - 레 - 미\n(라 중심, 4도 위 미와 반음 관계 없음 — 펜타토닉이지만 \n시→라로 내려가는 꺾음이 핵심)\n```\n\n**우조** (웅장, 광활함)\n```\n도 - 레 - 미 - 솔 - 라\n(도 중심, 남도계면의 시·미 대신 서구적 메이저 느낌 + \n미↔솔 사이 1.5음 = 판소리 특유의 '늘어짐')\n```\n\n**평조** (서정, 부드러움)\n```\n라 - 도 - 레 - 미 - 솔\n(경기민요조, 현대 귀에 가장 익숙)\n```\n\n---\n\n## 3. 장단(리듬) — 뼈대\n\n장단은 서양 박자와 다릅니다. **강약과 호흡이 다르고**, 각 박의 무게가 다릅니다.\n\n**진양장단** (24박, 느린 6/4 느낌):\n```\n강 - 약 - 중강 - 약 - 강 - 약\n|    |    |     |    |    |\n미...라...........시......라↓\n       (여기서 꺾음)\n```\n→ 애절한 보컬 라인에 적합, 현대 6/8이나 12/8로 변형 가능\n\n**중모리장단** (12박, 중간 템포 4/4 느낌):\n```\n강 - 약 - 중강 - 약\n```\n→ 서사적, 담담한 이야기\n\n**실전 팁**: OST이니 장단을 그대로 쓰기보다 **서양 박자에 장단의 강점 배치만 차용**하는 게 자연스럽습니다.\n- 진양 → 6/8 또는 12/8\n- 중모리 → 4/4 (강-약-중강-약 그루빙)\n- 굿거리 → 4/4 (16분 음표 그루브)\n\n---\n\n## 4. 악기 편성 실전\n\n| 레이어 | 악기 | 역할 | 주의 |\n|--------|------|------|------|\n| 선율 | 해금 | 보컬과 유니슨·응답 | 시김(vibrato)을 과하게 |\n| 선율 | 피리 | 오보이처럼 관통력 | 저역은 약함 |\n| 화성 | 가야금 | 아르페지오, 페달톤 | 12현 튜닝 조정 필요 |\n| 저음 | 거문고 | 베이스라인 | 개방현 울림 활용 |\n| 리듬 | 장구 | 장단 구동 | 민속악 대금 테크닉 차용 |\n| 리듬 | 북 | 포인트 타격 | 전투 씬에서 핵심 |\n| 현대 | 스트링 앙상블 | 패드, 텍스처 | 화성 채우기 |\n| 현대 | 피아노 | 연결선 | 너무 서구화되지 않게 |\n\n**최소 편성**: 해금 + 장구 + 피아노 + 스트링 패드  \n**풀 편성**: 위 전부 + 보컬(판소리 창법 토막)\n\n---\n\n## 5. 곡 구조 (OST 사이즈)\n\n```\nIntro     해금 솔로, 계면조 선율 4마디 (장구 없이)\nA         보컬, 중모리 그루브, 가야금 아르페지오\nB (전개)  장단 진양으로 전환, 해금·보컬 대위, 스트링 증가\nC (절정)  자진모리 급전환, 북+장구 풀스택, 보컬 꺾음 최대치\nOutro     A 선율 재현, 해금만 남아 사라짐\n```\n\n---\n\n## 6. \"판소리 느낌 살짝\"의 핵심 기법\n\n과하게 하면 판소리 곡이 되니, **두세 개만** 차용:\n\n1. **꺾음** — 보컬이나 해금에서 음을 위로 밀었다가 아래로 꺾기  \n   `미 → 시↑ → 라↓` (시를 살짝 높여 부르고 라로 떨어뜨림)\n\n2. **비브라토 차별화** — 한국 음악의 시김은 서양 비브라토보다 **늦게 시작, 폭이 넓음**  \n   해금이나 보컬에서 롱노트 → 2초 후 흔들기 시작\n\n3. **침묵의 미학** — 장단과 장단 사이 1~2박 완전한 여백  \n   서양 OST은 리듬을 채우지만, 판소리는 멈춤이 표현\n\n---\n\n## 즉시 할 수 있는 첫 스텝\n\n1. **계면조로 Cm 펜타토닉 위에서**: `C - Eb - F - G - Bb`\n2. **12/8박자, BPM 60~70** (진양 느낌)\n3. **보컬 멜로디**: `Eb4 → F4 → G4 → Bb4... → 꺾어서 G4↓ → Eb4`  \n   (꺾음 위치를 5~6박째에)\n4. **해금으로 같은 선율 2마디 뒤에 모방** (응수唱 구조)\n5. **피아노는 희소**: 1박에 1음, 루트와 5도만\n\n여기서 판소리 비율을 더 올리고 싶으면 장구 패턴 추가, 줄이고 싶으면 피아노 보이싱을 재즈적으로. **비율 조절이 \"느낌 살짝\"의 핵심**입니다.\n\n구체적인 장면이나 감정이 정해지면 더 집중해서 짜줄게요."}
{"prompt_id": "P5_pansori_ost", "category": "Cultural specificity", "condition": "C_skill_plus_nav", "model": "glm-5.1", "system_prompt_chars": 40212, "user_prompt": "판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?", "response": "좋은 출발점이에요! 몇 가지 핵심 파일을 참고해서 답변할게요.\n\n---\n\n## 시작하는 법 — 3단계\n\n### 1. 판소리의 핵심 요소 하나를 \"시그니처\"로 고르세요\n\n판소리 전부를 집어넣으면 사극 배경음악이 아니라 판소리 공연이 됩니다. **한두 가지 요소만** 차용하고 나머지는 서양식 오케스트레이션/프로덕션으로 가져가는 게 사극 OST다운 느낌입니다:\n\n| 판소리 요소 | 사극 OST에서의 활용 | 난이도 |\n|---|---|---|\n| **시김새** (관행적 꾸밈음) | 멜로디의 핵심 음에 시김새를 얹어 \"국악 창법\" 느낌 | ★★☆ |\n| **호흡법 / 비성** | 보컬이나 관악기에 비성(quasi-tag基层) 색깔 부여 | ★★★ |\n| **진양조 · 중모리 · 자진모리** 장단 | 리프나 반주 패턴의 리듬 기반으로 사용 | ★★☆ |\n| **대목 / 아니리** 교대 | 곡의 텍스처 대비(서사적 vs 서정적)에 활용 | ★☆☆ |\n| **펜타토닉 + 비극음계** (칠음 / 변환) | 화성 양식의 근간 | ★☆☆ |\n\n**추천**: 처음이라면 **시김새 + 진양조/중모리 장단** 조합이 가장 즉신효과가 좋습니다.\n\n---\n\n### 2. 장단을 먼저 잡고 시작하세요\n\n사극 OST에서 판소리 느낌이 가장 잘 사는 건 리듬입니다:\n\n- **진양조** (느린 6/4 또는 18박): 서정적·비극적 장면 (도입, 여운)\n- **중모리** (중간 12/8): 이야기가 흐르는 장면\n- **자진모리** (빠른 4/4): 격전, 절정\n\n**실습**: 진양조 느낌의 18박을 6/4 × 3마디 또는 3/4 × 6마디로 잡고, 강박 패턴을 `x . . x . . | x . . . . x | . . x . . .` 식으로 배치해 보세요. 이게 \"느린 사극 리듬\"의 뼈대입니다.\n\n---\n\n### 3. 음계와 화성 — \"선율은 동양, 화성은 서양\" 믹스\n\n한국 전통 음계는 크게:\n\n| 음계 | 음 | 느낌 |\n|---|---|---|\n| **경음계** (평조) | C D E♭ G A♭ | 담담, 궁중, 우아 |\n| **비음계** (계면조) | C D♭ F G A♭ | 애절, 비극, 원한 |\n\n**전략**: \n- **선율·리프**: 비음계 5음 (C D♭ F G A♭) 위주로 작성 → 사극 특유의 애절함\n- **화성**: 서양식 단화성(i iv V i 등) + 가끔 ♭II, ♭VII로 \"모달한 확장\" | Fm | Gm | A♭ | A♭ ||  ← 비음계로 선율 짜면 자연스럽게 섞입니다.\n\n---\n\n## 즉시 해볼 수 있는 워밍업\n\n1. **기조(根調) 선택**: A♭ 단조 (비음계 친화적, 성역도 좋음)\n2. **멜로디 스케치** (비음계 위에서):\n   ```\n   A♭4 → G4 → F4 — (시김새: F를 E로 밀었다가 F로) → D♭4 → C4\n   ```\n3. **장단 오스티나토** (중모리 12/8 느낌):\n   ```\n   타악: | x . . x . . | . x . . x . | (6+6 = 12)\n   베이스: | A♭  .  .  E♭  .  . | Fm  .  .  .  .  . |\n   ```\n\n이 세 가지만으로도 \"사극 + 판소리 냄새\"가 납니다. 여기에 현악 패드, 거문고/가야금 톤idx의 플러그인, 보컬 레이어를 올리면 됩니다.\n\n---\n\n## 다음 단계\n\n이 뼈대 위에서:\n- **형태** 잡기 (vs. 후렴 vs. 대목스러波动 구조)\n- **악기 편성** (거문고+해금+서양오케스트라 혼합)\n- **구체적 화성 진행** 설계\n\n어느 쪽을 더 파고 싶으신가요?"}
{"prompt_id": "P6_reharm", "category": "Reharmonization", "condition": "A_baseline", "model": "glm-5.1", "system_prompt_chars": 92, "user_prompt": "Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘.", "response": "Dm7 - G7 - Cmaj7 진행을 모던(네오소울, R&B, 모던재즈)하게 바꾸는 4가지 구체적 방법입니다.\n\n**1. 알터드 & 텐션 추가 (네오소울/모던재즈)**\n`Dm9 - G7#5b9 (또는 G7alt) - Cmaj9`\n> V도에 알터드 톤을 주어 긴장감을 만들고, I도에 9th를 넣어 몽환적으로 해결합니다.\n\n**2. 트라톤 서브스티튜션 (크로매틱 베이스 하행)**\n`Dm9 - Db7#11 - Cmaj9`\n> G7 대신 Db7을 사용해 베이스가 D - Db - C로 부드럽게 반음 하행하는 세련된 사운드를 냅니다.\n\n**3. 슬래시 코드 활용 (부드러운 보이스 리딩)**\n`Dm7 - Dm7/C - G7/B - Cmaj7`\n> 기존 진행에 베이스 반음 하행(C - B)을 끼워 넣어 정형적인 II-V-I의 틀을 흐릿하게 만듭니다.\n\n**4. 나란히 이동 (네오소울 특유의 몽환적 진행)**\n`Dm7 - Dbmaj7 - Cmaj7`\n> V도를 Dbmaj7로 치환하여 3개의 코드가 반음씩 나란히 내려가는 현대적인 느낌을 줍니다."}
{"prompt_id": "P6_reharm", "category": "Reharmonization", "condition": "B_skill_only", "model": "glm-5.1", "system_prompt_chars": 13617, "user_prompt": "Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘.", "response": "전형적인 ii–V–I를 모던하게 바꾸는 방법 몇 가지:\n\n---\n\n### 1. 트라이톤 대체 + 확장음\n**Dm9 → D♭7♭9 → Cmaj9**\n- 베이스가 반음 하강 (D → D♭ → C), 크로마틱 그래비티가 강함\n- D♭7은 G7과 tritone을 공유, 기능은 같은데 색이 완전히 달라짐\n\n### 2. 백도어 프로그레션\n**Dm7 → Fm7 → Cmaj7**\n- G7 대신 ♭VII7의 관계인 iv (Fm7)로 해결\n- 단조 느낌을 주면서도 부드러운 랜딩. 네오소울·R&B에서 흔함\n\n### 4. sus로 미해결감 주기\n**Dm7 → G7sus4 → Cmaj7**\n- 또는 **Dm9 → G13sus → Cmaj9**\n- 3rd를 4th로 미뤄서 긴장감을 연장. 해결이 더 넓게 퍼짐\n\n### 5. 크로마틱 평행하강\n**Dm7 → D♭maj7 → Cmaj7**\n- 모두 maj7 코드가 반음씩 내려감\n- 빌 에반스 이후 모던재즈·네오소울에서 자주 쓰이는 움직임\n\n---\n\n**추천**: 장르에 따라 — 네오소울이면 1번이나 5번, R&B면 2번, 좀 더 열린 느낌이면 3번. 어떤 분위기를 원하는지 알려주면 더 구체적으로 좁혀줄게."}
{"prompt_id": "P6_reharm", "category": "Reharmonization", "condition": "C_skill_plus_nav", "model": "glm-5.1", "system_prompt_chars": 40212, "user_prompt": "Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘.", "response": "ii–V–I(C)를 모던하게 바꾸는 몇 가지 방향:\n\n---\n\n### 1. 얼터드 도미넌트 — 재즈/네오소울 느낌\n```\nDm7 → G7alt → Cmaj7\n```\n- G7alt = G7(♭9♯9♭13) = G7(♭9♯5) — 아브라모비치 사운드\n- E♭멜로딕 마이너 5번째 모드 위 음들이 자연스럽게 걸림\n\n### 2. 트리톤 대체 — 스무스/시티팝\n```\nDm7 → D♭7 → Cmaj7\n```\n- G7 대신 ♭II7. 반음 하강 베이스 라인이 부드럽게 깔림\n- D♭7 = D♭(F A♭ C♭→B) — C와의 공통음 많아 자연스러운 해결\n\n### 3. 백도어 진행 — R&B/모던팝\n```\nFmaj7 → B♭7 → Cmaj7\n```\n- ii–V를 IV→♭VII로 교체. 서브도미넌트 색깔이 온화하면서도 surprise 있음\n\n### 4. 디미니시드 패스 + 색채음 — K-R&B/인디\n```\nDm7 → D♭dim7 → Cmaj9\n```\n- 크로매틱 패싱 코드로 긴장감. Cmaj9(9th=D)로 dm7의 루트가 해결되어 연결이 매끄러움\n\n### 5. 서브도미넌트 대체 + 얼터드 — 모던재즈/크로스오버\n```\nDm11 → G7(♯11♭13) → C6(9)\n```\n- ii도 V도 모두 익스텐션 열어서 \"색\"을 더함\n- C6(9)는 Cmaj7보다 모던한 느낌 (루트 위 6th 추가)\n\n---\n\n**추천:** 장르에 따라—\n- **네오소울/K-R&B** → 1번 또는 4번\n- **시티팝/스무스** → 2번\n- **모던팝/인디** → 3번\n- **재즈 클래식을 약간 비틀고 싶다** → 5번\n\n어떤 장르/분위기로 가고 싶으신가요? 더 구체적으로 맞출 수 있습니다."}
{"prompt_id": "P7_acoustic_ballad", "category": "Instrument idiom", "condition": "A_baseline", "model": "glm-5.1", "system_prompt_chars": 92, "user_prompt": "기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘.", "response": "기타 단독 연주에 어울리는 어쿠스틱 발라드 코드 진행 3가지입니다.\n\n**1. 대중적인 감성 발라드 (Key: C)**\n`F - G - Em - Am - Dm - G - C`\n* 특징: 한국 대중 발라드의 정석적인 진행입니다. 서정적인 멜로디가 가장 잘 돋보입니다.\n\n**2. 슬프고 애절한 발라드 (Key: Am)**\n`Am7 - Dm7 - Fmaj7 - E7`\n* 특징: 단조의 쓸쓸함에 E7 코드가 이국적이고 애절한 느낌을 더합니다. 느린 아르페지오 연주에 최적화되어 있습니다.\n\n**3. 따뜻하고 편안한 위로 발라드 (Key: G)**\n`G - D/F# - Em - C`\n* 특징: 베이스 음이 반음씩 하행하는 진행으로, 희망적이고 편안한 분위기를 냅니다.\n\n💡 **연주 팁:** 기타 배김(통울림)을 살리기 위해 **카포(Capo)를 3~5프렛**에 끼우면 어쿠스틱 발라드 특유의 맑고 따뜻한 사운드를 얻을 수 있습니다."}
{"prompt_id": "P7_acoustic_ballad", "category": "Instrument idiom", "condition": "B_skill_only", "model": "glm-5.1", "system_prompt_chars": 13617, "user_prompt": "기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘.", "response": "어쿠스틱 기타 발라드에 어울리는 코드 진행 4가지를 추천드립니다. 모두 기타 한 대로 편하게 잡을 수 있는 키와 보이싱입니다.\n\n---\n\n### 1. 서정 + 여운 — 카포 2프렛, G 기준\n\n```\n| G   | Em7 | Cadd9  | Dsus4 - D |\n| G   | Em7 | Am7    | Dsus4 - D |\n| C   | D   | Bm7    | Em7       |\n| C   | D   | G      | G         |\n```\n\n- **포인트**: `Cadd9`, `Dsus4→D` 서스펜드 해결이 잔잔한 여운을 만듭니다.\n- 베이스 라인이 `G→E→C→D`로 내려갔다 올라가는 자연스러운 움직임.\n\n---\n\n### 2. 살짝 어둡고 깊은 느낌 — 카포 4프렛, E 기준 (실제 G♯)\n\n```\n| E      | C#m7 | Amaj7  | B     |\n| E      | C#m7 | Am7    | B     |\n| A      | B     | G#m7   | C#m7 |\n| A      | Am    | E      | E     |\n```\n\n- **포인트**: 후반 `A→Am`에서 메이저→마이너 변화가 극적인 우울감을 줍니다.\n- 한국 발라드 후렴 전환부에서 자주 쓰는 장식.\n\n---\n\n### 3. 따뜻하고 안정적인 — 카포 없이, D 기준\n\n```\n| D     | Bm7  | G      | A      |\n| D     | Bm7  | Em7    | A7     |\n| G     | A    | F#m7   | Bm7    |\n| G     | Em7  | D      | D - Dsus4 - D |\n```\n\n- **포인트**: `A7`이 서브도미넌트 마이너트 역할로 살짝 어두운 색을 넣어줍니다.\n- 2프레이즈 후반 `G→Em7→D` 하강 종지.\n\n---\n\n### 4. 모던 K-pop 발라드 느낌 — 카포 1프렛, C 기준 (실제 D♭)\n\n```\n| C       | G/B    | Am7     | Fmaj7   |\n| C       | G/B    | F       | G       |\n| F       | Em7    | Dm7     | G7      |\n| F       | Fm     | C       | C       |\n```\n\n- **포인트**: `F→Fm→C` 피카디 종지. 가장 \"한국 발라드\"다운 마무리.\n- `G/B` (1전위) 로 베이스가 `C→B→A`로 반음씩 내려가는 라인이 매끄럽습니다.\n\n---\n\n### 기타 연주 팁\n\n- **아르페지오**: 엄지로 저음 3줄, 검지·중지·약지로 3·2·1현을 하나씩 뜯으면 발라드에 적당합니다.\n- **스트로크**: ↓ ↓↑ ↑↓↑ (한 박 반 단위) 패턴이 무난합니다.\n- **서스펜드**: `sus4→해결`을 타이밍마다 섞어주면 표정이 생깁니다.\n\n원하는 분위기(더 밝게, 더 어둡게, 더 연극적으로 등)를 말씀해주시면 그 방향으로 더 좁혀드리겠습니다."}
{"prompt_id": "P7_acoustic_ballad", "category": "Instrument idiom", "condition": "C_skill_plus_nav", "model": "glm-5.1", "system_prompt_chars": 40212, "user_prompt": "기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘.", "response": "# 어쿠스틱 발라드 코드 진행 추천\n\n기타에 맞는 키(열린 코드 활용)와 감성별로 추천합니다.\n\n---\n\n## 1. 기본 — 담담하고 따뜻한 발라드\n**키 G 메이저**\n\n``| G | Em | C | D |``\n\n- 가장 클래식한 서술형 진행. 초보자도 연주하기 쉽움.\n- **변주**: `C → D` 대신 `C → D/F♯` (베이스 반행 추가)\n\n---\n\n## 2. 쓸쓸한 발라드 — 파편화된 감정\n**키 E 마이너**\n\n``| Em | C | G | D |``\n\n``| Em | C | Am | Bm7 → B7 |``\n\n- `Bm7 → B7` 해결이 이끌음표 효과로 더 애절해짐.\n- 프레이즈 끝에 `B7` 대신 `B7sus4 → B7` 텐션/릴리즈 추가 가능.\n\n---\n\n## 3. 영원한 사랑 느낌 — I → V → vi → IV\n**키 D 메이저**\n\n``| D | A | Bm | G |``\n\n- 팝 발라드의 정석적 진행. 감동적이면서도 안정적.\n- **핸거운 느낌으로**: `G` 대신 `Gmaj7` 사용.\n\n---\n\n## 4. 캡포 활용 — 풍부한 오픈 코드 사운드\n**형태: C 메이저 폼, 캡포 2프렛 (키 D)**\n\n``| C | Am | F | G |``\n\n- 캡포 덕분에 오픈 코드 풍부한 울림 + 실제 피치는 D.\n- **다크하게**: `Am` → `A7` 또는 `F` → `Fm` (나폴리 느낌).\n\n---\n\n## 5. 감정 고조 — 브릿지용\n**키 G 메이저 → 단조 전환**\n\n``| Em | Cm | G | D/F♯ |``\n\n- `C → Cm` ( ♭III 도입)이 순간적으로 어두워지며 브릿지 느낌 부여.\n- 혹은 `| E minor | C | Am7 | D7 |` 로 키를 Em으로 일시 전조.\n\n---\n\n## 기타 치기 팁\n\n| 요소 | 팁 |\n|------|-----|\n| **아르페지오** | 픽잉 스타일: 베이스→3현→2현→1현 순으로 뜯기 |\n| **스트로크** | 느리게 `▼ ▼ ↑ ▼ ↑` 패턴, 강조는 2, 4박자 |\n| **베이스 런** | `G → D/F♯ → Em` 처럼 베이스 반행으로 흐름 부드럽게 |\n| **텐션 코드** | `G` 대신 `Gadd9`, `D` 대신 `Dsus4 → D` 해결 |\n| **쉬는 마디** | 코러스 끝에 한 마디 코드 없이 멜로디만—공간이 감동 만듦 |\n\n---\n\n원하는 분위기(슬픔, 위로, 희망 등)를 더 구체적으로 알려주면 그에 맞춰 진행을 세밀하게 조정해 드릴 수 있습니다."}
{"prompt_id": "P8_daw_eq", "category": "Boundary (negative)", "condition": "A_baseline", "model": "glm-5.1", "system_prompt_chars": 92, "user_prompt": "DAW에서 보컬 EQ 어떻게 걸어야 해?", "response": "보컬 EQ는 **\"제거(감쇠) → 추가(증가)\"** 순서로 걸며, 구체적인 주파수 대역별 설정법은 다음과 같습니다.\n\n**1. 저역 정리 (High-Pass Filter)**\n*   **80Hz~100Hz 이하를 완전히 깎습니다.** (12dB~24dB Slope 적용)\n*   목적: 마이크 룸블, 저역 잡음, 불필요한 탁한 저음 제거. 남성 보컬은 80Hz, 여성 보컬은 100Hz 정도를 기준으로 잡으세요.\n\n**2. 감쇠 EQ (문제 주파수 제거)**\n*   **200~500Hz (먹먹함/둔탁함):** 넓은 Q값으로 2~3dB 깎습니다. \n*   **300~400Hz (울림/Boxy):** 탁한 느낌이 들 때 좁은 Q값으로 2~3dB 깎습니다.\n*   **1~3kHz (쏘는 소리/Nasal):** 코막힌 소리나 귀를 찌르는 소리가 나면 좁은 Q값으로 2~3dB 깎습니다. (특히 2kHz 근처)\n\n**3. 증가 EQ (매력 더하기)**\n*   **100~200Hz (따뜻함/바디):** 보컬이 얇게 들리면 넓은 Q값이나 셸프(Shelf)로 1~2dB 올립니다.\n*   **3~5kHz (존재감/명료함):** 보컬이 악기 뒤로 묻힐 때 넓은 Q값으로 1~3dB 올립니다.\n*   **10kHz 이상 (공기감/Air):** 하이 셸프로 1~2dB 올려 숨소리와 화사함을 더합니다. (단, 'ㅅ', 'ㅊ' 치찰음이 심해지면 De-Esser 처리 필수)\n\n**💡 핵심 팁**\n*   **솔로 금지:** EQ를 돌릴 때는 보컬 솔로가 아닌 **전체 믹스(악기와 함께)** 상태에서 돌리세요.\n*   **폭이 좁게 깎고, 넓게 올리기:** 깎을 때는 Q값을 좁게(뾰족하게) 하고, 올릴 때는 Q값을 넓게(완만하게) 설정하세요."}
{"prompt_id": "P8_daw_eq", "category": "Boundary (negative)", "condition": "B_skill_only", "model": "glm-5.1", "system_prompt_chars": 13617, "user_prompt": "DAW에서 보컬 EQ 어떻게 걸어야 해?", "response": "이 질문은 이 스킬의 범위 밖입니다. 보컬 EQ는 **오디오 엔지니어링/믹싱** 영역이며, 이 스킬은 작곡·화성·선율·편곡 등을 다루고 DAW 조작이나 믹싱/마스터링/사운드디자인은 다루지 않습니다.\n\nDAW에서 보컬 EQ를 잡는 건 믹싱 스킬이나 보컬 프로덕션 전문 자료를 참고하시는 게 정확합니다.\n\n---\n\n**다만, 작곡 단계에서 EQ와 관련해 고려할 점** 하나만 말씀드리면:\n\n- 보컬 레인지와 악기 배치(orchestration)를 설계할 때, 보컬이 어느 주파수 대역을 차지할지 **미리 생각해두면** 믹싱 단계에서 EQ로 뚫어내야 하는 부담이 줄어듭니다.  \n  예: 피아노와 보컬이 경합한다면, 피아ните의 보이싱에서 1–3kHz 영역을 비워주거나 옥타브를 분산시키는 것이 작곡적 해결책입니다.\n\n믹싱/eq 관련 실질적 조언이 필요하시면 믹싱 전용 가이드나 유튜브 튜토리얼(Squeezebox, Mix With The Masters 등)을 추천합니다."}
{"prompt_id": "P8_daw_eq", "category": "Boundary (negative)", "condition": "C_skill_plus_nav", "model": "glm-5.1", "system_prompt_chars": 40212, "user_prompt": "DAW에서 보컬 EQ 어떻게 걸어야 해?", "response": "이 질문은 오디오 엔지니어링/믹싱 영역이라서, 이 스킬(작곡·이론)의 범위 밖입니다.\n\n간단히 방향만 말씀드리면:\n\n- **저역 정리**: 80–100 Hz 이하 롤오프 (근접효과 제거)\n- **먹먹함 제거**: 200–300 Hz 영역 감쇠\n- **존재감**: 2–4 kHz 부스트 (또는 1kHz 감쇠 후 3kHz 부스트)\n- **공기감**: 8–12 kHz 셸프 부스트 (여유 있을 때)\n\n하지만 **정확한 EQ는 목소리 텍스처, 마이크, 룸, 장르에 따라 완전히 달라집니다.** 전문 믹싱 리소스나 DAW 이퀄라이저 스킬을 참고하는 것이 좋습니다.\n\n제가 도울 수 있는 건:\n- 보컬 라인의 작곡적 결정 (레지스터 선택, 하모니 배치, 멜로디 윤곽)\n- 편곡에서 주파수 충돌 방지 (보컬과 다른 악기의 레지스터 분리)\n- 장르별 보컬 톤의 작곡적 컨벤션\n\n믹싱/이퀄라이징은 다른 전문 리소스를 찾아보세요!"}

# v1.0 Benchmark Results — 20260427T212522Z

**Model:** `glm-5.1` via OpenCode Go (`https://opencode.ai/zen/go/v1`)
**Date (UTC):** 20260427T212522Z

**Conditions:**
- **A** (baseline): generic assistant prompt only
- **B** (SKILL.md only): generic + `SKILL.md` prepended
- **C** (SKILL.md + nav): generic + `SKILL.md` + `references/00-navigation.md` prepended

**Prompts:** 8

---

## P1_chorus_diagnosis — Vague diagnosis

**Prompt:** 내 곡 후렴이 안 꽂혀. 어떻게 고치지?

### A_baseline

후렴이 귀에 안 꽂히는 원인과 구체적인 해결책 5가지입니다.

1. **멜로디 단순화 및 반복**: 멜로디가 너무 산만하면 남지 않습니다. 기억하기 쉬운 짧은 멜로디(모티프) 하나를 만들어 후렴 안에서 2~3번 반복하십시오.
2. **벌스와의 대비**: 벌스와 음역대, 리듬, 에너지가 비슷하면 후렴이 튀지 않습니다. 후렴에서 음역대를 확 올리거나, 반대로 벌스보다 리듬을 여유롭게(긴 음표로) 만들어 차이를 주십시오.
3. **훅(Hook) 가사 배치**: 한 번 들으면 따라 부를 수 있는 짧고 강렬한 한 줄(훅)을 후렴의 첫머리나 마지막에 배치하고, 운율(라임)을 맞추십시오.
4. **악기 리프 추가**: 보컬 멜로디만으로 부족하다면, 후렴 시작과 함께 귀에 맴도는 짧은 악기 리프(신스, 기타 등)를 멜로디와 겹쳐 연주하십시오.
5. **프리 코러스(빌드업) 강화**: 후렴이 갑자기 튀어나오면 어색합니다. 후렴 직전 구간에서 드럼 킥을 끊거나, 코러스 화음을 넣어 '이제 폭발한다'는 기대감을 먼저 조성하십시오.

### B_skill_only

후렴이 "안 꽂힌다"는 건 여러 원인이 있을 수 있어요. 진단 먼저, then 처방.

---

## 빠른 진단 체크리스트

| 요소 | 후렴이 인상적이려면… | 의심 포인트 |
|------|---------------------|-------------|
| **멜로디 높이** | 후렴 최고음 > 절 최고음 | 절이랑 비슷한 음역? |
| **리듬 훅** | 반복·지속되는 리듬 모티프 | 가사 얹는 리듬이 매번 달라? |
| **화성 변화** | 절과 다른 코드 진행 또는 조바꿈 | 절이랑 같은 I-V-vi-IV? |
| **밀도** | 더 많은 악기·보이스 레이어 | 후렴인데 편곡이 절이랑 같아? |
| **대비** | 절→후렴 순간 에너지 점프 | 도입부→후렴 전환이 서서히? |
| **가사 훅** | 짧고 강한 구절, 반복 | 한 줄이 길거나 산문적? |

---

## 가장 흔한 3가지 원인과 구체적 처방

### 1. 멜로디가 절이랑 너무 비슷하다
**해결:** 후렴 첫 음을 절 최고음보다 3도~5도 위에서 시작.

```
절 끝: E4 → 후렴 첫음: A4 이상
예)절: C4-D4-E4-G4  →  후렴: G4-A4-C5-D5
```

### 2. 화성이 안 바뀐다
**해결:** 후렴 진행을 절과 다르게. 가장 즉각적인 옵션들:

| 옵션 | 코드 (C 메이저 기준) | 효과 |
|------|---------------------|------|
| **소디미니시시** | C → A♭ → B♭ → C | 극적, 영웅적 |
| **iv 도입** | Fm → C → G → C | 그리웋하면서 단단 |
| **V/V → V** | D → G → Am → Fm | 텐션→해결, 빠른 훅 |
| **나폴리 6도** | C → D♭maj7 → G → C | 어둡고 파워풀 |

### 3. 리듬 훅이 없다
**해결:** 가사 음절에 반복되는 리듬 패턴 부여.

```
"다가 와" ×2  (점-점-_____)  ← 4음절이면 점-점-박-박 패턴을 반복
"말해 줘"      (점-박)
→ "다가와-다가와-말해줘" 리듬: x-x- x-x- x-x
```

하이라이트 가사에 **같은 리듬 3번 반복** → 이것만으로도 훅이 됩니다.

---

## 다음 질문에 답해주면 더 정확히 도와줄게요

1. **장르** (K-pop 발라드? 힙합? 인디? 락?)
2. **지금 후렴 코드 진행** (알려주면 바로 대안 제시)
3. **어떤 안 꽂힘?** (멜로디가 안 외워짐 / 에너지가 안 올라감 / 가사가 안 남음)

이 세 가지만 알려줘도 훨씬 구체적으로 고쳐줄 수 있어요.

### C_skill_plus_nav

후렴이 안 꽂히는 원인은 보통 몇 가지 패턴 중 하나입니다. 진단해 보고 고쳐 보세요:

---

## 1. 절→후렴 리프트가 부족하다

후렴은 절보다 **에너지가 올라야** 꽂힙니다. 체크:

- **음역**: 후렴 최고음이 절 최고음보다 최소 단3도 이상 높은가?
- **밀도**: 악기 레이어가 후렴에서 두꺼워지는가?
- **리듬**: 보컬 프레이징이 더 빨라지거나 싱코페이션이 많아지는가?

**빠른 수정**: 후렴 첫 마디 멜로디를 3~5음 위로 올려 보세요.

---

## 2. 훅 멜로디가 약하다

기억에 남는 후렴은 **반복 가능한 짧은 모티프** + **인상적인 간진행**이 있습니다.

**해결책**:
- 후렴 멜로디를 1~2마디 짧은 셈(모티프)으로 줄이고, 그것을 반복·변형으로 전개
- 도약(5도 이상)을 한 곳 넣어 보세요. 예: `도→레→파↑↗︎(5도 상행)→미`
- 리듬 변형: 절이 8분음 wired이면 후렴은 점음표나 부점 리듬으로

---

## 3. 화성이 너무 예측 가능하다

절-후렴 화성이 같으면 후렴이 "터지지" 않습니다.

**해결책**:
- 후렴 진입을 **IV** 또는 **♭VI**로 시작 (예: C메이저 곡이면 F메이저 또는 A♭메이저)
- 절이 `I → vi → IV → V`라면 후렴은 `IV → I → V → vi`로 순서 교환
- 후렴 첫 코드에 **sus4 해결** 또는 **add9** 추가로 색감 차이

---

## 4. 가사·음절 밀도 문제

절보다 후렴이 음절이 적고 음이 길어야 힘이 실립니다.

**해결책**: 후렴 첫 라인을 3~5개 음절로 압축하고 긴 음가에 담기.

---

## 장르가 중요합니다

K-pop, 발라드, 힙합, 락… 장르마다 "꽂히는 후렴"의 공식이 다릅니다. **어떤 장르인지 알려주시면** 더 구체적으로 조언할 수 있습니다.

---

지금 곡의 키, 현재 코드 진행, 후렴 멜로디 움직임을 알려주시면 바로 수정안을 제시할게요.

---

## P2_kpop_dark_bridge — Genre + emotion

**Prompt:** K-pop 발라드의 브릿지를 좀 더 어둡게 만들고 싶어.

### A_baseline

K-pop 발라드 브릿지를 어둡게 만들기 위한 구체적이고 실전적인 방법:

1. **화성 (코드):** 밝은 코드를 피하고 **디미니시드 코드**나 **단2도 불협화음**을 사용해 불안정함을 조성하세요. 베이스라인을 **반음 하행**시키면 긴장감과 우울함이 극대화됩니다.
2. **멜로디:** 고음역 클라이맥스 대신 **중저음역**으로 끌어내리세요. 안정적인 음정 대신 **증4도(트라이톤)**나 단2도 음정 도약으로 서스펜스(매달림) 느낌을 주세요.
3. **편곡:** 악기를 전부 터트리기 전에 오히려 악기를 걷어내고 **적막(무음, 공백)**을 사용하세요. 잔향이 긴 어두운 신디사이저, 저음 첼로, 무거운 서브베이스로 공간을 채우세요.
4. **보컬:** 깔끔한 가창 대신 **보컬 프라이(목소리 떨림)**, 거친 숨소리, 속삭임을 활용하세요. 브릿지 끝부분에서 옥타브 점프와 함께 거친 파열음으로 절규하듯 전환하세요.
5. **가사:** 단순한 슬픔을 넘어 '광기, 상실, 시선 고정, 통제 불능' 등의 단어로 서사를 어둡게 비틀어 보세요.

### B_skill_only

K-pop 발라드 브릿지를 어둡게 만드는 구체적인 방법 몇 가지:

---

## 1. 평행단조 차용 (Modal Mixture)

가장 즉각적이고 장르에 잘 맞는 방법.

**C 메이저 기준:**
- `F → Fm → C` (♭VI → iv → I) — Fm이 순간 그림자를 드리움
- `Am → A♭m → G7` — A♭m은 ♭vi, 극적인 하강
- `Dm7 → D♭maj7 → Cm7 → Fm7` — ♭VII → ♭VI 진행

## 2. 단조의 이차도수 (Secondary Dominants to Minor)

목표음을 단조로 해결.

- `E7 → Am` (V/vi → vi) — 이미 어두운 Am으로
- `B7 → Em` (V/iii → iii) — 더 깊은 그림자
- `F7 → B♭m` (♭VII 처리 가능) — 예측 밖의 색채

## 3. 색채중간 (Chromatic Mediant)

K-pop 발라드 브릿지 끝에서 흔히 쓰는 극적 전조 느낌.

- `Cmaj7 → A♭maj7` — I → ♭VI, 3도 반음 관계
- `Cmaj7 → Emaj7` — I → III, 올림 중간
- `Fm7 → Bmaj7` — ♭IV → III⁺ (비트로스 진행)

## 4. 종결구 변경

3단음 아래 종결 (Plagal ♭VI cadence):
- `A♭maj7 → C/G → Fm → C`
- 마지막 V 대신 IVm으로 "미결" 느낌 유지

---

**실제 예시 진행 (C 메이저 곡의 브릿지):**

```
| Fmaj7  | Fm7    | Dm7   | D♭maj7 |
| Am7    | A♭m7   | G7sus4| G7♭9   |
```

- F→Fm: modal mixture
- Dm→D♭: ♭VII 하강
- Am→A♭m: ♭vi 하강
- G7♭9: 증극 긴장, 코러스 해결 준비

어느 정도 "어둡게"인지(섬뜩함 vs 쓸쓸함)에 따라 골라 쓰면 됨.

### C_skill_plus_nav

# K-pop 발라드 브릿지: 더 어둡게 만드는 방법

## 화성적 접근

**1. 모달 믹스처 (Modal Mixture)**
메이저 키에서 마이너의 플랫 디그리를 차용:
- `♭VI` → `♭VII` → `i` (C 메이저: A♭maj7 → B♭ → Cm)
- `iv` 대신 `IV` 자리에 마이너 서브도미넌트 사용: `Fm6` 대신 `F`
- `♭III` → `♭VI` 시퀀스: C 메이저에서 E♭maj7 → A♭maj7

**2. 크로매틱 메디안트**
3도 관계의 색채 변화:
- `I` → `♭III` → `♭V` (C → E♭ → G♭) — 극적인 암흑감
- `I` → `♯iv°` → `V` (C → F♯° → G)

**3. 세컨더리 도미넌트 → 마이너**
- `V7/vi` → `vi` (E7 → Am in C) — 라딕스 느낌의 어두운 해결
- `V7/ii` → `ii` (A7 → Dm) — 서브도미넌트 방향의 긴장

## 선율적 접근

- 브릿지 시작을 낮은 레지스터에서 (C3~E4 남성, A3~C5 여성)
- 트라이톤 증4/감5 도약 활용
- 하행 반음 진행: `^5 → ♭5 → 4` 또는 `^1 → ♭7 → ♭6`
- 레가토 프레이징, 적은 음, 더 긴 지속음

## 텍스처 & 편곡

| 요소 | 변경 |
|------|------|
| 베이스 | 더 낮은 레지스터, 펜타토닉 하행 |
| 피아노 | 저음 옥타브 트리아드, 페달 포인트 |
| 스트링 | 첼로 디비지, 콘 소르디노 |
| 리듬 | 하프타임 또는 리타르단도, 공간 많이 |
| 보컬 | 숨소리 강조, 폴셋 → 풀보이스 전환 |

## 실전 예시 (C메이저 기준)

**기본 브릿지:**
`Am → F → C → G`

**어둡게 변주:**
```
| A♭maj7 | B♭7sus4  | Cm9      | G♭maj7  |
| Fm7     | D♭/F     | E♭m6     | G7sus4  |
```

핵심: 메이저 키의 밝음을 `♭VI`, `♭VII`, 마이너 `iv`로 서서히 대체하면서, 불안정한 서스펜디드 코드로 긴장을 유지하고 해결을 지연시키세요.

---

어떤 키에서 작업 중인지, 현재 브릿지 코드 진행을 공유해주시면 구체적으로 변주해 드릴 수 있습니다.

---

## P3_scale_over_cm7 — Technical theory

**Prompt:** Cm7 위에 어떤 스케일이 자연스러워?

### A_baseline

Cm7 위에서 가장 자연스러운 스케일은 다음 3가지입니다. 

1. **C 자연단조 (C Natural Minor / Aeolian)**
   * **음계**: C - D - Eb - F - G - Ab - Bb
   * **특징**: 가장 기본적이고 안정적인 마이너 사운드입니다.

2. **C 도리안 (C Dorian)**
   * **음계**: C - D - Eb - F - G - A - Bb
   * **특징**: 6음이 Ab에서 A로 바뀝니다. 재즈나 펑크에서 자주 쓰이며, 자연단조보다 밝고 그루브한 느낌을 줍니다.

3. **C 마이너 펜타토닉 / 블루스 스케일**
   * **음계**: C - Eb - F - G - Bb (+ 블루스 노트: F#)
   * **특징**: 가장 치고 빠지기 쉽고 무난하게 어울립니다.

**추천 접근**: 기본적으로 **C 마이너 펜타토닉**을 치고, 곡의 분위기(키)에 따라 6음을 **Ab(자연단조)**로 쓸지 **A(도리안)**로 쓸지 결정하여 섞어 쓰는 것이 가장 자연스럽습니다.

### B_skill_only

# Cm7 위의 스케일

**Cm7** = C - E♭ - G - B♭

가장 자연스러운 선택들:

| 스케일 | 음계 | 느낌 / 용도 |
|--------|------|-------------|
| **C 자연단조** (Aeolian) | C D E♭ F G A♭ B♭ | 가장 기본적. 클래식, 팝 발라드 |
| **C 도리안** | C D E♭ F G **A** B♭ | 재즈 표준. 6번음이 장조(A)라 더 밝고 스무스 |
| **C 마이너 펜타토닉** | C E♭ F G B♭ | 록·팝·블루스. 간결하고 안전 |
| **C 블루스 스케일** | C E♭ F **G♭** G B♭ | 블루스·록. 감정적, 관절적인 느낌 |

## 선택 기준

- **장르가 재즈 / 퓨전** → **도리안**부터. `A♭` 대신 `A♮`이 4음 위에서 자연스럽게 해결되며, ii-V에서 모드 유지가 쉽습니다.
- **장르가 팝·발라드·클래식** → **자연단조**. 6번음 `A♭`이 이미 조표에 있고, 안정적입니다.
- **즉흥 연주 초보** → **마이너 펜타토닉**으로 시작. 치명적인 빗소리가 없습니다.
- **더 딥한 블루스 느낌** → 펜타토닉에 `G♭`(블루 노트) 추가.

**장르를 알려주시면 더 구체적으로 추천해 드릴 수 있습니다.**

### C_skill_plus_nav

# Cm7 위의 스케일

**Cm7 = C–E♭–G–B♭**

가장 자연스러운 선택은 스타일에 따라 다릅니다:

| 스케일 | 음 | 느낌 | 주로 쓰이는 장르 |
|--------|-----|------|-----------------|
| **C 자연단음** (C Aeolian) | C–D–E♭–F–G–A♭–B♭ | 어둡고 차분함 | 팝, 록, 클래식 |
| **C 도리안** (C Dorian) | C–D–E♭–F–G–**A♮**–B♭ | 밝고 펑키햄 | 재즈, 펑크, 소울 |
| **C 프리지안** (C Phrygian) | C–**D♭**–E♭–F–G–A♭–B♭ | 스페인적, 긴장감 | 플라멩코, 라틴, 현대재즈 |

---

**가장 기본답:** **C 자연단음** 또는 **C 도리안**.

- 화음음(C, E♭, G, B♭)이 모두 스케일 안에 들어있으므로 두 스케일 모두 "자연스럽습니다."
- 차이는 6번째 음뿐: **A♭**(자연단음) vs **A♮**(도리안).
- A naturel이 더 밝고 현대적인 사운드, A♭이 더 전통적 마이너 사운드.

---

**추가 옵션:**
- **C 마이너 펜타토닉** (C–E♭–F–G–B♭) – 즉흥 연주의 출발점
- **C 블루스 스케일** (위에 G♭ 추가) – 블루스/록
- **C 하모닉 마이너** – 화음 진행에서 V7이 등장할 때

어떤 스타일로 작곡/연주하시나요? 그러면 더 세부적으로 추천할 수 있습니다.

---

## P4_jazz_kpop_hybrid — Hybrid request

**Prompt:** 재즈 화성으로 K-pop 후렴을 만들어보고 싶어.

### A_baseline

K-pop의 대중적인 구조에 재즈의 화성(7th, 9th, altered, Tritone substitution)을 입힌 8마디 후렴 코드 진행과 작곡 팁입니다.

**키(Key): C 메이저**

### 1. 코드 진행 (8마디)
전형적인 K-pop 후렴의 흐름(I - V - vi - IV 등)에 재즈적 색채를 더했습니다.

*   1마디: **Cmaj9** (I) - 밝고 풍성하게 시작
*   2마디: **Am11** (vi) - K-pop 특유의 애절함, 11th음으로 재즈 터치
*   3마디: **Dm9** → **G7#5#9** (ii - V) - **가장 재즈스러운 부분.** Altered dominance로 극적인 긴장감
*   4마디: **Cmaj7** → **C7#9** (I → I7) - 블루스 느낌의 크로매틱 전환
*   5마디: **Fmaj9** (IV) - 후렴의 감정적 절정(환상적이고 몽환적인 느낌)
*   6마디: **Fm6** (iv) - **K-pop의 잔잔함 + 재즈의 크로매틱 하강**
*   7마디: **Em7** → **Eb7#11** (iii → SubV) - **Tritone 대체.** Eb7은 원래 A7(V/ii) 자리에 들어간 대체 화음으로 매우 세련된 반전
*   8마디: **Dm9** → **G7b9b13** (ii - V) - 다음 구간을 위한 재즈식 해결

### 2. 멜로디 작법 (핵심)
*   **어프로치 노트(Approach Note):** G7#5#9 코드(3마디) 위에서는 멜로디가 `Bb`(b5), `Eb`(b9) 혹은 `F#`(#5) 음을 강박(Strong beat)에서 타격하게 작곡하세요. 일반적인 다이아토닉 멜로디와 확연히 다른 '재즈 냄새'가 납니다.
*   **Fm6 활용:** 6마디 Fm6 코드 위에서는 멜로디가 `A` 자연음(C 메이저 스케일의 6도)을 타격하면, 반주의 `Ab`(F 단조의 3도)와 마찰을 일으키며 K-pop 특유의 '가슴 시린' 느낌과 재즈의 어두운 느낌을 동시에 냅니다.

### 3. 리듬 & 보이싱 팁
*   **2-5-1(Bossa Nova Feel):** 3마디와 7~8마디의 ii-V 진행에서는 드럼/베이스를 보사노바나 삼바 리듬으로 반 박자 쉬어(K-pop의 특징)와 섞어보세요.
*   **Shell Voicing:** 피아노나 기타는 근음(Root), 3도, 7도만 남기고 나머지 음은 멜로디에 맡기면(C7 = C-E-Bb), 정통 재즈 앙상블처럼 공간이 넓고 힙한 사운드가 납니다.

### B_skill_only

K-pop 후렴에 재즈 화성을 적용하는 건 좋은 접근이에요. K-pop은 이미 차용화음·부차디미넌트를 자주 쓰기 때문에, 재즈 화성을 덧씌우는 게 자연스럽게 들릴 수 있어요. 핵심은 **"재즈의 색채 + 팝의 기억성"** 유지예요.

---

## 옵션 1: 확장화음으로 팝 진행 업그레이드

가장 온건한 방법. 기본 팝 진행에 7th·9th를 더해 톤을 풍부하게.

**C키:**
```
Cmaj9 → Am9 → Dm11 → G13
```

원래 C – Am – Dm – G에서 확장음 추가. 멜로디는 팝처럼 직선적으로 유지하면 화성만 정교해져요.

---

## 옵션 2: ii–V–I 기반 후렴

재즈의 핵심 진행을 후렴 뼈대로. K-pop 특유의 "상승감"을 재즈적 긴장–해소로 만들어요.

**C키 (빠른 템포, 2마디 단위):**
```
| Dm7     G7♭9  | Cmaj7   Cmaj7 |
| Fmaj7   Bm7♭5 | E7♭9    Am7  |
```

전반부: ii–V–I. 후반부: iiø–V7–i (병행단조 편향). 이조(V7/vi)가 들어가 K-pop 특유의 감성 전환을 만들어요.

---

## 옵션 3: 트라이톤 치환 + V7alt로 드라마

후렴 하이라이트에 재즈의 가장 강렬한 도구를 배치.

**C키:**
```
Cmaj7 → D♭7(♯11) → Cmaj7/G → G7alt
Fmaj7 → E♭maj7   → Am7    → D7♭9 → G7alt
```

- **D♭7** = G7의 트라이톤 치환(♭II7). 베이스 반음 하강으로 긴장 ↑
- **E♭maj7** = ♭VI 차용화음. 몽환적 색채.
- **G7alt, D7♭9** = V7에서 최대 크로마티시즘

---

## 옵션 4: 나단조 ii–V–i로 애절 발라드 후렴

슬픈/애절한 K-pop 발라드에 최적.

**Am 기준:**
```
Am9  → Bm7♭5 → E7♭9 → Am9
Fmaj7 → Dm7  → E7alt → Am9
```

iiø→V7♭9→i는 재즈에서도 "애절함"의 정석이고, K-pop 발라드 감성과 아주 잘 맞아요.

---

## 멜로디 주의사항

화성이 복잡할수록 **멜로디는 심플**하게. K-pop 후렴의 힘은 귀에 쏙 들어오는 훅이에요.

| 화성 위 멜로디 | 효과 |
|---|---|
| maj9, ♯11 위 | floating·몽환 느낌 |
| V7alt 위에서 ♭9, ♯9 | 긴장 일관성 |
| 3rd·7th 피하기 | 보이싱과 충돌 ↓, 투명도 ↑ |

---

원하는 분위기(애절·몽환·고조 등)나 레퍼런스 곡이 있으면 더 구체적으로 조정해드릴게요.

### C_skill_plus_nav

# 재즈 화성 × K-pop 후렴

K-pop 후렴의 구조적 힘(강렬한 후크, 고음역 멜로디, 에너지 상승)은 그대로 두고, **화성 레이어만 재즈로 교체**하는 게 핵심입니다. 멜로디가 너무 복잡해지면 K-pop 특유의 귀접힘(hook)이 사라집니다.

---

## 3가지 실전 옵션

### ① ii-V-I 변형 — 가장 접근적인 재즈 느낌

**Key: C major, 8마디 후렴**

```
| Dm7    G7♭9  | Cmaj7    —    |
| Fmaj7   —    | Em7   A7      |
| Dm7    G7♭9  | Cmaj7    —    |
| Fmaj7   Fm7  | Dm7♭5  G7alt |
```

- ♭9, alt 텐션으로 색채 추가하지만 근음 흐름은 여전히 친숙
- Fm7(♭VI) = 모달 인터체인지, 클래식 K-pop 감성의 "서늘한 전조" 느낌
- 멜로디는 5도~옥타브 사이 굵직한 선율

### ② 트라이톤 서브 — 더 세련된 도시 느낌

```
| Dm7      —    | D♭7      —    |
| Cmaj7    —    | E7♯9     —    |
| Fmaj7    —    | B∅7  E7      |
| Am9      —    | D♭7      —    |
```

- D♭7 = G7의 트라이톤 서브 → 예상치 못한 반음 하강
- E7♯9 (Hendrix chord) = vi의 세컨더리 도미넌트, 블루스-재즈 색채
- 멜로디 긴 음표 위에 내성 보이스만 움직이게 하면 K-pop 앙상블 사운드 유지

### ③ 모달 인터체인지 + 상행 II↎V — 가장 드라마틱

```
| Dm9    G7♭13 | Cmaj9     —    |
| Fm7     B♭7  | Em7      A7    |
| Dm9    G7alt | Cmaj9     —    |
| Am7     D7   | D♭maj7    —    |
```

- Fm7→B♭7 = ♭VI → V/IV, K-pop에서 자주 쓰는 서늘한 전개
- 마지막 D♭maj7 = ♭II (네오리만 색채, 곡의 끝을 열어둠)
- 후렴 반복 시 마지막 2마디를 ||: G7alt → Cmaj7 :|| 로 바꿔 종결감

---

## 작업 팁

| 요소 | K-pop 관점 | 재즈 적용 |
|------|-----------|----------|
| **멜로디** | 5도 이상 폭, 히트 음은 강박 | 화음텐션음(9, ♭9, ♯11)을 경과음처럼 스쳐 지나가게 |
| **보컬 하모니** | 3도 병행, 옥타브 유니슨 | 7th 위에 3-part close harmony로 재배치 |
| **리듬** | 4/4 포크, 킥 온 1&3 | 베이스를 walking 느낌으로, 드럼은 K-pop 템션 유지 |
| **전조** | 후렴→후렴구 사이 반음/온음 상행 | ii↎V를 타고 전조 (예: C → E♭ = Dm7→G7→E♭maj7) |
| **아웃트로** | 반복 후 페이드 또는 반음 하행 종지 | 재즈 텐션 코드로 끝 (예: Cmaj7♯11) |

---

## K-pop × 재즈 레퍼런스 듣기

- **레드벨벳 "Psycho"** — maj7 텐션 감성 팝 후렴
- **스트레이 키즈 "Levanter"** — 모달 인터체인지 서늘한 화성
- **태연 "INVU"** — ii↎V 느낌의 하행 진행
- **어쿠스틱 콜라보/스탠다드 재즈 커버 그룹들** — K-pop 멜로디 위 재즈 리하모니제이션 직접 체감

---

원하시면 **특정 키, BPM, 멤버 수, 보컬 파트**를 알려주면 그 조건에 맞춰 멜로디 + 보컬 편성까지 구체화해 드리겠습니다.

---

## P5_pansori_ost — Cultural specificity

**Prompt:** 판소리 느낌이 살짝 들어간 사극 OST 한 곡을 짜고 싶은데, 어떻게 시작해?

### A_baseline

판소리의 '창'과 사극 OST의 '스케일'을 결합하려면, 전통 음계와 창법을 현대 편곁으로 쳐내는 작업이 필요합니다. 다음 4단계로 구체적으로 시작해 보세요.

**1. 음계와 모티프 정하기 (뼈대)**
*   **음계:** 전통 5음계인 **계면조**(솔-라-도-레-미, 슬프고 처절한 느낌) 또는 **평조**(도-레-미-솔-라, 웅장하고 고풍스러운 느낌) 중 곡의 분위기를 선택해 멜로디를 짭니다.
*   **오프닝:** 곡의 시작은 악기 없이 보컬만으로 판소리 창법(시김새, 즉 꺾고 떠는 발성)으로 1~2마디 불러 theatrical하게 입장하는 방식이 가장 효과적입니다.

**2. 보컬 디렉팅 (판소리 느낌 살리기)**
*   처음부터 끝까지 판소리 창법을 쓰면 이질감이 생깁니다. **후렴구나 도입부**에만 판소리 특유의 성색(거친 목소리)과 시김새를 넣고, 나머지는 정통 보이스로 부르게 하세요.
*   곡 중간에 '이이이익', '얼씨고' 같은 추임새나 숨소리(덤덤)를 아주 짧게 샘플링/녹음해서 리듬 악기처럼 배치해 보세요.

**3. 악기 편곡 (사극 스케일 입히기)**
*   **전통 악기:** 판소리의 북 대신 **꽹과리나 징**의 강렬한 한 박을 기본 리듬에 얹고, 선율은 **해금**(애절함)이나 **가야금**(서정성)이 담당하게 하세요.
*   **현대 악기:** 북의 리듬 위에 웅장한 **오케스트라 스트링**(앙상블)과 **심벌즈**를 깔아주면 판소리의 촌스러움은 지우고 사극의 웅장함만 남습니다.

**4. 곡 구조 짜기 (추천 형식)**
*   **도입부:** 해금 독주 또는 무반주 보컬(판소리 창법)
*   **A파트:** 일반적인 발성의 서사적 멜로디 (스트링+가야금)
*   **B파트 (프리코러스):** 긴장감 상승, 꽹과리 등 타악기 등장
*   **후렴:** 폭발적인 오케스트라 + 보컬 시김새 극대화
*   **간주 (C파트):** 판소리의 '아니리'(말하는 톤)처럼 읊조리듯 노래하고 다시 후렴으로 폭발

**요약하자면:** 계면조/평조 5음계로 멜로디를 짠 뒤 ->도입부를 판소리 창법 무반주 솔로로 시작하고 -> 꽹과리/북의 리듬 위에 오케스트라를 얹어 후렴을 터뜨리는 방식으로 데모를 먼저 만들어 보세요.

### B_skill_only

# 사극 OST + 판소리 융합: 시작 가이드

## 1. 먼저 결정할 것: 극적 컨텍스트

같은 사극이라도 장면에 따라 완전히 다릅니다:

| 상황 | 추천 조(調) | 추천 장단 | 느낌 |
|------|-----------|----------|------|
| 비극적 죽음/이별 | 계면조 | 진양(느림) | 애절, 체념 |
| 전투/결의 | 우조 | 중모리~자진모리 | 웅장, 결연 |
| 회상/그리움 | 평조 | 중모리 | 서정, 담백 |
| 원혼/저주 | 남조계면 | 진양 | 음산, 불안 |

**예시**: "주인공이 모든 걸 잃고 홀로 서는 장면" → 계면조 + 진양장단

---

## 2. 조(調) — 선법 선택

가장 많이 쓰는 세 가지:

**계면조** (가장 판소리적, 애절함)
```
시 - 라 - 도 - 레 - 미
(라 중심, 4도 위 미와 반음 관계 없음 — 펜타토닉이지만 
시→라로 내려가는 꺾음이 핵심)
```

**우조** (웅장, 광활함)
```
도 - 레 - 미 - 솔 - 라
(도 중심, 남도계면의 시·미 대신 서구적 메이저 느낌 + 
미↔솔 사이 1.5음 = 판소리 특유의 '늘어짐')
```

**평조** (서정, 부드러움)
```
라 - 도 - 레 - 미 - 솔
(경기민요조, 현대 귀에 가장 익숙)
```

---

## 3. 장단(리듬) — 뼈대

장단은 서양 박자와 다릅니다. **강약과 호흡이 다르고**, 각 박의 무게가 다릅니다.

**진양장단** (24박, 느린 6/4 느낌):
```
강 - 약 - 중강 - 약 - 강 - 약
|    |    |     |    |    |
미...라...........시......라↓
       (여기서 꺾음)
```
→ 애절한 보컬 라인에 적합, 현대 6/8이나 12/8로 변형 가능

**중모리장단** (12박, 중간 템포 4/4 느낌):
```
강 - 약 - 중강 - 약
```
→ 서사적, 담담한 이야기

**실전 팁**: OST이니 장단을 그대로 쓰기보다 **서양 박자에 장단의 강점 배치만 차용**하는 게 자연스럽습니다.
- 진양 → 6/8 또는 12/8
- 중모리 → 4/4 (강-약-중강-약 그루빙)
- 굿거리 → 4/4 (16분 음표 그루브)

---

## 4. 악기 편성 실전

| 레이어 | 악기 | 역할 | 주의 |
|--------|------|------|------|
| 선율 | 해금 | 보컬과 유니슨·응답 | 시김(vibrato)을 과하게 |
| 선율 | 피리 | 오보이처럼 관통력 | 저역은 약함 |
| 화성 | 가야금 | 아르페지오, 페달톤 | 12현 튜닝 조정 필요 |
| 저음 | 거문고 | 베이스라인 | 개방현 울림 활용 |
| 리듬 | 장구 | 장단 구동 | 민속악 대금 테크닉 차용 |
| 리듬 | 북 | 포인트 타격 | 전투 씬에서 핵심 |
| 현대 | 스트링 앙상블 | 패드, 텍스처 | 화성 채우기 |
| 현대 | 피아노 | 연결선 | 너무 서구화되지 않게 |

**최소 편성**: 해금 + 장구 + 피아노 + 스트링 패드  
**풀 편성**: 위 전부 + 보컬(판소리 창법 토막)

---

## 5. 곡 구조 (OST 사이즈)

```
Intro     해금 솔로, 계면조 선율 4마디 (장구 없이)
A         보컬, 중모리 그루브, 가야금 아르페지오
B (전개)  장단 진양으로 전환, 해금·보컬 대위, 스트링 증가
C (절정)  자진모리 급전환, 북+장구 풀스택, 보컬 꺾음 최대치
Outro     A 선율 재현, 해금만 남아 사라짐
```

---

## 6. "판소리 느낌 살짝"의 핵심 기법

과하게 하면 판소리 곡이 되니, **두세 개만** 차용:

1. **꺾음** — 보컬이나 해금에서 음을 위로 밀었다가 아래로 꺾기  
   `미 → 시↑ → 라↓` (시를 살짝 높여 부르고 라로 떨어뜨림)

2. **비브라토 차별화** — 한국 음악의 시김은 서양 비브라토보다 **늦게 시작, 폭이 넓음**  
   해금이나 보컬에서 롱노트 → 2초 후 흔들기 시작

3. **침묵의 미학** — 장단과 장단 사이 1~2박 완전한 여백  
   서양 OST은 리듬을 채우지만, 판소리는 멈춤이 표현

---

## 즉시 할 수 있는 첫 스텝

1. **계면조로 Cm 펜타토닉 위에서**: `C - Eb - F - G - Bb`
2. **12/8박자, BPM 60~70** (진양 느낌)
3. **보컬 멜로디**: `Eb4 → F4 → G4 → Bb4... → 꺾어서 G4↓ → Eb4`  
   (꺾음 위치를 5~6박째에)
4. **해금으로 같은 선율 2마디 뒤에 모방** (응수唱 구조)
5. **피아노는 희소**: 1박에 1음, 루트와 5도만

여기서 판소리 비율을 더 올리고 싶으면 장구 패턴 추가, 줄이고 싶으면 피아노 보이싱을 재즈적으로. **비율 조절이 "느낌 살짝"의 핵심**입니다.

구체적인 장면이나 감정이 정해지면 더 집중해서 짜줄게요.

### C_skill_plus_nav

좋은 출발점이에요! 몇 가지 핵심 파일을 참고해서 답변할게요.

---

## 시작하는 법 — 3단계

### 1. 판소리의 핵심 요소 하나를 "시그니처"로 고르세요

판소리 전부를 집어넣으면 사극 배경음악이 아니라 판소리 공연이 됩니다. **한두 가지 요소만** 차용하고 나머지는 서양식 오케스트레이션/프로덕션으로 가져가는 게 사극 OST다운 느낌입니다:

| 판소리 요소 | 사극 OST에서의 활용 | 난이도 |
|---|---|---|
| **시김새** (관행적 꾸밈음) | 멜로디의 핵심 음에 시김새를 얹어 "국악 창법" 느낌 | ★★☆ |
| **호흡법 / 비성** | 보컬이나 관악기에 비성(quasi-tag基层) 색깔 부여 | ★★★ |
| **진양조 · 중모리 · 자진모리** 장단 | 리프나 반주 패턴의 리듬 기반으로 사용 | ★★☆ |
| **대목 / 아니리** 교대 | 곡의 텍스처 대비(서사적 vs 서정적)에 활용 | ★☆☆ |
| **펜타토닉 + 비극음계** (칠음 / 변환) | 화성 양식의 근간 | ★☆☆ |

**추천**: 처음이라면 **시김새 + 진양조/중모리 장단** 조합이 가장 즉신효과가 좋습니다.

---

### 2. 장단을 먼저 잡고 시작하세요

사극 OST에서 판소리 느낌이 가장 잘 사는 건 리듬입니다:

- **진양조** (느린 6/4 또는 18박): 서정적·비극적 장면 (도입, 여운)
- **중모리** (중간 12/8): 이야기가 흐르는 장면
- **자진모리** (빠른 4/4): 격전, 절정

**실습**: 진양조 느낌의 18박을 6/4 × 3마디 또는 3/4 × 6마디로 잡고, 강박 패턴을 `x . . x . . | x . . . . x | . . x . . .` 식으로 배치해 보세요. 이게 "느린 사극 리듬"의 뼈대입니다.

---

### 3. 음계와 화성 — "선율은 동양, 화성은 서양" 믹스

한국 전통 음계는 크게:

| 음계 | 음 | 느낌 |
|---|---|---|
| **경음계** (평조) | C D E♭ G A♭ | 담담, 궁중, 우아 |
| **비음계** (계면조) | C D♭ F G A♭ | 애절, 비극, 원한 |

**전략**: 
- **선율·리프**: 비음계 5음 (C D♭ F G A♭) 위주로 작성 → 사극 특유의 애절함
- **화성**: 서양식 단화성(i iv V i 등) + 가끔 ♭II, ♭VII로 "모달한 확장" | Fm | Gm | A♭ | A♭ ||  ← 비음계로 선율 짜면 자연스럽게 섞입니다.

---

## 즉시 해볼 수 있는 워밍업

1. **기조(根調) 선택**: A♭ 단조 (비음계 친화적, 성역도 좋음)
2. **멜로디 스케치** (비음계 위에서):
   ```
   A♭4 → G4 → F4 — (시김새: F를 E로 밀었다가 F로) → D♭4 → C4
   ```
3. **장단 오스티나토** (중모리 12/8 느낌):
   ```
   타악: | x . . x . . | . x . . x . | (6+6 = 12)
   베이스: | A♭  .  .  E♭  .  . | Fm  .  .  .  .  . |
   ```

이 세 가지만으로도 "사극 + 판소리 냄새"가 납니다. 여기에 현악 패드, 거문고/가야금 톤idx의 플러그인, 보컬 레이어를 올리면 됩니다.

---

## 다음 단계

이 뼈대 위에서:
- **형태** 잡기 (vs. 후렴 vs. 대목스러波动 구조)
- **악기 편성** (거문고+해금+서양오케스트라 혼합)
- **구체적 화성 진행** 설계

어느 쪽을 더 파고 싶으신가요?

---

## P6_reharm — Reharmonization

**Prompt:** Dm7 - G7 - Cmaj7 진행을 좀 더 모던하게 바꿔줘.

### A_baseline

Dm7 - G7 - Cmaj7 진행을 모던(네오소울, R&B, 모던재즈)하게 바꾸는 4가지 구체적 방법입니다.

**1. 알터드 & 텐션 추가 (네오소울/모던재즈)**
`Dm9 - G7#5b9 (또는 G7alt) - Cmaj9`
> V도에 알터드 톤을 주어 긴장감을 만들고, I도에 9th를 넣어 몽환적으로 해결합니다.

**2. 트라톤 서브스티튜션 (크로매틱 베이스 하행)**
`Dm9 - Db7#11 - Cmaj9`
> G7 대신 Db7을 사용해 베이스가 D - Db - C로 부드럽게 반음 하행하는 세련된 사운드를 냅니다.

**3. 슬래시 코드 활용 (부드러운 보이스 리딩)**
`Dm7 - Dm7/C - G7/B - Cmaj7`
> 기존 진행에 베이스 반음 하행(C - B)을 끼워 넣어 정형적인 II-V-I의 틀을 흐릿하게 만듭니다.

**4. 나란히 이동 (네오소울 특유의 몽환적 진행)**
`Dm7 - Dbmaj7 - Cmaj7`
> V도를 Dbmaj7로 치환하여 3개의 코드가 반음씩 나란히 내려가는 현대적인 느낌을 줍니다.

### B_skill_only

전형적인 ii–V–I를 모던하게 바꾸는 방법 몇 가지:

---

### 1. 트라이톤 대체 + 확장음
**Dm9 → D♭7♭9 → Cmaj9**
- 베이스가 반음 하강 (D → D♭ → C), 크로마틱 그래비티가 강함
- D♭7은 G7과 tritone을 공유, 기능은 같은데 색이 완전히 달라짐

### 2. 백도어 프로그레션
**Dm7 → Fm7 → Cmaj7**
- G7 대신 ♭VII7의 관계인 iv (Fm7)로 해결
- 단조 느낌을 주면서도 부드러운 랜딩. 네오소울·R&B에서 흔함

### 4. sus로 미해결감 주기
**Dm7 → G7sus4 → Cmaj7**
- 또는 **Dm9 → G13sus → Cmaj9**
- 3rd를 4th로 미뤄서 긴장감을 연장. 해결이 더 넓게 퍼짐

### 5. 크로마틱 평행하강
**Dm7 → D♭maj7 → Cmaj7**
- 모두 maj7 코드가 반음씩 내려감
- 빌 에반스 이후 모던재즈·네오소울에서 자주 쓰이는 움직임

---

**추천**: 장르에 따라 — 네오소울이면 1번이나 5번, R&B면 2번, 좀 더 열린 느낌이면 3번. 어떤 분위기를 원하는지 알려주면 더 구체적으로 좁혀줄게.

### C_skill_plus_nav

ii–V–I(C)를 모던하게 바꾸는 몇 가지 방향:

---

### 1. 얼터드 도미넌트 — 재즈/네오소울 느낌
```
Dm7 → G7alt → Cmaj7
```
- G7alt = G7(♭9♯9♭13) = G7(♭9♯5) — 아브라모비치 사운드
- E♭멜로딕 마이너 5번째 모드 위 음들이 자연스럽게 걸림

### 2. 트리톤 대체 — 스무스/시티팝
```
Dm7 → D♭7 → Cmaj7
```
- G7 대신 ♭II7. 반음 하강 베이스 라인이 부드럽게 깔림
- D♭7 = D♭(F A♭ C♭→B) — C와의 공통음 많아 자연스러운 해결

### 3. 백도어 진행 — R&B/모던팝
```
Fmaj7 → B♭7 → Cmaj7
```
- ii–V를 IV→♭VII로 교체. 서브도미넌트 색깔이 온화하면서도 surprise 있음

### 4. 디미니시드 패스 + 색채음 — K-R&B/인디
```
Dm7 → D♭dim7 → Cmaj9
```
- 크로매틱 패싱 코드로 긴장감. Cmaj9(9th=D)로 dm7의 루트가 해결되어 연결이 매끄러움

### 5. 서브도미넌트 대체 + 얼터드 — 모던재즈/크로스오버
```
Dm11 → G7(♯11♭13) → C6(9)
```
- ii도 V도 모두 익스텐션 열어서 "색"을 더함
- C6(9)는 Cmaj7보다 모던한 느낌 (루트 위 6th 추가)

---

**추천:** 장르에 따라—
- **네오소울/K-R&B** → 1번 또는 4번
- **시티팝/스무스** → 2번
- **모던팝/인디** → 3번
- **재즈 클래식을 약간 비틀고 싶다** → 5번

어떤 장르/분위기로 가고 싶으신가요? 더 구체적으로 맞출 수 있습니다.

---

## P7_acoustic_ballad — Instrument idiom

**Prompt:** 기타 한 대로 치는 어쿠스틱 발라드 코드 진행을 추천해줘.

### A_baseline

기타 단독 연주에 어울리는 어쿠스틱 발라드 코드 진행 3가지입니다.

**1. 대중적인 감성 발라드 (Key: C)**
`F - G - Em - Am - Dm - G - C`
* 특징: 한국 대중 발라드의 정석적인 진행입니다. 서정적인 멜로디가 가장 잘 돋보입니다.

**2. 슬프고 애절한 발라드 (Key: Am)**
`Am7 - Dm7 - Fmaj7 - E7`
* 특징: 단조의 쓸쓸함에 E7 코드가 이국적이고 애절한 느낌을 더합니다. 느린 아르페지오 연주에 최적화되어 있습니다.

**3. 따뜻하고 편안한 위로 발라드 (Key: G)**
`G - D/F# - Em - C`
* 특징: 베이스 음이 반음씩 하행하는 진행으로, 희망적이고 편안한 분위기를 냅니다.

💡 **연주 팁:** 기타 배김(통울림)을 살리기 위해 **카포(Capo)를 3~5프렛**에 끼우면 어쿠스틱 발라드 특유의 맑고 따뜻한 사운드를 얻을 수 있습니다.

### B_skill_only

어쿠스틱 기타 발라드에 어울리는 코드 진행 4가지를 추천드립니다. 모두 기타 한 대로 편하게 잡을 수 있는 키와 보이싱입니다.

---

### 1. 서정 + 여운 — 카포 2프렛, G 기준

```
| G   | Em7 | Cadd9  | Dsus4 - D |
| G   | Em7 | Am7    | Dsus4 - D |
| C   | D   | Bm7    | Em7       |
| C   | D   | G      | G         |
```

- **포인트**: `Cadd9`, `Dsus4→D` 서스펜드 해결이 잔잔한 여운을 만듭니다.
- 베이스 라인이 `G→E→C→D`로 내려갔다 올라가는 자연스러운 움직임.

---

### 2. 살짝 어둡고 깊은 느낌 — 카포 4프렛, E 기준 (실제 G♯)

```
| E      | C#m7 | Amaj7  | B     |
| E      | C#m7 | Am7    | B     |
| A      | B     | G#m7   | C#m7 |
| A      | Am    | E      | E     |
```

- **포인트**: 후반 `A→Am`에서 메이저→마이너 변화가 극적인 우울감을 줍니다.
- 한국 발라드 후렴 전환부에서 자주 쓰는 장식.

---

### 3. 따뜻하고 안정적인 — 카포 없이, D 기준

```
| D     | Bm7  | G      | A      |
| D     | Bm7  | Em7    | A7     |
| G     | A    | F#m7   | Bm7    |
| G     | Em7  | D      | D - Dsus4 - D |
```

- **포인트**: `A7`이 서브도미넌트 마이너트 역할로 살짝 어두운 색을 넣어줍니다.
- 2프레이즈 후반 `G→Em7→D` 하강 종지.

---

### 4. 모던 K-pop 발라드 느낌 — 카포 1프렛, C 기준 (실제 D♭)

```
| C       | G/B    | Am7     | Fmaj7   |
| C       | G/B    | F       | G       |
| F       | Em7    | Dm7     | G7      |
| F       | Fm     | C       | C       |
```

- **포인트**: `F→Fm→C` 피카디 종지. 가장 "한국 발라드"다운 마무리.
- `G/B` (1전위) 로 베이스가 `C→B→A`로 반음씩 내려가는 라인이 매끄럽습니다.

---

### 기타 연주 팁

- **아르페지오**: 엄지로 저음 3줄, 검지·중지·약지로 3·2·1현을 하나씩 뜯으면 발라드에 적당합니다.
- **스트로크**: ↓ ↓↑ ↑↓↑ (한 박 반 단위) 패턴이 무난합니다.
- **서스펜드**: `sus4→해결`을 타이밍마다 섞어주면 표정이 생깁니다.

원하는 분위기(더 밝게, 더 어둡게, 더 연극적으로 등)를 말씀해주시면 그 방향으로 더 좁혀드리겠습니다.

### C_skill_plus_nav

# 어쿠스틱 발라드 코드 진행 추천

기타에 맞는 키(열린 코드 활용)와 감성별로 추천합니다.

---

## 1. 기본 — 담담하고 따뜻한 발라드
**키 G 메이저**

``| G | Em | C | D |``

- 가장 클래식한 서술형 진행. 초보자도 연주하기 쉽움.
- **변주**: `C → D` 대신 `C → D/F♯` (베이스 반행 추가)

---

## 2. 쓸쓸한 발라드 — 파편화된 감정
**키 E 마이너**

``| Em | C | G | D |``

``| Em | C | Am | Bm7 → B7 |``

- `Bm7 → B7` 해결이 이끌음표 효과로 더 애절해짐.
- 프레이즈 끝에 `B7` 대신 `B7sus4 → B7` 텐션/릴리즈 추가 가능.

---

## 3. 영원한 사랑 느낌 — I → V → vi → IV
**키 D 메이저**

``| D | A | Bm | G |``

- 팝 발라드의 정석적 진행. 감동적이면서도 안정적.
- **핸거운 느낌으로**: `G` 대신 `Gmaj7` 사용.

---

## 4. 캡포 활용 — 풍부한 오픈 코드 사운드
**형태: C 메이저 폼, 캡포 2프렛 (키 D)**

``| C | Am | F | G |``

- 캡포 덕분에 오픈 코드 풍부한 울림 + 실제 피치는 D.
- **다크하게**: `Am` → `A7` 또는 `F` → `Fm` (나폴리 느낌).

---

## 5. 감정 고조 — 브릿지용
**키 G 메이저 → 단조 전환**

``| Em | Cm | G | D/F♯ |``

- `C → Cm` ( ♭III 도입)이 순간적으로 어두워지며 브릿지 느낌 부여.
- 혹은 `| E minor | C | Am7 | D7 |` 로 키를 Em으로 일시 전조.

---

## 기타 치기 팁

| 요소 | 팁 |
|------|-----|
| **아르페지오** | 픽잉 스타일: 베이스→3현→2현→1현 순으로 뜯기 |
| **스트로크** | 느리게 `▼ ▼ ↑ ▼ ↑` 패턴, 강조는 2, 4박자 |
| **베이스 런** | `G → D/F♯ → Em` 처럼 베이스 반행으로 흐름 부드럽게 |
| **텐션 코드** | `G` 대신 `Gadd9`, `D` 대신 `Dsus4 → D` 해결 |
| **쉬는 마디** | 코러스 끝에 한 마디 코드 없이 멜로디만—공간이 감동 만듦 |

---

원하는 분위기(슬픔, 위로, 희망 등)를 더 구체적으로 알려주면 그에 맞춰 진행을 세밀하게 조정해 드릴 수 있습니다.

---

## P8_daw_eq — Boundary (negative)

**Prompt:** DAW에서 보컬 EQ 어떻게 걸어야 해?

### A_baseline

보컬 EQ는 **"제거(감쇠) → 추가(증가)"** 순서로 걸며, 구체적인 주파수 대역별 설정법은 다음과 같습니다.

**1. 저역 정리 (High-Pass Filter)**
*   **80Hz~100Hz 이하를 완전히 깎습니다.** (12dB~24dB Slope 적용)
*   목적: 마이크 룸블, 저역 잡음, 불필요한 탁한 저음 제거. 남성 보컬은 80Hz, 여성 보컬은 100Hz 정도를 기준으로 잡으세요.

**2. 감쇠 EQ (문제 주파수 제거)**
*   **200~500Hz (먹먹함/둔탁함):** 넓은 Q값으로 2~3dB 깎습니다. 
*   **300~400Hz (울림/Boxy):** 탁한 느낌이 들 때 좁은 Q값으로 2~3dB 깎습니다.
*   **1~3kHz (쏘는 소리/Nasal):** 코막힌 소리나 귀를 찌르는 소리가 나면 좁은 Q값으로 2~3dB 깎습니다. (특히 2kHz 근처)

**3. 증가 EQ (매력 더하기)**
*   **100~200Hz (따뜻함/바디):** 보컬이 얇게 들리면 넓은 Q값이나 셸프(Shelf)로 1~2dB 올립니다.
*   **3~5kHz (존재감/명료함):** 보컬이 악기 뒤로 묻힐 때 넓은 Q값으로 1~3dB 올립니다.
*   **10kHz 이상 (공기감/Air):** 하이 셸프로 1~2dB 올려 숨소리와 화사함을 더합니다. (단, 'ㅅ', 'ㅊ' 치찰음이 심해지면 De-Esser 처리 필수)

**💡 핵심 팁**
*   **솔로 금지:** EQ를 돌릴 때는 보컬 솔로가 아닌 **전체 믹스(악기와 함께)** 상태에서 돌리세요.
*   **폭이 좁게 깎고, 넓게 올리기:** 깎을 때는 Q값을 좁게(뾰족하게) 하고, 올릴 때는 Q값을 넓게(완만하게) 설정하세요.

### B_skill_only

이 질문은 이 스킬의 범위 밖입니다. 보컬 EQ는 **오디오 엔지니어링/믹싱** 영역이며, 이 스킬은 작곡·화성·선율·편곡 등을 다루고 DAW 조작이나 믹싱/마스터링/사운드디자인은 다루지 않습니다.

DAW에서 보컬 EQ를 잡는 건 믹싱 스킬이나 보컬 프로덕션 전문 자료를 참고하시는 게 정확합니다.

---

**다만, 작곡 단계에서 EQ와 관련해 고려할 점** 하나만 말씀드리면:

- 보컬 레인지와 악기 배치(orchestration)를 설계할 때, 보컬이 어느 주파수 대역을 차지할지 **미리 생각해두면** 믹싱 단계에서 EQ로 뚫어내야 하는 부담이 줄어듭니다.  
  예: 피아노와 보컬이 경합한다면, 피아ните의 보이싱에서 1–3kHz 영역을 비워주거나 옥타브를 분산시키는 것이 작곡적 해결책입니다.

믹싱/eq 관련 실질적 조언이 필요하시면 믹싱 전용 가이드나 유튜브 튜토리얼(Squeezebox, Mix With The Masters 등)을 추천합니다.

### C_skill_plus_nav

이 질문은 오디오 엔지니어링/믹싱 영역이라서, 이 스킬(작곡·이론)의 범위 밖입니다.

간단히 방향만 말씀드리면:

- **저역 정리**: 80–100 Hz 이하 롤오프 (근접효과 제거)
- **먹먹함 제거**: 200–300 Hz 영역 감쇠
- **존재감**: 2–4 kHz 부스트 (또는 1kHz 감쇠 후 3kHz 부스트)
- **공기감**: 8–12 kHz 셸프 부스트 (여유 있을 때)

하지만 **정확한 EQ는 목소리 텍스처, 마이크, 룸, 장르에 따라 완전히 달라집니다.** 전문 믹싱 리소스나 DAW 이퀄라이저 스킬을 참고하는 것이 좋습니다.

제가 도울 수 있는 건:
- 보컬 라인의 작곡적 결정 (레지스터 선택, 하모니 배치, 멜로디 윤곽)
- 편곡에서 주파수 충돌 방지 (보컬과 다른 악기의 레지스터 분리)
- 장르별 보컬 톤의 작곡적 컨벤션

믹싱/이퀄라이징은 다른 전문 리소스를 찾아보세요!

---

# User-Agent Collaboration Patterns

This file describes how the agent should collaborate with the user across an iterative composition session. It is about conversation design: keeping momentum, preserving decisions, making revisions legible, and turning the user's taste into musical choices.

Use this file when:

- the user wants to build a piece over multiple turns
- the user asks for brainstorming, then revision, then critique
- the user wants the agent to remember the current sketch inside the conversation
- the user is comparing options and needs a decision process
- the user asks for "workflow", "process", "let's develop this", or "help me finish it"

For idea generation specifically, use `musical-brainstorming.md`. For quick state templates, use `../../assets/session-brief-and-decision-log.md`.

## Collaboration stance

The agent should behave like a composition partner, not a replacement composer.

Good collaboration:

- keeps the user's aesthetic intent visible
- offers options before locking decisions
- explains trade-offs briefly
- preserves rejected ideas when they might be useful later
- tracks the current version of the sketch
- turns vague reactions into musical variables

Bad collaboration:

- repeatedly asks for the same information
- rewrites the entire piece when the user asked for a small change
- makes hidden assumptions about genre, vocal range, or instrumentation
- treats the first generated idea as final
- loses track of key, tempo, form, or the user's chosen option

## Session modes

Identify the current mode and answer accordingly.

| Mode | User signal | Agent behavior |
|---|---|---|
| Discovery | "I don't know what I want yet" | offer lanes, ask one anchor, avoid finality |
| Ideation | "Give me ideas" | generate option board, no long critique |
| Drafting | "Write a verse / chorus / cue" | provide concrete material and rationale |
| Revision | "Make it more X" | change one or two variables and show before/after |
| Diagnosis | "This doesn't work" | identify likely causes, prioritize fixes |
| Selection | "Which is better?" | compare against the user's stated goal |
| Polish | "Make it tighter" | small edits, idiom, prosody, playability, clarity |
| Learning | "Explain why" | teach the principle after the example |

If the mode changes, name it briefly:

```text
Cool — we're moving from ideation into revision. I'll keep the chorus concept and only change the harmonic rhythm.
```

## The project card

For multi-turn work, maintain a compact project card inside the conversation when useful.

```md
**Project card**
- Working title:
- Goal / mood:
- Genre frame:
- References:
- Key / tempo / meter:
- Current form:
- Chosen seed:
- Do-not-change:
- Open decisions:
```

Do not force the user to fill every field. Use it only when it helps the session stay coherent.

## The decision log

When the user chooses a direction, record it in plain language.

```md
**Decisions so far**
1. Keep the chorus in D minor, not F major.
2. Use the syncopated hook rhythm from Option B.
3. Avoid copying the reference's bassline; keep only its halftime energy.
```

This prevents the agent from undoing user choices later.

## Revision granularity

Before editing, decide the scale of change.

| Scale | When to use | Example |
|---|---|---|
| Micro | one chord, one note, one lyric stress | change bar 4 from G to G7sus4 |
| Phrase | 2–4 bars | rewrite the second half of the melody |
| Section | verse/chorus/bridge | create a contrasting chorus |
| Arrangement | instrumentation/density | keep chords, change energy arc |
| Concept | mood/genre/reference | reframe the whole song |

If the user asks for a small change, do not perform a concept rewrite.

## Before/after revision format

Use this for concrete edits.

```md
**Original**
| Am | F | C | G |

**Revision: more lift into chorus**
| Am | F | C/E | Gsus4 G |

**What changed**
- `C/E` makes the bass climb instead of reset.
- `Gsus4 G` creates a clearer pre-chorus release.
```

## A/B/C comparison format

Use this when the user is deciding.

```md
| Version | Strength | Risk | Best if you want... |
|---|---|---|---|
| A | immediate, familiar | predictable | radio-pop clarity |
| B | darker, more distinctive | less singalong | emotional tension |
| C | bold texture | may distract from lyric | a signature moment |
```

End by asking which trade-off matters most, not which is "correct".

## Taste extraction

When the user reacts emotionally, translate the reaction into variables.

| User phrase | Possible variable |
|---|---|
| "More expensive" | wider voicings, cleaner voice leading, fewer busy fills |
| "More intimate" | lower register, fewer layers, closer vocal rhythm |
| "More anthemic" | higher tessitura, longer notes, group-vocal-friendly hook |
| "More modern" | shorter intro, earlier hook, leaner loop, current groove/reference check |
| "More classic" | longer setup, live-band idiom, stable form, less abrupt editing |
| "More weird" | asymmetry, modal color, unexpected timbre, phrase displacement |
| "More emotional" | appoggiaturas, registral peak, stronger cadence or withheld cadence |

Always confirm the interpretation when ambiguity matters.

## Handling conflict between user goals

Users often want mutually tense things:

- familiar but not generic
- complex but catchy
- modern but timeless
- sad but uplifting
- genre-authentic but fresh
- reference-like but original

Turn the conflict into a design rule.

```text
Let's make the groove familiar and the harmony fresh.
```

```text
Let's keep the chorus catchy, but make the bridge harmonically stranger.
```

```text
Let's use the reference for energy only, not melody or bassline.
```

## Branching without chaos

When exploring multiple directions, label branches.

```md
**Branch A: intimate ballad**
- 72 BPM, piano, close vocal, vi-IV-I-V

**Branch B: glossy midtempo**
- 96 BPM, syncopated bass, brighter chorus lift

**Branch C: darker cinematic**
- 68 BPM, i-♭VI-iv-V, strings/pad texture
```

When the user chooses, retire or archive the others:

```text
We'll continue Branch B. I'll keep A and C as fallback ideas if the chorus needs contrast later.
```

## The "do-not-change" list

Before revising heavily, protect what the user likes.

```md
**Keep**
- the title phrase
- the descending bass in the verse
- the halftime chorus feel

**Change**
- pre-chorus length
- second line of melody
- final cadence
```

This is especially important when the user says, "I like this, but..."

## How to ask useful follow-up questions

Ask questions that unlock composition decisions.

Weak:

```text
What exactly do you want?
```

Better:

```text
Should the chorus feel bigger because the melody rises, because the chords brighten, or because the arrangement opens up?
```

Weak:

```text
What genre?
```

Better:

```text
Should the listener hear this as pop first, R&B first, or cinematic first?
```

Weak:

```text
Any references?
```

Better:

```text
Give me one reference for energy and one for vocal phrasing. They don't need to be the same genre.
```

## When not to ask a question

Do not ask a clarifying question if the user clearly wants momentum and the missing detail can be assumed safely.

Instead:

```text
I'll assume 4/4, midtempo, contemporary pop for this sketch. If you want jazz/EDM/film later, we can reframe it.
```

Good assumptions are reversible. Bad assumptions silently determine the whole piece.

## Keeping outputs playable

For every generated idea, provide at least one of:

- chord symbols
- Roman numerals
- scale degrees
- rhythm grid
- instrumentation/density instruction
- lyric line with stress marks
- form map
- register/tessitura note

Avoid abstract labels without material.

## User momentum protocols

### When the user is excited

Move forward. Do not slow down with a lecture.

```text
Great, let's develop that. I'll turn Option 3 into an 8-bar chorus and keep the pre-chorus unresolved.
```

### When the user is unsure

Reduce choices.

```text
Pick between these two: A is more singable, B is more distinctive.
```

### When the user is blocked

Use constraints.

```text
Let's write a bad version in 4 bars, then fix one thing at a time.
```

### When the user dislikes everything

Change the axis, not just the notes.

```text
The harmonic lane may be wrong. Let's try rhythm-first and texture-first ideas instead.
```

## Cross-references

- Musical brainstorming → `musical-brainstorming.md`
- Revision and feedback loops → `revision-and-feedback-loops.md`
- Session templates → `../../assets/session-brief-and-decision-log.md`
- Response formats → `../../assets/response-templates.md`
- Diagnostic tables → `../../assets/diagnostic-checklists.md`
- General workflow → `../workflow.md`
- Critique mode → `../critique-and-feedback.md`
- Reference digging → `../research/reference-track-digging.md`
- Style/copyright guardrails → `../research/style-reference-and-copyright.md`

# Critique and Feedback

This file is about how the agent gives critique and feedback on a user's existing work — a finished or in-progress piece, a sketch, a recording, a lyric. Critique is a different mode from generative help (where the agent suggests what to write); it requires listening / reading first, identifying what's working and what isn't, and delivering observations in a way that respects the user's vision and supports their decision-making.

This file complements `workflow.md` (the agent's general advisory framework) and `analysis.md` (the methodology for analytically understanding a piece). Critique builds on analysis but adds an *evaluative* layer — and, more importantly, a *communicative* layer that determines whether feedback actually helps the user.

## When to consult this file

- The user shares a piece (audio, score, lyric, chart) and asks for feedback
- "What do you think?" / "Is this good?" / "What's wrong?"
- Comparing the user's work to a reference track they're aiming at
- Diagnosing why a piece "isn't working" without a specific symptom
- Helping the user evaluate their own work
- Late-stage critique (mixing review, mastering review — though mostly out of scope; see `production-aware/`)
- The user wants the agent to argue the case for or against a creative decision

## What kind of feedback is the user asking for

Different requests look similar but want different responses. The agent should identify which type before responding.

| User says | They probably want |
|-----------|---------------------|
| "What do you think?" | Open-ended response; usually validation + light critique |
| "Is this good?" | Quality judgment; often validation-seeking but might want honest assessment |
| "What's wrong with this?" | Specific problem identification; user has noticed something off |
| "How do I make this better?" | Generative critique with concrete options |
| "Critique this honestly" | Permission to be direct; user is ready for sharp observations |
| "Compare this to [reference]" | Benchmark assessment; specific comparative analysis |
| "Does this work for [genre/purpose]?" | Genre-fit assessment, not absolute quality |
| "I'm not sure about [specific element]" | Targeted critique on that element only |
| "This part feels weird but I can't tell why" | Diagnostic; user has the symptom, needs the cause |
| "My [collaborator/audience] said [criticism]" | Validate or refute the external feedback; meta-critique |

If unclear, ask: "Are you looking for honest critique, brainstorming for improvements, or just wanting to share?"

## Listen / read first, evaluate second

The most important rule of critique: don't critique what isn't there yet. Spend time engaging with the actual material before forming evaluative opinions.

For audio: listen multiple times. Try to feel its intent before naming its flaws. Then map the form, identify the strongest moments, identify what catches your ear (positively or negatively).

For a score / chart: read through, ideally hearing it internally. Identify the form, key plan, melodic identity, motivic structure.

For a lyric: read it as a poem first, then as a sung text. Notice imagery, structure, flow, emotional arc.

Forming critique without engaging fully with the material produces generic or wrong feedback. The agent should resist the temptation to start critiquing as soon as the user shares material.

## Lead with what's working

Genuine, specific praise before critique serves several purposes:

- **Calibrates the user's reading**: they know you've engaged with the actual work, not just pattern-matched to a generic critique
- **Identifies what to preserve**: improvements should not break what's already strong
- **Builds trust**: the user accepts critique more easily after seeing you appreciate the work
- **Models specific observation**: showing you can name a strong moment teaches the user how to identify them themselves

Avoid generic praise ("nice melody", "good vibe"). Specific praise: "the descending line in bar 12 lands the resolution beautifully" or "the contrast between the dense pre-chorus and the sparser chorus is a strong move".

If you genuinely can't find something that works, pause. Either the work is unusually weak (in which case ask the user what their goals are before critiquing) or you haven't engaged with it deeply enough yet.

## Critique frameworks

Different aspects of a piece can be evaluated against different criteria. The agent's job is to identify which framework applies to which observation, and to communicate that.

### Compositional dimensions

A piece can be examined along these axes (drawing from the rest of this skill):

- **Form**: section pacing, transitions, energy arc, length
- **Harmony**: progression choices, chromatic interest, voice leading, cadential weight
- **Melody**: contour, range, climax placement, motivic identity, memorability
- **Rhythm**: groove, syncopation, density, meter use
- **Lyrics** (if present): prosody, imagery, structure, point of view
- **Arrangement / orchestration**: instrument choices, voicing, density, layering
- **Production / mix** (if applicable): frequency balance, dynamic range, sonic identity
- **Performance**: vocal expression, instrumental phrasing, ensemble cohesion

Strong critique identifies *which dimension* is the problem, not just that "something feels off". A piece can have strong harmony but weak form; a piece can have a great hook but a weak bridge. Naming the dimension is half the work.

### Genre fit vs. absolute quality

A piece that doesn't work for its target genre might be excellent for a different genre. Distinguish:

- **Genre-fit critique**: "This isn't quite working as K-pop because [specific genre convention]"
- **Absolute critique**: "This doesn't work because [universal compositional issue]"
- **Cross-genre observation**: "This isn't K-pop, but it might work as something else"

Genre-fit critique requires identifying the target genre. If the user hasn't named it, ask. If they have, evaluate by that genre's conventions, not by some abstract standard.

### Intent vs. execution

A piece can be poorly executed in pursuit of a clear intent, or well executed but with unclear intent. The diagnosis matters:

- **Unclear intent + good execution**: the piece sounds professional but you can't tell what it's trying to do. Critique points at communication / focus.
- **Clear intent + poor execution**: you understand what they were going for but specific moments don't deliver. Critique points at technique.
- **Unclear intent + poor execution**: the broader, harder problem. Critique should start with intent before tackling execution.

When critique seems hard to formulate, this is often the underlying issue — you're trying to evaluate execution before knowing what it's executing.

### Personal artistic choice vs. compositional issue

Some apparent "problems" are deliberate artistic choices the agent should respect, not fix. Distinguish:

- **A choice that sounds wrong because it's wrong**: voice leading errors, prosody misalignment, technically unintentional dissonance
- **A choice that sounds wrong because it's unconventional**: deliberate dissonance, unusual structure, anti-genre move
- **A choice that sounds wrong because of the listener's frame**: the listener expected genre A; the user is making genre B

Only the first deserves "fix" framing. The other two deserve "this is a choice; here's what it costs and what it gains" framing.

When uncertain, ask: "Was the [unusual element] intentional?" If yes, treat it as a choice; if no, treat it as an issue.

## Communicating critique well

Beyond identifying what's wrong, the agent needs to deliver the observation in a way that helps the user act on it.

### Specific over general

- **Bad**: "The chorus could be stronger"
- **Better**: "The chorus melody peaks early — bar 2 is the highest note, and bars 3–4 don't have anywhere to go. Try saving the high note for bar 3 or 4."

Generic critique tells the user something is wrong but doesn't help them fix it. Specific critique points at the moment and suggests a direction.

### Options, not commands

- **Bad**: "Change the bridge to be in the relative minor"
- **Better**: "The bridge feels like it's not contrasting enough with the chorus. A few options: modulate to the relative minor, drop to half-time, strip the arrangement back to voice + piano. Pick the one that fits the song's emotional arc."

The user is the composer; the agent supplies options and rationale, not directives.

### Diagnose, then prescribe

Critique is most useful when the diagnostic frame and the prescriptive frame are connected. "The chorus doesn't hit because the verse already used the same harmony and density" is a diagnosis; "raise the chorus tessitura by a fourth and add a layered chorus vocal" is a prescription. Both serve.

The diagnostic frame matters because:
- It teaches the user to spot similar issues in future work
- It explains *why* the prescription should help
- It exposes alternative prescriptions the user can choose from

### Calibrate the dose

Some pieces need broad-strokes feedback; some need fine detail. Match:

- **Early sketch**: focus on broad structural / directional questions; don't critique production details that haven't been finalized
- **Working draft**: detailed compositional critique appropriate; production critique still premature
- **Near-final**: fine-grained critique at the level of specific bars, transitions, vocal lines
- **Released or about-to-release**: only critique on request; the piece is committed; offer perspective rather than prescription

### Order matters

When delivering multiple observations, sequence them well:

1. Lead with strengths (genuine, specific)
2. Address the most important issue next
3. Address smaller issues after
4. Close with options or a question that invites the user to direct further work

A critique that opens with "the form is off" and then mentions five smaller things buries the most important point. A critique that ends "what do you want to focus on next?" hands control back to the user.

## Defensive responses — when the user pushes back

Users sometimes resist critique. Reasons vary:

- They want validation, not critique
- The critique exposes anxiety about a section they were already worried about
- They disagree on artistic grounds
- The critique misunderstood the piece's intent
- The critique was technically correct but poorly framed

The agent should:

- **Acknowledge the disagreement without retreating from honest observation**. "I hear you — your intent there was X, and that does change my reading of that bar" is different from "you're right, I was wrong" or doubling down.
- **Distinguish technical issues from artistic choices**. If the disagreement is about a deliberate choice, that's fine — the user is the composer. If it's about a technical issue (voice leading error, prosody mismatch), the agent can hold the position while explaining why.
- **Offer to drop the topic**. If the user clearly doesn't want to engage, "let's leave that aside; what else can I help with?" respects their autonomy.
- **Don't get sycophantic when challenged**. Backing down from a well-founded observation in response to mild disagreement leaves the user without honest information. Hold positions calmly when they're well-founded; release positions when they're not.

## Critique in different stages

The right kind of critique depends on where the user is.

### Idea / concept stage

Critique here is mostly about whether the idea has compositional weight, whether it could go somewhere, and what the next move might be. Don't critique unwritten music; critique the seed.

Useful: "This concept is strong because [reason]; it suggests [direction]; the risk is [pitfall]."

Not useful: "The chorus might not hit." (No chorus exists yet.)

### Sketch / draft stage

Critique focused on form, key choices, structural decisions. Production-level critique premature.

Useful: "The verse-chorus relationship is clear; the bridge feels tacked on. Try connecting it to the chorus by [specific motivic / harmonic move]."

Not useful: "The kick isn't punchy enough." (Production isn't finalized.)

### Working draft

Most generative critique fits here. Compositional issues, motivic development, transition quality, lyric specificity all in scope.

### Near-final / mix-ready

Fine-grained critique. Specific bars, specific phrases, specific production decisions. Be careful — late-stage changes are expensive (see `production-aware/pre-production-decisions.md`).

### Released / committed work

Critique becomes retrospective rather than prescriptive. "This works because X; if I were doing this from scratch I might try Y" rather than "fix this".

## Pitfalls in critique

- **Don't critique what you didn't engage with.** Forming opinions before listening / reading thoroughly produces generic or wrong critique.
- **Don't impose your taste as truth.** "I would have done X" is your taste; "this doesn't work because Y" requires Y to be supported by something more than preference.
- **Don't critique unfinished elements.** A demo doesn't need mix critique; a sketch doesn't need polish critique. Match the depth of your critique to the stage of the work.
- **Don't blow up an entire piece for one issue.** "This whole piece doesn't work" is rarely true and rarely actionable. Identify specific issues; the piece probably has redeemable elements.
- **Don't avoid critique to spare feelings.** Sycophancy disguised as kindness leaves the user without honest information. Direct critique delivered respectfully is the kindness.
- **Don't pile critique high.** Five specific observations are more useful than fifteen. Prioritize.
- **Don't forget to check the user's goal.** Critique that ignores what the user was trying to do can be technically correct and entirely useless.
- **Don't moralize.** "This should be better" is moralizing; "this could be more X if Y" is critique. Stay in the descriptive-practical register.
- **Don't validate when you mean to critique, and don't critique when validation is warranted.** Both errors damage trust.

## When critique should turn into something else

Sometimes a "critique request" really wants something different:

- **The user is anxious and needs to be heard before being critiqued.** Listen / read; acknowledge the work's effort; ask what they want before evaluating.
- **The user is stuck and needs new options, not evaluation.** Switch to generative mode: brainstorm extensions, alternatives, fresh directions.
- **The user is comparing to a reference unrealistically.** Reframe: "this isn't [reference] because [reason]; that's a different goal — your piece has its own goal".
- **The user is processing failure.** Sometimes a piece they shared was rejected, didn't land with audiences, or didn't get them what they hoped. Critique here can sting. Acknowledge the context; offer perspective rather than dissection.

## Quick decision matrix

| User signal | Response mode |
|-------------|---------------|
| Excited share, fresh from finishing | Lead with what's working; offer light critique only if invited |
| "What do you think?" with low specificity | Validation-leaning; offer one or two specific observations; ask if they want more |
| "Is this good?" | Honest evaluative response with caveats about genre/goal/intent |
| "Critique this honestly" | Direct; pull punches less; still respectful, still specific |
| "What's wrong with [specific section]?" | Targeted diagnosis; don't drift to other sections unless invited |
| "How can I make this better?" | Generative critique: identify issues + concrete options |
| Sharing late-stage work | Match critique depth to stage; don't critique early-stage elements that have been finalized |
| Sharing released work | Retrospective framing; comment on what works and what was an interesting choice |
| Disagreement / pushback | Hold defensible positions; offer to drop the topic; don't get sycophantic |
| Visible distress about the work | Acknowledge effort; ask what they want before continuing critique |

## Cross-references

- The agent's general advisory framework → `workflow.md`
- Methodology for analyzing existing music → `analysis.md`
- Compositional dimensions to evaluate against → harmony files, melody files, form files, etc.
- Pre-production stage and what's appropriate to critique when → `production-aware/pre-production-decisions.md`
- Genre-specific evaluation criteria → relevant `genres/*.md`
- Pedagogical critique (when teaching, not when collaborating) → `teaching-composition.md`
- Production critique (frequency, mix) → `production-aware/arrangement-for-mix.md`, `production-aware/energy-and-dynamics.md`
- Lyric-specific critique → `songwriting/lyric-writing.md`
- Hook-specific critique → `songwriting/hooks-and-memorability.md`

# Classical Forms

This file covers the standard formal types of Western art music — binary, ternary, sonata, rondo, theme-and-variations, fugue, and the larger composite forms (sonata cycle, suite). For popular song forms, see `popular-song-forms.md`. For narrative arcs and transitions across forms, see `narrative-and-transitions.md`. For the techniques used to organize these forms (motif development, modulation, phrase grammar), see `melody/motivic-development.md`, `harmony/modulation.md`, `melody/phrase-structure.md`.

## When to consult this file

- The user asks about sonata form, rondo, variations, or any classical formal type
- Analyzing a classical piece's structure
- Composing in a classical-form-based idiom (modern or pastiche)
- Understanding how film scores or longer pop pieces borrow from classical forms

## Why form matters

Form is the architecture of musical time. A 30-second pop hook works on memorability; a 30-minute symphonic movement works on architecture. Classical forms developed over centuries to handle longer durations — they manage tension, return, contrast, and resolution at scales that pure repetition can't.

When advising:

- Form provides expectations the music can fulfill or subvert.
- Form gives structural meaning to local events (a cadence at "the right moment" feels different than the same cadence elsewhere).
- Form is genre-bound: sonata form belongs to the Classical-Romantic period; using it now is either pastiche or specific homage.
- Modern composers often blend formal types or invent custom forms. Understanding the classical defaults clarifies what's being modified.

## Binary form (AB)

The simplest closed form: two sections, both repeated. Both sections often end with cadences in their respective key areas.

Standard pattern:
- **A**: starts in tonic, modulates to a related key (V or relative major), cadences there.
- **B**: starts in the new key, returns to tonic, cadences there.

In notation: `||: A :||: B :||` (each section repeated).

Used in Baroque dances (gigue, courante, sarabande, gavotte), folk songs, and many short Classical pieces.

### Rounded binary

A binary form where the second half ends with a return of the A section's opening material. Sometimes notated `||: A :||: B-A' :||` or just felt as a rounded structure.

This blurs into ternary form (ABA). The distinction is one of proportion — a "rounded binary" has B and A' as a single B section (the return is brief), while ternary has B and A as separate, balanced sections.

## Ternary form (ABA)

Three sections: a statement, a contrasting middle, and a return.

The classic pattern:
- **A**: complete statement in tonic.
- **B**: contrasting section, usually in a related key (relative or parallel minor/major, dominant, mediant). May be more developmental.
- **A**: return of A. Often shortened ("A'") or ornamented.

The return of A is the structural payoff. It feels like "coming home" after the contrast of B.

Used in:

- The slow movements of many Classical sonatas and symphonies.
- Many opera arias (the *da capo* aria is ABA, with the return ornamented by the singer).
- Minuet-and-trio movements (Minuet is A, Trio is B, return of Minuet is A — actually a ternary at the movement level).
- Many short character pieces (Chopin nocturnes, Brahms intermezzi).

### Compound ternary

Each section is itself a smaller form (often binary). Common in minuet-trio movements: Minuet is itself rounded binary, Trio is rounded binary, the whole movement is ABA.

## Sonata form

The dominant form of classical instrumental music from roughly 1770 to 1900. A movement built around the contrast and reconciliation of two key areas (and, by extension, two thematic groups). The structural engine is *modulation* — the form is fundamentally about leaving home and coming back.

Three main sections:

### 1. Exposition

States the thematic material in two distinct key areas.

- **First theme group** (in the tonic key)
- **Transition (modulating)**: travels from tonic to the second key, usually through sequence and tonicization
- **Second theme group** (in a related key — typically V for major-mode pieces, III for minor-mode)
- **Closing material**: cadences strongly in the second key, sometimes with new material called the "codetta"

The exposition is usually repeated (`||: Exposition :||` in older scores; less often in modern performance).

### 2. Development

Takes the exposition's themes (and motifs from them) through various keys, fragmenting, sequencing, transforming. The development creates instability and tension by:

- Modulating constantly (often to distant keys)
- Fragmenting themes into smaller motifs
- Combining themes contrapuntally
- Using sequences to drive momentum
- Building toward a structural dominant (a long passage on V of the home key, called the *retransition*) before the recapitulation

### 3. Recapitulation

Returns to the tonic and restates the exposition's themes — but now both theme groups are in the tonic key (rather than tonic and dominant). The transition is rewritten to stay in tonic instead of modulating away.

This resolution — second theme group "brought home" — is the structural climax of sonata form.

### Optional sections

- **Slow introduction** (before the exposition): common in Classical and Romantic symphonies.
- **Coda**: extended ending after the recapitulation. In late Beethoven and Romantic sonata-form, the coda often functions as a "second development" or final expansion.

### Sonata form variations

- **Sonata-rondo**: combines sonata form with rondo's refrain principle. Structure: ABACABA, where the first AB pair is the exposition, the second AC is the development (with C being a contrasting section in a new key), and the return is the recapitulation.
- **Sonatina**: a sonata-form movement without a development section (or with a very short one). Common in early Classical and pedagogical pieces.
- **Slow-movement sonata**: typically has a brief or absent development. Movement structure may be closer to ternary with sonata-form lighting.

### Where to find sonata form

Almost every first movement (and many last movements) of:

- Haydn, Mozart, Beethoven sonatas, symphonies, string quartets, concertos.
- Schubert, Mendelssohn, Schumann, Brahms — Classical-influenced Romantic.
- Late Romantic (Bruckner, Mahler, Tchaikovsky) — sonata form expanded to enormous scale.
- 20th century — Shostakovich, Prokofiev, Britten — sonata form adapted to modernist language.

## Rondo

A form where a refrain (A) alternates with contrasting episodes. The simplest pattern is `ABACADA...` but most rondos are more constrained.

### Five-part rondo (ABACA)

The most common Classical rondo. Three statements of A separated by two contrasting episodes (B and C).

- **A**: refrain in tonic
- **B**: first episode, often in dominant or relative
- **A**: refrain in tonic
- **C**: second episode, often in a more distant key (subdominant, parallel mode)
- **A**: refrain in tonic

### Seven-part rondo (ABACABA)

Adds another A and another B, making it more symmetrical. Common in finale movements of Classical sonatas and concertos.

### Sonata-rondo

Hybrid of rondo and sonata form. ABACABA where the first ABA serves as exposition, the C functions as development, and the second ABA is recapitulation.

### Where to find rondo

Often the finale of:

- Classical concertos (Mozart, Beethoven)
- Haydn and Mozart symphonies and sonatas
- Some string quartets

Rondos are typically lively, dance-like, and easy to follow. The repeated refrain gives the movement a sense of fun.

## Theme and variations

A form where a theme is stated, then varied multiple times. Each variation transforms the theme but keeps something invariant — usually the harmonic structure (the *chord progression* of the theme) and/or the phrase structure.

### What gets varied

| Variation type | What changes |
|----------------|--------------|
| **Melodic / figural** | The melody is decorated with passing tones, ornaments, faster rhythmic figures |
| **Rhythmic** | The theme's rhythm is changed (e.g., from quarters to triplets) |
| **Harmonic** | The theme's chords are reharmonized (e.g., minor mode for a major-mode theme) |
| **Modal** (parallel mode) | The theme is recast in major/minor of the same tonic |
| **Contrapuntal** | The theme is combined with a counter-melody, or stated in canon |
| **Textural** | The theme's accompaniment is changed (e.g., piano LH from chords to arpeggios) |
| **Character / topical** | The theme is recast in a new style (march, dance, hymn) |

### Famous variation sets

- Bach's *Goldberg Variations* — 30 variations on a theme, organized in groups of three (canons at progressively wider intervals).
- Beethoven's *Diabelli Variations* — 33 variations ranging from elaborate to bizarre.
- Brahms's *Variations on a Theme by Haydn*, *Variations on a Theme by Paganini*.
- Rachmaninoff's *Rhapsody on a Theme of Paganini* — a variation set rather than a true rhapsody.

### Variations within other forms

Variation technique appears within sonata-form movements, in slow movements (often theme + 2-3 variations), and in jazz (where soloists effectively play variations on the changes). The technique is a building block of many forms.

For details on transformation operations, see `techniques/theme-and-variation.md`.

## Fugue

Already covered in detail in `counterpoint.md`. Briefly:

A fugue is both a contrapuntal *technique* (multiple voices imitating a subject) and a *form* (a piece organized around fugal procedures).

Structural sections:

- **Exposition**: voices enter one by one with the subject (or its tonal answer)
- **Episodes**: contrapuntal interludes between subject entries
- **Middle entries**: the subject reappears in related keys
- **Stretto** (optional): subject entries overlap before they finish
- **Final entries**: subject returns in the home key, often with augmentation or pedal point
- **Cadence**

Fugue is a rare formal type today but was the supreme form of Baroque instrumental music. Bach's *Well-Tempered Clavier* and *Art of Fugue* are the canonical studies.

## Through-composed form

Music that doesn't repeat sections — each part is new. Used when the music's narrative or text demands continuous development without return.

Common in:

- Late Romantic art songs (Schubert's "Erlkönig" is famous).
- Wagnerian opera scenes (entire scenes flow without sectional returns).
- 20th-century classical, especially when the music follows a non-musical structure (text, narrative, program).
- Some film cues.

The challenge of through-composed form: without return, what creates coherence? Usually it's motivic development, harmonic logic, or a dramatic narrative.

## The sonata cycle (multi-movement structure)

A *cycle* is a multi-movement piece (sonata, symphony, string quartet, concerto). Standard Classical-Romantic patterns:

### Three-movement structure (typical of concertos and most Classical sonatas)

1. **First movement** (sonata form, fast, in tonic)
2. **Slow movement** (ternary or theme-and-variations, slow, in a related key — often subdominant or relative minor)
3. **Finale** (rondo or sonata-rondo, fast, in tonic)

### Four-movement structure (typical of symphonies and string quartets)

1. **First movement** (sonata form, fast, in tonic; often with slow introduction)
2. **Slow movement** (ternary or theme-and-variations, slow, in a related key)
3. **Dance movement** (minuet-and-trio in Classical; scherzo-and-trio in Beethoven onward; ternary, in tonic, moderately fast or fast)
4. **Finale** (rondo, sonata-rondo, sonata, or rarely theme-and-variations; fast, in tonic)

The order can be modified — Beethoven's 9th puts the scherzo second and the slow movement third; Brahms's 4th uses a passacaglia (variation form on a ground bass) for the finale.

### Suite

A multi-movement piece based on dance forms, common in Baroque era (Bach's *English Suites*, *French Suites*, *Partitas*, *Cello Suites*). Each movement is a different dance type (allemande, courante, sarabande, gigue, optional minuet/gavotte/bourrée), all in the same key.

### Concerto

A piece for soloist(s) and orchestra, typically three movements. The first movement is in *concerto sonata form*, a variant where the orchestra states the themes once, then the soloist enters and the structure proceeds in a modified sonata form. The slow movement is often ternary; the finale typically rondo.

## Other classical forms (briefly)

| Form | Description | Where used |
|------|-------------|------------|
| **Passacaglia / chaconne** | Variations over a repeating bass line (ground bass) | Baroque (Bach, Buxtehude); occasional revivals (Brahms's 4th finale) |
| **Toccata** | Free, virtuosic showpiece | Baroque keyboard music |
| **Prelude** | Often improvisatory, free; introduces a fugue or stands alone | Bach's WTC; Chopin's preludes |
| **Etude** | Study piece focused on a technical challenge | Chopin, Liszt, Debussy |
| **Nocturne** | Lyrical, often in compound meter, evocative of night | Field, Chopin, Fauré |
| **Scherzo** | Fast, often humorous; replaces minuet in 19th-century cycles | Beethoven, Chopin, Brahms |
| **Symphonic poem** | Single-movement orchestral work with extramusical content | Liszt, R. Strauss, Sibelius |
| **Fantasy / fantasia** | Free form, often improvisatory | Bach, Mozart, Schubert, Schumann |
| **Rhapsody** | Free form, sectional, often based on folk material | Liszt, Brahms, Gershwin |

## Modern uses of classical forms

Classical forms haven't disappeared:

- **Film scoring** uses sonata-like structures for major themes; theme-and-variations for character development; ternary form for individual cues.
- **Progressive rock** explicitly borrows classical forms — multi-section structures, returns, formal contrast. Bands like Yes, Genesis, ELP draw on classical templates.
- **Jazz** uses ABAB (32-bar form) and AABA — neither is "classical" but both share the ternary-like return principle.
- **Pop** rarely uses classical forms intact, but borrows the principle of return (verse-chorus-verse-chorus).
- **Contemporary classical** uses both classical forms (in neoclassical or pastiche modes) and through-composed structures; minimalists like Reich and Glass often use process-based forms that look unlike classical templates but share some functional ideas.

## Common pitfalls when advising

- **Don't impose sonata form on every long piece.** Sonata form is a specific technique with specific requirements (modulation, development, recap with key resolution). Many long pieces aren't sonata form and don't want to be.
- **Don't mistake variation for development.** Theme-and-variations keeps the theme intact (with surface variation); sonata form fragments and transforms motifs. Different techniques.
- **Don't reduce form to letter labels.** ABACABA is a label, not a form. The actual form depends on what happens within each section. A pop song's ABACABA is structurally different from a Classical rondo's ABACABA.
- **Don't assume listeners track form consciously.** Most listeners feel form rather than analyze it. The structural logic should serve the listener's experience, not impress them with cleverness.
- **Don't blend pastiche and modernism without intent.** Writing a "sonata" today is either pastiche, neoclassical homage, or modernist subversion. Be clear which.

## Quick decision matrix

| Want | Try |
|------|-----|
| Short, balanced piece with contrast | Ternary (ABA) |
| Story-like long piece with return | Sonata form |
| Refrain-with-episodes lively piece | Rondo |
| Variations on a single idea | Theme and variations |
| Continuous narrative, no return | Through-composed |
| Multi-section work (movement set) | Sonata cycle (3 or 4 movements) |
| Suite-style dance collection | Suite (multiple stylized dances in one key) |
| Contrapuntal showpiece | Fugue |
| Ground-bass variations | Passacaglia / chaconne |
| Narrative orchestral work | Symphonic poem (programmatic) |

## Cross-references

- Phrase structure (the building blocks of these forms) → `melody/phrase-structure.md`
- Motivic development (the engine of sonata form and fugue) → `melody/motivic-development.md`
- Modulation (the structural principle of sonata form) → `harmony/modulation.md`
- Counterpoint (the technique of fugue) → `counterpoint.md`
- Theme-and-variation as a technique → `techniques/theme-and-variation.md`
- 20th-century formal innovations → `techniques/20th-century-techniques.md`
- Popular song forms (a different formal world) → `popular-song-forms.md`
- Narrative arcs across forms → `narrative-and-transitions.md`
- Classical period conventions → `genres/classical-periods.md`
- Form analysis → `analysis.md`
- Form templates for quick reference → `assets/form-templates.md`

# Afrobeats and Amapiano

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and regional scenes with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition-facing conventions for Afrobeats/Afropop and amapiano-related crossover writing. It is not a complete guide to African popular music, West African traditions, South African club music, or any local scene.

## When to consult this file

- The user asks for Afrobeats, Afropop, Naija pop, Ghanaian pop influence, amapiano, or popiano energy
- The user wants a danceable global-pop groove with African/Caribbean rhythmic influence
- The user wants to blend pop/R&B with log-drum or Afrobeats-style percussion
- The user asks what is current in African pop; pair with active web research

## Important distinctions

- **Afrobeat** without the final "s" usually refers to the Fela Kuti lineage: long-form Nigerian music blending highlife, Yoruba music, funk, jazz, horns, political vocals, and extended grooves.
- **Afrobeats** with "s" is a broad 21st-century West African popular-music umbrella, especially associated with Nigeria and Ghana, mixing local pop, highlife, dancehall, hip-hop, R&B, soca, and more.
- **Amapiano** is a South African dance-music family associated with piano/keys, deep-house/kwaito influence, shakers, spacious grooves, and log-drum basslines.

Do not use these terms interchangeably.

## Afrobeats / Afropop composition priorities

**Core priorities**:

- groove before chord complexity
- syncopated percussion and bass interplay
- relaxed but danceable pocket
- short, memorable vocal phrases
- call-response or ad-lib-friendly hook design
- warm major/minor loops, often with guitar/keys ostinati
- space for vocal rhythm and language prosody

## Afrobeats groove, composition view

A useful writing mindset:

```text
Steady 4/4 dance grid + syncopated percussion + bassline that talks back to the vocal
```

Practical moves:

- Let guitar/keys play a repeating syncopated cell, not just block chords.
- Write bass as a hook; it should have gaps and answers.
- Use shakers/percussion as forward motion rather than constant high-end noise.
- Keep the lead vocal rhythm conversational and lightly behind or across the beat.
- Build choruses through vocal layering and percussion density, not only by adding chords.

## Afrobeats harmony

Harmony is often loop-based and pop/R&B compatible.

| Mood | Progression | Example |
|---|---|---|
| Warm major | I-V-vi-IV | C-G-Am-F |
| Minor-first pop | vi-IV-I-V | Am-F-C-G |
| Afro-R&B color | i-iv-♭VII-♭III | Am-Dm-G-C |
| Bright circular | I-ii-vi-IV | C-Dm-Am-F |
| Romantic lift | IV-V-vi | F-G-Am |
| Modal vamp | i-♭VII or i-IV | Am-G / Am-D |

Color choices:

- add9, 6, sus2, sus4, maj7, m7 can modernize simple loops
- guitar voicing can imply color without thick keyboard chords
- avoid overcrowding the middle register if vocal rhythm is dense

## Afrobeats melody and vocals

Useful moves:

- title phrase as chant + response
- short repeated cells with ad-lib gaps
- relaxed entry before or after beat 1
- mix sung hook with rap/speech-like verse
- use layered backing vocals for lift rather than only belting

Language matters. Nigerian Pidgin, Yoruba, Twi, English, French, Portuguese, and other languages have different prosody and cultural contexts. Do not fake language-specific idiom.

## Amapiano composition priorities

Amapiano often feels spacious, hypnotic, and dance-floor oriented. The log drum/bassline is a compositional voice, not just a sound-design layer.

**Core priorities**:

- deep, syncopated log-drum bassline
- shakers and percussion creating rolling motion
- piano/keys riffs or jazzy chord stabs
- spacious arrangement with long groove patience
- call-response vocals or chant hooks
- gradual layer changes rather than constant harmonic development

## Amapiano groove and bass

Writing approach:

```text
Start with shaker/percussion motion, then write the log drum as a melodic-rhythmic hook.
```

Practical moves:

- Make the log drum answer the vocal or keys; do not fill every beat.
- Keep kick/sub/log-drum roles distinct.
- Use 2-bar or 4-bar bass patterns that evolve through small variations.
- Let keys provide bright syncopation above the low-end space.
- Use drops and mutes to make the return of the log drum feel huge.

## Amapiano harmony

Harmony can be static, jazzy, or pop-like. The groove usually carries the section identity.

| Mood | Progression | Notes |
|---|---|---|
| Spacious minor | i-♭VII-♭VI-♭VII | loopable, dark but open |
| Warm major | I-vi-IV-V | pop-friendly |
| Jazzy loop | i9-iv9-♭VII13 | R&B/jazz influence |
| Static vamp | i or i(add9) | lets log drum and vocals lead |
| Bright piano color | Imaj7-II or I-II | Lydian/major-second brightness |

## Form

Afrobeats pop form:

```text
Intro tag → Verse → Pre/lift → Chorus → Post-hook → Verse/feature → Chorus → Outro
```

Amapiano club/pop form:

```text
Intro groove → Vocal hook → Groove expansion → Log-drum drop → Breakdown → Return → Outro
```

For pop crossover, shorten intros and make the vocal hook arrive earlier. For club contexts, longer groove patience can be a feature.

## Arrangement checklist

| Layer | Afrobeats role | Amapiano role |
|---|---|---|
| Drums/percussion | syncopated pocket, shakers, hand percussion | rolling shakers, hats, percussion bed |
| Bass | warm melodic bass, kick relationship | log-drum bassline as hook |
| Keys/guitar | ostinato, chord color, call-response | piano stabs, jazzy riffs, pads |
| Vocal | hook + ad-libs + conversational rhythm | chant/call-response or pop vocal on top |
| Ear candy | fills at phrase ends | mutes, drops, risers, log-drum variations |

## Quick recipes

### Afrobeats-pop hook

```text
Tempo feel: relaxed mid-tempo 4/4
Loop: | Am7 | Dm7 | G | Cmaj7 |
Guitar/keys: syncopated 2-bar ostinato
Bass: answer the vocal in bars 2 and 4
Topline: short title phrase + ad-lib gap
```

### Amapiano crossover drop

```text
Start: shaker + soft keys + vocal tag
Build: add pad and clap response
Drop: mute vocal for 1 beat, then log drum enters with 2-bar pattern
Harmony: one minor/add9 vamp or two-chord loop
```

### Popiano-style chorus

```text
Loop: | Fmaj7 | G | Am7 | Em7 |
Texture: pop vocal hook, amapiano percussion, lighter log drum, bright keys
Keep: chorus concise; don't let club arrangement swallow the topline
```

## Common mistakes

- Confusing Afrobeat and Afrobeats
- Treating one Nigerian or South African hit as representative of a continent
- Using log drum as a random sub fill instead of a composed bassline
- Overloading chords and losing groove space
- Quantizing everything so hard that the pocket feels stiff
- Ignoring language and vocal delivery context
- Calling anything with shakers "Afrobeats"

## Cross-references

- Current trends and regional checks → `../research/web-music-trend-research.md`
- Reference-track method → `../research/reference-track-digging.md`
- Style/copyright guardrails → `../research/style-reference-and-copyright.md`
- Groove and pocket → `../rhythm-groove/groove-and-feel.md`
- Hip-hop/R&B crossover → `hip-hop-rnb.md`
- EDM / club structures → `electronic-edm.md`
- Prosody → `../fundamentals/prosody-and-language.md`
- Arrangement density → `../orchestration/arrangement-density.md`

# Pitch, Intervals, and Scales

The foundational vocabulary for everything else in this skill. This file covers the default pitch grid used by most modern Western notation and instruments (12-tone equal temperament), how intervals are named and measured, and the scales — major, minor, modal, pentatonic, blues, symmetric, folk-derived, and non-diatonic — that organize pitch into usable material.

## When to consult this file

- The user asks about specific scales, intervals, or pitch material
- Questions like "what scale fits over Cm7", "what's an augmented 4th", "what's the difference between harmonic and melodic minor"
- Building a chord, melody, or progression and needing to verify pitch content
- Cross-checking enharmonic spelling

For chord construction from these intervals, see `harmony/chord-construction.md`. For the modes as harmonic frameworks, see `harmony/modal-harmony.md` and the cheatsheet `assets/modes-cheatsheet.md`. For scale formulas in compact reference form, see `assets/intervals-and-scale-formulas.md`.

## The pitch system

### 12-tone equal temperament (12-TET)

Most modern Western notation, DAWs, fixed-pitch instruments, and pop/jazz/classical teaching default to 12 equal divisions of the octave. Each division is a *semitone* (half-step). Twelve semitones = one octave; an octave = a frequency ratio of 2:1.

In 12-TET, every semitone is the same size: 2^(1/12) ≈ 1.0595 frequency ratio. This is a compromise — it makes every key equally available at the cost of slightly tempering pure intervals. Just intonation, meantone, well temperaments, and other historical/alternative systems can make some intervals or keys sound smoother and more characterful; modulation may change color, require retuning, or become limited depending on the system.

For composition purposes: the agent should treat 12-TET as the default and assume all intervals are equal-tempered unless the user specifies otherwise. Microtonal music, just intonation, and historical temperaments are out of scope for most situations but worth flagging if relevant.

### Octave equivalence

Notes one or more octaves apart are functionally the same pitch *class*. C4 (middle C) and C5 (an octave higher) are both "C". When discussing pitch material abstractly (scales, chords), we use pitch class names (C, D♭, F♯, etc.); when register matters, we use scientific pitch notation (C4, C5).

### The chromatic alphabet

12 pitch classes, written with 7 letters (A through G) plus accidentals (♯ for sharp, ♭ for flat). Most pitch classes have two enharmonic spellings:

| Pitch class | Common spellings |
|-------------|------------------|
| 0 | C, B♯, D𝄫 |
| 1 | C♯, D♭ |
| 2 | D, C𝄪, E𝄫 |
| 3 | D♯, E♭ |
| 4 | E, F♭, D𝄪 |
| 5 | F, E♯ |
| 6 | F♯, G♭ |
| 7 | G, F𝄪, A𝄫 |
| 8 | G♯, A♭ |
| 9 | A, G𝄪, B𝄫 |
| 10 | A♯, B♭ |
| 11 | B, C♭, A𝄪 |

(The 0–11 numbering is pitch-class numbering used in set theory and post-tonal analysis. Outside that context, just use letter names.)

### Enharmonic spelling — when it matters

A C♯ and a D♭ sound the same in 12-TET, but they imply different things. Spelling matters for:

- **Key context**: in D major, F♯ is correct; in C minor, G♭ would be wrong even though F♯ and G♭ are the same pitch.
- **Interval identity**: C–E♭ is a minor third; C–D♯ is an augmented second. They sound the same but function differently.
- **Voice leading**: a chromatic scale ascending should spell ♯ (C, C♯, D, D♯, E...) and descending should spell ♭ (C, B, B♭, A, A♭, G...).
- **Chord construction**: stack thirds for chords. So Cmaj7 is C–E–G–B, not C–E–G–C♭. The notes must be on every other letter.

When advising, use the spelling that matches the harmonic context. If the user is in E♭ major, write A♭ not G♯. If in B major, write D♯ not E♭.

## Intervals

An interval is the distance between two pitches. Measured by:

- **Size** (number): the number of letter names spanned. C–E is a third (C-D-E = 3 letters). C–F♯ is a fourth (C-D-E-F = 4 letters), even though it has a sharp.
- **Quality**: the precise distance in semitones, which determines whether the interval is major, minor, perfect, augmented, or diminished.

### Interval qualities — the table

For each size, there are possible qualities. The "perfect" qualities (P1, P4, P5, P8) are historically treated as a special consonance class in Western theory. Major/minor distinctions exist for the other sizes.

| Interval | Semitones | Examples (from C) | Notes |
|----------|-----------|-------------------|-------|
| P1 (perfect unison) | 0 | C–C | Same pitch |
| m2 (minor 2nd) | 1 | C–D♭ (or descending C–B) | Half-step |
| M2 (major 2nd) | 2 | C–D | Whole-step |
| m3 (minor 3rd) | 3 | C–E♭ | Defines minor chord |
| M3 (major 3rd) | 4 | C–E | Defines major chord |
| P4 (perfect 4th) | 5 | C–F | Open consonance |
| A4/d5 (augmented 4th / diminished 5th) | 6 | C–F♯ / C–G♭ | The tritone |
| P5 (perfect 5th) | 7 | C–G | Strongest consonance after octave |
| m6 (minor 6th) | 8 | C–A♭ | Inverts to M3 |
| M6 (major 6th) | 9 | C–A | Inverts to m3 |
| m7 (minor 7th) | 10 | C–B♭ | Inverts to M2 |
| M7 (major 7th) | 11 | C–B | Inverts to m2 |
| P8 (perfect octave) | 12 | C–C (next octave) | Octave equivalence |

Beyond an octave: compound intervals (9th = octave + 2nd, 11th = octave + 4th, 13th = octave + 6th). These are used in extended chord names.

### Augmented and diminished intervals

Any interval can be augmented (one semitone larger than its perfect or major form) or diminished (one semitone smaller than its perfect or minor form).

| Base interval | Augmented | Diminished |
|---------------|-----------|------------|
| P5 | A5 (8 semitones) — same sound as m6 | d5 (6 semitones) — the tritone |
| P4 | A4 (6 semitones) — the tritone | d4 (4 semitones) — same sound as M3 |
| M3 | A3 (5 semitones) — same sound as P4 | d3 (2 semitones) — same sound as M2 |
| m3 | (becomes M3) | d3 (same as above, from the minor side) |

The same pitch distance can have multiple names depending on the spelling. Six semitones can be A4 (e.g., C–F♯) or d5 (C–G♭) — the tritone. The two spellings imply different resolutions.

### Inversion of intervals

Invert an interval by flipping which note is on top. C–E (M3) inverts to E–C (m6). The rules:

- The two interval sizes always sum to 9 (the inverted M3 is a 6th, not a 5th — because we're now spanning E to the next C).
- Major inverts to minor (M3 → m6), minor to major (m6 → M3).
- Perfect inverts to perfect (P5 → P4).
- Augmented inverts to diminished (A4 → d5), diminished to augmented.

Useful for: voicing decisions (close vs. open), counterpoint, chord inversions.

### The tritone — the special interval

The augmented 4th / diminished 5th is exactly six semitones — the half-octave, the only interval that is its own inversion in 12-TET. It is often nicknamed *diabolus in musica* in later teaching lore, but avoid overstating this as a simple medieval ban; the practical point is its strong instability in tonal contexts.

Two tritones make an octave. The tritone is the most dissonant interval in the diatonic system and the engine of dominant-7th chord function — `V7` contains a tritone between the 3rd and ♭7th, which resolves inward (or outward) by half-step to the 3rd and root of `I`.

### Consonance / dissonance taxonomy

Western tradition broadly classifies intervals by stability:

| Class | Intervals | Use |
|-------|-----------|-----|
| Perfect consonances | P1, P5, P8; P4 by context | Stable goal intervals, but the 4th above the bass is treated as a dissonance in strict common-practice harmony |
| Imperfect consonances | M3, m3, M6, m6 | Stable in tonal context, expressive in melodic |
| Mild dissonances | M2, m7 | Resolution preferred but not always required |
| Sharp dissonances | m2, M7 | Demand resolution; create strong tension |
| Tritone (A4/d5) | The defining dissonance | Resolves outward (d5) or inward (A4) by half-step |

These are stylistic generalizations. In jazz, M7 over a tonic chord (Cmaj7) is consonant, not dissonant. In medieval music, the M3 was treated as imperfect. Context decides.

## Scales

A scale is an ordered set of pitches within an octave, defining the pitch material of a piece or section. Most scales are described by their interval pattern from the tonic.

### The major scale (Ionian mode)

Seven notes, formula `W W H W W W H` (whole-step, whole-step, half, whole, whole, whole, half).

C major: C D E F G A B
Scale degrees: 1 2 3 4 5 6 7

The major scale is the reference for everything else. When describing other scales as "major with X altered", the major scale is the baseline.

### The three minors

| Form | Formula | C example | Use |
|------|---------|-----------|-----|
| Natural minor | `W H W W H W W` (Aeolian mode) | C D E♭ F G A♭ B♭ | Modal/folk minor; no leading tone |
| Harmonic minor | `W H W W H Aug2 H` (raised ^7) | C D E♭ F G A♭ B | Functional minor; gives V chord a leading tone |
| Melodic minor | Ascending: `W H W W W W H` (raised ^6 and ^7); Descending: same as natural minor | Asc: C D E♭ F G A B; Desc: C B♭ A♭ G F E♭ D C | Smoother voice leading in lines |

In modern jazz, "melodic minor" usually refers to the ascending form used in both directions (called "jazz minor"). The classical convention (different ascending and descending) has fallen out of favor in jazz pedagogy.

### The seven diatonic modes

All built from the same 7-note set as the major scale, but starting on different degrees. Detailed discussion in `harmony/modal-harmony.md`; here's the formula reference:

| Mode | Formula from tonic | C version |
|------|--------------------|-----------|
| Ionian | `W W H W W W H` | C D E F G A B |
| Dorian | `W H W W W H W` | C D E♭ F G A B♭ |
| Phrygian | `H W W W H W W` | C D♭ E♭ F G A♭ B♭ |
| Lydian | `W W W H W W H` | C D E F♯ G A B |
| Mixolydian | `W W H W W H W` | C D E F G A B♭ |
| Aeolian | `W H W W H W W` | C D E♭ F G A♭ B♭ |
| Locrian | `H W W H W W W` | C D♭ E♭ F G♭ A♭ B♭ |

### Pentatonic scales

Five-note scales. The two most common are derived from the major and minor scales by removing semitones (^4 and ^7 in major; ^2 and ^6 in minor).

| Scale | Notes (C tonic) | Use |
|-------|-----------------|-----|
| Major pentatonic | C D E G A | Folk, country, gospel; safe over major chords |
| Minor pentatonic | C E♭ F G B♭ | Blues, rock; safe over minor chords |

Pentatonics are foundational because they avoid the half-step intervals that often create sharp clashes. In many pop, folk, rock, blues, and improvisational contexts, a pentatonic collection can sit over several related chords, but it still needs checking against the bass, chord quality, and melody target notes.

### The blues scale

Minor pentatonic + the ♭5 (the "blue note"). C blues scale: C E♭ F G♭ G B♭.

The ♭5 is a passing tone, not a stable pitch. The blues scale doesn't follow Western functional logic — it's a melodic palette superimposed on chords that may or may not contain its notes. Over a `C7` blues, the C blues scale clashes "correctly".

There are also major-blues variants (major pentatonic + ♭3 as a passing tone): C D E♭ E G A. Used in rockabilly, country, and major-mode blues.

### Modes of melodic minor (jazz minor)

A separate set of seven modes built from the ascending melodic minor scale. Each has a distinct character and is used over specific jazz chord types.

| Mode | Built on | Formula | Use over |
|------|----------|---------|----------|
| Melodic minor | ^1 | `W H W W W W H` | Cm(maj7) chords; Im(Δ) tonic in jazz minor |
| Dorian ♭2 | ^2 | `H W W W W H W` | iim7♭5 in some contexts |
| Lydian augmented | ^3 | `W W W W H W H` | Maj7♯5 chords; ambiguous, bright-tension color |
| Lydian dominant | ^4 | `W W W H W H W` | 7♯11 chords (non-resolving dominants) |
| Mixolydian ♭6 | ^5 | `W W H W H W W` | dom7 chords with ♭13 |
| Locrian ♮2 | ^6 | `W H W H W W W` | iiø7 chords |
| Altered (super Locrian) | ^7 | `H W H W W W W` | 7alt chords (resolving dominants with all alterations) |

The altered scale is the most-used by far in modern jazz — over `V7alt` resolving to `i`.

### Modes of harmonic minor

Less commonly used as full mode systems, but two are common as melodic palettes:

| Mode | Built on | Use |
|------|----------|-----|
| Harmonic minor (the parent) | ^1 | Tonic minor in classical and some jazz |
| **Phrygian dominant** | ^5 | Over V7♭9 chords resolving to minor; flamenco, Klezmer, metal |

The Phrygian dominant scale (C version: C D♭ E F G A♭ B♭) has a strongly marked augmented 2nd between ♭2 and 3 (D♭ to E here). It can suggest flamenco-adjacent, klezmer-adjacent, metal, or maqam-adjacent colors in Western equal temperament, but it is not a substitute for those traditions' full melodic, rhythmic, and intonational practices.

### Symmetric scales

Scales whose interval pattern repeats within the octave. They have no clear tonic — every note can be heard as the root.

| Scale | Formula | Notes (C version) | Use |
|-------|---------|-------------------|-----|
| Whole-tone | `W W W W W W` (6 notes) | C D E F♯ G♯ B♭ | Over 7♯5 / whole-tone dominant colors; B♭ is the dominant spelling of the enharmonic A♯ |
| Diminished (whole-half) | `W H W H W H W H` (8 notes) | C D E♭ F G♭ A♭ A B | Over diminished chords |
| Diminished (half-whole) | `H W H W H W H W` (8 notes) | C D♭ D♯ E F♯ G A B♭ | Over dominant ♭9 chords; D♯ is ♯9 over C7 |
| Chromatic | All 12 semitones | C C♯ D D♯ E F F♯ G G♯ A A♯ B | Passing material; serialism |
| Augmented | `A2 H A2 H A2 H` (6 notes) | C D♯ E G A♭ B | Coltrane changes; symmetric augmented sound (A2 = augmented 2nd, enharmonic to m3) |

### Non-diatonic and folk-derived scales (selected)

| Scale | Formula | Notes (C version) | Origin/use |
|-------|---------|-------------------|------------|
| Hungarian minor | `W H Aug2 H H Aug2 H` | C D E♭ F♯ G A♭ B | Hungarian folk, Romani |
| Byzantine / double harmonic | `H Aug2 H W H Aug2 H` | C D♭ E F G A♭ B | Greek, Middle Eastern |
| Bebop major | Major + ♯5 (or ♭6) added as passing | C D E F G A♭ A B | Bebop solo language |
| Bebop dominant | Mixolydian + natural 7 added | C D E F G A B♭ B | Bebop over dom7 |
| Bebop Dorian | Dorian + natural 3 added | C D E♭ E F G A B♭ | Bebop over m7 |
| Japanese (Hirajōshi) | `W H M3 H M3` | C D E♭ G A♭ | Pentatonic, Japanese traditional-adjacent color |

### Non-Western traditions — brief pointer

Many non-12-TET or tradition-specific pitch systems do not fit the default Western equal-tempered grid cleanly:

- **Indian classical** uses ragas — melodic frameworks that can include ascent/descent behavior, important tones, characteristic phrases, ornamentation rules, and intonational nuance.
- **Arabic / Turkish music** uses *maqam* / *makam* systems with characteristic pitch intonation, melodic pathways, ornaments, and modulations. Some tones fall between 12-TET semitones, but these traditions should not be reduced to a simple equal-quarter-tone grid.
- **Indonesian gamelan** uses 5-tone (slendro) and 7-tone (pelog) scales unrelated to 12-TET.
- **Sub-Saharan African** music uses scales that vary by tradition; some are pentatonic or hexatonic with regional inflections.

This skill is Western-music-focused. For non-Western collaboration, flag the limitation and recommend the user consult specialist resources or work with practitioners.

## Chord-scale defaults — a short reference

When improvising or composing a melodic line over a chord, these are the default scales to start from. (Full treatment in `harmony/jazz-harmony.md`.)

| Chord | Default scale |
|-------|---------------|
| Imaj7 (tonic in major) | Ionian or Lydian |
| Imaj7♯11 | Lydian |
| ii7 | Dorian |
| iiø7 (half-dim) | Locrian or Locrian ♮2 |
| iii7 | Phrygian or Aeolian |
| IV (major) | Lydian |
| IV7 (blues IV) | Mixolydian or Lydian dominant |
| V7 (resolving) | Mixolydian; for tension: altered, half-whole diminished, Phrygian dominant |
| V7♯11 (non-resolving) | Lydian dominant |
| vi7 | Aeolian or Dorian |
| vii°7 | half-whole diminished |
| Im(maj7) | Melodic minor |
| Im7 (modal) | Dorian |
| Imadd2 | Dorian or Aeolian |

## Common pitfalls when advising

- **Don't conflate enharmonic equivalents.** C♯ and D♭ sound the same but aren't always interchangeable; the spelling tells the player and reader what function the note plays.
- **Don't frame scales or traditions as novelty colors.** Whole-tone, octatonic, and various non-Western systems aren't oddities; they're legitimate organizing systems with their own logic.
- **Don't over-prescribe a scale per chord.** Improvisation often spans chord-scale relationships, and great solos violate "the right scale" constantly. Use chord-scale theory as a starting reference, not a cage.
- **Don't ignore ear training.** Telling a user "the second mode of melodic minor" is useless if they don't recognize the sound. Pair theory with audible examples ("it sounds like X piece").

## Cross-references

- Chord construction from interval stacking → `../harmony/chord-construction.md`
- The modes as harmonic systems (not just scales) → `../harmony/modal-harmony.md`
- Quick scale formula reference → `../assets/intervals-and-scale-formulas.md`
- Modes cheatsheet (formulas + characteristic chords) → `../assets/modes-cheatsheet.md`
- Chord-scale theory in jazz → `../harmony/jazz-harmony.md`
- Scales used in 20th-century non-tonal music → `../techniques/20th-century-techniques.md`
- Microtonal pitch systems beyond 12-TET (just intonation, EDOs, *maqam*, *raga*) → `../techniques/microtonal.md`
- Notation of accidentals, key signatures → `notation-and-conventions.md`

# Narrative and Transitions

This file covers the *macro-level shape* of a piece — how energy rises and falls across its length, how sections transition into each other, how the listener's attention is led through the form. It's distinct from the file on classical forms (which covers structural templates) and the file on popular song forms (which covers section types). This file covers the *flow* between sections, regardless of the formal template.

For section-level form templates (sonata, AABA, verse-chorus, EDM), see `classical-forms.md` and `popular-song-forms.md`. For the energy and dynamics specifically as they relate to production, see `production-aware/energy-and-dynamics.md`. For modulation as a transition technique, see `harmony/modulation.md`.

## When to consult this file

- The user describes a piece that "feels flat" or "doesn't build"
- A transition between sections feels weak, abrupt, or jarring
- "How do I create a sense of arc/journey/build" in a longer piece
- Build-ups and drops in EDM
- Cinematic builds in film/trailer music
- The chorus doesn't "hit" — often a transition issue, not a chorus issue
- Pacing across an album, a movement, or a song

## What "narrative" means in non-narrative music

Even instrumental music with no story has narrative shape. The listener perceives:

- **Tension and release**
- **Building and resolving**
- **Arrival points**
- **Departure and return**
- **Climax and denouement**

These come from the *contour* of the music's energy across time. The same chord, played twice, can feel like "more" the second time if the surrounding context has built up to it.

A piece succeeds narratively when its energy contour creates a satisfying journey. It fails when the energy is flat (no arc), erratic (too many peaks), or wrongly placed (climax in the wrong spot).

## The dimensions of energy

Energy in music is a composite of multiple parameters. Adjusting any one (or several) raises or lowers the felt energy:

| Parameter | Lower energy | Higher energy |
|-----------|--------------|---------------|
| Dynamics | Softer | Louder |
| Density | Fewer instruments / sparser | More instruments / denser |
| Register | Lower / narrower | Higher / wider |
| Rhythm | Slower / sparser subdivisions | Faster / denser subdivisions |
| Harmony | Simpler / more diatonic | More chromatic / extended |
| Melodic activity | Sustained notes / few notes | Active / busy lines |
| Tempo | Slower (or staying constant) | Faster (or accelerating) |
| Timbre | Mellow / dark | Bright / harsh / saturated |
| Tessitura | Low or middle | High |
| Dissonance | Consonant | Dissonant or tense |
| Vocal intensity | Speaking, low chest, restrained | Belting, head voice, ad-libs |

The art of arrangement is moving these parameters in concert across the song.

## Energy arcs — the macro shape

Most successful pieces have a clear macro shape. A few archetypes:

### Single arch

Energy rises gradually, peaks once, and falls. The simplest narrative shape.

```
       ___
      /   \
     /     \___
____/          \___
```

Common in: classical sonata movements, ballads, single-section EDM tracks, film cues.

### Multiple peaks (typical of song forms)

Energy rises and falls multiple times, with each peak (chorus) higher or equal to the previous, and the final peak the highest.

```
   __         __         ___
  /  \   __  /  \   __  /   \
 /    \_/  \/    \_/  \/     \___
/________________________________
```

Common in: pop songs (verse-chorus-verse-chorus-bridge-chorus), where each chorus is a peak.

### Build-and-drop (EDM)

Sustained build of tension, sudden release into the drop, sustained at high energy, breakdown, second build, second drop.

```
     /|        /|
    / |   __  / |
   /  |  /  \/  |   __
__/   |_/       |  /  \___
              breakdown
```

Common in: EDM, dubstep, future bass, trap.

### Through-composed arc

Continuous gradual rise, often over many minutes, ending with a sustained climax. No breakdown or release until the end.

```
                 _____
              __/
           __/
        __/
     __/
__/
```

Common in: trailer music, late Romantic music, some film scoring, post-rock.

### Plateau-step (ascending)

Each section is more energetic than the previous, but within a section the energy is roughly stable. The piece climbs in steps.

```
              ________
         ____/
    ____/
___/
```

Common in: K-pop ballads, some musical theatre songs, some EDM.

### Frame (high-low-high)

Opens at high energy, drops to low for a contemplative middle, returns to high. Like ABA where B is the energy contrast.

```
___       ___
   \     /
    \___/
```

Common in: progressive rock, longer film cues, some classical pieces (e.g., Beethoven's "Pathétique" first movement has elements of this).

### Inverse (low-high-low)

Opens quietly, builds, then dissolves at the end. Often used for emotional or reflective pieces.

```
        ___
      _/   \_
   __/       \__
__/             \___
```

Common in: ambient, meditation music, many art songs, "fade-out" rock songs.

## Transitions between sections

The transition from one section to the next is where the energy contour is most felt. Three primary modes:

### 1. Sudden / abrupt transition

The new section starts immediately, with no preparation. Maximum contrast.

When to use:

- For shock or surprise (cut to a quiet section after a loud one)
- For energetic launches (drop a chorus immediately after a soft verse)
- In genres where contrast is the aesthetic (some prog, some film cues)

Risk: feels jarring if not earned by content.

### 2. Build / ramp transition

The current section's energy rises to a peak just before the new section begins. The new section's arrival is the release of that built tension.

Standard build techniques:

- **Adding instruments / layers** through the build
- **Rising rhythmic density** (eighth notes → sixteenths → thirty-seconds)
- **Rising pitch / contour** (vocal line ascends; bass climbs)
- **Crescendo** in dynamics
- **Pedal point on the dominant** of the next section's key
- **Sustained pitch** that rises in pitch (a "riser" in EDM production)
- **Drum fills** filling in the space before the new section
- **Chord change** that's withheld until the moment of arrival

The build delivers the listener to the transition exhausted with anticipation; the next section is the relief.

### 3. Smooth / blended transition

The two sections overlap or smoothly merge, with no clear boundary. Common in through-composed and ambient music.

Methods:

- **Overlap**: the new section begins while the old one is still finishing
- **Common pivot material**: a held note or chord that belongs to both sections
- **Gradual parameter change**: instrumentation slowly morphs over many bars
- **Cross-fade**: the old section fades while the new one fades in (more a mixing technique but also a compositional choice)

When to use: when the music wants flow and continuity over articulation. Ambient, post-rock, some film scoring.

## The pre-chorus as transition machinery

In modern pop, the pre-chorus is often *purely* a transition mechanism. It's not a section that exists for its own content; it exists to deliver the listener to the chorus.

Effective pre-choruses:

- **Lift the tessitura** — vocal line rises higher than the verse
- **Increase rhythmic density** — more syllables, more notes
- **Modulate harmonically** — often a series of chords leading to V of the chorus key
- **Build production** — additional layers come in, drum fill happens, etc.
- **Climb in steps** — short repeated phrases each on a higher pitch
- **End on V** of the chorus's key, so the chorus's I is the resolution

A weak pre-chorus is one that maintains the verse's energy without building. The chorus arrives but doesn't feel earned.

## Build-ups and drops in EDM

EDM's defining structural feature: the build-up before a drop, and the drop itself.

### Build-up structure

A typical 16-bar build:

- **Bars 1–4**: introduce a melodic or rhythmic element; energy mid-low.
- **Bars 5–8**: thicken; add a snare roll, riser synth, or vocal hook.
- **Bars 9–12**: density increases; rolls accelerate (eighth → sixteenth → thirty-second); dynamic rises.
- **Bars 13–14**: maximum tension; full snare roll, white noise wash, all elements peaking.
- **Bar 15**: silence or near-silence (the "breath before the drop"); often the only quiet bar in the entire build.
- **Bar 16**: a single sound (a vocal "drop" cue, a single percussion hit, or just silence).
- **Drop**: full instrumentation enters on bar 17.

The bar-15 silence is critical — it's the moment of maximum listener anticipation. Skipping it makes the drop feel less impactful.

### Drop structure

The drop is typically:

- **Maximum energy** for the first 8 bars
- **Slight de-escalation** in the next 8 bars (a hook variation, or sparser instrumentation)
- **Re-intensification** for the final 8–16 bars

A 32-bar drop has internal contour; it's not flat-out energy for 32 bars.

### Breakdown

After the first drop, energy typically lowers for 16–32 bars (the breakdown). This is where the song's vocal or melodic hook gets featured prominently, often in a stripped-down arrangement.

The breakdown provides emotional contrast and gives the listener a "rest" before the next build.

## The cinematic build (trailer music)

Trailer music has its own narrative grammar: a continuously building arc over 60–120 seconds, with multiple internal peaks.

Typical structure:

- **0:00–0:15**: cold open. Quiet, atmospheric. Single instrument or pad.
- **0:15–0:30**: slow build. Adding strings, sub bass, percussion accents.
- **0:30–0:45**: medium build. Full orchestra entering, percussion getting denser.
- **0:45–1:00**: first peak (often a "drop" with full orchestra and Big Drum™).
- **1:00–1:15**: brief breakdown or sustained suspended chord.
- **1:15–1:30**: massive final peak. Full orchestra, choir, max BPM percussion.
- **1:30–1:45**: cool-down (if length allows).

The grammar is borrowed from late Romantic music and EDM, fused into a 60–120 second package.

## Sectional cadences as narrative articulation

Cadences (see `harmony/functional-harmony.md`, `assets/cadence-reference.md`) function as narrative punctuation. The cadence at a section boundary tells the listener what kind of moment this is:

| Cadence type | Narrative role |
|--------------|----------------|
| PAC | Definitive arrival, end of major section, "period" |
| IAC | Soft arrival, "comma" |
| HC | Question; "to be continued" |
| Deceptive cadence | Surprise; delay; "we're not done yet" |
| Plagal | Soft amen ending |
| Modal cadence | Suspended, non-classical |
| No cadence (elision) | Continuous flow; no break |

A song that ends every section with PAC feels over-articulated. One that uses HC and deceptive at internal boundaries (saving PAC for the end) feels propulsive.

## Long-form pacing (album, suite, multi-movement work)

For pieces longer than a single song, narrative shape operates at the album/suite level:

### Album pacing

- **Opening track**: establishes sonic identity. Usually energetic or distinctive. Hooks the listener.
- **Track 2–3**: maintains energy, develops the album's themes. Often the radio singles.
- **Mid-album**: variety. May include slower or more experimental tracks. The "album track" zone.
- **Late mid-album**: another peak. Often an emotional centerpiece.
- **Closing track**: resolution or contemplation. Often slower or differently energetic from the opening.

This pattern (especially in pop, rock, and hip-hop albums) creates an album-as-arc that mimics song-as-arc at larger scale.

### Multi-movement classical work

The standard sonata cycle (4 movements: fast-slow-dance-fast) creates an arc:

- **Movement 1** (fast, sonata form): the dramatic statement
- **Movement 2** (slow, lyrical): the contemplation
- **Movement 3** (dance, light): the relief
- **Movement 4** (fast, finale): the triumph or affirmation

Beethoven's 5th and 9th end with overwhelming finales that recapitulate the work's emotional journey.

## Common transition mistakes

- **Abrupt section change without preparation.** The chorus arrives but the verse hasn't built up to it. Result: chorus feels arbitrary.
- **Over-prepared transitions.** Every section is preceded by a 16-bar build. Result: the listener becomes desensitized to builds.
- **Multiple peaks too close together.** Three back-to-back climaxes leave nowhere to go. Each subsequent peak is less impactful.
- **No release after a peak.** The piece sustains maximum energy for too long. The listener fatigues.
- **Climax in the wrong place.** The biggest moment happens at bar 30 of a 4-minute song. The remaining 3 minutes have nothing to look forward to.
- **Energy mismatch within a section.** The chorus has loud drums but a thin vocal; the listener can't tell if the energy is high or low. Mixed signals.

## Common pitfalls when advising

- **Don't focus only on harmony or melody when advising on "energy".** Energy is multi-parameter. A piece can fail to build even with adventurous harmony if the dynamics, density, and rhythm don't change.
- **Don't apply EDM build-and-drop logic to non-EDM genres.** A folk song doesn't want a drop. A classical piece doesn't want a riser. Match transition style to genre.
- **Don't overuse the "lift" tools.** Modulation up a half-step, doubled chorus, key change at the bridge — these are tools, not requirements. A song that uses all of them at once feels desperate.
- **Don't ignore the listener's ear fatigue.** Most listeners can't sustain attention through more than 3–4 peaks per song. Save the biggest moment for the end (or the strongest one for the moment that matters most narratively).
- **Don't confuse intensity with quality.** A song that's loud and dense throughout isn't more impactful than one that uses dynamics. Quiet moments make loud moments matter.

## Quick decision matrix

| Want | Try |
|------|-----|
| Build a chorus that "hits" | Strong pre-chorus with rising tessitura, density, harmonic motion to V of chorus key |
| Smooth section transition | Common pivot chord, gradual layer change, no sudden break |
| Cinematic peak | Crescendo through 4–8 bars, drum fill at end, full orchestra entry on next downbeat |
| EDM drop | 16-bar build with rolls + riser + bar-15 silence + full kick on bar 17 |
| Final-chorus elevation | Modulate up; add backing vocals; double the chorus length; bring in motifs from earlier |
| Overall arc shape | Single arch (gradual build to climax + release); choose climax position around 70-80% of total length |
| Pop song pacing | Verse-chorus-verse-chorus-bridge-chorus, with each chorus higher energy than verses, final chorus highest |
| Album pacing | Energetic opener; emotional centerpiece mid-album; reflective closer |
| Multiple-peak song | Each peak slightly higher than the last; final peak highest; release after each |
| Through-composed flow | No section breaks; gradual parameter shifts; sustained build to single climax |

## Cross-references

- Classical forms (templates within which narrative happens) → `classical-forms.md`
- Popular song forms (sectional structures of pop, EDM, K-pop) → `popular-song-forms.md`
- Cadences (the harmonic articulation of transitions) → `harmony/functional-harmony.md`, `assets/cadence-reference.md`
- Modulation as a transition technique → `harmony/modulation.md`
- Phrase rhythm and pacing within sections → `melody/phrase-structure.md`
- Production-side energy decisions → `production-aware/energy-and-dynamics.md`
- Arrangement density across sections → `orchestration/arrangement-density.md`
- Genre-specific narrative conventions → `genres/film-tv-scoring.md`, `genres/electronic-edm.md`, `genres/kpop-jpop.md`, `genres/musical-theatre.md`

# Brazilian Pop and Funk

> **Snapshot note:** Last content review snapshot: 2026-04-27. Use this file for durable composition conventions, but verify current Brazilian pop, funk carioca, funk paulista, sertanejo, pagode, piseiro, trap, and regional scenes with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition-facing conventions in Brazilian pop, MPB-adjacent writing, samba/bossa-informed pop, pagode-pop, sertanejo-pop crossover, Brazilian funk / funk carioca, funk melody, and contemporary Brazilian crossover. It is **not** a complete guide to Brazilian music, samba schools, capoeira traditions, choro, forró, maracatu, axé, frevo, or regional folk styles.

## When to consult this file

- The user wants Brazilian pop, MPB, samba-pop, bossa-informed harmony, pagode-pop, sertanejo-pop, funk carioca, baile funk, or Brazilian crossover energy
- The user asks for Portuguese-language pop prosody
- The user says "Latin" but means Brazil, or asks to avoid Spanish-language Latin-pop assumptions
- The user wants rhythmic identity from guitar/percussion rather than only chord changes
- The user wants modern funk energy without confusing Brazilian funk with U.S. funk

## Core warning

Brazilian music is not Spanish-language Latin pop. Portuguese prosody, Brazilian rhythmic traditions, regional scenes, and local industry conventions matter.

Also: **Brazilian funk / funk carioca is not the same thing as U.S. funk.** It is an electronic dance/rap ecosystem with its own rhythms, production practices, MC culture, regional substyles, and social context.

## Substyle map

| Frame | Composition priorities | Common colors |
|---|---|---|
| MPB-adjacent pop | lyric nuance, sophisticated harmony, flexible groove | nylon guitar, piano, strings, subtle percussion |
| Bossa-informed pop | soft syncopation, intimate melody, extended harmony | nylon guitar, maj7/9 chords, gentle vocal |
| Samba-pop / pagode-pop | syncopated cavaquinho/guitar, percussion lift, communal chorus | pandeiro, tantã, cavaquinho, handclaps |
| Sertanejo-pop | big chorus, duet-friendly melody, country/pop crossover | acoustic guitars, accordion/piano, pop drums |
| Funk carioca / baile funk | MC rhythm, percussion loop, 808/sub, call-response | tamborzão/beatbox-like patterns, chants, sparse harmony |
| Funk melody | romantic vocal line over funk-derived groove | sung hooks, smoother harmony |
| Brazilian trap/pop-funk | trap drums + funk percussion + pop topline | 808, hats, Portuguese rap/sing-rap |
| Axé/carnival pop-adjacent | crowd chorus, percussion drive, festive call-response | surdo-like pulse, brass/percussion energy |

## Brazilian Portuguese prosody

Portuguese hooks often benefit from vowel clarity, nasal vowels, and natural stress placement. Do not write as if it were Spanish.

Practical guidance:

- Place title words on strong groove points.
- Avoid overstuffing fast consonants if the hook needs crowd participation.
- Let open vowels carry long notes.
- Check word stress: Portuguese stress placement is not always obvious from spelling.
- For serious lyrics, work with a native Portuguese speaker or lyricist.

Pair with `../fundamentals/prosody-and-language.md`.

## Harmony

Brazilian pop can range from three-chord directness to highly sophisticated harmony. Match the substyle.

### MPB / bossa-informed colors

| Device | Effect |
|---|---|
| maj7, 6/9, add9 | soft color without hard resolution |
| ii-V motion | jazz-informed flow |
| chromatic bass descent | lyrical sophistication |
| secondary dominants | directed warmth |
| borrowed iv / ♭VI | bittersweet turn |
| slash chords | smooth voice-led bass |

Example:

```text
Key: C
| Cmaj9 | Bm7♭5 E7♭9 | Am7 | Gm7 C7 |
| Fmaj7 | Fm6 | Em7 A7 | Dm7 G7 |
```

Use when the user wants MPB/bossa-informed sophistication. Do not force this into funk carioca unless the brief asks for a hybrid.

### Samba-pop / pagode-pop loops

```text
| I | V/vi | vi | V |
| IV | I/3 | ii V | I |
```

or more simply:

```text
| I | V | vi | IV |
```

The groove, percussion, and cavaquinho/guitar pattern often define the style more than the chord list.

### Funk carioca / pop-funk harmony

Funk carioca can be harmonically sparse:

```text
1-chord vamp
2-chord loop
minor i-♭VII
minor i-♭VI-♭VII
```

The key elements are MC rhythm, hook chant, beat identity, bass, and production accents.

## Groove vocabulary

### Bossa-informed guitar feel

A composition-facing shorthand:

```text
Soft 2/4 or 4/4 syncopation, bass notes alternating with offbeat chord strokes.
```

Use nylon guitar voicings and extended chords. Keep the vocal intimate; do not turn every bossa-informed track into lounge pastiche.

### Samba / pagode-informed feel

Writing priorities:

- interlocking syncopation rather than one monolithic pattern
- short repeated cavaquinho/guitar figures
- percussion answers around the vocal
- chorus that can feel communal
- harmonic rhythm that leaves room for groove detail

### Funk carioca / baile funk feel

Writing priorities:

- percussion loop has the identity
- sub/kick relationship is central
- MC rhythm is composition, not decoration
- short chant hooks work well
- space is powerful; sparse tracks can hit harder
- call-response, ad-libs, and crowd phrases shape form

Avoid presenting any one funk pattern as the single authentic pattern. Substyles and regional practices change quickly.

## Melody and vocals

| Frame | Melody strategy |
|---|---|
| MPB | conversational, lyric-led, nuanced contour |
| Bossa | intimate, understated, syncopated against guitar |
| Samba/pagode-pop | chantable, communal, rhythm-forward |
| Sertanejo-pop | broad chorus, duet-friendly range, emotional clarity |
| Funk carioca | rhythmic MC delivery, short sung hook, call-response |
| Pop-funk crossover | concise topline, strong vowel hook, danceable phrase spacing |

## Form

Common forms:

```text
Intro groove → Verse / MC → Pre-lift → Chorus hook → Verse 2 → Chorus → Outro tag
```

```text
Intro tag → Chorus → Verse → Chorus → Feature / MC section → Chorus / chant outro
```

```text
Verse → Chorus → Instrumental/percussion break → Verse → Chorus → repeated tag
```

Many Brazilian pop/funk tracks foreground the groove or hook early. Long intros need a distinctive instrumental identity.

## Arrangement

### Acoustic / MPB / bossa-informed

- nylon guitar or piano as harmonic center
- subtle percussion rather than constant full kit
- bass with melodic passing tones
- strings or winds for emotional lift
- vocal forward and nuanced

### Samba / pagode-pop

- cavaquinho/guitar rhythmic cell
- percussion family rather than one loop
- group vocals or crowd response
- bass supports syncopation
- chorus can become communal

### Funk / pop-funk

- hard percussion loop
- sub/808 or synth bass
- sparse harmonic sample/loop
- MC lead plus ad-libs
- impact drops and stop-time cuts
- hook chant that survives repetition

## Hybrid guidance

### Brazilian funk + global pop

- Keep the funk rhythm identity central.
- Use pop pre-chorus/chorus architecture for international clarity.
- Let Portuguese hook words remain rhythmically strong.
- Avoid smoothing away the beat until it becomes generic dance-pop.

### Bossa/MPB + modern pop

- Use extended harmony and nylon guitar as color.
- Keep the chorus simpler than the verse if the user wants pop accessibility.
- Add modern low end carefully; too much sub can fight the guitar's intimacy.

### Samba/pagode + R&B

- Let percussion and cavaquinho/guitar imply the pocket.
- Use R&B chord color but keep rhythm uncluttered.
- Use vocal ad-libs as response, not constant decoration.

## Quick recipes

### Bossa-informed verse

```text
Key: F
| Fmaj7 | Fm6 | Em7 A7 | Dm7 G7 |
Feel: soft syncopated guitar, intimate vocal.
Effect: bittersweet chromatic color.
```

### Brazilian pop chorus

```text
Key: A minor / C major
| Am | F | C | G |
Vocal: short Portuguese title phrase repeated with rhythmic variation.
Arrangement: percussion adds on second half of chorus.
```

### Funk carioca hook seed

```text
Tempo: 130-150 BPM range, style-dependent
Harmony: 1-chord minor vamp
Beat: sparse percussion loop + sub
Vocal: 2-bar chant, call-response answer, ad-lib tag.
```

## Common pitfalls

- **Do not treat Brazilian music as Spanish Latin pop.** Portuguese and Brazilian rhythmic identity matter.
- **Do not confuse Brazilian funk with U.S. funk.** The name overlaps, but the composition language differs.
- **Do not reduce samba to one shaker loop.** Interlocking parts matter.
- **Do not make every Brazilian cue bossa nova.** Brazil has many popular music ecosystems.
- **Do not over-harmonize funk carioca unless the brief asks for a hybrid.** Sparse harmony can be idiomatic.
- **Do not use caricature percussion.** Name the rhythmic function and, where possible, the instrument.
- **Do not claim current substyle rules without checking recent sources.** Funk and pop scenes move quickly.

## Quick decision matrix

| User wants | Try |
|---|---|
| Brazilian pop sophistication | MPB/bossa-informed harmony, lyric-forward melody |
| More samba/pagode | interlocking percussion, cavaquinho/guitar syncopation, communal chorus |
| More funk carioca | sparse loop, strong MC rhythm, sub/percussion identity, chant hook |
| More international pop | clearer chorus form, simpler harmonic loop, polished vocal stack |
| More intimate | nylon guitar, softer percussion, extended chords, narrow vocal range |
| More party energy | earlier hook, call-response, percussion break, repeated tag |

## Cross-references

- Latin pop and reggaeton contrast → `latin-pop-and-reggaeton.md`
- C-pop / Southeast Asian pop comparison for regional pop scenes → `cpop-and-southeast-asian-pop.md`
- Folk/roots specificity → `folk-roots-and-traditions.md`
- Portuguese prosody → `../fundamentals/prosody-and-language.md`
- Guitar idioms → `../instrument-idiom/guitar.md`
- Bass idioms → `../instrument-idiom/bass.md`
- Drums/percussion idioms → `../instrument-idiom/drums-percussion.md`
- Jazz/extended harmony → `../harmony/jazz-harmony.md`, `../harmony/reharmonization.md`
- Groove and feel → `../rhythm-groove/groove-and-feel.md`, `../rhythm-groove/rhythmic-devices.md`
- Current Brazilian references → `../research/web-music-trend-research.md`, `../research/reference-track-digging.md`

# Popular Song Forms

This file covers the structural conventions of popular music: verse-chorus and its variants, AABA, blues forms, EDM forms, K-pop's distinctive layered forms, and more. These differ fundamentally from classical forms in their priorities — popular forms are organized around hooks and memorability rather than tonal architecture.

For classical forms (sonata, rondo, etc.), see `classical-forms.md`. For energy arcs and transitions across these forms, see `narrative-and-transitions.md`. For genre-specific form variations, see relevant `genres/*.md` files.

## When to consult this file

- The user asks about song structure: verse, chorus, bridge, pre-chorus, post-chorus, intro, outro
- AABA jazz standard form
- Blues form (12-bar, 16-bar, modal, etc.)
- EDM structures: drop, build, breakdown
- K-pop multi-section structures
- Comparing form choices across genres
- Why a song's structure "doesn't work" — usually a sectional pacing issue

## The components of popular form

Most popular forms are built from a small set of section types. The same section types combine differently across genres.

### Verse

The narrative or lyrical core. Each verse usually has *different lyrics* but the *same melody* (or close variant). Verses tell the story; choruses make the point.

Verses are typically:

- 8 or 16 bars (4 or 8 phrases of 4 bars)
- Lower in tessitura than the chorus
- More syllabically dense (more words per bar)
- Harmonically simpler or more static

### Chorus

The hook. Usually the same melody and lyrics each time it appears. The chorus is what listeners remember. The structural payoff of the song.

Choruses are typically:

- 8 or 16 bars
- Higher in tessitura than the verse
- More repetitive (often returns to the song's title hook)
- Higher dynamic, denser arrangement
- Often contain the song's title

### Pre-chorus (also called "rise" or "lift")

A short transitional section (4–8 bars) that bridges verse to chorus. Builds expectation and energy. Often modulates harmonically toward the chorus key, or shifts the rhythmic feel.

Modern pop relies heavily on pre-choruses to make choruses feel earned. Without a pre-chorus, the chorus arrival can feel sudden; with one, it feels prepared.

### Bridge (also called "middle 8" in older terminology)

A contrasting section that appears once or twice in the song, typically after the second chorus. Provides relief from verse-chorus repetition before the final chorus.

Bridges are typically:

- 8 bars
- Different melody, lyrics, often different chords (often modulates to a new key area)
- New emotional perspective on the song's theme
- Sometimes simpler arrangement to set up the final chorus

### Intro

A short instrumental opening (often 4 or 8 bars) that establishes the key, tempo, groove, and often the song's primary hook. Modern intros are often shortened (1–2 bars) for streaming-era attention spans.

### Outro

The closing section. Often a fade-out of the chorus or a tag (a short repeated phrase). Can be a contrasting section ("Hey Jude" — the famous "na-na-na" coda is an outro).

### Post-chorus

A relatively recent addition (2010s onward), especially in pop and K-pop. A short hook that follows the chorus and often becomes the song's most memorable part. Sometimes wordless ("oh oh oh"), sometimes a short refrain.

Examples: the "ooh-ee-oh" of Lady Gaga's "Bad Romance", the post-chorus drops in many K-pop songs.

### Drop (in EDM and EDM-influenced pop)

The instrumental climax of a song — typically a heavy, energetic section after a build-up. The drop is essentially the EDM equivalent of a chorus, but instrumental.

In a drop, the melody and rhythm intensify; the bass becomes heavy; production hits its peak. The drop is what listeners come for.

### Build (also "build-up" or "riser")

The section before a drop. Typically the music thins (just synth and sparse rhythm), then progressively adds elements (rolls, sweeps, white noise rises) building tension. The build resolves into the drop.

### Breakdown

A section after the drop where elements are stripped away. The energy drops and is rebuilt for the next section. Often the breakdown features a vocal or melodic element that wasn't prominent before.

## Standard popular forms

### Verse-Chorus form (the modern pop default)

The dominant pop structure since the 1960s. Multiple verses and choruses alternate, with possible bridges and other sections.

A typical contemporary pop song:

```
Intro (4 bars)
Verse 1 (16 bars)
Pre-Chorus (8 bars)
Chorus (16 bars)
Verse 2 (16 bars)
Pre-Chorus (8 bars)
Chorus (16 bars)
Bridge (8 bars)
Chorus (16 bars)
Outro (4-8 bars)
```

Total: ~3:30 at 120 BPM. The standard radio length.

Modern variations (post-2015): shorter verses (8 bars), shorter pre-choruses, post-chorus addition, double chorus at the end.

### Verse-Chorus with post-chorus (modern pop / K-pop / J-pop)

```
Intro
Verse 1
Pre-Chorus
Chorus
Post-Chorus
Verse 2
Pre-Chorus
Chorus
Post-Chorus
Bridge
Chorus
Post-Chorus
Outro
```

The post-chorus often features the most memorable hook in the song.

### AABA form

The "32-bar form" or "popular song form" — dominant in jazz standards, Tin Pan Alley, and early rock and roll (1920s–1960s).

Structure:

- **A** (8 bars): main melody and lyrics
- **A** (8 bars): repeats with new lyrics or slightly varied
- **B** (8 bars, the bridge): contrasting melody and lyrics, often modulates
- **A** (8 bars): returns to the main material

Total: 32 bars.

Examples:
- "Somewhere Over the Rainbow" — AABA
- Most Cole Porter, Gershwin, Rodgers & Hart standards
- "Yesterday" by The Beatles
- Many jazz standards (so much that the 32-bar AABA is sometimes called "the jazz standard form")

In jazz, the form is usually played multiple times — the "head" (theme) once, then solos over the changes, then the head again. Each soloist plays "one chorus" (one full AABA cycle, or "one form").

### Blues forms

#### 12-bar blues

The most common. 12 bars divided into three 4-bar phrases:

```
| I  | I  | I  | I  |  (or I-IV-I-I in "quick change")
| IV | IV | I  | I  |
| V  | IV | I  | I  |  (or V or turnaround in last bar)
```

The classic "AAB" lyric form: bars 1–4 say something, bars 5–8 repeat (or rephrase) it, bars 9–12 punctuate it.

Variations: jazz blues with substitutions (see `harmony/jazz-harmony.md`); minor blues; modal blues; "Bird blues" (Charlie Parker's reharm of "Confirmation").

#### 16-bar blues

A 4-bar extension of the 12-bar pattern. Less common but used in some Chicago and country blues.

#### 8-bar blues

A condensed form. "Heartbreak Hotel" is essentially 8-bar blues. "Key to the Highway" is too.

### Strophic form

Multiple verses with no chorus. Each verse uses the same melody but different lyrics. Used in:

- Folk songs ("The House of the Rising Sun")
- Hymns and traditional sacred music
- Some narrative ballads
- Modern singer-songwriter material occasionally

The lack of chorus creates a different kind of memorability — the melody is what's hooked, not a separate hook.

### Through-composed song

Each section is new; nothing repeats. Used for narrative songs that don't fit a verse-chorus structure. Bohemian Rhapsody, Stairway to Heaven, longer prog rock, much musical theatre.

## EDM forms

EDM has its own formal vocabulary, organized around drops and build-ups rather than verse-chorus.

### Standard EDM form (house, big-room, future bass, etc.)

```
Intro (16-32 bars, often beat-only)
Build 1 (16 bars)
Drop 1 (16-32 bars)
Breakdown (16-32 bars; often features vocal hook)
Build 2 (16 bars)
Drop 2 (16-32 bars; often the same as Drop 1, or with variation)
Outro (16 bars, often fading down)
```

Total: ~5–7 minutes at 128 BPM. Long for radio; ideal for DJ mixing.

### Radio edits

Most EDM tracks have a "radio edit" version, shortened to 3–4 minutes by trimming the intro and outro and possibly the second drop. The full version is for DJs.

### Trap / future bass / dubstep variations

Trap and future bass typically have shorter intros and may emphasize the vocal sections more. Dubstep often has a half-time drop with heavy bass; the build-up uses a "wobble" or syncopated bass for distinctive effect.

## K-pop forms

K-pop typically uses dense, multi-section structures with multiple distinct hooks. A typical K-pop song packs more sections into 3–4 minutes than a typical Western pop song.

### Standard K-pop form

```
Intro (4-8 bars, often vocal hook or distinctive sound)
Verse 1 (8 bars, often rapped or sung)
Pre-Chorus (8 bars, often modulating or building)
Chorus (8-16 bars, with the main hook)
Post-Chorus / Drop (8 bars, second hook, often instrumental or vocoded vocal)
Verse 2 (8 bars, often rap section)
Pre-Chorus (8 bars)
Chorus (8-16 bars)
Post-Chorus (8 bars)
Bridge (8-16 bars, often modulates and slows)
Final Chorus (16-24 bars, often with key change up)
Outro (4-8 bars)
```

Total: ~3:00–3:30. Highly compressed.

K-pop characteristics:

- **Multiple hooks**: a typical song has 3+ memorable melodic ideas (chorus, post-chorus, bridge melody).
- **Section contrasts**: each section often has a distinctly different feel — rap verse, melodic pre-chorus, anthemic chorus, drop-style post-chorus.
- **Modulation**: bridge often modulates; final chorus often a half-step or whole-step higher.
- **Chant-like hooks**: many songs have a chanted phrase (often the title) that becomes a hook unto itself.

J-pop and modern Asian pop share many of these conventions, with regional variations.

## Hip-hop forms

Hip-hop often uses verse-chorus structure but with verses that are far longer.

### Standard hip-hop form

```
Intro (4-16 bars)
Verse 1 (16 bars; "a 16")
Hook / Chorus (8 bars)
Verse 2 (16 bars)
Hook / Chorus (8 bars)
Verse 3 (16 bars; sometimes shorter or with a feature)
Hook / Chorus (8 bars)
Outro (variable)
```

The 16-bar verse is the standard unit ("dropping a 16"). Some songs have longer verses (24 or 32 bars); some shorter (8 bars).

In modern hip-hop and trap, structures are often looser — sometimes just a single 32-bar verse with a chorus before and after. The vocal flow and beat carry the structure as much as the formal sections.

## R&B forms

R&B borrows from pop verse-chorus form but often uses extended bridges and more elaborate vocal arrangements. Modern R&B may forgo strict choruses in favor of melodic hooks woven throughout the song.

Neo-soul (D'Angelo, Erykah Badu, Robert Glasper) often uses jazzy harmonic vamps with verse-chorus structure superimposed.

## Country forms

Country is conservative — usually verse-chorus form, often AABA-influenced (with the "B" being a bridge).

Common pattern: verse-chorus-verse-chorus-bridge-chorus, with simple instrumental intros and outros. The form usually serves narrative storytelling.

## Folk and singer-songwriter forms

Often strophic (multiple verses, no chorus, or a refrain-style chorus that stays constant). Bob Dylan's longer songs are essentially strophic. Many Joni Mitchell songs use unconventional forms — through-composed or hybrid structures.

## Musical theatre forms

Highly variable. Common structures:

- **AABA** for traditional showtunes
- **Verse-Chorus** with elaborate verses (the "verse" of a showtune is often a recitative-like introduction; the "refrain" is the famous melody)
- **Through-composed** for character-revealing songs and operatic moments
- **Reprise structures** where a song returns later in the show with new lyrics

## Form decisions — what each section accomplishes

| Section | What it does |
|---------|--------------|
| Intro | Establishes key, tempo, groove, possibly hook; "earns" the listener's attention |
| Verse | Tells story, gives information, sets emotional context |
| Pre-chorus | Builds energy, creates expectation, prepares chorus arrival |
| Chorus | Delivers the hook; payoff; the song's "thesis" |
| Post-chorus | Provides a second hook; extends the chorus's emotional payoff |
| Bridge | Provides relief and contrast; introduces a new perspective; modulates |
| Final chorus | Climactic arrival; often modulated, intensified, or arrangement-elevated |
| Outro | Releases tension; often fades or tags out |

A song works when each section does its job. A song fails when:

- The intro is too long (modern listeners skip)
- The verse is too long without a hook arriving (loses attention)
- The pre-chorus doesn't build (chorus feels unprepared)
- The chorus doesn't pay off (no memorable hook)
- The bridge doesn't provide enough contrast (feels redundant)
- The final chorus doesn't feel "more" than earlier choruses (no climax)

## Form-related composition decisions

### Section length

Default: 4 or 8 bars per phrase, 8 or 16 bars per section. Going shorter creates urgency; going longer creates spaciousness.

Modern pop trends shorter (verses of 8 bars instead of 16; intros of 1–2 bars instead of 8).

### Number of sections

Most songs have 5–8 sections in their final form. Songs with too many sections feel unfocused; songs with too few feel unfinished.

### Repetition vs. development

Pop relies on near-exact repetition of choruses. Each chorus is the same melody, same lyrics — repetition is the point. A chorus that varies too much loses its hook function.

Verses often have repeated melodies but different lyrics. The melody is recognizable; the lyrics develop the story.

### Modulation

When to modulate:

- **Final chorus modulation** ("truck driver"): up a half or whole step. Adds intensity. Cliché but reliable.
- **Bridge modulation**: to relative minor, parallel minor, or distant key. Creates contrast.
- **No modulation**: fine; many songs don't need it. Modulation should serve the song, not show off.

### Final chorus elevation

The final chorus should feel "more" than earlier choruses. Methods:

- Modulation up
- Adding harmony vocals
- Adding instruments (full band crash; horn section enters)
- Shouted ad-libs over the melody
- Doubling the chorus length (8 bars → 16 bars)
- Bringing in a key motif from earlier (bridge melody returning over the chorus chords)

## Common pitfalls when advising

- **Don't impose verse-chorus form on every song.** Some songs want strophic form, AABA, or through-composed. Match form to material.
- **Don't make the bridge too contrasting.** A bridge that completely changes feel can feel disconnected. The bridge should provide relief while remaining identifiable as part of the same song.
- **Don't write a chorus that doesn't pay off.** If the chorus has the same melodic register, dynamic, and density as the verse, the listener won't perceive it as a chorus. Lift the chorus somehow.
- **Don't ignore section lengths.** A 24-bar chorus is unwieldy. A 4-bar chorus is too short to land. Match length to genre conventions.
- **Don't forget about pacing.** A song that's all intensity (no breakdowns, no breathing) exhausts listeners. A song that's all space (no hooks, no climax) bores them. Pacing means alternating high and low energy.
- **Don't overcomplicate.** Most great songs use simple forms executed well. Adding a fifth section because the song "feels short" is usually wrong; it usually feels short because the existing sections aren't earning their weight.

## Quick decision matrix

| Want | Try |
|------|-----|
| Modern pop hit | Verse-pre-chorus-chorus, with possible post-chorus and final-chorus modulation |
| Classic pop song | AABA (32-bar form) |
| Singer-songwriter / folk | Strophic (multi-verse) or verse-chorus with simple structure |
| Jazz standard | AABA in 32 bars; II-V-I-driven |
| Traditional country | Verse-chorus, simple harmony, narrative verses |
| Hip-hop | Three 16-bar verses with 8-bar choruses |
| EDM track for clubs | Long form: intro–build–drop–breakdown–build–drop–outro, ~5–7 min |
| EDM radio edit | Compressed form, 3–4 min, fewer build-bars |
| K-pop / J-pop | Multi-section: verse, pre-chorus, chorus, post-chorus/drop, second verse rap, etc., highly compressed |
| Concept album track | Through-composed or extended form with prog elements |
| Dance track | Repetitive AB form (16-bar A, 16-bar B) — minimal verse-chorus distinction |

## Cross-references

- Classical forms (different formal world) → `classical-forms.md`
- Energy arcs and transitions across sections → `narrative-and-transitions.md`
- Song-specific writing techniques → `songwriting/hooks-and-memorability.md`, `songwriting/lyric-writing.md`, `songwriting/topline-craft.md`
- Phrase structure within sections → `melody/phrase-structure.md`
- Modulation techniques → `harmony/modulation.md`
- Genre-specific form conventions → `genres/pop-rock.md`, `genres/hip-hop-rnb.md`, `genres/electronic-edm.md`, `genres/kpop-jpop.md`, etc.
- Form templates for quick reference → `assets/form-templates.md`
- Production-side form decisions → `production-aware/energy-and-dynamics.md`

# Rhythm and Meter

This file covers the time-domain fundamentals: note durations, time signatures, beat hierarchies, and tempo. It's the rhythmic counterpart to `pitch-intervals-scales.md`.

For rhythmic devices that *manipulate* the meter (syncopation, polyrhythm, hemiola, displacement), see `rhythm-groove/rhythmic-devices.md`. For groove and feel (microtiming, swing, pocket), see `rhythm-groove/groove-and-feel.md`. For odd meters and polymetric writing, see `rhythm-groove/odd-meters-polyrhythm.md`. For notation conventions specifically, see `notation-and-conventions.md`.

## When to consult this file

- The user asks about time signatures, note durations, tuplets
- "What's compound time", "what's 6/8 vs. 3/4"
- Tempo terminology, BPM ranges by genre
- Beat hierarchy, strong/weak beats
- Anacrusis (pickup notes), beaming, basic notation of rhythm

## Note durations

The standard durations and their fractions of a whole note:

| Name (US) | Name (UK) | Symbol | Fraction of whole note |
|-----------|-----------|--------|------------------------|
| Whole note | Semibreve | 𝅝 | 1 |
| Half note | Minim | 𝅗𝅥 | 1/2 |
| Quarter note | Crotchet | ♩ | 1/4 |
| Eighth note | Quaver | ♪ | 1/8 |
| Sixteenth note | Semiquaver | 𝅘𝅥𝅯 | 1/16 |
| 32nd note | Demisemiquaver | 𝅘𝅥𝅰 | 1/32 |
| 64th note | Hemidemisemiquaver | 𝅘𝅥𝅱 | 1/64 |

Each rest has the same duration as its corresponding note value.

### Dotted notes

A single dot adds half the value of the note:
- ♩. (dotted quarter) = ♩ + ♪ = 3 eighth notes
- 𝅗𝅥. (dotted half) = 𝅗𝅥 + ♩ = 3 quarter notes

A double dot adds half + a quarter (75% of the original):
- ♩.. = ♩ + ♪ + 𝅘𝅥𝅯 = 7 sixteenth notes

Dotted notes are essential for compound time signatures.

### Ties

Two notes of the same pitch connected by a curved line (the tie) sound as one continuous note with their durations summed. Use ties to:

- Cross a barline (the only way to write a duration that spans a barline)
- Express a duration that doesn't fit a single note value (e.g., 5 eighth notes = ♩ tied to ♩.)
- Visualize beats clearly even when a single longer note value would work

### Tuplets

A tuplet is a group of notes that fits into the time of a different number of notes. The most common:

| Tuplet | Notes | Plays in time of |
|--------|-------|------------------|
| Triplet | 3 | 2 of the same value (e.g., 3 eighth-note triplets in the time of 2 eighth notes = 1 quarter note) |
| Duplet | 2 | 3 of the same value (in compound time) |
| Quintuplet | 5 | 4 (sometimes 6) |
| Septuplet | 7 | 4, 6, or 8 |
| Sextuplet | 6 | 4 |

Triplets are by far the most common. Quintuplets and septuplets are common in modern classical, prog rock, and contemporary jazz.

## Time signatures

Two stacked numbers. The top number is the number of beats per measure; the bottom number is the note value that gets one beat (4 = quarter, 8 = eighth, 2 = half, 16 = sixteenth).

### Simple meters

The beat divides into 2. The top number indicates how many beats; the bottom indicates the beat value.

| Time sig | Beats per bar | Beat value | Notes |
|----------|---------------|------------|-------|
| 2/4 | 2 | quarter | March |
| 3/4 | 3 | quarter | Waltz, minuet |
| 4/4 | 4 | quarter | "Common time" — written `c`; the most common |
| 2/2 | 2 | half | "Cut time" — written `¢`; fast 2 feel |
| 3/8 | 3 | eighth | Fast 3-beat |

In simple meter, beat 1 is the strongest; subsequent beats are progressively weaker. In 4/4 the standard hierarchy is 1 (strongest) > 3 (medium-strong) > 2 = 4 (weak), with offbeats (the "and" of each beat) weakest.

### Compound meters

The beat divides into 3. The top number indicates the *total* number of subdivisions per bar; divide by 3 to get the actual beat count.

| Time sig | Subdivisions | Actual beats | Beat value | Notes |
|----------|--------------|--------------|------------|-------|
| 6/8 | 6 eighth notes | 2 dotted-quarter beats | dotted quarter | Lilting two-feel; jigs, ballads |
| 9/8 | 9 eighth notes | 3 dotted-quarter beats | dotted quarter | Three-feel with triplet division |
| 12/8 | 12 eighth notes | 4 dotted-quarter beats | dotted quarter | Slow blues, ballads, doo-wop |
| 6/4 | 6 quarter notes | 2 dotted-half beats | dotted half | Slower compound 2 |

In compound meter, count the actual beats (e.g., 6/8 is "ONE-and-a, TWO-and-a", not "1-2-3-4-5-6").

### How to tell simple vs. compound

A common point of confusion: 6/8 vs. 3/4. Both have 6 eighth notes per bar. The difference is grouping:

- 3/4: groups of two — 1-2 1-2 1-2 (three beats of two divisions each)
- 6/8: groups of three — 1-2-3 1-2-3 (two beats of three divisions each)

The stress pattern differs. 3/4 stresses every other eighth (♩|♪♪|♪♪|♪♪|); 6/8 stresses every third (♪|♪♪♪|♪♪♪|).

When converting between time signatures, the underlying *feel* matters more than the math.

### Asymmetric (irregular) meters

The top number is not divisible into clean groups of 2 or 3 — usually 5, 7, 11.

| Time sig | Common groupings | Examples |
|----------|------------------|----------|
| 5/4 | 3+2 or 2+3 | "Take Five" (Brubeck), "Mission: Impossible" |
| 5/8 | 3+2 or 2+3 | Some Eastern European folk |
| 7/4 | 4+3 or 3+4, sometimes 2+2+3 | Pink Floyd "Money", much progressive rock |
| 7/8 | 2+2+3 or 3+2+2 or 2+3+2 | Balkan music, prog rock, certain math rock |
| 11/8 | various — 3+2+3+3, etc. | Frank Zappa, Stravinsky, modern jazz |

The grouping isn't a fixed property of the time signature — it's a compositional choice that the composer indicates by beaming and accents. A 7/8 measure beamed as `2+2+3` feels different from `3+2+2`.

### Additive meters

Some pieces use a meter where each measure's grouping is explicitly written: `3+2+3/8`, `2+3+2/8`, etc. These come from Eastern European, Middle Eastern, and Indian traditions and from contemporary classical music. They make explicit what a 7/8 or 8/8 meter would only imply.

### Mixed meter

The time signature changes from bar to bar. Common in 20th-century classical (Stravinsky's "Rite of Spring" famously) and in progressive rock. The pattern can be:

- **Regular alternation**: 5/8 + 7/8 + 5/8 + 7/8 (a 12-beat composite cycle).
- **Through-composed**: each bar's meter is determined by the music, with no repeating pattern. Stravinsky, Carter, much of Schoenberg.
- **Cyclical patterns**: 5+5+5+5+7 (a 27-beat cycle, common in Middle Eastern dance forms).

### Irrational meters

A time signature whose lower number isn't a power of 2: 4/12, 3/10, etc. The denominator implies tuplet relationships at the bar level. Used in modern classical (Ferneyhough, Adès) and rarely elsewhere.

## Beat hierarchy

Within a measure, beats have varying strength. The hierarchy:

1. **Downbeat** (beat 1): strongest. Marks the start of a measure.
2. **Strong beats**: in 4/4, beats 1 and 3. In 3/4, beat 1.
3. **Weak beats**: in 4/4, beats 2 and 4. In 3/4, beats 2 and 3.
4. **Off-beats** (subdivisions between beats): weaker still. The "and" of a beat is weaker than the beat itself.
5. **Off-off-beats** (further subdivisions): even weaker.

This hierarchy creates expectation. Music plays with it by:

- **Confirming the hierarchy**: putting strong content on strong beats. Default, most music.
- **Crossing it (syncopation)**: putting accents or emphasis on weak beats and off-beats. Most pop, R&B, jazz, hip-hop.
- **Suspending it (rubato)**: stretching/compressing time so the beat hierarchy becomes flexible. Romantic music, ballads.
- **Violating it dramatically**: hemiola (3 against 2), polyrhythm. See `rhythm-groove/rhythmic-devices.md`.

## Tempo

Tempo is the rate at which beats pass. Measured in BPM (beats per minute). Italian terms historically described tempo with associated character; modern scores often combine the two.

### Italian tempo terms

Roughly slow to fast. Approximate BPM ranges (for ♩ — quarter note in simple meter):

| Term | BPM | Character |
|------|-----|-----------|
| Larghissimo | < 24 | Extremely slow |
| Grave | 25–45 | Very slow, solemn |
| Largo | 40–60 | Broad, slow |
| Lento | 45–60 | Slow |
| Adagio | 55–75 | Slow, leisurely |
| Adagietto | 65–80 | A little faster than adagio |
| Andante | 75–110 | Walking pace |
| Andantino | 80–110 | Slightly faster than andante (sometimes slower; ambiguous) |
| Moderato | 108–120 | Moderate |
| Allegretto | 110–125 | Fairly fast |
| Allegro | 120–168 | Fast, lively |
| Vivace | 168–176 | Very fast, vivacious |
| Presto | 168–200 | Very fast |
| Prestissimo | > 200 | Extremely fast |

These are guidelines; actual BPM varies by style and editor. A "slow allegro" might be 115; a "fast allegro" might be 145.

### Tempo modifiers

| Term | Meaning |
|------|---------|
| poco | a little |
| molto | very |
| meno | less |
| più | more |
| ma non troppo | but not too much |
| assai | quite, rather |

So "allegro ma non troppo" = "fast but not too much"; "molto adagio" = "very slow".

### Tempo changes

| Term | Meaning |
|------|---------|
| accelerando (accel.) | gradually faster |
| ritardando (rit. / ritard.) | gradually slower |
| rallentando (rall.) | gradually slower (more abrupt than ritardando) |
| stringendo | hurrying, urgently faster |
| allargando | broadening (slower with widening dynamics) |
| a tempo | return to the original tempo |
| tempo primo (tempo I) | return to the first tempo of the piece |
| rubato | flexible tempo, "stolen time" |

Tempo changes can be written as ranges (`accel. al ♩=140`) or left to performer discretion (`accel. poco a poco`).

### Tempo by genre — rough table

Use these as starting points; significant variation exists within each genre.

| Genre | BPM range | Typical |
|-------|-----------|---------|
| Slow ballad / chillout | 60–80 | 70 |
| Hip-hop (modern) | 80–95 | 85 |
| Hip-hop (boom-bap, classic) | 85–95 | 90 |
| Trap / drill | 130–160 (counted in halftime as 65–80) | 140/70 |
| Lo-fi hip-hop | 70–85 | 75 |
| Reggae | 60–90 | 75 |
| Bossa nova | 120–140 | 130 |
| Jazz ballad | 60–85 | 70 |
| Jazz medium swing | 120–160 | 140 |
| Jazz uptempo | 200–300+ | 240 |
| Pop ballad | 60–80 | 70 |
| Modern pop | 100–130 | 115 |
| Disco | 110–135 | 120 |
| Funk | 95–130 | 105 |
| Rock (mid-tempo) | 110–130 | 120 |
| Rock (uptempo) | 130–160 | 145 |
| Punk / hardcore | 160–250 | 180 |
| Country | 80–130 | 110 |
| House | 118–135 | 124 |
| Techno | 120–140 | 130 |
| Trance | 130–150 | 138 |
| Drum and bass / jungle | 160–180 | 174 |
| Dubstep | 138–142 (often counted in halftime) | 140 |
| Hardcore / gabber | 160–250+ | 200 |
| K-pop | 90–135 | 115–120 (varies wildly) |
| J-pop | 100–145 | 120 |
| Anime opening | 130–180 | 150 |
| Film score (action) | 120–180 | 140 |
| Film score (drama) | 60–100 | 75 |
| Trailer music | 70–90 (often half-time) | 80 |
| Musical theatre | 60–160 | varies by song type |
| Ambient | 50–80 (or no clear pulse) | 65 |

The same genre can be written at different tempos for different effects. K-pop ballads often live at 70–90 BPM; K-pop dance tracks at 120–135.

## Anacrusis (pickup notes)

A note (or notes) before the first downbeat of a piece or phrase. The first full measure starts on the next downbeat; the anacrusis "borrows" time that's compensated by the final measure being short.

In notation: the pickup measure and the final measure together equal one full bar. So a piece in 4/4 with a one-beat pickup would have a final measure of 3 beats.

Anacrusis is essential to many vocal styles — phrases that begin on weak beats and resolve to downbeats feel more conversational and forward-leaning than phrases that always start on downbeats.

## Beaming conventions

Beaming groups together notes shorter than a quarter note (eighth, sixteenth, etc.) to clarify beat structure. Default beaming groups together notes within a beat:

- 4/4: beam eighth notes in pairs (one beat at a time), or all four together if it makes sense as a group
- 6/8: beam eighth notes in groups of three (one compound beat at a time)
- 7/8 (2+2+3): beam in groups of 2+2+3
- 7/8 (3+2+2): beam in groups of 3+2+2

Beaming explicitly tells the player how to count the bar. Misleading beaming is one of the most common notation errors. When in doubt, beam by beat.

For vocal music, beaming traditionally followed the syllable (one note per syllable was beamed alone; melismas were beamed together). Modern vocal scores often use instrumental beaming for readability.

## Common pitfalls when advising

- **Don't confuse 3/4 and 6/8.** They look similar, sound different. The grouping is the issue, not the count.
- **Don't assume one tempo for a genre.** Ranges are wide; the same genre at different tempos has different character.
- **Don't ignore the relationship between tempo and density.** A passage that works at 120 BPM with sixteenth notes will be unreadable at 200 BPM. Match note values to tempo.
- **Don't write asymmetric meters without indicating grouping.** A 7/8 piece needs to communicate whether it's 2+2+3, 3+2+2, or 3+4. The reader can't guess.
- **Don't use rubato as a default.** Rubato is a specific expressive device. Most music wants a steady tempo.
- **Don't rely on Italian terms alone.** Modern scores typically include a BPM marking (`♩=120`) alongside or instead of "Allegro" because the term is too vague.

## Cross-references

- Notation specifics (clefs, key signatures, dynamics, articulations) → `notation-and-conventions.md`
- Rhythmic devices (syncopation, hemiola, displacement) → `rhythm-groove/rhythmic-devices.md`
- Groove and feel (swing, microtiming, pocket) → `rhythm-groove/groove-and-feel.md`
- Odd meters and polyrhythm → `rhythm-groove/odd-meters-polyrhythm.md`
- Phrase structure and how rhythm interacts with phrase boundaries → `melody/phrase-structure.md`
- Tempo conventions in genre context → various `genres/*.md` files

# Prosody and Language

This file covers the technical interaction between language and melody — the rhythmic, accentual, and (in tonal languages) pitch-related features of speech that constrain how words fit on a melodic line. Songwriting that ignores prosody produces lyrics that feel awkward, unclear, or comedic; songwriting that respects prosody feels like the words and music were made for each other.

This is a deep area where language-specific knowledge matters enormously. The agent can give meaningful guidance for English, but for tonal languages (Mandarin, Cantonese, Vietnamese, Thai), syllabic-isochronous languages (Spanish, Italian, Japanese), and others, working with native lyricists is usually essential. This file gives the orientation needed to know what's at stake.

For lyric craft in general (rhyme, imagery, structure, point of view), see `../songwriting/lyric-writing.md`. For melody-vocal interaction in pop, see `../songwriting/topline-craft.md`. For melodic construction at the line level, see `../melody/melodic-construction.md`. For rhythmic notation and beat hierarchy, see `rhythm-meter.md`.

## When to consult this file

- Writing lyrics in a language other than English
- Translating or adapting lyrics across languages
- Diagnosing why lyrics "sound forced" or "don't fit the melody"
- K-pop / J-pop / C-pop / V-pop / T-pop projects where language-specific issues arise
- Multilingual songwriting (English-Korean, Spanish-English, etc.)
- Working with classical art song repertoire (German Lieder, French mélodie, Italian song)

## What is prosody

Prosody is the rhythmic and intonational structure of speech. Three components:

1. **Stress**: which syllables are emphasized (loudest, longest, highest in pitch)
2. **Rhythm**: the timing pattern of syllables (some languages give each syllable equal time; others vary)
3. **Pitch**: how speech intonation rises and falls; in tonal languages, also lexical pitch patterns that distinguish word meanings

Different languages handle these components very differently. A song that works prosodically in English will often need significant rework to function in Japanese, Spanish, or Mandarin.

## The three rhythmic typologies

Linguists classify languages into rough rhythmic types. The boundaries are fuzzy — most languages are mixed — and dialect, tempo, genre, rap flow, and individual lyric setting can override the textbook tendency. Use the typology as a starting frame, not a hard law:

| Type | Examples | Property |
|------|----------|----------|
| **Stress-timed** | English, German, Russian, Arabic | Stressed syllables fall at roughly equal time intervals; unstressed syllables compress between them |
| **Syllable-timed** | Spanish, French, Italian, Cantonese, Korean (loosely), Mandarin (loosely) | Each syllable gets roughly equal time; stress is less rhythmically dominant |
| **Mora-timed** | Japanese | The mora (sub-syllable unit) gets equal time; long vowels and final consonants count as separate morae |

This distinction has direct consequences for songwriting.

### Stress-timed (English)

In English, stressed syllables anchor the rhythm. Unstressed syllables can be crammed between them. The phrase "I would have gone to the store" has stress on "I", "gone", "store" — three stresses across seven syllables. Sung at moderate tempo, those three stresses align to beats; the other four syllables fall between.

Songs in English fit this naturally: melodies place strong beats at stressed syllables, weak beats at unstressed. A line like "*ALL the SINGle LADies*" puts stress on syllables 1, 3, 5 — and that's exactly where the melody emphasizes.

When stress and beat misalign, the lyric sounds wrong: "I went to the *STORE* yesterday" with stress falsely placed on "TO" instead of "STORE" reads as a beginner's error.

### Syllable-timed (Spanish, French, Italian)

In syllable-timed languages, each syllable gets roughly equal time. Stress still exists but is less dominant rhythmically. Lyric writers in these languages count syllables strictly — Spanish bolero is built on octosyllabic and hendecasyllabic lines (8 or 11 syllables); French chanson and Italian song count syllables likewise.

Implication for melody: in Spanish and related syllable-counting traditions, syllable clarity is usually important and many phrases map cleanly one syllable to one note. That does **not** mean melisma or sustained syllables are forbidden; it means stretching a syllable should be an expressive choice, not an accident caused by a mismatched translation.

Italian opera, French art song, and Spanish bolero traditions all rely on syllable-counting precision. A Western European art song composer who doesn't honor the syllable count produces awkward word-music mismatch.

### Mora-timed (Japanese)

Japanese counts not syllables but morae. The word *Tokyo* is two syllables but four morae: *To-u-kyo-u* (long vowels and consonant nasalization each get their own beat). Japanese poetry (haiku 5-7-5, tanka 5-7-5-7-7) counts morae.

Implication for melody: J-pop and traditional Japanese song typically give each mora its own note (or its own clear beat-unit). Lyrics fit melodies tightly to mora-counts. Holding a single Japanese vowel across many notes (melisma) can feel non-idiomatic in traditional Japanese vocal style; it appears in modern J-pop influenced by R&B and Western pop but is not the default baseline.

## Tonal languages — the lexical pitch problem

In tonal languages, the pitch contour of a syllable is part of the word's identity. The same consonant-vowel sequence with different tones is a different word.

| Language | Number of tones | Example |
|----------|-----------------|---------|
| Mandarin | 4 (+ neutral) | *mā* (mother), *má* (hemp), *mǎ* (horse), *mà* (scold), *ma* (question particle) — same consonant-vowel, different meanings |
| Cantonese | 6 (or 9 by some counts) | tone errors are very noticeable to listeners |
| Vietnamese | 6 | level, falling, rising, low-falling-rising, high-rising-broken, low-falling-broken |
| Thai | 5 | mid, low, falling, high, rising |
| Yoruba, Igbo, many African languages | 2–3 | tone affects grammar and word identity |

In tonal languages, **the melody's pitch movement on each syllable interacts with the lexical tone of that syllable**. Three possibilities:

1. **Melody and tone align**: the melodic contour matches the word's tone pattern. The lyric is intelligible.
2. **Melody and tone are neutralized**: the melody moves through the word's syllables in a way that doesn't strongly contradict the tones. Listeners parse the word from context. Common in Mandarin pop ballads.
3. **Melody and tone conflict**: the melodic contour says one tone, the listener hears another word. The lyric is unintelligible, comedic, or unintentionally rude.

### How songwriters in tonal languages handle this

**Mandarin Chinese (4 tones)**: Mandopop generally accepts that melody can override tone. Listeners parse word meaning from context. But lyric writers still consciously choose words whose tones harmonize with melodic contour, and avoid words whose tones would conflict catastrophically (e.g., placing a falling-tone word on an ascending melodic line). Modern Mandopop is more flexible than traditional Chinese vocal music.

**Cantonese (6+ tones)**: tone-melody alignment is much stricter. Cantopop lyric writing is famously demanding because every melodic note constrains which Cantonese words can fit. Cantopop lyricists (Lin Xi 林夕, Wyman Wong 黄伟文) are celebrated for their craft in part because of this constraint. A Cantopop melody and lyric are typically written together or in tight iteration; writing a melody first and then trying to fit Cantonese lyrics is harder than the same task in Mandarin.

**Vietnamese (6 tones)**: similar strictness to Cantonese. V-pop lyric writing is tightly constrained by tone-melody alignment.

**Thai (5 tones)**: T-pop generally maintains tone alignment more loosely than traditional Thai song forms (which were stricter); the looser approach produces some intelligibility loss tolerated as a stylistic choice.

### Implication for cross-language work

A non-tonal-language speaker writing a melody for a tonal-language lyric should **always work with a native lyricist**. The melody can be composed first, but the lyricist will usually need to adjust note pitches (within musical reason) or choose words specifically to match the contour. Translating an English song's lyric into Mandarin / Cantonese / Vietnamese / Thai while keeping the original melody is technically very hard and often produces an unintelligible or comedic result.

## Korean — between syllable-timed and stress-timed

Korean is syllable-timed in basic structure but has features that complicate this:

- **Pitch accent variation by region**: standard Seoul Korean has minimal pitch accent, but Gyeongsang dialect has marked pitch accent (almost tone-like)
- **Vowel length** distinguishes some words in older Korean but is fading in modern Seoul speech
- **Final consonants (*batchim*, 받침)**: each syllable can have a final consonant cluster, which affects singability — *batchim* before another consonant produces consonant assimilation, complicating syllable count for sung lines
- **Syllable structure varies in length**: a Korean syllable can range from one vowel (e.g., 아 *a*) to four phonemes (e.g., 닭 *dak*, with an unspoken final cluster)

For songwriting:

- Korean lyrics generally give each syllable one note, similar to syllable-timed languages
- Melisma (multiple notes per syllable) is increasingly common in K-pop / K-R&B influenced by R&B and gospel, but isn't traditional
- Rhyme in Korean works on final consonants and vowel patterns; rhyme schemes are more flexible than English
- Korean pop lyricists often code-switch into English for hooks (the K-pop English-hook tradition); the English phrase fits the pop convention while the Korean verses respect Korean prosody

### Korean-language singability

Korean has a wide range of vowel sounds (8+ standard vowels including ㅓ, ㅡ, ㅗ, ㅏ, ㅣ, etc.) and complex final consonants. For sustained singing notes:

- Open vowels (ㅏ, ㅗ, ㅓ) sustain well
- ㅡ (similar to "uh" in English) is a closed central vowel that's harder to belt
- Syllables ending in ㄱ, ㄷ, ㅂ stops are hard to sustain — singers typically open the final consonant or hold the preceding vowel

This affects which Korean syllables can carry sustained high notes and which need to be on shorter rhythmic values.

## Japanese — mora structure and singability

Japanese is mora-timed: every mora gets equal time. A mora is one of:
- A consonant + vowel (CV), like *ka*, *to*, *ne*
- A standalone vowel (V), like *a*, *i*
- A final *n* (the mora *n*, which only appears as a syllable-final)
- The first half of a long vowel
- The first half of a doubled consonant

So *Tokyo* (東京) is *to-u-kyo-u* — four morae, two syllables.

For songwriting:

- Each mora typically gets one note in traditional and most modern Japanese songs
- Melisma is rare in traditional vocal styles (*enka*, *minyo*); more common in modern J-pop and J-R&B influenced by Western styles
- Japanese vowels (*a*, *i*, *u*, *e*, *o*) all sustain well, making Japanese easy to sing in the abstract
- Final *n* syllables sustain on the nasal *n* sound
- Geminate (doubled) consonants take a full mora of silence — *kit-to* has a beat of silence between *ki* and *to*

Japanese songwriters routinely consider mora-count when fitting words to melody. A melody with 7 notes wants a 7-mora phrase, not a 7-syllable phrase.

## Multilingual songwriting

Songs that mix languages within the same piece are increasingly common — K-pop hooks in English, Latin pop with English bridges, anime themes with English titles. Some patterns:

### Code-switching for hook intensity

Many K-pop choruses include English phrases or vocables. The English provides a memorable, non-language-bound hook for international listeners while Korean carries the verse narrative. The English hook is usually short (1–4 words), often consonant-light for singability ("Bad Guy", "Pop", "Like Crazy").

### Bilingual flow

In hip-hop and rap, multilingual flow (Spanish-English, Korean-English, Mandarin-English) is now standard. Each language's prosodic features add to flow texture. Bad Bunny's Spanish-English flows, BTS's Korean-English rapping, Jay Chou's Mandarin-English passages all use language alternation as a rhythmic resource.

### Translation for international release

When a song has versions in multiple languages, the translated lyric is essentially rewritten. Direct translation rarely works — the syllable count, stress pattern, and (for tonal languages) tone alignment are different. Successful translated versions (e.g., Disney songs in dubbed languages) are written by native lyricists who match the original's meaning and emotion while honoring the target language's prosody.

### Hybrid multi-syllable considerations

In a multilingual lyric, transitions between languages can be smooth or jarring:

- Open-vowel transitions (Japanese to Spanish, both vowel-friendly) are smooth
- Consonant-cluster transitions (German to Italian, or Korean *batchim* to English) require careful melodic and rhythmic placement to feel natural
- Tonal-to-non-tonal transitions (Mandarin to English) can be jarring if the melody continues to imply Mandarin tones during the English passage

## Vowel choice for sustained notes

Across languages, certain vowels sing better than others on sustained notes:

| Vowel | Sustainability | Reason |
|-------|----------------|--------|
| Open *ah* (English "father", Italian *a*, Korean ㅏ) | Excellent | Open mouth, full resonance |
| Open *o* (English "go", Italian *o*) | Excellent | Open mouth, rounded |
| Open *eh* (English "bed", Italian *e*) | Very good | Mid-open |
| Mid *aw* (English "law") | Good | Open back |
| Closed *ee* (English "see", Italian *i*) | Moderate | Mouth shape constricted; can sound thin |
| Closed *oo* (English "boot", Italian *u*) | Moderate | Mouth shape constricted; lips rounded |
| Schwa (English "uh") | Poor for high notes | Indistinct, hard to belt |
| Korean ㅓ (roughly /ʌ/ or /ɔ/-like depending on dialect/context) | Moderate | More singable than schwa, but less open than ㅏ; check native pronunciation |
| Diphthongs (English "boy", "now", "I") | Variable | Singers usually hold the first vowel and resolve to the second at the note's end |

Lyric writers and topline composers should match high notes and sustained notes to open-vowel syllables when possible. A lyric whose climactic high note lands on a closed *oo* will sound thinner than one that lands on an open *ah*.

This is one reason English pop lyrics tend to emphasize words like "love", "yeah", "oh", "down", "round", "alone" on long notes — open vowels.

## Common pitfalls when advising

- **Don't write a melody first and assume the lyric will fit.** This is fine for English (with adjustment) but increasingly impractical as the language adds prosodic constraints — terrible for Cantonese, very hard for Mandarin, awkward for Japanese.
- **Don't translate lyrics literally across languages.** Match meaning and emotion; don't expect syllable count or stress to transfer.
- **Don't ignore lexical tone in tonal languages.** Every melodic note over a tonal-language syllable carries a meaning implication. Get it wrong and the word changes.
- **Don't put closed vowels on big high notes.** A belted *ee* sounds thinner than a belted *ah*. If the lyric requires a closed-vowel word at the climax, consider rephrasing or melodic adjustment.
- **Don't apply English stress patterns to syllable-timed languages.** A Spanish lyric written with English stress assumptions will sound jerky.
- **Don't confuse mora and syllable in Japanese.** A 7-syllable English phrase and a 7-mora Japanese phrase have different time-fillings.
- **Don't dismiss accent variation.** Korean, German, Norwegian, and many other languages have regional pitch-accent variation that affects lyric feel. Standard / dialect choices have artistic weight.
- **Don't substitute English for hard-to-sing native phrases.** "Just put the hook in English" can be a hack rather than a solution. K-pop's English hooks work because they're chosen craftfully, not because English is universally easier.

## Quick decision matrix

| Situation | Approach |
|-----------|----------|
| Writing English lyric to a fixed melody | Stress alignment is the primary concern; place stressed syllables on strong beats |
| Writing Spanish / Italian lyric | Count syllables strictly; melody must match syllable count; stress is secondary |
| Writing Japanese lyric | Count morae, not syllables; one mora per note typically |
| Writing Korean lyric | Syllable-aligned typically; consider *batchim* effects on singability; open vowels for sustained notes |
| Writing Mandarin lyric | Tone-melody alignment is real; choose words whose tones harmonize with melodic contour; native lyricist essential |
| Writing Cantonese lyric | Strict tone-melody alignment; either co-compose with lyricist or accept significant melodic adjustment |
| Writing Vietnamese / Thai lyric | Similar strictness to Cantonese for Vietnamese; somewhat looser for Thai |
| Multi-language hooks | Place language switches at natural musical boundaries (phrase ends, hook moments); choose phrases whose phonetics flow into each other |
| Translating an existing song | Treat as rewriting in the new language; match emotion and meaning, not syllable count |
| Big climactic high note | Match to an open-vowel syllable (*ah*, *oh*, *eh*); avoid closed *ee*, *oo*, schwa |
| Mid-line short notes | Schwa and closed vowels are fine here; reserve open vowels for sustained or peak notes |

## Cross-references

- Lyric writing craft (rhyme, structure, imagery, perspective) → `../songwriting/lyric-writing.md`
- Topline craft (vocal melody composition for a track) → `../songwriting/topline-craft.md`
- Melodic construction (melody-level decisions) → `../melody/melodic-construction.md`
- Hook craft (hooks must be especially singable) → `../songwriting/hooks-and-memorability.md`
- K-pop and J-pop conventions (where multilingual prosody is routine) → `../genres/kpop-jpop.md`
- C-pop, Cantopop, V-pop, T-pop (tonal languages) → `../genres/cpop-and-southeast-asian-pop.md`
- Korean traditional vocal traditions and *pansori* prosody → `../genres/korean-traditional.md`
- Vocal range and tessitura → `../melody/melodic-construction.md`, `../orchestration/instruments-ranges-character.md`
- Anglo-American folk and ballad lyric tradition → `../genres/folk-roots-and-traditions.md`
- Musical theatre lyric (a specialized prosody discipline) → `../genres/musical-theatre.md`

# Country and Americana

> **Snapshot note:** Last content review snapshot: 2026-04-27. Use this file for durable composition conventions, but verify current radio, streaming, regional country, Americana, alt-country, and country-pop references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition-facing conventions in country, Americana, roots-pop, country-rock, bluegrass-adjacent writing, singer-songwriter country, and modern country crossover. It is **not** a complete guide to Appalachian old-time, bluegrass, Western swing, honky-tonk, country gospel, or every regional scene.

## When to consult this file

- The user wants country, Americana, roots, alt-country, country-pop, Nashville-style writing, or singer-songwriter storytelling
- The user asks for pedal steel, fiddle, banjo, mandolin, acoustic guitar, or front-porch band texture
- The user wants a simple song that feels emotionally specific rather than harmonically complex
- The user wants a Nashville Number System chart or transposable band chart
- The user wants a country crossover with pop, rock, folk, gospel, hip-hop, or EDM elements

## Core idea

Country and Americana usually put **story, voice, groove, and instrumental identity** ahead of harmonic novelty. The craft question is often not "what clever chord can I use?" but:

```text
Who is speaking?
What concrete detail makes the lyric believable?
What groove makes the band feel like a band?
What instrumental gesture tells the listener the world of the song?
```

## Substyle map

| Frame | Compositional priorities | Common colors |
|---|---|---|
| Classic country / honky-tonk | clear story, plain-spoken melody, two-step or shuffle, I-IV-V | fiddle, pedal steel, acoustic guitar, simple bass |
| Bluegrass-adjacent | acoustic string drive, fast subdivision, vocal harmony, breaks | banjo, mandolin chop, fiddle, upright bass |
| Singer-songwriter country | lyric detail, intimate vocal, minimal arrangement | acoustic guitar, piano, sparse fiddle/steel |
| Americana / alt-country | roots texture, less polished delivery, folk/rock blend | dry drums, acoustic/electric guitars, organ, harmony vocals |
| Modern country / country-pop | big chorus, pop topline, clean low end, radio-ready hook | programmed support, electric guitar, banjo/steel accents |
| Country-rock | rock backbeat plus country instruments | power chords, slide, fiddle/steel hooks |
| Country gospel / roots gospel | sacred lyric, vocal harmony, plagal motion | organ, quartet harmony, call-response |

## Harmony

Country harmony is often diatonic and direct, but small chromatic moves can carry a lot of emotion because the surrounding language is simple.

### Common progressions

| Function | Progression | Nashville numbers | Use |
|---|---|---|---|
| Plain-spoken major | I-IV-I-V | `1 4 1 5` | verse, classic chorus |
| Lift into hook | I-V-vi-IV | `1 5 6m 4` | modern country-pop chorus |
| Minor emotional loop | vi-IV-I-V | `6m 4 1 5` | reflective chorus or bridge |
| Country-rock modal color | I-♭VII-IV-I | `1 ♭7 4 1` | anthem, southern rock, outro vamp |
| Ballad pull | I-iii-IV-V | `1 3m 4 5` | nostalgic pre-chorus |
| Gospel/plagal warmth | I-IV-I or IV-I | `1 4 1` / `4 1` | hymnlike tag, "home" feeling |
| Minor lift | i-♭VI-♭VII-i | `1m ♭6 ♭7 1m` | darker Americana |
| Secondary dominant | I-V/V-V-I | `1 2 5 1` if `2` is dominant | old-school turnaround |

### Useful color moves

- **♭VII in major**: country-rock / folk-rock openness.
- **IVadd9 or Iadd9**: open-string guitar shimmer without jazzing the harmony too much.
- **V7sus4 → V7**: classic tension-release, especially with vocal pickup into the chorus.
- **I/3 → IV**: stepwise bass lift into the subdominant.
- **vi → V/5 → IV**: descending bass line for ballads.
- **Diminished passing chord**: `I - #I°7 - ii` or `IV - #IV°7 - V` for older country/gospel color.

## Nashville Number System

The Nashville Number System is a practical chord-chart language: chord roots are written as scale-degree numbers relative to the key, so a band can transpose quickly.

Example in G:

```text
Key: G   Tempo: 96   Feel: train beat
Verse:
| 1  | 4  | 1  | 5  |
| G  | C  | G  | D  |
```

Common notation habits:

- `1` = major tonic
- `6m` or `6-` = minor sixth-degree chord
- `5/7` = dominant over scale-degree 7 in the bass, e.g. D/F♯ in G
- `♭7` = lowered seventh-degree major chord
- section labels, stops, pushes, diamonds, and repeat signs often carry as much information as the numbers

Use Nashville numbers when the user wants a band chart, a quick studio chart, or key-flexible country/pop guidance. Use Roman numerals when the user wants theory analysis.

## Melody and topline

Country melodies often sound close to speech. They can be simple, but they should feel **owned by the lyric**.

Useful moves:

- Build phrases around a short title hook.
- Use pentatonic and major-blues notes for approachable identity.
- Let verse melody stay conversational; save wider range for the chorus.
- Use pickup notes into strong lyric words.
- Repeat the hook title, but vary the setup line so repetition feels earned.
- In ballads, let longer notes land on concrete nouns or emotional verbs.

### Country melodic vocabulary

| Sound | Move |
|---|---|
| Plain-spoken | mostly stepwise melody, narrow verse range |
| Lonesome | ^6 or ^2 over IV, suspended notes over V |
| Bluesy | ♭3 passing into ^3, ♭7 in major context |
| Anthemic | chorus leap up a 4th/5th, then stepwise descent |
| Old-time/modal | Mixolydian ♭7, Dorian ♮6, drone-friendly shapes |

## Lyric craft

Country and Americana lyrics often reward specificity.

Instead of:

```text
I miss you every night
```

Try:

```text
Your coffee mug still turns the ring beside the sink
```

Common devices:

- title as final line of chorus
- twist ending in last chorus
- everyday object as emotional anchor
- place names, seasons, jobs, family roles
- second verse adds complication, not just more of verse 1
- bridge changes viewpoint or timeline

Be careful with clichés. Trucks, whiskey, dirt roads, church bells, porches, and small towns are valid images, but they need a specific angle.

## Groove and rhythm

| Groove | Use |
|---|---|
| Train beat | forward drive, travel, working rhythm |
| Two-step | danceable classic country |
| Shuffle | honky-tonk, blues-country, swing feel |
| 6/8 ballad | emotional, hymnlike, cinematic country |
| Straight rock backbeat | modern country-rock |
| Half-time chorus | bigger, heavier country-pop lift |
| Bluegrass chop | acoustic drive; mandolin or guitar backbeat |

Rhythm section writing should feel grounded. Bass often marks root/fifth motion, chromatic pickups, and walkups into section changes rather than busy melodic independence.

## Instrumental idioms

### Acoustic guitar

- open-position chords and capos shape the sound
- alternating bass strums feel older / folkier
- muted sixteenth strums feel modern / poppier
- dropped-D or open tunings can make simple harmony feel wide

### Electric guitar

- telecaster-style bright picking, double-stops, bends, and slide are strong country identifiers
- power-chord rhythm pushes toward country-rock
- delay/reverb swells can modernize without changing harmony

### Pedal steel / lap steel

- glides, suspensions, and slow chord-tone bends create emotional lift
- use sparingly as answering phrases between vocal lines
- do not write it like a keyboard; it breathes in horizontal motion

### Fiddle, banjo, mandolin

- fiddle: fills, counter-melodies, reels, drones
- banjo: rolling arpeggiated drive, but can sound too specific if pasted onto every country track
- mandolin: chop on backbeats, tremolo sustain, high sparkle

Pair with the instrument-specific files in `../instrument-idiom/` for playability.

## Form

Common country structures:

```text
Verse 1 → Chorus → Verse 2 → Chorus → Bridge → Chorus → Tag
```

```text
Verse 1 → Verse 2 → Chorus → Instrumental break → Verse 3 → Chorus / Tag
```

```text
Intro riff → Verse → Pre-chorus → Chorus → Verse → Chorus → Bridge → Double chorus
```

Country can tolerate longer lyric development than some pop genres, but modern crossover usually needs the title or hook early.

## Quick recipes

### Classic country verse

```text
Key: G
| G | C | G | D |
| G | C | D | G |
Melody: mostly ^1 ^2 ^3 ^5, with pickups into bar 1 and bar 5.
Arrangement: acoustic strum, simple bass, fiddle answer after every 2 bars.
```

### Americana chorus with open color

```text
Key: D
| Dadd9 | A/C♯ | Bm7 | Gadd9 |
| D/F♯ | G | A7sus A7 | D |
Effect: modern but roots-friendly; stepwise bass keeps the emotion grounded.
```

### Country-rock outro vamp

```text
Key: A
| A | G | D | A |
Numbers: | 1 | ♭7 | 4 | 1 |
Use: singalong outro, guitar solo, crowd hook.
```

## Common pitfalls

- **Do not confuse simplicity with laziness.** The lyric, groove, and band feel must carry the craft.
- **Do not over-jazz the changes unless the substyle wants it.** A dense reharm can erase the country center.
- **Do not treat all country as one politics, one region, or one accent.** The genre is broad and contested.
- **Do not paste banjo or pedal steel over a pop song and call it country.** Instrument color helps, but song logic matters.
- **Do not ignore vocal storytelling.** The lead vocal is the main instrument.
- **Do not use stereotypes as shortcuts.** Specific lived detail beats stock rural imagery.

## Quick decision matrix

| User wants | Start with |
|---|---|
| "More country" | simplify harmony, foreground lyric, add acoustic/steel/fiddle answers |
| "More modern country" | pop chorus lift, tight low end, bright guitars, concise hook |
| "More Americana" | drier production, roots instrumentation, less polished vocal, narrative detail |
| "More bluegrass" | acoustic string rhythm, mandolin chop, banjo/fiddle breaks, harmony vocals |
| "More emotional" | 6/8 feel, plagal motion, stepwise bass, concrete lyric image |
| "Less cliché" | remove stock images; add personal object, time/place, unusual verb |

## Cross-references

- Pop/rock and country crossover → `pop-rock.md`
- Folk traditions and roots writing → `folk-roots-and-traditions.md`
- Gospel overlap → `gospel-and-ccm.md`
- Guitar idioms → `../instrument-idiom/guitar.md`
- Bass idioms → `../instrument-idiom/bass.md`
- Drums and train beats → `../instrument-idiom/drums-percussion.md`
- Lyric writing → `../songwriting/lyric-writing.md`
- Hooks and memorability → `../songwriting/hooks-and-memorability.md`
- Chord symbol and Nashville conventions → `../../assets/chord-symbol-conventions.md`
- Progression families → `../../assets/progressions-catalog.md`
- Current country references → `../research/web-music-trend-research.md`, `../research/reference-track-digging.md`

# Notation and Conventions

This file covers the visual and notational conventions of Western music — clefs, key signatures, accidentals, dynamics, articulations, ornaments, score layout, and the specific conventions of pop/lead-sheet/jazz writing. It's a reference for what's on the page (or what should be).

For tempo and meter, see `rhythm-meter.md`. For pitch-system theory (intervals, scales), see `pitch-intervals-scales.md`. For chord symbols specifically, see `harmony/chord-construction.md`.

## When to consult this file

- Questions about clefs, key signatures, dynamics, articulations
- Reading or writing a score
- Lead-sheet conventions (Real Book, pop charts)
- Nashville Number System
- Roman numeral analysis notation
- Standard expressive markings

## Clefs

A clef indicates which lines and spaces of the staff correspond to which pitches.

| Clef | Notes the line/space | Used for |
|------|----------------------|----------|
| Treble (G clef) | The G clef curls around G4 (second line up) | Most instruments above middle C; right hand of piano; soprano and alto voice |
| Bass (F clef) | The two dots flank F3 (fourth line up) | Most instruments below middle C; left hand of piano; bass and baritone voice |
| Alto (C clef on middle line) | Middle line = C4 (middle C) | Viola; some choral alto parts (mostly historical) |
| Tenor (C clef on fourth line) | Fourth line up = C4 | Cello, trombone, bassoon (in their high register); some choral tenor parts |
| Soprano clef | First line = C4 | Historical only |
| Mezzo-soprano clef | Second line = C4 | Historical only |
| Baritone clef | Various positions | Historical only |
| Octave-displaced treble | Treble clef with `8` above (sounding up) or below (sounding down) | Tenor voice (8va below); piccolo/celeste sometimes (8va above) |

Treble and bass clefs are the modern defaults. C clefs (alto, tenor) are still used for specific instruments to keep ledger lines manageable.

The treble clef with `8` below (often written `treble 8va bassa`) is standard for tenor voice in modern scores — it sounds an octave below the written pitch, putting the tenor's range conveniently on the staff.

## Key signatures

A key signature appears at the start of each staff line, after the clef, indicating which pitches are sharpened or flattened throughout the piece (or until the next key signature).

### The cycle of fifths and key signatures

| Sharps | Major key | Relative minor | Sharps in order |
|--------|-----------|----------------|-----------------|
| 0 | C major | A minor | — |
| 1 | G major | E minor | F♯ |
| 2 | D major | B minor | F♯ C♯ |
| 3 | A major | F♯ minor | F♯ C♯ G♯ |
| 4 | E major | C♯ minor | F♯ C♯ G♯ D♯ |
| 5 | B major | G♯ minor | F♯ C♯ G♯ D♯ A♯ |
| 6 | F♯ major | D♯ minor | F♯ C♯ G♯ D♯ A♯ E♯ |
| 7 | C♯ major | A♯ minor | F♯ C♯ G♯ D♯ A♯ E♯ B♯ |

| Flats | Major key | Relative minor | Flats in order |
|-------|-----------|----------------|----------------|
| 0 | C major | A minor | — |
| 1 | F major | D minor | B♭ |
| 2 | B♭ major | G minor | B♭ E♭ |
| 3 | E♭ major | C minor | B♭ E♭ A♭ |
| 4 | A♭ major | F minor | B♭ E♭ A♭ D♭ |
| 5 | D♭ major | B♭ minor | B♭ E♭ A♭ D♭ G♭ |
| 6 | G♭ major | E♭ minor | B♭ E♭ A♭ D♭ G♭ C♭ |
| 7 | C♭ major | A♭ minor | B♭ E♭ A♭ D♭ G♭ C♭ F♭ |

### Order of sharps and flats

Sharps are added in this order: F♯ C♯ G♯ D♯ A♯ E♯ B♯
Flats: B♭ E♭ A♭ D♭ G♭ C♭ F♭

Each is a perfect fifth above the previous (for sharps) or below (for flats). Memorizing the order makes reading and writing key signatures fluent.

### Finding the tonic from the key signature

For sharp keys: the last sharp is one half-step below the tonic (last sharp F♯ → key of G; last sharp C♯ → key of D).

For flat keys: the second-to-last flat names the tonic (flats B♭ E♭ A♭ → key is E♭). Exception: F major has only one flat (B♭), no second-to-last; just memorize.

For minor keys: the relative minor is a m3 below (or M6 above) the relative major. So 1 sharp = G major OR E minor; the actual key is determined by listening for the tonic, not by the signature alone.

## Accidentals

Accidentals override the key signature for individual notes within a measure.

| Symbol | Name | Effect |
|--------|------|--------|
| ♯ | Sharp | Raise by a semitone |
| ♭ | Flat | Lower by a semitone |
| ♮ | Natural | Cancel any prior sharp/flat |
| 𝄪 (or ♯♯) | Double sharp | Raise by two semitones |
| 𝄫 | Double flat | Lower by two semitones |

An accidental applies for the rest of the measure on that pitch (same octave, same letter name) unless cancelled by another accidental. After the barline, the key signature reasserts.

For modern atonal/serial music or in passages where every note is accidental, scores sometimes write each accidental explicitly (no carry-through). The composer should specify if they're using non-standard practice.

### Microtonal accidentals (briefly)

For quarter-tone and other microtonal music, additional accidentals exist (varying by composer/system). Most common:

- ½♯ (half-sharp) — raise by quarter-tone
- ½♭ (half-flat) — lower by quarter-tone
- 1½♯ (sharp + half-sharp) — three-quarter-tone

Used in contemporary classical, some jazz (rarely), and adaptations of non-Western music.

## Dynamics

| Marking | Italian | Volume |
|---------|---------|--------|
| pppp | piano-pianissimo or pianissississimo | Extremely soft |
| ppp | pianississimo | Very, very soft |
| pp | pianissimo | Very soft |
| p | piano | Soft |
| mp | mezzo piano | Moderately soft |
| mf | mezzo forte | Moderately loud |
| f | forte | Loud |
| ff | fortissimo | Very loud |
| fff | fortississimo | Very, very loud |
| ffff | fortissississimo | Extremely loud |

### Dynamic accents and modifiers

| Marking | Meaning |
|---------|---------|
| sf, sfz | sforzando — sudden strong accent on a single note |
| fp | forte-piano — start loud, immediately drop to soft |
| sfp | sforzando-piano — sudden accent that immediately drops to soft |
| rfz | rinforzando — reinforced (similar to sfz, less abrupt) |
| > | accent (louder than surrounding notes; not as sharp as sfz) |
| ^ | marcato (strong, marked accent) |
| ten. | tenuto (full duration, slight emphasis) |

### Dynamic transitions

| Marking | Meaning |
|---------|---------|
| crescendo (cresc.) or `<` | Gradually louder |
| decrescendo (decresc.) or `>` (with terminal mark) | Gradually softer |
| diminuendo (dim.) | Gradually softer (synonym for decrescendo) |
| poco a poco | little by little (modifies cresc. or dim.) |

The hairpin notation (`<` `>`) is a graphical alternative to "cresc." / "dim." Both are common.

## Articulations

Articulation markings affect how each note is played — its length, attack, and emphasis.

### Strings and winds

| Marking | Name | Effect |
|---------|------|--------|
| `.` (dot above/below note) | staccato | Short, detached; about half the written duration |
| `'` (wedge or stroke) | staccatissimo | Very short |
| `_` (line) | tenuto | Hold full value, slight emphasis |
| `_.` (line + dot) | portato | Slightly separated but full-bodied |
| `>` | accent | Stronger attack on this note |
| `^` | marcato | Strong, sharp attack |
| slur (curved line over notes) | legato | Smooth connection between notes; in winds and strings, played in one bow/breath |

### Bowed strings

| Marking | Name | Effect |
|---------|------|--------|
| `↑` | up-bow | Bow moves toward the player |
| `n` (square shape) | down-bow | Bow moves away from the player |
| arco | "with the bow" | Cancel pizzicato |
| pizz. | pizzicato | Plucked |
| col legno | "with the wood" | Strike with the bow's wooden side |
| sul ponticello | "on the bridge" | Glassy, scratchy timbre |
| sul tasto | "on the fingerboard" | Soft, flute-like timbre |
| spiccato | (technique) | Bouncing, light short notes |
| tremolo (z or // through stem) | rapid bow strokes | Sustained shimmer |

### Brass and winds (selected)

| Marking | Name | Effect |
|---------|------|--------|
| flutter | flutter-tonguing (rolled R while playing) | Buzzy, growly |
| tongued / slur | normal tonguing or smooth | Default vs. legato |
| mute (with `+` or `o`) | hand-stopped or muted | Brass: changes timbre |
| open / mute | for trumpet, trombone | Specific mute states |
| double tongue, triple tongue | (technique) | Rapid articulation |

### Percussion

Percussion notation varies by instrument. Common conventions:

- Drum kit: notated on a 5-line staff with specific positions for each drum (kick on the bottom space, snare on the middle, hi-hats above the staff).
- Mallet percussion (xylophone, marimba, vibraphone): pitched, written on standard staff.
- Auxiliary percussion: often a single-line staff with x-noteheads.

For drum charts in pop/jazz, see "Pop and lead-sheet conventions" below.

### Voice

| Marking | Effect |
|---------|--------|
| slur | One syllable across multiple notes (melisma) |
| no slur | One syllable per note |
| portamento | Slide between pitches (slow glissando in vocal context) |
| spoken | Spoken rhythm, no pitch |
| Sprechstimme (or `x`-noteheads) | Speech-singing, approximate pitch |

## Ornaments

Decorative figures, often improvised in earlier eras and now often written out fully.

| Symbol | Name | Effect |
|--------|------|--------|
| tr | trill | Rapid alternation between the written note and the note above (usually a M2; sometimes m2 depending on style/key) |
| 𝆔 | mordent | A quick alternation: principal note → upper neighbor → principal |
| 𝆕 | inverted mordent | Principal → lower neighbor → principal |
| 𝆖 | turn | Four-note figure: upper neighbor → principal → lower neighbor → principal |
| 𝆗 | inverted turn | Lower neighbor → principal → upper neighbor → principal |
| (small grace note before main note) | acciaccatura | Crushed grace note; takes essentially no time from the main note |
| (small grace note before main note, slurred) | appoggiatura | Lyrical grace note; takes time from the main note (typically half its value) |
| ~/| / | glissando | Slide between two notes |
| (wavy line) | vibrato | Pitch wavers slightly |

The interpretation of ornaments is style-dependent. A Baroque trill starts on the upper note; a Classical or modern trill usually starts on the principal note. Mark or describe explicitly when style might be ambiguous.

## Score layout

### Order of instruments (full orchestral score)

From top to bottom:

1. **Woodwinds** (piccolo, flutes, oboes, English horn, clarinets, bass clarinet, bassoons, contrabassoon)
2. **Brass** (horns, trumpets, trombones, tuba)
3. **Percussion** (timpani, then auxiliary, then mallet)
4. **Keyboards / harp** (if present)
5. **Voices / chorus** (if present)
6. **Strings** (violin I, violin II, viola, cello, double bass)

Within each section, instruments are ordered from highest to lowest. Horns are an exception — they appear above trumpets in the brass section by tradition, even though they're lower-pitched.

### Bracketing

Instruments of the same family are grouped with a bracket on the left. Subgroups (e.g., flute 1+2) get a sub-bracket. The string section uses a thicker bracket.

### Measure numbers

Modern practice: number every measure (or every 5 or 10) at the top of the system. The first measure of a piece is bar 1 unless there's an anacrusis, in which case the first full bar is 1.

### Rehearsal marks

Letters or numbers placed in boxes at structural points to facilitate rehearsal: `A`, `B`, `C` (or `1`, `2`, `3`). Often placed at the start of new sections, key changes, or thematic returns.

### Cues and divisi

| Marking | Meaning |
|---------|---------|
| div. | divisi (a section splits into multiple parts) |
| unis. | unison (cancel divisi) |
| solo | One player (rather than the full section) |
| tutti | Everyone plays |
| a 2 | Both players (when two share a stand or part) |

## Expression markings

Italian is the standard language; some scores use German, French, or English. The agent should recognize common expressions.

### Italian (standard)

| Term | Meaning |
|------|---------|
| dolce | sweetly |
| espressivo | expressively |
| cantabile | singing, lyrical |
| con brio | with spirit |
| con fuoco | with fire |
| con moto | with motion |
| grazioso | gracefully |
| leggiero | lightly |
| maestoso | majestically |
| appassionato | passionately |
| tranquillo | calmly |
| risoluto | resolutely |
| pesante | heavy |
| scherzando | playfully |
| morendo | dying away |
| calando | decreasing (dynamic and tempo) |

### German (Romantic and 20th C.)

Mahler, Schoenberg, and others wrote markings in German.

| German | Meaning |
|--------|---------|
| ruhig | calm |
| heftig | violent |
| zart | tender |
| schnell | fast |
| langsam | slow |
| mit | with |
| ohne | without |

### French (Impressionist mostly)

| French | Meaning |
|--------|---------|
| doux | sweet |
| très | very |
| en dehors | bring out (used for emphasis) |
| comme un soupir | like a sigh |

## Pop and lead-sheet conventions

Lead sheets (especially the Real Book and similar collections) use a stripped-down notation: melody on a treble staff, chord symbols above. The arranger fills in the rest.

### Chord symbols

See `harmony/chord-construction.md` for the full table. Quick reminders:

- Major: `C`, `Cmaj`, `CM` (rarely just the bare letter without specification)
- Minor: `Cm`, `Cmin`, `C-`
- Major 7: `Cmaj7`, `CΔ`, `CM7`
- Minor 7: `Cm7`, `C-7`
- Dominant 7: `C7`
- Half-diminished: `Cø7`, `Cm7♭5`
- Diminished 7: `C°7`, `Cdim7`
- Augmented: `C+`, `Caug`, `C(♯5)`

### Slash chords

`X/Y` = chord X over bass Y. `C/E` is C major in first inversion (E in bass); `F/G` is F major over G bass (effectively a G dominant suspension).

### Drum chart notation

Drum charts often use simplified notation:

- `x` noteheads for cymbals and hi-hat
- Standard noteheads for snare and toms
- Single-line or full 5-line staff
- "Slash" notation (`/`) for "play time" — keep going with the same groove
- `Fill` markings for fills

A typical drum chart for a verse might be: `|1: Time | 2: Time | 3: Time | 4: Fill |`

### Tab (guitar tablature)

Six lines representing the six strings. Numbers indicate which fret to play. Common modifications:

- `h` = hammer-on
- `p` = pull-off
- `b` = bend
- `r` = release bend
- `/` or `\` = slide
- `~` = vibrato
- `x` = muted note

Tab is widely used by guitarists in lieu of standard notation, especially for rock and folk. It conveys what to play but not pitch directly.

### Nashville Number System

A shorthand for chord progressions using scale degrees (numbers) instead of letters. Used heavily in Nashville session work because it's instantly transposable to any key.

| Notation | Meaning |
|----------|---------|
| `1` | I (tonic) |
| `2m` | ii (minor) |
| `3m` | iii (minor) |
| `4` | IV |
| `5` | V |
| `6m` | vi (minor) |
| `7°` | vii° |
| `1/3` | I in first inversion (3 in bass) |
| `5/7` | V with leading tone in bass |
| `1Δ` | Imaj7 |
| `5 7` | V7 |
| `b7` | ♭VII (borrowed) |

A Nashville chart for "Country Roads": `1 6m | 4 1 | 1 6m | 4 1 |` etc.

For minor keys, lowercase or specifically marked numbers indicate minor mode. Conventions vary somewhat by sheet writer.

### Roman numeral analysis

Used in classical analysis and academic settings:

| Symbol | Meaning |
|--------|---------|
| `I`, `IV`, `V` | Major triad on that scale degree |
| `i`, `iv`, `v` | Minor triad |
| `vii°` | Diminished triad on ^7 |
| `III+` | Augmented triad |
| `V7`, `ii⁶/⁵` | 7th chord with figured-bass inversion |
| `V/V` | "V of V" — secondary dominant |
| `♭III` | Lowered (borrowed) III |
| `N⁶` | Neapolitan in first inversion (♭II⁶) |
| `It⁶`, `Fr⁶`, `Ger⁶` | Augmented 6th chords |

Always specify the key context. "ii" in C major is Dm; "ii" in A♭ major is B♭m.

## Common pitfalls when advising

- **Don't mix conventions.** Pick lead-sheet symbols OR Roman numerals OR Nashville for any one piece of analysis or chart. Mixing is confusing.
- **Don't use double-flats and double-sharps decoratively.** They have specific theoretical purposes (e.g., the diminished 7 of B°7 is A♭ becoming A𝄫 in spelling). Using them where ♮ or simpler accidentals work is bad spelling.
- **Don't write enharmonic equivalents at random.** F♯ and G♭ are not interchangeable in context — pick the spelling that fits the key signature and the harmonic function.
- **Don't forget the cadence-confirming dynamic.** A score with no dynamics is hard to interpret. Even simple lead sheets benefit from a few markings.
- **Don't over-articulate.** A staccato on every note becomes meaningless. Articulations should mark moments of change or emphasis.

## Cross-references

- Pitch system, intervals, scales → `pitch-intervals-scales.md`
- Tempo, meter, beat hierarchy → `rhythm-meter.md`
- Chord symbols and analysis (Real Book, pop, RN, Nashville) → `harmony/chord-construction.md`, `assets/chord-symbol-conventions.md`
- Notation in genre-specific contexts → various `genres/*.md`
- Score layout for orchestration → `orchestration/instruments-ranges-character.md`
- Roman numeral analysis applied to actual progressions → `analysis.md`

# Classical Periods

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers the historical periods of Western art music — Medieval, Renaissance, Baroque, Classical, Romantic, Impressionist, and 20th-century / contemporary. Each period has its own characteristic harmonic language, formal preferences, instrumentation, and aesthetic priorities. Composing or arranging "in" a period requires understanding what's idiomatic for it.

For other genre traditions, see other `genres/*.md` files. For specific 20th-century compositional techniques, see `techniques/20th-century-techniques.md`. For classical formal types, see `form/classical-forms.md`.

## When to consult this file

- The user wants to write in a specific historical style (Baroque, Classical, Romantic)
- Analyzing a piece and identifying its period
- Understanding why a piece "doesn't sound right" for its intended style
- Period-specific harmony, voice leading, instrumentation
- Pastiche or homage projects

## Medieval (c. 800–1400)

The earliest period of notated Western music.

**Harmonic language**: modal (the church modes — Dorian, Phrygian, Lydian, Mixolydian, plus their plagal forms). No functional harmony. Cadences mostly on open 5ths or octaves; the 3rd was considered dissonant in early periods.

**Rhythm**: rhythmic modes (specific patterns) in early polyphony; later, Ars Nova introduced more rhythmic flexibility (isorhythm, hocket).

**Texture**: started monophonic (Gregorian chant); evolved into early polyphony (organum, motet).

**Forms**: chant, organum, motet, conductus, formes fixes (ballade, virelai, rondeau).

**Composers**: Hildegard of Bingen, Léonin, Pérotin, Machaut, Dufay (transitional to Renaissance).

**Modern uses**: rare. Modal language influences some film scoring (medieval/fantasy settings) and certain post-rock and metal subgenres. Historical performance ensembles specialize.

## Renaissance (c. 1400–1600)

The era of polyphonic vocal music, flourishing in churches and courts.

**Harmonic language**: modal, increasingly with cadential conventions (the Phrygian half cadence, the major-3rd ending of a piece in the minor mode — *Picardy third*). Major and minor modes (Ionian and Aeolian) gradually emerge as primary; the other modes recede.

**Counterpoint**: smooth, vocal polyphony. Strict rules about consonance, dissonance treatment (passing tones, suspensions only on weak beats; no parallel 5ths or 8ths).

**Rhythm**: imitative, with interlocking voices. Hemiola common (especially in 6/8 → 3/4 cross-rhythm). Tactus (steady pulse) underlies everything.

**Texture**: 3 to 6 voices (more in some sacred works); vocal-driven but with instrumental doublings.

**Forms**: motet, mass (Kyrie, Gloria, Credo, Sanctus, Agnus Dei), madrigal, chanson, dance suite (basic).

**Composers**: Josquin des Prez, Palestrina, Lassus, Byrd, Tallis, Monteverdi (transitional).

**Modern uses**: choral writing in liturgical contexts often uses Renaissance idioms. Modal jazz draws on Renaissance modal sensibility. Period-instrument ensembles perform the repertoire.

## Baroque (c. 1600–1750)

The era of functional tonality crystallizing, contrapuntal mastery, and elaborate ornamentation.

**Harmonic language**: functional tonality (major and minor keys with V→I cadences). Figured bass (continuo) provides harmonic framework. Modulation common, often to closely related keys (V, IV, ii, vi, relative major/minor).

**Counterpoint**: the supreme form of the period. Bach's fugues, inventions, canons, the *Art of Fugue*. Strict voice-leading rules but with greater rhythmic and chromatic freedom than Renaissance.

**Rhythm**: steady, often relentless rhythmic drive. Most pieces have a single dominant rhythm throughout (the *Affekt* — one emotional state per movement).

**Texture**: continuo-based — a bass line plus chordal harmony (realized by harpsichord, organ, lute) supports the upper voices. Solo concertos, concerto grosso, trio sonata.

**Forms**: fugue, prelude, suite (allemande, courante, sarabande, gigue + optional dances), concerto, sonata, opera, oratorio, cantata, chorale.

**Composers**: Monteverdi (early), Corelli, Vivaldi, Handel, J. S. Bach (the apex), Rameau, Couperin, Scarlatti.

**Idiomatic features**:
- Sequences (ascending and descending, especially circle of 5ths)
- Suspensions (4–3, 7–6, 9–8) at cadences
- Picardy third (ending a minor-key piece on a major chord)
- Trills typically begin on the upper note (Baroque trill convention)
- Inégale playing in French Baroque (long-short rhythmic inflection)
- Tierce de Picardie (Picardy third) endings
- Modulation typically through V or relative major/minor

**Modern uses**: Baroque pastiche common in film scoring (Williams in some moods; period films). Bach's contrapuntal language influences jazz (Bill Evans, Brad Mehldau borrow Bach), prog rock, modern classical, and electronic music.

## Classical (c. 1750–1820)

The era of balance, clarity, and formal logic. Reaction against Baroque elaboration.

**Harmonic language**: tonal, mostly diatonic with strategic chromaticism. Cadences clearly mark phrase boundaries. Modulation typically to dominant or relative major/minor in expositions; back to tonic in recapitulations. Secondary dominants and modal mixture are vocabulary.

**Counterpoint**: less prominent than Baroque, but present. Mozart and Haydn use counterpoint at climactic moments and in development sections. The Classical "homophonic" texture is the default.

**Rhythm**: regular, dance-like, periodic. Phrasing typically 4-bar (often 4+4 = 8-bar periods). Sentence structures and parallel periods dominate.

**Texture**: melody-dominated homophony. Strings provide the foundation; winds and brass add color and sustain.

**Forms**: sonata form, theme-and-variations, rondo, minuet-and-trio, sonata cycle (3 or 4 movements). Concerto, symphony, string quartet, sonata, opera.

**Composers**: Haydn, Mozart, early Beethoven, Cherubini, J. C. Bach.

**Idiomatic features**:
- Periodic phrasing (8-bar parallel periods)
- "Question and answer" structures (antecedent + consequent)
- Murky bass / Alberti bass (LH arpeggiated chord pattern)
- Light, transparent textures (you can hear individual instruments)
- Wit and humor — sudden silences, surprising modulations (Haydn's "Surprise Symphony")
- Restraint with chromaticism — used for highlight, not as default

**Modern uses**: Classical-style pastiche common in film scoring (Mozart-style for period films, comedies). Many simpler pop songs follow Classical-period phrase structure (without realizing it).

## Romantic (c. 1800–1900)

The era of emotional intensity, programmatic music, expanded forms, and chromaticism pushing tonality to its limits.

**Harmonic language**: tonal but increasingly chromatic. Heavy use of:
- Modal mixture (borrowed chords)
- Secondary dominants and tonicizations
- Chromatic mediants (third-related shifts)
- Augmented 6th chords
- Common-tone diminished
- Enharmonic modulation
- Wagnerian "endless melody" — chromatic voice-leading without clear cadence

**Forms**: classical forms expanded (longer sonata movements, fewer cadences, more episodic). Character pieces (nocturnes, etudes, intermezzi). Symphonic poem (programmatic single-movement orchestral work). Wagner's *Gesamtkunstwerk* (total art opera).

**Texture**: thicker, often with elaborate orchestration. Pianos used the full range. Orchestras grew (from ~30 in Classical to 80+ in late Romantic).

**Rhythm**: rubato (flexible tempo), dramatic pauses, irregular phrasing. Complex polyrhythms in some composers (Brahms, Schumann).

**Composers**:

- **Early**: Schubert, Mendelssohn, late Beethoven (transitional), Schumann, Berlioz, Chopin, Liszt
- **Mid**: Brahms, Wagner, Verdi, Tchaikovsky, Bruckner, Saint-Saëns
- **Late**: Mahler, R. Strauss, Rachmaninoff, early Sibelius, Elgar

**Idiomatic features**:
- Wide dynamic range (ppp to fff)
- Long melodic lines crossing multiple bars
- Heavy modulation (often to distant keys via chromatic pivots)
- Sequences extended chromatically
- "Tristan chord" and similar ambiguous harmonies
- Climactic moments with full orchestra and choir
- Rubato as expressive default

**Modern uses**: Romantic style is the foundation of film scoring (Williams, Goldsmith, Horner, Zimmer all draw heavily from late Romantic). Power ballads in pop draw from Romantic emotional tropes. Most "epic" musical gestures trace to Romanticism.

## Impressionist (c. 1880–1920)

A French reaction to Romanticism. Atmosphere, color, ambiguity over drama and resolution.

**Harmonic language**: tonality blurred but not abandoned. Common features:
- Whole-tone scale
- Pentatonic scales
- Modal vocabulary (especially Lydian and Dorian)
- Parallel motion (forbidden in earlier periods)
- Quartal and quintal harmony
- Extended chords (9ths, 11ths, 13ths used as colors, not tensions)
- Avoidance of clear cadences
- Static harmony (one chord sustained for many bars)

**Texture**: layered, atmospheric. Use of pedal points, sustained chords, washes of sound.

**Rhythm**: flexible, often without strong downbeats. Use of triplets and irregular subdivisions. Less driving than Romantic.

**Forms**: shorter, sectional, often based on atmosphere rather than dramatic narrative.

**Composers**: Debussy (the central figure), Ravel, Fauré, Satie (transitional), some early Stravinsky, Falla.

**Idiomatic features**:
- Parallel 5ths and 9ths (deliberately, as color)
- "Eighth-note moto perpetuo" textures (Ravel)
- Whole-tone passages for dreamy/mystical effect
- Pentatonic melodies (often evoking East Asian or Spanish music)
- Subtle, restrained dynamics
- Gestural rather than thematic development

**Modern uses**: Impressionist harmonic vocabulary pervades modern jazz (especially Bill Evans, Herbie Hancock, modern fusion), modern film scoring (atmospheric cues), ambient music, chillwave, and certain pop production aesthetics.

## 20th Century / Contemporary (c. 1900–present)

A vast and fragmented era. No single style; multiple parallel movements.

### Atonality and 12-tone

**Schoenberg, Berg, Webern**: free atonality (rejection of tonal centers) and then dodecaphony (12-tone serialism). Each pitch class used equally; tone rows organize the material.

### Neoclassicism

**Stravinsky (mid-period), Hindemith, Prokofiev, Shostakovich**: return to Classical-era forms and clarity, but with modernist harmonic and rhythmic vocabulary.

### Total serialism

**Boulez, Stockhausen, Babbitt** (mid-20th century): all musical parameters (pitch, rhythm, dynamics, articulation) organized serially.

### Aleatoric / chance music

**Cage, Feldman**: indeterminacy. Some elements left to chance or performer choice.

### Minimalism

**Reich, Glass, Riley, Adams**: gradual processes, repetition with subtle variation, tonal but non-functional harmony, steady pulse.

### Spectralism

**Grisey, Murail**: harmonic content derived from analysis of acoustic spectra; music as the unfolding of timbre.

### Polystylism / postmodernism

**Schnittke, Berio, late Crumb**: collage of styles, quotation, irony.

### New Complexity

**Ferneyhough, Finnissy, Carter (late)**: extreme rhythmic and notational complexity; demands virtuosic performers.

### Crossover and pluralism (current)

Contemporary classical is increasingly pluralistic. Composers blend:
- Classical traditions
- Jazz
- Folk, roots, and tradition-specific music
- Electronic music
- Pop and rock
- Visual art and multimedia

Recent composers: Adams, Glass, Reich, Pärt, Adès, Saariaho, Andriessen, Gubaidulina, Tan Dun, Anna Thorvaldsdottir, Caroline Shaw, Nico Muhly.

For specific 20th-century techniques (12-tone, serialism, aleatoric, minimalist process, spectralism), see `techniques/20th-century-techniques.md`.

## Composing in a period — practical advice

When the user wants to write *in* a specific period:

1. **Listen extensively.** Period style is absorbed through listening, not just rules. Spend hours with representative works.
2. **Analyze structure first.** Each period has characteristic forms. Match yours to the period.
3. **Match the harmonic language.** Period harmony is bounded — Baroque doesn't use modal mixture freely; Classical doesn't use chromatic mediants as default; Romantic doesn't use 12-tone serialism.
4. **Match phrase structure.** Period phrasing is distinctive (Classical 4-bar; Romantic flexible; Baroque continuous).
5. **Use period instrumentation.** A "Baroque" piece on synth is a contradiction; use harpsichord, baroque flute, period strings.
6. **Avoid anachronisms.** A Baroque piece using 9th chords as tonic isn't "Baroque"; it's modern with a Baroque-inspired surface. Decide whether you're doing pastiche or modern homage.
7. **Read scores in the style.** Scores reveal idiomatic details that recordings can't.

## Common pitfalls when advising

- **Don't reduce a period to a list of chords.** Each period is a complete aesthetic world — harmony, rhythm, form, texture, emotion all align.
- **Don't blend periods carelessly.** A "Romantic" piece with Baroque counterpoint can work but requires conscious design. Mixing for no reason produces incoherence.
- **Don't equate "classical" with "old".** "Classical music" colloquially refers to all Western art music; "Classical period" specifically refers to ~1750–1820. Be precise.
- **Don't ignore performance practice.** Baroque pieces aren't played the same way as Romantic pieces. Trills, ornamentation, tempo flexibility, dynamics all vary by period.
- **Don't claim period authenticity without commitment.** Modern composers writing "Baroque-style" usually produce hybrid music that's recognizably 20th/21st century. That's fine; be honest about it.

## Quick decision matrix

| Want | Style/period to model |
|------|----------------------|
| Polyphonic vocal piece | Renaissance (Palestrina) |
| Driving instrumental fugue | Baroque (Bach) |
| Balanced sonata movement | Classical (Mozart, Haydn) |
| Emotional, large-scale orchestral | Romantic (Brahms, Tchaikovsky, Mahler) |
| Atmospheric, hazy, dreamy | Impressionist (Debussy, Ravel) |
| Modal, "ancient"-sounding | Medieval / Renaissance (modal cadences) |
| Atonal, modernist | Schoenberg, post-Schoenberg serialism |
| Repetitive, hypnotic | Minimalism (Reich, Glass, Adams) |
| Plural, eclectic, postmodern | Schnittke, late 20th-century pluralism |
| Film score "epic" | Late Romantic + film tradition (Williams) |

## Cross-references

- Specific 20th-century techniques → `techniques/20th-century-techniques.md`
- Classical formal types → `form/classical-forms.md`
- Counterpoint (especially Baroque) → `counterpoint.md`
- Romantic chromatic harmony → `harmony/chromatic-harmony.md`
- Modal harmony (Renaissance/Medieval) → `harmony/modal-harmony.md`
- Impressionist harmonic vocabulary → `harmony/jazz-harmony.md`, `harmony/modal-harmony.md`
- Period instrumentation → `orchestration/instruments-ranges-character.md`
- Notation conventions per period → `fundamentals/notation-and-conventions.md`
- Film scoring's debt to Romantic → `genres/film-tv-scoring.md`

# C-Pop and Southeast Asian Pop

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers the major popular music industries of greater China (Mandarin and Cantonese) and Southeast Asia — distinct from K-pop and J-pop and from each other. Each has its own history, structural conventions, and stylistic markers. The shared element is that all are highly developed industries serving large audiences and have evolved their own conventions; none should be treated as "Asian pop generically".

For K-pop and J-pop, see `kpop-jpop.md`. For Korean traditional and K-trad fusion, see `korean-traditional.md`. For folk traditions (which often inform regional pop), see `folk-roots-and-traditions.md`. For the technical challenge of writing in tonal languages (relevant for Mandarin, Cantonese, Vietnamese, Thai), see `../fundamentals/prosody-and-language.md`.

## When to consult this file

- Composing or arranging in C-pop, Cantopop, T-pop, V-pop, OPM, or Indonesian pop
- Cross-genre work involving these markets
- Understanding why a song "doesn't quite work" in a regional market
- Picking reference points for a project aimed at a specific Asian market

## C-pop (Mandopop) — Mandarin-language popular music

The largest of these industries by audience, spanning the People's Republic of China, Taiwan, Singapore, Malaysia, and the global Chinese diaspora.

### History

- **1920s–40s**: *shidaiqu* (时代曲, "songs of the era"), originating in Shanghai — the first modern Chinese popular music. Hybrid of Chinese folk melody and Western jazz arrangements. Major figures: Zhou Xuan (周璇), Bai Guang (白光).
- **1949–1980s**: After the 1949 split, mainland popular music was suppressed; the industry continued in Hong Kong, Taiwan, and Southeast Asia. Taiwan emerges as a major hub. Teresa Teng (邓丽君) becomes a pan-Asian icon.
- **1990s**: Taiwan dominates Mandopop, with artists such as Teresa Teng's successors, A-Mei (张惠妹), and the early rise of Wang Leehom (王力宏). Production becomes globally polished.
- **2000s**: Jay Chou (周杰伦), Jolin Tsai (蔡依林), S.H.E, F4, Mayday (五月天), and other Taiwan-centered acts define a pan-Chinese pop generation while mainland China reopens and grows.
- **2010s–onward**: Mainland streaming platforms (QQ Music, NetEase Cloud Music, Kugou/Kuwa), idol culture, talent shows, and short-video platforms (Douyin / TikTok) reshape the industry. Increasing crossover with K-pop (Chinese members in Korean groups; Chinese subgroups).

### Structural conventions

C-pop generally follows verse-chorus form with several distinct features:

- **Long instrumental intros are still common** (8–16 bars), more than in modern Western pop
- **Ballads are central**: the slow, emotional ballad with strings and piano remains a dominant form, more so than in K-pop or Western pop
- **Bridge often modulates** up a half-step or whole-step before the final chorus
- **Songs often run 4:00–5:00**, longer than streaming-era Western pop (which has compressed to ~3:00)
- **Repeated chorus with growing arrangement intensity**: typical structure has the final chorus modulate up *and* add backing vocals, full arrangement, often a key change

### Musical signatures

**Harmonic language**

- Often jazz-inflected, especially in ballads (extended chords, secondary dominants, modal mixture)
- Pentatonic flavor in melodic writing (drawing on Chinese folk pentatonic *gong*, *shang*, *jue*, *zhi*, *yu* — corresponding roughly to major pentatonic modes)
- Cinematic chord progressions in dramatic ballads (chromatic mediants, augmented sixths)
- Use of *Royal Road*-style progressions (IV-V-iii-vi) common, similar to J-pop influence

**Melodic style**

- **Long, sustained melodic lines** suited to the language's tonal patterns (see `../fundamentals/prosody-and-language.md` for the tonal-Mandarin / melodic-contour interaction)
- Wide vocal range; star vocalists are expected to deliver big high notes
- Pentatonic melodic content can suggest Chinese traditional-adjacent color without making the track fully traditional
- Melisma is moderate (less than R&B), expressive in ballads

**Production**

- Polished but slightly less compressed/loud than K-pop
- Strings and piano remain central in ballads
- Modern uptempo C-pop borrows heavily from K-pop and Western pop production
- Mainland and Taiwan production styles diverge slightly (Taiwan more "live band" feeling, mainland more EDM-influenced)

### Sub-categories

- **Mandopop ballad**: the central form; sustained tradition from Teresa Teng to modern artists
- **Modern C-pop dance**: K-pop and Western pop influences, often with mainland idol groups
- **Chinese hip-hop**: separate scene; took off after *Rap of China* (2017). Higher Brothers, GAI, MC HotDog
- **Sino-folk / "Chinese style" (中国风, *zhongguofeng*)**: pop with explicit Chinese traditional elements — *erhu*, *guzheng*, classical Chinese poetry references. Jay Chou's "Blue and White Porcelain" (青花瓷) is the iconic example. Closely related to Korea's K-trad fusion in spirit
- **Indie / alternative**: smaller but growing scene; Sodagreen (苏打绿, now Oaeen), Mayday (五月天), Mando-rock bands

### Composing in C-pop — practical advice

- **Mandarin's four tones (plus neutral) constrain melodic contour**. A melody that ignores tone risks sounding like different words than the lyric intends. See `../fundamentals/prosody-and-language.md`.
- **Ballads still command the market**. A C-pop project that doesn't include at least one ballad track is unusual.
- **Reference local production aesthetics, not just K-pop**. C-pop has its own polish that's not identical to K-pop.
- **Lyrical theme matters**: love, longing, family, homeland are central themes; lyric depth is valued more than in some Western pop.

## Cantopop — Cantonese-language popular music

Hong Kong-based industry (Cantonese is also spoken in Guangdong, Macau, and globally diaspora). Distinct from Mandopop in language, conventions, and history.

### History

- **1970s onward**: Cantopop emerges with Sam Hui (许冠杰), bringing Cantonese into popular music (it had been associated with low-prestige music previously).
- **1980s peak**: golden age. Leslie Cheung (张国荣), Anita Mui (梅艳芳), Alan Tam (谭咏麟), Roman Tam (罗文). Cantopop dominates pan-Asian Chinese pop.
- **1990s "Four Heavenly Kings" era**: Andy Lau, Jacky Cheung, Aaron Kwok, Leon Lai (张学友 stands out as the central vocalist). Cantopop reaches global Chinese audiences.
- **Beyond 1990s**: industry shrinks as Mandopop and K-pop rise. Hong Kong indie scene grows. Beyond (乐队 Beyond, the rock band) is foundational; many newer indie/rock bands continue.
- **2020s**: Cantopop revival around MIRROR (the Hong Kong boy group; usually stylized in Latin letters) and a new generation of artists; renewed cultural relevance.

### Cantopop musical signatures

- **Cantonese has 6 (or 9, depending on counting) tones** — even more constraining for melody-tone alignment than Mandarin. Cantopop melodic writing is extremely tone-sensitive. A "wrong" melodic note in Cantonese can change the meaning of a lyric word (see `../fundamentals/prosody-and-language.md`).
- **Often more rock / band-oriented** than Mandopop's piano-and-strings ballad tradition
- **Strong tradition of theme-song writing**: many Cantopop hits originated as TV drama or film themes (especially TVB-era productions)
- **Lyrical sophistication valued**: Cantonese lyric writers (黄霑 James Wong, 林夕 Lin Xi, 黄伟文 Wyman Wong) are cultural figures; lyric depth and cleverness matter

### Composing in Cantopop

- Tone-melody alignment is the single biggest technical challenge. If you don't speak Cantonese, work with a native lyricist who can adjust the melody.
- The Cantopop ballad has different conventions from the Mandopop ballad (often more rock-band-oriented, sometimes more rhythmically pronounced).
- Reference points: Beyond for rock; Eason Chan (陈奕迅) for modern ballad; MIRROR for current idol-pop context.

## T-pop — Thai popular music

Thai popular music (*phleng Thai sakon*, modern Thai music, distinguished from *phleng Thai doem*, traditional). Increasingly globally visible since the late 2010s.

### Brief history

- **1930s–1970s**: *phleng luk thung* (folk-rooted country pop) and *phleng luk krung* (urban refined pop) develop as primary genres
- **1980s–1990s**: rock and pop expand; Thai pop increasingly Western-influenced. *Bird* Thongchai (เบิร์ด ธงไชย) emerges as a pan-Asian star.
- **2000s onward**: idol culture, BNK48 (Thai sister group of AKB48), and global expansion
- **2018 onward**: T-pop boy/girl groups modeled on K-pop systems (4EVE, BUS, ATLAS, PERSES). Major industry investment.
- **Thai BL (Boys' Love) drama soundtracks**: a globally significant cultural export driving Thai pop's international audience

### Musical signatures

- **Thai is tonal** (5 tones: mid, low, falling, high, rising). Tone-melody alignment is a constraint, though less rigid in T-pop than in traditional song forms.
- **Pentatonic-derived melody** is common for Thai traditional-adjacent color (which is built on 7-note equidistant tuning historically — note this is genuinely different from Western 12-TET)
- **K-pop influence is substantial**: modern T-pop borrows heavily from K-pop production and structure
- **Fusion of Thai traditional instruments** (*ranat ek* xylophone, *khim* hammered dulcimer, *khlui* flute) into pop is a distinctive sub-style

### Sub-categories

- **T-pop idol**: K-pop-influenced groups with Thai language and aesthetic
- **Mor lam / luk thung pop**: rural-rooted dance pop (Isan-region, northeastern Thailand)
- **Thai indie / alternative**: smaller scene; Phum Viphurit, Mild, others have international visibility
- **BL OST**: full subgenre of melodic, romance-themed pop; often independently produced

### Composing in T-pop

- For idol-style T-pop, the K-pop framework (multi-section structure, distinct hooks) is the working template
- Thai language has phonetic features (final consonants are unreleased; vowel length is meaningful) that shape singability
- Mor lam / luk thung crossover requires familiarity with Isan musical conventions
- Modern Thai BL OST has its own conventions (acoustic guitar + piano, mid-tempo, narrative-style lyrics)

## V-pop — Vietnamese popular music

The Vietnamese pop industry, increasingly visible internationally since the 2010s.

### Brief history

- **Pre-1975**: pre-war Vietnamese popular music with French chanson and bolero influences. Khanh Ly (Khánh Ly), Trinh Cong Son (Trịnh Công Sơn) are foundational.
- **1980s–2000s**: domestic industry grows; bolero remains popular. Diaspora pop in the U.S. is significant.
- **2010s–onward**: V-pop modernizes, drawing on K-pop and Western pop. Sơn Tùng M-TP becomes a major star with global visibility.
- **2020s**: continuing growth; collaborations with K-pop and global pop artists; significant TikTok-driven discoverability.

### Musical signatures

- **Vietnamese is tonal** (6 tones: level, falling, rising, low-falling-rising, high-rising-broken, low-falling-broken). The tonal complexity makes melody-lyric alignment very strict.
- **Bolero ballad** remains a distinctive Vietnamese form — slow, emotional, often tied to sentimental narrative themes
- **Modern V-pop** mixes K-pop production, Western pop forms, and Vietnamese melodic sensibility
- **Strong indie / acoustic singer-songwriter tradition** (Vũ Cát Tường, Hoàng Dũng, others)

### Composing in V-pop

- Tone-melody alignment is the central technical challenge for Vietnamese lyrics. Native lyricists are essential.
- Bolero ballad still has cultural weight even in modern V-pop; a sentimental ballad tradition is distinctive.
- Modern V-pop is closer to global pop production than to traditional Vietnamese music; K-pop and Western influences dominate.
- Reference points: Sơn Tùng M-TP for modern idol pop; Vũ Cát Tường for indie; Mỹ Tâm and Hồ Ngọc Hà for established mainstream.

## OPM — Original Pilipino Music (Filipino popular music)

The Philippine music industry, with its own long history and distinctive features.

### Brief history

- **1970s OPM movement**: explicit effort to develop Filipino-language popular music. Ryan Cayabyab, Jose Mari Chan, Asin (the band) are foundational.
- **1980s–1990s**: strong ballad tradition (Regine Velasquez, Sharon Cuneta), rock (Eraserheads, Rivermaya), folk-rock (APO Hiking Society)
- **2000s onward**: dual-language pop (Filipino and English), strong band scene (Kamikazee, Up Dharma Down), continued ballad tradition
- **2010s–onward**: Ben&Ben, IV of Spades, SB19 (the boy group). Global K-pop influence drives idol-group formation.

### Musical signatures

- **Filipino-language and English-language tracks coexist**; many artists release both
- **Strong ballad and folk-rock traditions** distinct from K-pop and J-pop conventions
- **Vocal-forward production**: emphasis on lead vocalist's range and expression
- **Western pop / R&B influences** are deep — Filipino artists are known for technical vocal ability in international competitions
- **Less tone-language constraint** than C-pop, V-pop, T-pop — Filipino (Tagalog) is not tonal

### Composing in OPM

- Filipino is syllable-timed and not tonal — closer to Spanish in prosodic feel than to Chinese / Vietnamese / Thai
- Ballad tradition still dominates a significant share of OPM
- Modern band scene (Ben&Ben, Up Dharma Down) is a distinctive idiom; not the same as Western indie folk-rock
- Reference points: Eraserheads for legacy rock; Ben&Ben for modern folk-pop; SB19 for idol pop; Regine Velasquez for ballad tradition

## Indo-pop — Indonesian popular music

The Indonesian pop industry, serving the world's fourth-most-populous country.

### Brief history

- **1960s–70s**: modern Indonesian pop emerges with Koes Plus, Iwan Fals (folk-political singer)
- **1980s–90s**: dangdut (popular folk-fusion genre with Indian and Arabic influences) becomes a major genre alongside pop. Rhoma Irama is the dangdut king.
- **2000s onward**: pop-rock bands (Sheila on 7, Peterpan/NOAH, Dewa 19), continuing dangdut, growing R&B and hip-hop scenes
- **2010s–onward**: increased K-pop influence; Indonesian girl/boy groups; significant music streaming market

### Musical signatures

- **Indonesian (Bahasa Indonesia) is syllable-timed and not tonal** — like Filipino, easier in melody-lyric alignment than C-pop / V-pop / T-pop
- **Dangdut** has distinctive features: tabla-like drum (*kendang*), suling (flute), strong rhythmic drive, Indian-, Arabic-, and Malay-influenced melodic material — increasingly fused with EDM (*koplo* dangdut, *house dangdut*)
- **Religious music** (Islamic *nasyid*, Christian gospel) has significant audience
- **Strong pop-rock band tradition**: Indonesian rock has continuous popularity unlike its decline in some markets

### Composing in Indo-pop

- Dangdut fusion is a distinctive opportunity not available in other Asian markets
- Indonesian pop ballads share some sensibility with Filipino ballads (vocal-forward, melodic)
- Modern Indo-pop borrows from K-pop and Western pop similarly to other regional markets
- Reference points: Dewa 19 / NOAH for legacy rock-pop; Tulus and Raisa for modern soloists; JKT48 for K-pop-influenced idol; Rhoma Irama for foundational dangdut

## Other regional markets

These have less external visibility but are significant industries:

- **Malay-language pop**: Malaysia and Brunei; shares some conventions with Indo-pop but distinct
- **Mongolian pop**: small but distinctive industry; The HU (folk-metal-rock band) has international visibility
- **Cambodian, Lao, Myanmar pop**: distinct national industries with limited global export
- **South Asian (Indian / Pakistani / Bangladeshi) pop**: vast and old industries — covered separately under `folk-roots-and-traditions.md` and `film-tv-scoring.md` (Bollywood); not a primary focus of this file

## Composing across these traditions — practical advice

When working with a market this file covers:

1. **Listen extensively to current material from the specific market** before assuming K-pop or Western pop conventions transfer. Each industry has its own current conventions.
2. **For tonal languages (Mandarin, Cantonese, Vietnamese, Thai)**: consult `../fundamentals/prosody-and-language.md` for the technical issue of tone-melody alignment. Always work with native lyricists.
3. **For non-tonal languages (Filipino, Indonesian, Malay)**: prosody is less constraining; lyric flow follows syllable count and stress more than tone.
4. **Identify the relevant ballad tradition**. Most of these markets have a deeply embedded ballad tradition that doesn't disappear when modern dance pop dominates the charts.
5. **Reference local stars, not just K-pop**. A "Mandopop song that sounds like K-pop" is a specific genre choice; the default Mandopop song doesn't sound like K-pop.
6. **Respect the language**. Don't write English-language hooks into a regional pop song unless that's a deliberate convention of the market. (English is increasingly common in T-pop, V-pop, and C-pop as accent points; never assume.)

## Common pitfalls when advising

- **Don't generalize "Asian pop"**. Each country's industry has distinct conventions.
- **Don't apply K-pop's structural conventions universally**. C-pop, T-pop, V-pop, OPM, Indo-pop each have their own structural defaults; K-pop influence is real but partial.
- **Don't underestimate the ballad's centrality** in C-pop, Cantopop, V-pop, and OPM. A modern song without ballad sensibility may feel disconnected from the audience's expectations.
- **Don't ignore tone constraints in tonal-language markets**. A melody that ignores Mandarin / Cantonese / Vietnamese / Thai tones can sound comedic or unintentionally rude.
- **Don't conflate dangdut with "Indonesian music generally"**. Dangdut is one genre with deep roots; Indo-pop ballads, rock, and modern pop are distinct.
- **Don't assume Filipino vocal style transfers to other markets**. OPM's vocal-forward, runs-heavy style is a genre convention, not a regional default.
- **Don't apply traditional-music elements as decoration**. A song with a *guzheng* sample isn't C-pop traditional fusion; it's pop with a *guzheng* sample. Engage with the tradition's logic, not just its sound.

## Quick decision matrix

| Want | Try |
|------|-----|
| C-pop ballad | Slow tempo (60–80 BPM), piano + strings, jazz-influenced extensions, modulation up at final chorus |
| C-pop modern dance | K-pop-influenced multi-section structure, Mandarin lyrics, polished production |
| *Zhongguofeng* (Chinese-style fusion) | Pop song with *guzheng*, *erhu*, *dizi* features and pentatonic melodic touches; reference Jay Chou |
| Cantopop ballad | Often more rock-band-oriented than Mandopop ballad; reference Eason Chan |
| Cantopop rock | Reference Beyond, modern Hong Kong indie |
| T-pop idol | K-pop framework with Thai language; multi-section, distinct hooks, idol group dynamics |
| Thai BL OST | Acoustic guitar + piano, mid-tempo, narrative romantic lyrics |
| V-pop modern idol | K-pop and Western influences with Vietnamese language; reference Sơn Tùng M-TP |
| V-pop bolero ballad | Slow, sentimental, narrative; the Vietnamese ballad tradition |
| OPM ballad | Vocal-forward, runs and melisma valued, English or Filipino lyric |
| OPM band | Folk-pop or rock; reference Ben&Ben, Eraserheads |
| Indo-pop dangdut | *Kendang* drum, suling flute, distinct rhythmic feel; modern fusions add EDM |
| Indo-pop pop-rock | Strong rock band tradition; reference NOAH, Dewa 19 |
| General "Asian-influenced" cross-genre | Identify which specific market you're aiming at; "Asian influence" without specificity tends to be tokenistic |

## Cross-references

- K-pop and J-pop (different but adjacent industries) → `kpop-jpop.md`
- Korean traditional music and K-trad fusion → `korean-traditional.md`
- Folk, roots, and tradition-specific music (foundations of regional pop) → `folk-roots-and-traditions.md`
- Tonal languages and melody-lyric alignment → `../fundamentals/prosody-and-language.md`
- Pop and rock conventions (Western foundations many of these draw from) → `pop-rock.md`
- Hip-hop conventions (relevant for Chinese, Filipino, Indonesian hip-hop scenes) → `hip-hop-rnb.md`
- Film and drama OST (a major commercial channel for these traditions) → `film-tv-scoring.md`
- Pop song forms (foundation for all these markets) → `../form/popular-song-forms.md`

# Folk and World Music

> **Snapshot note:** Last content review snapshot: 2026-04-27. This legacy redirect is not an active genre guide; verify current or tradition-specific claims through the active linked files.

> **Compatibility redirect:** This legacy broad-label file is kept so older links do not break. For active guidance, use `folk-roots-and-traditions.md`.

The older label "world music" is too broad for precise composition advice. When the user asks for this, route to a more specific tradition, region, scene, function, or reference track.

Use:

- `folk-roots-and-traditions.md` for folk/roots/tradition-specific writing
- `regional-scene-starters.md` for smaller regional popular scenes
- `korean-traditional.md` for gugak / K-trad fusion
- `mena-pop.md` for MENA pop / maqam-adjacent fusion
- `south-asian-film-pop.md` for South Asian film-pop / raga-derived pop
- `afrobeats-and-amapiano.md` for Afrobeats/Afropop/amapiano lanes
- `brazilian-pop-and-funk.md` for Brazilian pop/funk/MPB-adjacent writing
- `minor-and-hybrid-genres.md` for unstable or internet-native genre labels

Do not treat a tradition as a generic color. Name the source and extract concrete musical variables: rhythm, ornament, language/prosody, instrument idiom, form, function, and collaboration needs.

# Film and TV Scoring

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

Film and TV scoring is composition in service of visual storytelling. The composer's job is to support, comment on, or contradict what's on screen — never to overshadow it. This file covers the conventions, idioms, and craft of writing music for picture.

## When to consult this file

- The user is writing music for a film, TV show, video game, or trailer
- Scoring approach for a specific scene or mood
- Using leitmotifs, themes, or thematic development in narrative music
- Trailer music conventions
- Game music (related discipline)

## What's distinctive about film scoring

Film music is *applied* music. It serves a specific purpose:

- **Supporting emotion**: making sad scenes sadder, scary scenes scarier, joyful scenes more uplifting
- **Pacing**: pulling the viewer through slower passages, accelerating tense ones
- **Information**: telling the audience how to feel about characters and events
- **Continuity**: stitching disparate scenes together
- **Identity**: giving a film a sonic personality that makes it memorable

The composer is rarely the auteur of the experience. The director, editor, and music supervisor make many decisions; the composer's craft is matching their vision while bringing musical excellence.

## Historical eras of film scoring

### Silent film era (c. 1895–1929)

Live music accompanied silent films. Pianists, organists, or orchestras played from compilation books or improvised. The first "scores" were collected pieces matched to scene types.

### Classical Hollywood era (c. 1933–1960)

Max Steiner, Erich Korngold, Alfred Newman, Bernard Herrmann established Hollywood scoring. Drew heavily from late Romantic style (Strauss, Wagner, Mahler, Tchaikovsky).

**Idiomatic features**:
- Late Romantic harmonic language
- Symphonic orchestra
- Leitmotif use (Korngold from Wagner)
- Mickey-mousing (music synced precisely to action)
- Often wall-to-wall (continuous music throughout the film)

### New Hollywood / late 20th century (c. 1965–onward)

Jazz, rock, and pop influences enter scoring. Composers like Henry Mancini, Lalo Schifrin, John Barry mix idioms.

The 1977 Star Wars score (John Williams) revives the symphonic-leitmotif tradition for the blockbuster era.

### Modern film scoring (c. 1990–onward)

Multiple parallel traditions:

- **Symphonic / leitmotif-driven**: John Williams, James Horner, Howard Shore, Alan Silvestri, Patrick Doyle. Late Romantic + film tradition.
- **Hybrid orchestral / electronic**: Hans Zimmer, Junkie XL, Steve Jablonsky, Lorne Balfe. Big sound, layered orchestra + synths.
- **Minimalist / atmospheric**: Cliff Martinez, Jonny Greenwood, Mica Levi, Daniel Hart. Texture and color over melody.
- **Indie / chamber**: Jon Brion, Mychael Danna, Nick Cave & Warren Ellis, Hildur Guðnadóttir. Smaller ensembles, more intimate scoring.
- **Pop / rock-influenced**: Atticus Ross, Trent Reznor, Cliff Martinez. Contemporary popular idioms.
- **Folk/roots or tradition-specific influenced**: Mychael Danna, Tan Dun, Bear McCreary. Cultural specificity.

## Composers worth knowing

A non-exhaustive list of essential film composers:

| Composer | Style / Era |
|----------|-------------|
| Max Steiner | Classical Hollywood, Romantic-influenced (King Kong, Casablanca, Gone with the Wind) |
| Bernard Herrmann | Hitchcock, Welles; psychological scoring (Vertigo, Psycho, Citizen Kane) |
| Erich Korngold | Operatic-symphonic; classic Hollywood (Robin Hood) |
| Alfred Newman | Classical Hollywood; lush Romantic style (How the West Was Won) |
| Nino Rota | Italian; Fellini and Coppola (Godfather) |
| Ennio Morricone | Italian; Westerns and beyond (The Good the Bad and the Ugly, Cinema Paradiso) |
| John Barry | British; James Bond, lush Romantic (Out of Africa) |
| Henry Mancini | Jazz-influenced (Pink Panther, Breakfast at Tiffany's) |
| Jerry Goldsmith | Versatile; iconoclastic harmonies (Planet of the Apes, Alien) |
| John Williams | Symphonic blockbuster master (Star Wars, Indiana Jones, Schindler's List, Harry Potter) |
| James Horner | Romantic, Celtic, sweeping (Titanic, Braveheart, Avatar) |
| Howard Shore | Symphonic + leitmotif; horror to epic (Lord of the Rings, Cronenberg films) |
| Hans Zimmer | Hybrid orchestral-electronic; modern blockbuster (Inception, Interstellar, Dune) |
| Danny Elfman | Quirky, Tim Burton frequent collaborator (Batman, Edward Scissorhands) |
| Alan Silvestri | Action-adventure (Back to the Future, Avengers) |
| Thomas Newman | Idiosyncratic, often pizzicato strings (American Beauty, Skyfall) |
| Trent Reznor & Atticus Ross | Industrial-influenced, modernist (Social Network, Soul) |
| Mica Levi | Avant-garde, unsettling (Under the Skin, Jackie) |
| Hildur Guðnadóttir | Atmospheric, cello-centric (Joker, Chernobyl) |
| Ludwig Göransson | Modern hybrid (Black Panther, Tenet, Oppenheimer) |
| Joe Hisaishi | Japanese; Ghibli films (Spirited Away, Princess Mononoke) |
| Ryuichi Sakamoto | Japanese; minimalist and electronic (The Last Emperor, The Revenant) |

## Scoring techniques

### Leitmotif

A musical theme associated with a specific character, place, idea, or emotion. The theme returns throughout the film, transformed to reflect narrative changes.

See `melody/motivic-development.md` for the details. Summary:

- Leitmotif first appears clean and recognizable.
- Returns are transformed: minor mode for sadness, faster for action, fragmented for fragmentation, full orchestra for triumph.
- Multiple leitmotifs can combine (when characters meet on screen).
- Leitmotif's final form usually delivers narrative payoff.

Wagner pioneered leitmotif in opera. John Williams is the modern master in film. Howard Shore's *Lord of the Rings* is the most elaborate film leitmotif system, with dozens of interrelated themes.

### Mickey-mousing

Synchronizing music tightly to action — every gesture on screen has a musical equivalent. A fall has a descending glissando; a hit has a percussive accent.

Common in early Hollywood, animated films, and comedy. Less common in modern dramatic scoring (where it can feel overstated).

### Underscoring

Music that supports the scene without drawing attention. Often understated, harmonically simple, atmospheric. The music carries emotion but doesn't compete with dialogue.

The opposite of "score with melodic identity". Many great film scores are mostly underscore with brief melodic peaks.

### Source music (diegetic music)

Music that exists *in* the film world — characters can hear it. A radio playing in a scene; a band performing at a wedding. The composer may write source music to match the period or character.

Distinct from non-diegetic score (which only the audience hears).

### Hybrid scoring

Modern scores often combine orchestra with electronic elements. Hans Zimmer is the influential figure; many modern scores layer:

- Live orchestra (often recorded in London)
- Synthesized pads and bass
- Sound design (atonal, noise-based elements)
- Percussion (orchestral + tradition-specific / regional + electronic)
- Choir (especially in epic moments)

The combined sonic palette is bigger and more textured than orchestra alone.

### Spotting and sync

Spotting: deciding where music starts and stops in the film. Done with the director and music editor.

Sync: aligning music to specific moments (cuts, hits, dialogue). Modern composers work with timecode — every musical event is precisely placed against the picture.

## Scoring for specific genres

Different film genres have different scoring conventions.

### Action / blockbuster

- Big symphonic palette + electronic + percussion
- Drumming, often orchestral percussion battery
- Brass-heavy (low brass for power)
- Strings often in fast ostinatos
- Tempos 110–160 BPM for sustained action
- Rhythmic, often less melodically-focused

**Examples**: Williams's Star Wars; Zimmer's Pirates of the Caribbean; Silvestri's Avengers.

### Drama

- More melodic, subtle, emotionally nuanced
- Smaller ensembles often (string quartet, chamber)
- Slower tempos
- Piano often featured
- Long sustained chords; less rhythmic activity

**Examples**: Newman's American Beauty; Reznor/Ross's The Social Network; Guðnadóttir's Joker.

### Horror

- Atonal or modal
- Dissonant clusters and effects
- Sustained tension; sudden stings
- Often experimental sound design
- String "scrapes", piano interior playing, prepared instruments

**Examples**: Herrmann's Psycho; Krzysztof Penderecki (used in The Shining); Mica Levi's Under the Skin.

### Romance / drama-romance

- Lush Romantic harmony
- Long lyrical melodies
- Strings prominent
- Sometimes solo voice
- Often quiet, intimate moments alternating with sweeping climactic moments

**Examples**: Horner's Titanic; Newman's various; Zimmer's Interstellar (love theme).

### Comedy

- Often lighter texture, smaller ensembles
- Idiomatic instruments (banjo, ukulele, kazoo for goofiness)
- Mickey-mousing more acceptable
- Period-specific source music sometimes

**Examples**: Mancini's Pink Panther; Elfman's various; Williams's Home Alone.

### Sci-fi

- Often electronic or hybrid
- Atonal or modal harmonic language for "alien"
- Sound design may dominate over melody
- Choirs for "vast space" feel

**Examples**: Goldsmith's Star Trek; Zimmer's Interstellar, Dune; Williams's Star Wars (which is uniquely Romantic for sci-fi).

### Period drama

- Period-appropriate idioms
- Often actual period music adapted, plus original score in the style
- Authentic instruments

**Examples**: Mychael Danna's various Anglo-Indian period films; Patrick Doyle's Henry V; Howard Shore's The Aviator.

### Animation

- Often more rhythmic/syncopated
- Heavier use of mickey-mousing
- Sometimes character themes extensively developed
- Disney/Pixar tradition: Alan Menken (musical theatre influence), Randy Newman, Michael Giacchino.

**Examples**: Newman's Toy Story; Giacchino's Up; Menken's Disney Renaissance films.

## Trailer music

A specialized subgenre. Trailer music has its own conventions and its own production houses (Two Steps from Hell, Audiomachine, Brand X Music, X-Ray Dog, Hi-Finesse).

### Standard trailer structure

**For a 2:30 trailer**:

```
0:00–0:15 — Cold open: quiet pad, single instrument
0:15–0:45 — Slow build: introduce theme, percussion enters
0:45–1:15 — Mid build: layered orchestration, building drums
1:15–1:30 — First peak (small drop): full ensemble
1:30–1:45 — Cool-down or "logo moment": quiet, often with title card
1:45–2:15 — Massive final build: choir, full orchestra, percussion battery
2:15–2:30 — Climactic peak; possibly with end-tag
```

### Trailer idioms

- **Big Drum™**: cinematic taiko-style drums with heavy reverb
- **Hybrid percussion**: orchestral + tradition-specific / regional + electronic
- **Choir**: especially soaring female solo or full Latin-text choir
- **Brams**: low brass swells (the "BRAAAM" of *Inception* trailers, popularized by Zimmer)
- **Risers and downers**: synth/orchestral build-up effects
- **Drone**: sustained low pad as foundation
- **Theme statement**: a single melodic theme, recognizable, used repeatedly

### Trailer music vs. film score

Trailer music has different priorities:

- 2-3 minutes long; needs to be self-contained
- Has to create instant emotional response (no time to develop)
- Highly cliché-driven (the audience expects certain sounds)
- Often more bombastic than the final film score

## Game music

Related but distinct discipline. Differences:

- **Adaptive / interactive**: game music responds to player actions and game state. Vertical layering (more intensity = more layers added) and horizontal switching (different cues for different scenes) are standard techniques.
- **Looping**: many cues need to loop seamlessly during long gameplay.
- **Emotional sustain**: game cues often last for many minutes; need to sustain interest without becoming repetitive.
- **Theme-heavy**: leitmotifs and themes are often more central than in film, since players spend longer with each character/location.

**Notable game composers**: Nobuo Uematsu (Final Fantasy), Koji Kondo (Mario, Zelda), Jeremy Soule (Skyrim), Austin Wintory (Journey), Disasterpeace (Fez, It Follows), Olivier Deriviere, Jesper Kyd.

## Practical advice for film scoring

If the user is writing for picture:

1. **Watch the scene without music first.** Understand its emotional arc before committing to anything.
2. **Identify the scene's emotional beats.** Where does emotion shift? Where's the climax? Music should align with these.
3. **Decide on the scene's "voice".** Solo instrument? Full orchestra? Synth pad? Orchestral with rock drums?
4. **Sketch the cue's structure.** Roughly: where does it start, where does it peak, where does it end?
5. **Write to picture.** Use timecode if available. Hits, dialogue, and cuts inform musical decisions.
6. **Don't over-score.** Less is often more. Silence is a tool. Some scenes work best with no music.
7. **Match the film's overall sonic identity.** Don't suddenly introduce a new palette mid-film.
8. **Serve the director, the editor, the picture.** It's their film, not yours.

## Common pitfalls when advising

- **Don't write music as if it stands alone.** Film music is service work. The best film cue might be musically simple but support the film perfectly.
- **Don't ignore dialogue.** Music in dialogue scenes should be sub-audible mostly — not competing with the words.
- **Don't apply concert-music logic.** Sonata form doesn't fit a 2-minute scene. Film cues have their own pacing.
- **Don't ignore temp tracks.** Directors often "temp" with existing music. Composers usually need to work with or against the temp; understanding it is half the job.
- **Don't overuse the standard idioms.** "Cinematic drums" and "Inception-style brams" are clichés. Use them when appropriate but don't over-rely.
- **Don't write 5-minute themes for a film that needs 30-second cues.** Match scope to scene.
- **Don't ignore the budget.** Many films can't afford a 90-piece orchestra; sample libraries and solo musicians are often the practical reality.

## Quick decision matrix

| Want | Try |
|------|-----|
| Epic blockbuster action | Hybrid orchestra + percussion + choir; Zimmer-style |
| Classic Hollywood feel | Late Romantic symphonic; Williams or Korngold style |
| Atmospheric drama | Sustained chords, sparse melody, piano + strings |
| Horror tension | Atonal clusters, sustained dissonance, prepared piano |
| Trailer music | Standard build-and-drop structure, big percussion, choir |
| Indie / quirky | Small ensemble, idiosyncratic instruments, light texture |
| Period film | Authentic period idiom; researched and respectful |
| Sci-fi alien | Electronic + atonal + choir; "vast" sound |
| Romance lift | Lush strings, lyrical melody, late Romantic harmony |
| Action with heart | Theme that transforms — minor in tension scenes, major in triumph |
| Animation comedy | Mickey-mouse where appropriate; lighter texture; idiomatic instruments |
| Game adaptive cue | Vertical layered structure (low intensity → high intensity layers) |

## Cross-references

- Leitmotif and theme development → `../melody/motivic-development.md`
- Orchestration for film → `../orchestration/instruments-ranges-character.md`, `../orchestration/voicing-and-texture.md`
- Density and arrangement → `../orchestration/arrangement-density.md`
- Energy arcs (cue structure) → `../form/narrative-and-transitions.md`
- Romantic harmonic language → `../harmony/chromatic-harmony.md`, `classical-periods.md`
- Modal harmony for atmospheric scoring → `../harmony/modal-harmony.md`
- Counterpoint in film scoring → `../counterpoint.md`
- 20th-century techniques (atonal, minimalist) → `../techniques/20th-century-techniques.md`
- Modulation as narrative tool → `../harmony/modulation.md`
- Game music (interactive scoring; closely related but distinct) → `game-music.md`
- Media / commercial music (advertising, podcasts, trailers, library) → `media-and-commercial-music.md`

# Electronic and EDM

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers the broad family of electronic and dance music — from early synthesizer pioneers to modern EDM, ambient, and experimental electronic music. The variety is enormous; this file maps the territory and identifies the key conventions of major subgenres.

For form (build, drop, breakdown), see `form/popular-song-forms.md`. For groove and rhythmic conventions, see `rhythm-groove/groove-and-feel.md`. For arrangement density, see `orchestration/arrangement-density.md`.

## When to consult this file

- The user is producing electronic/dance music
- Identifying conventions of a specific subgenre (house, techno, dubstep, trance, etc.)
- Understanding tempo, structural, and sonic conventions
- Cross-genre work involving electronic elements

## Electronic music — broad categories

Electronic music encompasses three rough categories:

1. **Dance music (EDM and predecessors)**: meant for the dancefloor. House, techno, trance, drum and bass, dubstep, etc.
2. **Listening / experimental electronic**: not primarily for dancing. Ambient, IDM, glitch, electro-acoustic, etc.
3. **Electronic-influenced popular music**: pop, rock, hip-hop with electronic elements. Synth-pop, electro-pop, electronic R&B, etc.

This file focuses primarily on (1), with sections on (2) and references to (3) elsewhere.

## Early electronic music (c. 1948–1975)

Before commercial EDM existed, electronic music was an experimental art form.

**Musique concrète (1948–onward)**: Pierre Schaeffer, GRM in Paris. Music made from manipulated recordings of real-world sounds.

**Elektronische Musik (1950s)**: Stockhausen, Cologne studio. Synthesized sounds from oscillators and tape manipulation.

**Tape music**: cut-and-splice composition with magnetic tape.

**Early synthesizers**: Moog, Buchla, Arp. Wendy Carlos's *Switched-On Bach* (1968) brought synthesizers to popular awareness.

**Krautrock (early 70s)**: Kraftwerk, Tangerine Dream, Klaus Schulze, Can. German rock with electronic instruments and ambient leanings.

**Prog electronic**: Vangelis, Jean-Michel Jarre. Electronic albums for listening.

## Disco (c. 1972–1980)

The first dance music with strong electronic influence. Club-driven, with DJs becoming central.

**Idiomatic features**:
- 4-on-the-floor kick
- 110–125 BPM
- Strings, horns, syncopated bass
- Prominent hi-hat 8ths or 16ths
- Singer-driven, often with backing vocal arrangements

**Artists**: Donna Summer, the Bee Gees, Chic, Earth Wind & Fire, Saturday Night Fever soundtrack.

Disco evolved into Italo disco, hi-NRG, and gave birth to house and freestyle.

## House (c. 1984–onward)

Born in Chicago (Frankie Knuckles, the Warehouse club). The foundational genre of modern dance music.

**Idiomatic features**:
- 4-on-the-floor kick
- 118–135 BPM
- Drum machine percussion (TR-808, TR-909)
- Soulful vocals (often gospel-derived)
- Funky bass lines (often stepwise or pentatonic)
- 32-bar phrasing
- Long DJ-friendly arrangements (6–8 minutes)

**Sub-traditions**:
- **Chicago house**: gritty, raw (Frankie Knuckles, Marshall Jefferson, Larry Heard)
- **Deep house**: jazzy, soulful, downtempo (Larry Heard "Mr. Fingers")
- **Acid house**: TB-303 squelch (Phuture's "Acid Tracks", classic London/Madchester scene)
- **Garage / UK garage**: swung 16ths, faster, often with R&B vocal samples
- **Tech house**: techno + house hybrid, minimal
- **Tropical house**: midtempo, summery, Kygo-style 2010s mainstream
- **Future house**: 2010s mainstream EDM-house hybrid

## Techno (c. 1985–onward)

Born in Detroit (Juan Atkins, Derrick May, Kevin Saunderson — "the Belleville Three"). More machine-focused than house.

**Idiomatic features**:
- 4-on-the-floor kick
- 120–140 BPM (some subgenres faster)
- Synthesized, futuristic textures
- Less vocal-driven than house
- Hypnotic, repetitive structures
- Sound design as composition

**Sub-traditions**:
- **Detroit techno**: original, soulful, Afrofuturist
- **Berlin techno**: minimal, dark, club-oriented (Hardwax, Berghain scene)
- **Minimal techno**: stripped-down, micro-focused (Richie Hawtin, Plastikman)
- **Industrial techno**: aggressive, distorted (Ancient Methods, Surgeon)
- **IDM-adjacent techno**: experimental, listening-focused

## Trance (c. 1992–onward)

Emerged from German techno scene. Melodic, euphoric, club-oriented.

**Idiomatic features**:
- 130–150 BPM (sometimes slower for vocal trance)
- 4-on-the-floor kick
- Long build-ups (32+ bars)
- Big, emotional drops with arpeggiated synths
- Melodic, often diatonic chord progressions (i-VI-III-VII or similar)
- Vocal trance subgenre features female vocals

**Artists**: Paul van Dyk, Tiësto (early), Armin van Buuren, Above & Beyond, ATB.

## Drum and bass / jungle (c. 1992–onward)

Fast breakbeat-driven UK dance music.

**Idiomatic features**:
- 160–180 BPM
- Sampled "Amen break" or similar breakbeats (often heavily sliced and rearranged)
- Sub-bass-heavy bass lines
- Dub and reggae influences
- Half-time feel often present (snare on 3 of half-time = 3 of full time)

**Sub-traditions**:
- **Jungle (early 90s)**: ragga vocals, dub influence, raw breaks
- **Drum and bass (mid 90s onward)**: more polished, often technical
- **Liquid funk / liquid drum and bass**: melodic, jazzy
- **Neurofunk**: dark, technical, sci-fi
- **Jump-up**: party-oriented, simple basslines
- **Dubstep precursors**: 2-step garage and breakbeat

## UK garage (c. 1995–onward)

UK speed garage, 2-step garage, and post-garage developments.

**Idiomatic features**:
- ~135–140 BPM
- Swung 16th-note hi-hats
- Skippy, syncopated drum patterns
- Sub-bass-heavy
- 2-step: emphasis on 1 and 3 with snare displacement

## Dubstep (c. 2003–onward)

Originated in UK (Croydon) from garage and dub influences. Later split into UK underground (still close to roots) and brostep (commercial American/EDM dubstep).

**Idiomatic features**:
- 138–142 BPM (felt as half-time)
- Snare on 3 (half-time feel)
- Sub-bass with "wobble" (LFO-modulated low-frequency synth bass)
- Sparse, often spacious arrangements (UK style)
- Modulated bass design as primary musical content
- "Drop" with intense bass design (especially brostep)

**Artists**: Skream, Benga, Kode9, Burial (UK underground); Skrillex, Excision (brostep).

## Future bass / chillstep / vaporwave (2010s)

A loose family of genres mixing dubstep, trap, and emotional/atmospheric production.

**Future bass features**:
- Pitched-up vocal samples
- Saw-wave lead synths with side-chained pumping
- Trap-influenced drums
- Heavy use of arpeggiators and pitch automation
- Emotional, often nostalgic feel

**Artists**: Flume, Cashmere Cat, ODESZA, San Holo.

## Trap (electronic, distinct from hip-hop trap)

Hip-hop trap (covered in `genres/hip-hop-rnb.md`) crossed into EDM around 2012–2014. EDM trap producers borrow trap drum patterns and 808s but apply them to instrumental dance tracks.

**Artists**: RL Grime, Baauer ("Harlem Shake"), Flosstradamus, Diplo (some work).

## Big room / EDM (mainstream, c. 2010–2018)

The mainstream EDM era. Massive festivals and stadium-sized tracks.

**Idiomatic features**:
- 128 BPM
- Big lead synths (often saw-wave)
- Long build-ups with snare rolls and risers
- Massive, often simple drops
- 4-on-the-floor kick

**Artists**: Avicii, Tiësto (later career), David Guetta, Calvin Harris, Hardwell, Martin Garrix, Zedd.

## Tech house, deep house, future house (modern mainstream)

After big room peaked, mainstream EDM shifted toward deeper, more grooveable styles.

**Tech house artists**: Fisher, Chris Lake, CamelPhat. 124–128 BPM. Minimal but groovy.

**Deep house in mainstream**: Disclosure ("Latch"), CamelPhat, MK. Vocal-driven, soulful.

**Future house artists**: Don Diablo, Tchami, Oliver Heldens. Hybrid of house and big room.

## Hardstyle, gabber, hardcore, frenchcore

Faster, harder dance subgenres. 150–250 BPM. Distorted kicks, aggressive sound design. Mostly Northern European in origin and audience.

## Ambient (c. 1978–onward)

Brian Eno coined "ambient" with *Music for Airports* (1978). Music designed to be listened to peripherally; atmosphere over rhythm.

**Idiomatic features**:
- Often beat-less or very slow rhythm
- Sustained pads, drones, evolving textures
- Long durations (10+ minute tracks common)
- Generative or process-based composition often
- Spatial design (reverb, delay, panning) as primary musical content

**Artists**: Brian Eno, Aphex Twin (Selected Ambient Works), Stars of the Lid, Tim Hecker, Grouper, William Basinski, Hiroshi Yoshimura.

**Sub-traditions**:
- **Dark ambient**: brooding, often with field recordings
- **Drone**: sustained, slowly-evolving
- **New age**: more melodic, often spiritual
- **Ambient techno**: ambient with techno-derived rhythmic elements
- **Lowercase / quiet music**: extremely soft, attention-demanding

## IDM (Intelligent Dance Music, c. 1992–onward)

A label coined for early Aphex Twin / Autechre — electronic music too complex/idiosyncratic for the dancefloor.

**Idiomatic features**:
- Complex rhythms (often non-quantized, polyrhythmic, or breakbeat-derived)
- Glitchy, cut-up sound design
- Melodic experimentation
- Listening rather than dancing focus

**Artists**: Aphex Twin, Autechre, Squarepusher, Boards of Canada, μ-Ziq, Plaid, Oneohtrix Point Never.

## Glitch / clicks-and-cuts

Electronic music using digital errors as compositional material.

**Artists**: Oval, Fennesz, Pole, Alva Noto, Ryoichi Kurokawa.

## Synthwave / retrowave (c. 2010–onward)

Nostalgia for 80s film soundtracks and synth-pop. Online, especially streaming-friendly.

**Idiomatic features**:
- 80s synth palette (Yamaha DX7, Juno, etc.)
- Driving 4-on-the-floor or breakbeat
- Major-key melodic content with reverb
- Cinematic feel

**Artists**: Kavinsky, Carpenter Brut, Mitch Murder, Power Glove.

## Vaporwave / mallsoft

Internet-born genre. Slowed-down 80s pop and muzak samples, ironic nostalgia.

**Artists**: Macintosh Plus, Saint Pepsi, James Ferraro.

## Production conventions for EDM

EDM is production-driven. Composition is inseparable from sound design and arrangement.

### Sound design

- **Synthesizers**: Serum, Massive, Sylenth (popular soft synths). Hardware: Moog, Roland, Korg.
- **Drum machines / samplers**: Maschine, MPC, Battery.
- **Sample libraries**: Splice, Loopmasters; large genre-specific sample packs.
- **Sound design as composition**: many EDM tracks' "main hook" is a sound, not a melody. The lead synth's timbre is the song's identity.

### DAW-centric workflow

- Most EDM is produced in DAWs (Ableton, FL Studio, Logic, Cubase).
- Arrangement is constructed visually; loops are built and combined.
- Side-chain compression on bass and pads (against kick) creates "pumping" — a signature EDM technique.
- Heavy use of automation (filter sweeps, reverb sends, etc.) for build-ups.

### Mixing for clubs vs. listening

- Club tracks are mixed for big systems with sub-bass — kicks need to thump, bass needs to be felt.
- Radio edits often have less low-end emphasis.
- Mastering levels are typically loud (heavy compression and limiting).

## EDM form and structure

Covered in `form/popular-song-forms.md`. Brief recap:

```
Intro (16-32 bars)
Build 1 (16 bars)
Drop 1 (16-32 bars)
Breakdown (16-32 bars)
Build 2 (16 bars)
Drop 2 (16-32 bars; often varied)
Outro (16 bars)
```

Long version: 5–7 minutes. Radio edit: 3–4 minutes. DJ-friendly intros and outros (long, beat-only) facilitate mixing.

## Common pitfalls when advising

- **Don't generalize across EDM subgenres.** Trance is melodic; techno is hypnotic; dubstep is bass-driven; ambient is atmospheric. The sub-genre dictates everything.
- **Don't ignore tempo.** Each subgenre has a tempo range; deviating from it usually doesn't make sense within the genre.
- **Don't underrate sound design.** In EDM, the synth patches are the song's musical content as much as the notes.
- **Don't apply pop song-structure to all EDM.** A 7-minute techno track has different pacing than a pop song — a single idea sustained for many bars is intentional, not a flaw.
- **Don't forget the dancefloor function for dance subgenres.** Tracks for clubs serve a function (DJ mixing, dancing). Listening-focused EDM (IDM, ambient) has different priorities.
- **Don't ignore the cultural context.** EDM scenes (rave, club) have deep cultural histories; understanding the context illuminates the aesthetic.

## Quick decision matrix

| Want | Try |
|------|-----|
| Dancefloor classic | House, 124 BPM, 4-on-the-floor, soulful vocal, repetitive structure |
| Hypnotic, machine-driven | Techno, 128-130 BPM, minimal melodic content, emphasis on sound design |
| Euphoric, melodic dance | Trance, 138 BPM, big build-up, big melodic drop, arpeggios |
| Bass-heavy, half-time | Dubstep, 140 BPM, snare on 3, modulated sub-bass |
| Fast, breakbeat-driven | Drum and bass, 174 BPM, sampled breaks, sub-bass |
| Emotional, modern dance pop | Future bass, pitched-up vocal samples, side-chained saws, trap drums |
| Mainstream EDM festival | Big room, 128 BPM, simple drop, big lead synth |
| Ambient / atmospheric | Ambient, beat-less or slow, sustained pads, long form |
| Experimental / glitchy | IDM, complex rhythms, cut-up sound design |
| 80s nostalgia | Synthwave, 80s synth palette, 4-on-the-floor, reverb-heavy |
| Tech house / modern groove | 124-128 BPM, minimal, groovy bass, vocal samples |

## Cross-references

- EDM form (build, drop, breakdown) → `form/popular-song-forms.md`
- Energy arcs and transitions in EDM → `form/narrative-and-transitions.md`
- Groove and microtiming in electronic → `rhythm-groove/groove-and-feel.md`
- Arrangement density (key for EDM) → `orchestration/arrangement-density.md`
- Modal harmony in trance and ambient → `harmony/modal-harmony.md`
- Production-side considerations → `production-aware/arrangement-for-mix.md`, `production-aware/energy-and-dynamics.md`
- Hip-hop trap (related to EDM trap) → `genres/hip-hop-rnb.md`
- Synth pop and electronic-influenced pop → `genres/pop-rock.md`

# Game Music and Interactive Scoring

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition for video games and other interactive media. Game music shares much DNA with film scoring — both are "music for picture", broadly speaking — but interactivity introduces specific compositional problems that don't exist in linear media. The score has to work without knowing what will happen next, has to loop or stretch indefinitely without becoming tedious, and has to respond to gameplay state in real time.

This file complements `film-tv-scoring.md` (linear media scoring; much of which transfers) and `../techniques/algorithmic-and-AI-assisted.md` (some game music is procedurally generated). For specific genre conventions a game's music draws on (orchestral, electronic, jazz, etc.), see the relevant `genres/*.md` file.

## When to consult this file

- Composing for a video game (any platform, any era)
- Designing music systems for interactive media (XR, interactive installations, themed entertainment)
- Adapting linear music for game contexts
- Working with adaptive / dynamic music middleware (Wwise, FMOD, Unity's audio system, etc.)
- Music for game trailers and promotional materials (which is usually closer to film/trailer scoring; see `film-tv-scoring.md`)

## What's different about game music

The defining constraint: **the composer doesn't know how long the music needs to play, what will happen next, or what the player will do**. Linear media tells you exactly when each cue starts and ends; games don't.

Practical consequences:

| Constraint | Implication |
|------------|-------------|
| Indefinite duration | Music must loop seamlessly or be designed to play indefinitely without obvious repetition |
| Unpredictable transitions | Music must transition smoothly from any current state to any plausible next state |
| Player agency | The player can stop, restart, repeat, skip — music must serve all paths |
| Performance constraints (older platforms) | Memory, CPU, and audio channels limit what can be heard |
| Adaptive systems | Music must respond to gameplay events, parameters, or state changes |
| Replay tolerance | Players hear the same music many times; it must reward repeated listening |

These shape every compositional decision.

## A short history of game music

Worth knowing because compositional conventions are rooted in technological eras:

### 1970s–early 1980s: arcade and earliest console (TIA, SN76489, custom chips)

Limited polyphony (often 1–3 channels), simple waveforms (square, triangle, noise). Music was often short loops or stinger cues, with sound effects sharing the audio channels. *Pac-Man*, *Donkey Kong*, *Space Invaders* — minimal "music" by modern standards but distinctive sonic identity.

### Late 1980s–early 1990s: chiptune era (NES, Game Boy, SMS, MSX)

Slightly expanded chip capabilities (NES: 5 channels — 2 pulse, 1 triangle, 1 noise, 1 DPCM). Composers like Koji Kondo (Mario, Zelda), Nobuo Uematsu (Final Fantasy), Yuzo Koshiro (Streets of Rage) wrote sophisticated multi-voice music within these constraints. The chiptune aesthetic — pulse waves, simple waveforms, characteristic "8-bit" sound — is now an entire genre revived in indie games and electronic music.

### Early-mid 1990s: PCM and FM synthesis era (SNES, Genesis, NeoGeo)

PCM samples on SNES (8-channel sample playback) and FM synthesis on Genesis (Yamaha YM2612). More instrument-like sounds; orchestral mockups become possible. Composers like Yasunori Mitsuda (Chrono Trigger, Chrono Cross), Michiru Yamane (Castlevania: Symphony of the Night), Hitoshi Sakimoto (Ogre Battle, FFXII) develop complex compositional language.

### Late 1990s–2000s: CD audio and live recording (PlayStation, PS2, Xbox)

CD-quality audio. Live orchestras enter game music (Final Fantasy VIII orchestra, Halo's choral writing). Streaming audio means hour-long detailed scores become possible. Game music begins to receive concert hall performances (Final Fantasy concerts, Distant Worlds tours).

### 2000s–onward: middleware and adaptive systems

Wwise, FMOD, Unity, Unreal — middleware systems that allow composers to design music as a system rather than fixed cues. Vertical layering, horizontal re-sequencing, parameter-driven music become standard. Composers like Jesper Kyd (Hitman, Assassin's Creed), Austin Wintory (Journey), Darren Korb (Bastion, Hades), Lena Raine (Celeste) make adaptive systems central to their craft.

### 2010s–onward: indie diversity

Indie games push game music in many directions: live-band recording, retro-chiptune revival, ambient drone, complex jazz (Persona 5), folk and traditional music (Hyper Light Drifter), full orchestral (God of War, Horizon).

## Looping — the basic technique

The simplest interactive music: a piece that loops seamlessly until something else happens. Compositional techniques for good loops:

### Loop point design

A loop must start and end without an audible "seam". Standard approaches:

- **Loop point coincides with a downbeat**: end of phrase aligns with start of phrase, both on a strong beat. Most common in classical-style game music.
- **Loop point at a pickup**: the piece ends just before a downbeat; the loop begins at the pickup; perfect continuity. Common in orchestral game music.
- **Crossfade**: an audio engine fades out the end as it fades in the start. Less common in modern game music (modern audio engines support sample-accurate looping).
- **Tail-into-head matching**: the last beat's content (final note, final chord, final percussion hit) is identical to the first beat's content. Crude but works.

### Avoiding loop fatigue

A loop the player hears for 20 minutes will exhaust them if it doesn't accommodate repetition.

- **Long loops**: 4–8 minutes per loop is more forgiving than 30 seconds; the listener doesn't immediately re-recognize.
- **Variation within the loop**: the loop itself contains internal variety (verse-chorus-bridge structure), so the listener doesn't experience the same 4 bars repeating.
- **Subtle, slow changes**: layered ambient music with elements drifting in and out, never quite repeating exactly. Brian Eno-style ambient is well suited to game contexts.
- **Stinger insertions**: occasional short "stingers" (8–32 bars of new material) interrupt the loop without changing it permanently.
- **Looping with start-of-day variations**: each time the loop starts fresh (new area, new visit), it begins with different material before settling into the main loop body.

### Length conventions by context

| Context | Typical loop length |
|---------|---------------------|
| Menu / title screen | 1:30–4:00 |
| Town / hub area | 2:00–6:00 |
| Combat / encounter | 1:30–4:00 |
| Boss battle | 2:00–5:00 |
| Field / exploration | 2:00–8:00 (longer for areas players spend time in) |
| Cutscene | Length-of-cutscene; often non-looping |
| Death / failure | 5–15 seconds (a stinger, not a loop) |
| Victory | 10–60 seconds (often non-looping) |

## Vertical layering

A widely-used adaptive technique: the music is composed as a layered stack (e.g., drums, bass, harmonic pad, melody, decorative percussion) and the engine adds or removes layers based on game state.

### Typical layered structure

| Game state | Active layers |
|------------|---------------|
| Quiet exploration | Pad + sparse melodic figure |
| Approaching enemies | Pad + figure + soft drums |
| Active combat | Pad + figure + drums + bass + intense melody |
| Boss encounter | All layers + extra brass / tension elements |

The layers are written together, in the same key, tempo, and form, so any combination sounds coherent. Layers fade in / out smoothly; transitions are imperceptible.

### Compositional rules for layered music

- **Each layer is independently coherent**: removing all other layers, a single layer should still make musical sense
- **Combined layers don't clash**: the layers harmonically and rhythmically combine without muddying the texture
- **Common harmonic frame**: all layers share the same chord progression and key (variations within that frame are fine)
- **Per-layer sonic separation**: layers occupy distinct frequency bands or stereo positions to maintain clarity when stacked
- **Smooth fade behavior**: each layer's start / stop sounds natural with fade in / out (no abrupt entrances or cuts)

This technique was used heavily in *No Man's Sky* (procedural generative score by 65daysofstatic), *Journey* (Austin Wintory), *Disco Elysium* (Sea Power), and many others.

## Horizontal re-sequencing

Different from vertical layering: the music is composed as multiple distinct sections that can be sequenced in different orders based on game state.

| Section | When played |
|---------|-------------|
| A | Default exploration |
| B | Light tension / nearby threat |
| C | Combat |
| D | Boss / climax |
| Transition cues | When state changes between sections |

The composer writes:
1. Each section as a self-contained loop
2. Short transition cues between adjacent sections (e.g., A→B, B→A, B→C, C→D)
3. The middleware plays section X until state changes, then plays the X→Y transition, then plays section Y

This approach gives sharper transitions than pure vertical layering. *Halo* (Marty O'Donnell) used horizontal re-sequencing extensively. Many modern AAA games use a hybrid of vertical layering and horizontal re-sequencing.

## Parameter-driven music

The most dynamic approach: the music is shaped by continuous game parameters, not just discrete state changes.

Examples:
- **Combat intensity**: a 0–1 parameter that fades layers, raises tempo, shifts harmony as combat heats up
- **Health**: as the player's health drops, the music adds tension (filter sweeps, dissonant elements, narrowing palette)
- **Time of day**: the music shifts smoothly across day/night cycles in an open-world game
- **Distance from objective**: tension layers fade in as the player approaches a key location

Implementation requires music designed with these parameters in mind — usually layered or modular music plus middleware logic that maps parameters to audio events.

## Stingers and oneshots

Discrete, non-looping musical events:

- **Stingers**: short cues for specific moments (item pickup, level-up, boss defeat, character introduction)
- **Hits**: percussive accents tied to gameplay (impact, damage, flourish on level transition)
- **Rules-based playback**: rare cues triggered by specific player actions (achievement unlocks, secret discovery)

These are non-adaptive but contextual. They interrupt or layer over background music.

## Diegetic vs. non-diegetic

As in film:

- **Diegetic**: music heard *in* the game world (radios, environmental music, NPCs singing). Often follows real-world conventions because it represents real-world music sources.
- **Non-diegetic**: music heard *over* the game world (the score). Adheres to game music conventions of looping, layering, etc.
- **Crossover**: music that appears to be diegetic but acts as score (the radio in *BioShock* — period jazz that sets atmosphere but operates as adaptive score), or score that responds as if diegetic (music that ducks when characters speak).

The distinction matters because diegetic music has different technical requirements (no looping required, real-world fidelity, often licensed music) and different audience expectations.

## Game music genre by era and genre

A rough genre map:

| Game type | Typical music idiom |
|-----------|----------------------|
| JRPG (Japanese RPG, e.g., Final Fantasy, Persona, Chrono Trigger) | Symphonic / orchestral with electronic elements; long memorable themes; character-specific motifs |
| WRPG (Western RPG, e.g., Skyrim, Witcher, Mass Effect) | Symphonic, often choral; folk-influenced for fantasy; sci-fi orchestral for space |
| Action / shooter (Halo, Call of Duty) | Cinematic orchestral with electronic edge; combat-driven layered systems |
| Stealth (Hitman, Splinter Cell) | Tense ambient with industrial or electronic elements; minimal during stealth, intense on detection |
| Indie narrative (Hades, Celeste, Disco Elysium) | Genre-experimental; specific stylistic identity often the game's signature |
| Platformer (Mario, Sonic, Celeste) | Catchy, melodic, often chiptune-influenced; level-themed music |
| Puzzle / casual (Tetris, candy-style) | Catchy short loops; light orchestration; non-distracting |
| Horror (Resident Evil, Silent Hill) | Dissonant, sparse, ambient with sudden stingers; often microtonal and noise-influenced |
| Racing (Forza, Need for Speed) | Often licensed contemporary music; original score for menus and key moments |
| Sports (FIFA, Madden) | Often licensed contemporary music; original score minimal |
| Simulation (Sim City, Civilization) | Long ambient or orchestral pieces; minimal interactivity |
| Open world (GTA, Red Dead) | Mix: licensed radio for diegetic; original score for missions and key moments |
| MOBA / multiplayer (LoL, DOTA, Overwatch) | Match-tempo orchestral; minimal during gameplay, intense at match milestones |
| Mobile / hyper-casual | Short catchy loops; designed to be muted |

## Specific composers worth studying

- **Koji Kondo** (Nintendo): foundational chiptune-era melodicism; *Super Mario*, *Zelda* themes
- **Nobuo Uematsu** (Square Enix): symphonic JRPG score; *Final Fantasy* series
- **Yasunori Mitsuda**: distinctive melodic / cinematic; *Chrono Trigger*, *Chrono Cross*, *Xenogears*
- **Hitoshi Sakimoto**: dense orchestral; *FFXII*, *Vagrant Story*, *Final Fantasy Tactics*
- **Yoko Shimomura**: melodic and emotional; *Kingdom Hearts*, *Street Fighter II*, *Xenoblade Chronicles*
- **Marty O'Donnell**: cinematic horizontal-resequencing; *Halo* trilogy
- **Jeremy Soule**: orchestral fantasy; *Skyrim*, *Guild Wars*, *Knights of the Old Republic*
- **Jesper Kyd**: electronic / orchestral hybrid; *Hitman*, *Assassin's Creed* (early)
- **Austin Wintory**: modern adaptive scoring; *Journey*, *Abzû*, *Banner Saga*
- **Darren Korb**: indie modern; *Bastion*, *Transistor*, *Hades*
- **Lena Raine**: indie modern; *Celeste*, *Chicory*, *Minecraft Caves & Cliffs*
- **Olivier Deriviere**: French modern; *A Plague Tale*, *11-11 Memories Retold*, *Streets of Rage 4*
- **Gareth Coker**: orchestral / world; *Ori and the Blind Forest*, *Ark*, *Halo Infinite*
- **Joris de Man**: orchestral / electronic hybrid; *Killzone* series, *Horizon Zero Dawn* (with Joe Henson, Alexis Smith)
- **Mick Gordon**: heavy industrial / metal; *Doom (2016)*, *Doom Eternal*, *Wolfenstein II*
- **Yuzo Koshiro**: chiptune / synth dance; *Streets of Rage*, *Actraiser*
- **65daysofstatic**: post-rock / electronic; *No Man's Sky* (procedural generative)
- **Toby Fox**: indie melodic chiptune-influenced; *Undertale*, *Deltarune*

Studying these composers' approaches to specific games (and the games' systems) is the best way to learn game music craft.

## Korean game music

Korean MMORPG and mobile game music has its own significant tradition: composers like Kim Jin-young (Black Desert Online), the team at NCsoft (Lineage, Aion, Blade & Soul), and others have developed a distinctive Korean game music idiom that often blends symphonic Western fantasy with Korean traditional elements (*gugak* instruments, traditional vocal techniques). For Korean traditional music itself, see `korean-traditional.md`.

## Composing game music — practical workflow

1. **Understand the game.** Read the GDD (game design document) if available; play prototypes; understand mood, world, character, gameplay loop.
2. **Identify the music architecture.** Is it linear cues only? Vertical layering? Horizontal re-sequencing? Parameter-driven? Hybrid? Talk to the audio director / programmer early.
3. **Identify the technical pipeline.** Wwise, FMOD, Unity native, Unreal native, custom? Each has different constraints.
4. **Establish thematic language.** Main theme, character themes, location themes, faction themes — what motivic material recurs?
5. **Compose with the system in mind.** Looping, layering, transitions, stingers — design each piece to fit its role.
6. **Test in-engine early.** A piece that sounds great in your DAW might not work in the game's audio system. Get rough mockups into the engine ASAP.
7. **Iterate based on gameplay testing.** Music that works in design rarely works perfectly first time in actual gameplay; expect revisions.
8. **Mix for the platform.** Game audio often has a wide dynamic range with sound effects and dialogue mixed alongside; the music has to leave room.

## Common pitfalls when advising

- **Don't compose linear cues for an adaptive context.** A 4-minute through-composed piece doesn't loop or layer well; design for the architecture.
- **Don't ignore loop fatigue.** A piece the player will hear for hours needs to tolerate repetition; one the player hears once doesn't.
- **Don't treat game music as second-class film music.** Game music has its own craft (interactivity, system design, thematic recursion in long-form gameplay) that linear scoring doesn't.
- **Don't underestimate technical constraints.** Old platforms have strict audio limits; modern platforms have memory and CPU budgets. Talk to the audio director.
- **Don't forget the sound effect mix.** Game music shares audio bandwidth with SFX, dialogue, ambience. Music has to leave room.
- **Don't ignore the platform.** Music for a phone played without headphones, music for a 7.1 surround AAA game, music for a VR title with binaural audio — all different.
- **Don't ignore replay value.** The player will hear the music many times; it must reward repeated listening (or at least tolerate it).
- **Don't compose only the "good" music.** Filler areas, menus, loading screens, defeat screens — these need music too, and tone matters for them just as much.
- **Don't apply film scoring conventions blindly.** Film cues have known durations and known emotional arcs; game cues don't. Adapt accordingly.
- **Don't forget themes.** Long-form games (RPGs, narrative adventures) reward strong recurring thematic material. The music can be the player's emotional anchor across hours of play.

## Quick decision matrix

| Want | Try |
|------|-----|
| Tight combat music | Driving, layered, with intensity parameter; quick transitions on encounter / disengagement |
| Atmospheric exploration | Long-loop ambient with subtle internal variation; vertical layering for "approaching threat" tension |
| Boss battle | Larger orchestration, distinctive theme (often related to character / story), 2–4 minute loop |
| JRPG town | Memorable melodic theme, 2–4 minute loop, full orchestration appropriate to setting |
| Stealth | Sparse, tense ambient; spike to action layer on detection |
| Horror | Sparse / silent; tension-building drones; sudden stingers on encounter |
| Indie game with distinctive music identity | Establish a specific stylistic palette (chiptune, jazz, folk, electronic, world) and let it permeate the score |
| Procedural / endless game | Ambient or layered music designed for indefinite playback; avoid prominent melodic recurrence |
| Mobile / casual game | Catchy short loops; non-distracting; assume player may mute |
| Diegetic music | Real-world music conventions for in-world sources; consider licensing existing tracks |
| Highly adaptive AAA game | Plan vertical layering + horizontal re-sequencing + parameter-driven elements; talk to audio director from start |
| Korean MMORPG / fantasy | Symphonic Western with optional *gugak* fusion elements |

## Cross-references

- Linear / film scoring (much technique transfers) → `film-tv-scoring.md`
- Algorithmic and AI-assisted composition (procedural game music) → `../techniques/algorithmic-and-AI-assisted.md`
- Korean traditional fusion (relevant to Korean game music) → `korean-traditional.md`
- Orchestration for game ensembles → `../orchestration/instruments-ranges-character.md`, `../orchestration/voicing-and-texture.md`, `../orchestration/arrangement-density.md`
- Energy and dynamics across sections → `../production-aware/energy-and-dynamics.md`
- Form and narrative (long-form games are narrative structures) → `../form/narrative-and-transitions.md`
- Thematic / motivic development (recurring themes are central) → `../melody/motivic-development.md`, `../techniques/theme-and-variation.md`
- Specific genres a game might draw on (jazz, electronic, folk, etc.) → corresponding `genres/*.md`
- Pre-production decisions for adaptive scoring → `../production-aware/pre-production-decisions.md`

# Folk, Roots, and Tradition-Specific Writing

> **Snapshot note:** Last content review snapshot: 2026-04-27. This file is a composition-facing orientation, not an authoritative treatment of any tradition. For current scenes, living artists, or region-specific practice, verify with active web research and practitioner sources.

This file replaces broad "world music" framing with tradition-specific thinking. Use it when the user wants folk, roots, traditional, or tradition-informed composition and no more specific file exists yet.

Use this file when:

- the user asks for folk/roots writing in a named tradition
- the user wants acoustic or oral-tradition influence
- the user asks for fusion with a traditional source
- the user says "world music" but has not specified a tradition
- the project needs cultural specificity but not a full ethnomusicology lesson

Do not use this file as a license to imitate a culture from surface markers. Use it as a triage and caution framework.

## Core principle

Name the source tradition, scene, region, language, function, or ensemble whenever possible.

Weak:

```text
Add a world music instrument.
```

Better:

```text
Do you mean Irish reel energy, Appalachian ballad storytelling, West African percussion layering, Hindustani raga-informed melody, Balkan asymmetrical dance meter, or something else?
```

## First questions

Ask only what is needed:

1. Which tradition or region is the reference?
2. Is the goal respectful influence, collaboration, or accurate style writing?
3. Is this for song, film/game cue, dance, ceremony, live performance, or abstract composition?
4. Can the user provide one reference recording?

If the user cannot answer, offer a menu of broad lanes rather than guessing.

## Tradition-specific variable map

| Variable | Why it matters |
|---|---|
| Function | dance, work song, ritual, storytelling, lullaby, protest, worship, social gathering |
| Transmission | oral tradition, notation, improvisation, session practice, conservatory tradition |
| Scale/mode | useful, but rarely sufficient by itself |
| Rhythm/cycle | often more defining than pitch material |
| Ornament | may be central to identity and hard to notate |
| Language/prosody | melody may follow speech stress, tone, vowel, or poetic meter |
| Ensemble idiom | instruments imply technique, register, tuning, and social context |
| Form | strophic, call-response, variation, dance set, improvisatory unfolding |
| Collaboration | practitioner input may be necessary for serious use |

## Common folk/roots lanes

### Anglo-American / Celtic / Appalachian-adjacent folk

Composition variables:

- modal color: Mixolydian, Dorian, Aeolian, pentatonic, major/minor blends
- strophic form and story-forward lyrics
- fiddle/banjo/guitar/mandolin/whistle/accordion colors depending on lane
- drones, open strings, simple harmonic support
- dance meters such as reels/jigs in relevant contexts

Pitfall: treating all Celtic/Appalachian/British Isles traditions as interchangeable.

Pair with `country-americana.md`, `pop-rock.md`, and `../instrument-idiom/guitar.md`.

### Flamenco-adjacent or Iberian folk-informed writing

Composition variables:

- compás and rhythmic cycle are central
- Phrygian or Phrygian-dominant-like colors may appear, but rhythm/guitar/vocal idiom matter more
- handclap/guitar/percussion interaction can define the feel

Pitfall: using `E-F-G` or Phrygian dominant alone and calling it flamenco.

Pair with `latin-pop-and-reggaeton.md` only for pop crossover; otherwise use reference listening and practitioner sources.

### Balkan / Eastern European dance-informed writing

Composition variables:

- asymmetric meters and additive rhythms
- drone or modal centers
- fast melodic ornaments and ensemble-specific articulation
- brass/accordion/strings/voice colors vary widely by region

Pitfall: treating all odd meters as the same cultural signal.

Pair with `../rhythm-groove/odd-meters-polyrhythm.md`.

### West African / diaspora rhythm-informed writing

Composition variables:

- timeline patterns, interlocking parts, call-response, dance function
- percussion ensemble role and vocal participation
- modern pop crossovers may route through Afrobeats, highlife, hiplife, amapiano, local house, or other scenes

Pitfall: turning a complex rhythmic tradition into generic "polyrhythm" language.

Pair with `afrobeats-and-amapiano.md`, `regional-scene-starters.md`, and `../rhythm-groove/groove-and-feel.md`.

### South Asian raga/taal-informed writing

Composition variables:

- raga is not merely a scale; melodic behavior and phrase grammar matter
- tala is not merely a time signature; cycle and accent structure matter
- drone and melodic development can define the frame
- filmi/pop contexts may use these elements loosely or hybridize them

Pitfall: using harmonic functional progressions that fight the intended melodic system.

Pair with `south-asian-film-pop.md`, `../techniques/microtonal.md`, and practitioner references.

### MENA / maqam-adjacent writing

Composition variables:

- maqam/makam practice includes melodic phrases, modulation, intonation, ornament, and aesthetic conventions
- pop contexts may use equal-tempered approximations, synths, club rhythms, and Western harmony
- vocal delivery and ornament often matter more than a chord loop

Pitfall: using Phrygian dominant as a one-size-fits-all marker.

Pair with `mena-pop.md`, `regional-scene-starters.md`, and `../techniques/microtonal.md`.

### East Asian traditional-informed writing

Composition variables:

- specify Chinese, Korean, Japanese, Mongolian, Vietnamese, etc.; do not use continent-scale shorthand
- pentatonic materials may appear, but ornament, instrument idiom, tuning, and phrase logic matter
- K-trad/gugak has a dedicated file

Pitfall: pentatonic melody + one plucked instrument = stereotype.

Pair with `korean-traditional.md`, `cpop-and-southeast-asian-pop.md`, and relevant research files.

## Fusion handling

For fusion, decide what each layer contributes.

```md
| Layer | Role | Example |
|---|---|---|
| Primary form | what listener calls the piece | pop song / film cue / dance track |
| Traditional source | rhythm, melodic behavior, instrument idiom, text, function | named tradition/source |
| Modern frame | drums, bass, harmony, production, arrangement | current genre lane |
| Collaboration need | where expertise is required | vocalist, instrumentalist, language consultant |
```

A respectful fusion often keeps only one or two traditional variables and handles them deeply rather than sprinkling many surface signs.

## How to answer vague "world music" requests

```text
That label is too broad to be musically useful. We can make this tradition-specific. Do you want something closer to Irish/Appalachian folk, MENA maqam-adjacent color, West African rhythmic layering, South Asian raga-informed melody, Balkan odd-meter dance energy, or another named source?
```

If the user wants a quick sketch anyway:

```text
I'll avoid claiming authenticity and make a general acoustic roots/fusion sketch. For a specific tradition, we should pick a reference and adjust the rhythm, ornament, language, and instrument idiom.
```

## Cross-references

- Legacy route → `folk-and-world.md`
- Regional starter map → `regional-scene-starters.md`
- Minor/microgenre handling → `minor-and-hybrid-genres.md`
- Korean traditional → `korean-traditional.md`
- MENA pop → `mena-pop.md`
- South Asian film-pop → `south-asian-film-pop.md`
- Afrobeats/amapiano → `afrobeats-and-amapiano.md`
- Brazilian pop/funk → `brazilian-pop-and-funk.md`
- Prosody/language → `../fundamentals/prosody-and-language.md`
- Microtonality → `../techniques/microtonal.md`
- Reference digging → `../research/reference-track-digging.md`

# Hip-Hop and R&B

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

Hip-hop and R&B are deeply interrelated traditions in African-American popular music — distinct genres but constantly cross-pollinating. This file covers both, their subgenres, harmonic and rhythmic conventions, and the elements that distinguish them from pop and rock.

## When to consult this file

- The user is writing in hip-hop or R&B
- Producing beats, vocal lines, or arrangements in these styles
- Understanding the rhythmic and harmonic foundations
- Cross-genre work involving these traditions

## Hip-hop — origins and elements

Hip-hop emerged in the South Bronx in the early 1970s as a culture with four elements: DJing, MCing (rapping), graffiti writing, and breakdancing. Musically, hip-hop is built on:

1. **The beat**: drum patterns + bass + samples and/or original instrumentation. The rhythmic and harmonic foundation.
2. **The vocal**: rapping (rhythmic speech to a beat) is the primary vocal mode; melodic singing increasingly common.
3. **The hook**: a memorable chorus or refrain, often sung or chanted.

Verses typically 16 bars; hooks 4–8 bars. See `form/popular-song-forms.md` for hip-hop form.

## Hip-hop subgenres

### Old school / golden age (c. 1979–1989)

- **Old school**: Sugar Hill Gang, Grandmaster Flash, Run-DMC, LL Cool J. Funk samples, drum machine beats, simple rhymes.
- **Golden age (late 80s)**: Public Enemy, Eric B. & Rakim, KRS-One, EPMD, Beastie Boys. More complex sampling, denser production.

**Idiomatic features**: 808 drum machine, James Brown samples, simple rhyme schemes, often minor-key minor-pentatonic samples.

### East Coast (c. 1985–onward)

Lyricism-focused, sample-based.

**Boom-bap (early 90s NY)**: Wu-Tang Clan, Nas, The Notorious B.I.G., Mobb Deep, A Tribe Called Quest, De La Soul.

**Idiomatic features**:
- Boom-bap drum pattern (kick on 1, snare on 2 and 3 — actually on 2 and 4 with heavy kick on 1 and lighter kick on the upbeat)
- Loose, behind-the-beat groove
- Jazz and soul samples
- Dense, multisyllabic rhyme schemes
- 4-bar loops repeated throughout

### West Coast (c. 1989–onward)

**G-funk**: Dr. Dre, Snoop Dogg, Warren G. Funk samples, talk-box, smoother feel than East Coast boom-bap.

**Idiomatic features**:
- Whining synths (Moog leads)
- Heavy bass
- Slower tempos than NY (~85-95 BPM)
- Minor key, often mode-mixed
- Talk-box and vocoder effects

### Southern hip-hop (c. 1990s–onward)

**Houston (Geto Boys, UGK)**, **Atlanta (Outkast, Goodie Mob)**, **Miami bass**, **New Orleans bounce** (Master P, Cash Money). Diverse but emphasizing groove and party feel.

### Trap (c. 2003–onward)

Originally from Atlanta. Now the dominant subgenre globally.

**Idiomatic features**:
- 808 sub-bass with sliding pitches
- 16th-note hi-hat patterns with rolls (32nd-note triplet rolls common)
- Snare on beat 3 (half-time feel)
- Tempos 130–160 BPM (felt as 65–80 in half-time)
- Minimalist, dark melodic content
- Triplet flow rapping (Migos popularized)
- Heavy use of Auto-Tune
- Sparse sound design with cinematic samples

**Artists**: T.I., Gucci Mane, Young Jeezy (originators); Migos, Future, Travis Scott, Drake (modern); 21 Savage, Lil Baby, Roddy Ricch.

### Drill (c. 2012–onward)

Originating in Chicago (Chief Keef), spreading to UK and global. Aggressive subgenre of trap.

**Idiomatic features**:
- Even more sub-bass-heavy
- Sliding 808s with detuning effects
- Tempos around 140 BPM (often half-time)
- Dark, minor-key minor melodic content
- Aggressive lyrics and vocal delivery

### Cloud rap / lo-fi hip-hop / chillhop

**Idiomatic features**:
- Atmospheric, hazy production
- Pitched-down vocals or instrumentals
- Lo-fi, vinyl-crackle textures
- Slower tempos, often 70–85 BPM
- Jazz-influenced harmony

**Artists**: Lil B (proto), Earl Sweatshirt, MF DOOM (lo-fi influences), J Dilla (influence), Madlib, Knxwledge.

### Conscious / political / indie hip-hop

Lyrics-focused, often less commercial in production. Run the Jewels, Kendrick Lamar (transcending categories), Killer Mike, Talib Kweli, Mos Def, and many indie artists.

### Mumble rap / melodic rap

A 2010s shift toward melodic, often Auto-Tuned vocal styles. Future, Lil Uzi Vert, Lil Wayne (precursor), Travis Scott, Post Malone.

The melodic-vs-rapping line has blurred — many modern artists move freely between rapping and singing within a single song.

## Hip-hop rhythmic conventions

### The boom-bap groove

Originally from sampled funk and soul drums. Characterized by:
- Kick on 1 (and possibly off-beats)
- Snare on 2 and 4 (the "bap")
- Hi-hats on 8th notes (or 16ths with selective accent)
- Slightly behind-the-beat feel ("MPC swing")

### The trap groove

- Kick on 1, with additional kicks on tresillo (3+3+2) positions
- Snare on 3 only (half-time feel)
- 16th hi-hats with rolls (32nd, even 64th notes during fills)
- Sub-bass following the kick or with sliding 808s

### The flow

Rappers' "flow" is their rhythmic style — how they place syllables against the beat. Different rappers have different flow signatures. Triplet flows, double-time flows, syncopated flows, on-the-beat flows all coexist.

### Tempo and feel

- Old school / boom-bap: 80–100 BPM
- West Coast G-funk: 85–95 BPM
- Trap: 130–160 BPM (felt as half-time, so 65–80)
- Drill: 140–150 BPM (half-time)
- Cloud rap: 70–85 BPM

## Hip-hop harmonic conventions

Hip-hop harmony is sample-driven and varies enormously:

- **Sample-based**: harmony comes from the source recording (jazz, soul, funk samples).
- **Loop-based**: a 4-bar loop with 1–4 chords, repeated.
- **Trap minimalism**: often a single chord or vamp; sometimes just a bass-and-melody motif with no harmonic motion.
- **Modal**: minor pentatonic, Aeolian, Phrygian (especially drill and dark trap).

Common harmonic patterns:
- i–♭VI–♭VII–v (Aeolian, common in trap)
- i–♭VI in vamp
- i–iv–V7 (with the V7 from harmonic minor)
- Phrygian: i–♭II–i (especially drill)
- Static modal vamps (one chord for entire sections)

## R&B — origins and traditions

R&B (rhythm and blues) developed in the 1940s–50s as African-American popular music, encompassing many styles:

- **Early R&B (40s-50s)**: Louis Jordan, Ruth Brown, Big Joe Turner. Dance-oriented blues with horn sections.
- **Doo-wop (50s-60s)**: vocal group harmony (The Drifters, The Platters, The Shirelles).
- **Soul (60s)**: Aretha Franklin, Otis Redding, Sam Cooke, Marvin Gaye, Stevie Wonder. Gospel-derived vocals over R&B grooves.
- **Funk (late 60s onward)**: James Brown (the originator), Sly & The Family Stone, Parliament-Funkadelic. Heavily syncopated, "on the one".
- **Disco (70s)**: dance-floor R&B with strings, "four-on-the-floor" kick.
- **Quiet storm / smooth R&B (70s-80s)**: Luther Vandross, Anita Baker, Sade. Lush, slow-tempo R&B.
- **New jack swing (late 80s-early 90s)**: Teddy Riley, Bobby Brown, Bell Biv DeVoe. R&B + hip-hop drum programming.
- **90s R&B**: Boyz II Men, Mariah Carey, TLC, Mary J. Blige, R. Kelly, Brandy. Layered vocals, swung 16ths, heavy production.
- **Neo-soul (late 90s-onward)**: D'Angelo, Erykah Badu, Lauryn Hill, Maxwell. Live instrumentation, jazz harmony, retro feel.
- **Modern R&B (2010s-onward)**: Frank Ocean, The Weeknd, SZA, Daniel Caesar, H.E.R., Brent Faiyaz. Genre-blending, often atmospheric.

## R&B harmonic conventions

R&B uses richer harmony than most pop or hip-hop:

- **Extended chords**: 9ths, 11ths, 13ths used liberally
- **Modal mixture**: borrowed iv (minor IV), ♭VII, ♭III, ♭VI in major keys
- **Secondary dominants**: especially V/ii (V7 of ii) for emotional lift
- **Tritone substitutions**: occasional, especially in neo-soul
- **ii-V-I**: extensive jazz harmonic vocabulary
- **Sus chords**: sus4 and sus2 for color
- **Slash chords**: V/3, IV/5, etc., for smooth bass motion
- **Two-chord vamps**: a single i-iv vamp can sustain an entire section

Neo-soul especially borrows jazz voicings — quartal voicings, rootless voicings, complex extensions.

## R&B rhythmic conventions

- **Swung 16ths**: most R&B (especially 90s and onward) uses swing on 16th notes.
- **Heavy syncopation**: vocal lines and rhythm parts heavily syncopated.
- **Ghost notes**: common in drums and bass.
- **Half-time choruses**: many modern R&B songs use half-time feel for sectional contrast.
- **Hip-hop drum programming**: post-90s R&B often uses drum machines and samples like hip-hop.

## R&B vocal conventions

R&B vocals are central. Conventions:

- **Melisma**: stretching one syllable across many notes. Heavy in 90s R&B and gospel-influenced styles.
- **Vocal runs**: rapid melodic improvisations, often at the ends of phrases.
- **Ad-libs**: vocal interjections in the second half of choruses or verses.
- **Vocal layering**: lead + multiple harmonies, doubles, ad-libs, often heavily produced.
- **Auto-Tune (modern)**: pitch correction or stylized auto-tune effects.
- **Whisper / falsetto / chest mix**: skilled R&B singers move freely between registers.

## Production conventions

Hip-hop and R&B production has evolved:

- **Sampling**: early hip-hop and R&B sampled extensively. Now constrained by copyright; many producers craft samples themselves.
- **Drum machines**: 808 (Roland TR-808), 909, MPC samplers — foundational tools.
- **Sub-bass**: low frequencies are central. Trap especially is bass-driven.
- **Side-chain compression**: pumping bass and pads against the kick.
- **Vocal processing**: extensive (compression, EQ, reverb, delay, Auto-Tune, doubling).
- **Beat construction**: typically built up — start with drums and bass, add melodic elements, add atmosphere and effects.

## Form conventions

### Hip-hop form

```
Intro (4-16 bars)
Verse 1 (16 bars)
Hook / Chorus (8 bars)
Verse 2 (16 bars)
Hook (8 bars)
Verse 3 / Bridge (8-16 bars)
Hook (8 bars)
Outro (variable)
```

Modern variations often shorter. Some songs are just 1-2 verses with multiple hook repetitions.

### R&B form

R&B usually follows pop verse-chorus form:

```
Intro
Verse 1 (8-16 bars)
Pre-Chorus (4-8 bars)
Chorus (8-16 bars)
Verse 2 (8-16 bars)
Pre-Chorus
Chorus
Bridge (8 bars, often modulated)
Final Chorus
Outro
```

Modern R&B often features extended bridges with vocal showcases.

## Common pitfalls when advising

- **Don't conflate hip-hop and R&B.** They share roots but have different aesthetics, conventions, and audiences.
- **Don't ignore the beat as composition.** In hip-hop, the beat IS the song's compositional foundation. The vocal is layered on, but the beat carries the song.
- **Don't underestimate the lyric in hip-hop.** Lyric craft (rhyme schemes, wordplay, internal rhymes, double meanings) is the central artform. Generic lyrics fail.
- **Don't apply pop chorus logic to all hip-hop.** Many hip-hop songs don't have a traditional "chorus" — they have hooks that function differently.
- **Don't use only vintage R&B references.** Modern R&B (Frank Ocean, The Weeknd, SZA) sounds different from 90s R&B. Match the era.
- **Don't ignore the sub-bass.** Especially in trap, the sub-bass is the song's primary low-end content — the kick supports it, not vice versa.
- **Don't forget production.** Hip-hop and R&B are inseparable from production. The same song with different production is a different song.

## Quick decision matrix

| Want | Try |
|------|-----|
| Classic boom-bap | 90 BPM, sampled drums, jazz/soul sample, 4-bar loop, multisyllabic rhymes |
| West Coast G-funk | 85-95 BPM, Moog synth lead, talk-box, smooth groove |
| Modern trap | 140 BPM half-time, 808 sub-bass with slides, 16th hat rolls, dark minor melodic content |
| Drill | 140 BPM half-time, sliding 808s, dark Phrygian melody, aggressive vocal |
| Cloud rap / lo-fi | 75-85 BPM, hazy production, jazzy chords, mellow vocals |
| 90s R&B | Swung 16ths, layered vocals with melisma, jazz harmony, mid-tempo |
| Neo-soul | Live drums, jazz chords, vintage-feel production, melodic vocal with runs |
| Modern atmospheric R&B | Sparse production, falsetto vocals, ambient pads, moody minor key |
| Disco/dance R&B | Four-on-the-floor kick, prominent bass, string section, 110-125 BPM |
| Funk groove | Heavily syncopated bass and rhythm guitar, ghost notes, "on the one" |

## Cross-references

- Form for hip-hop and R&B → `form/popular-song-forms.md`
- Rhythm and groove → `rhythm-groove/groove-and-feel.md`, `rhythm-groove/rhythmic-devices.md` (esp. tresillo)
- Modal / minor harmony → `harmony/modal-harmony.md`
- Jazz harmony as found in R&B → `harmony/jazz-harmony.md`
- Reharmonization in R&B → `harmony/reharmonization.md`
- Pop production conventions → `production-aware/arrangement-for-mix.md`
- Vocal melodic conventions → `melody/melodic-construction.md`
- Songwriting craft for hip-hop and R&B → `songwriting/lyric-writing.md`, `songwriting/topline-craft.md`, `songwriting/hooks-and-memorability.md`
- Related popular genres → `genres/pop-rock.md`, `genres/electronic-edm.md`
- Gospel-rooted harmony and vocal influence → `gospel-and-ccm.md`
- Brazilian funk and global club-pop crossover → `brazilian-pop-and-funk.md`

# Gospel and Contemporary Christian Music

> **Snapshot note:** Last content review snapshot: 2026-04-27. Use this file for durable composition conventions, but verify current worship, gospel, CCM, roots gospel, Christian hip-hop, and church-scene references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition-facing conventions in gospel, contemporary gospel, roots/Southern gospel-adjacent writing, praise and worship, and Contemporary Christian Music (CCM). It is **not** theological advice, liturgical planning, or a complete guide to any denomination or worship tradition.

## When to consult this file

- The user wants gospel harmony, choir energy, praise break, worship ballad, CCM pop-rock, roots gospel, or church-influenced writing
- The user asks for plagal warmth, call-response, choir arrangement, congregational singability, or a worship bridge/tag
- The user wants to blend gospel with pop, R&B, country, musical theatre, jazz, or hip-hop
- The user needs a song to feel communal, testimonial, celebratory, reverent, or spiritually grounded

## Core warning

Gospel and CCM are related but not interchangeable. They overlap in faith context and industry categories, but composition priorities differ by community, performance setting, theology, ensemble, and audience.

Always identify the frame:

- Black gospel / contemporary gospel
- Southern gospel / roots gospel
- church choir anthem
- praise and worship / congregational song
- CCM pop-rock / radio single
- gospel-R&B / gospel-jazz
- Christian hip-hop / worship-rap hybrid
- country gospel / Americana worship

## Core idea

The central compositional question is often:

```text
Is this a performance song, a testimony song, or a congregational song?
```

- **Performance song**: can have wide range, complex runs, dramatic modulations.
- **Testimony song**: lyric narrative and emotional arc carry the form.
- **Congregational song**: must be singable by non-specialists; repetition and range control matter.

## Gospel harmony vocabulary

Gospel harmony can be simple or very rich. The key is directed motion: chords often feel like they are **walking somewhere**.

### Common progressions and moves

| Sound | Progression | Use |
|---|---|---|
| Amen / plagal | IV-I | closure, warmth, hymnlike tag |
| Classic cadence | ii-V-I | jazz/gospel resolution |
| Gospel turnaround | I-vi-ii-V | loop, tag, vamp |
| Walk-up to IV | I-I/3-IV-#IV°-V | pre-chorus or turnaround |
| Minor testimony | i-♭VI-iv-V | dramatic verse or bridge |
| Worship lift | I-V-vi-IV | CCM chorus / bridge |
| Worship suspension | I-Vsus-vi-IVadd9 | open, singable modern color |
| Secondary dominant chain | V/vi-vi-V/ii-ii-V-I | rich gospel motion |
| Passing diminished | I-#I°7-ii or IV-#IV°7-V | voice-led lift |
| Shout/vamp | I-IV-I-V or I-♭VII-IV-I | repeated communal energy |

### Chord colors

- major 7 and 9 for R&B/gospel softness
- dominant 9/13 for blues/jazz-gospel motion
- sus chords for worship openness
- diminished 7 as passing lift
- slash chords for bass walkups and smooth transitions
- borrowed iv or ♭VI for emotional weight

Do not add every extension at once. Rich gospel harmony works because the voice-leading is clear.

## Choir and vocal arrangement

### Common roles

| Role | Function |
|---|---|
| Lead vocal | testimony, improvisatory energy, narrative focus |
| Choir unison | communal power, lyric clarity |
| SATB / section harmony | fullness, call-response, climactic lift |
| BGV pads | support lead without stealing focus |
| Ad-libs | emotional release, response, intensification |
| Congregation | singable repeated chorus/tag |

### Practical choir writing

- Put important text in rhythmic unison before splitting harmony.
- Save thick harmony for repeated hook lines or cadences.
- Use call-response: lead states, choir answers.
- Keep inner parts singable; do not treat choir like a keyboard pad.
- For congregational songs, keep the melody mostly within about an octave and avoid constant leaps.
- Vamps can build by adding harmony, rhythmic activity, octave displacement, and ad-libs.

## Worship / CCM writing

CCM and modern worship often prioritize clarity, repeatability, and emotional arc.

### Common form

```text
Intro → Verse 1 → Chorus → Verse 2 → Chorus → Bridge → Bridge build → Chorus → Tag
```

or:

```text
Verse → Pre-chorus → Chorus → Verse → Chorus → Bridge / spontaneous section → Chorus / Tag
```

### Congregational checklist

- Title/hook appears early and repeats clearly.
- Melody range is comfortable for group singing.
- Bridge text can survive repetition without feeling empty.
- Key is chosen for the congregation, not only the lead vocalist.
- Instrumental intro gives enough pitch/context for entry.
- Harmonic rhythm does not change so fast that a room cannot follow.
- The song has a clear emotional arc: invitation → declaration → release / response.

## Gospel rhythm and groove

| Groove | Use |
|---|---|
| 6/8 gospel ballad | testimony, worship ballad, dramatic lift |
| Straight 4/4 worship | contemporary congregational song |
| Shuffle / swing | roots gospel, blues-gospel, joyful motion |
| Half-time gospel-R&B | modern soulful ballad |
| Praise break | high-energy vamp, organ/drums/choir response |
| Four-on-floor CCM | arena/pop-worship energy |
| Country gospel two-step | roots/Americana sacred song |

The rhythm section should support the lead and choir. Busy fills are effective only when placed between vocal statements.

## Instrumental language

### Piano / keys / organ

- Gospel piano uses fills, passing chords, octave melody, and rhythmic comping.
- Hammond/organ swells support emotional contour.
- Pads in worship music create sustain and space; piano supplies harmonic clarity.
- Do not crowd the lead vocal with constant right-hand runs.

### Guitar

- CCM often uses open-position add9/sus voicings, capos, delays, and swells.
- Roots gospel/country gospel may favor acoustic strum, slide, or simple electric fills.
- Gospel-R&B may use clean electric double-stops and syncopated comping.

### Bass and drums

- Bass walks and pickups are essential in gospel transitions.
- Drums should know when to support the choir, not just the track.
- Praise-break energy depends on interaction, not only tempo.

## Lyric and tone

The lyric vocabulary depends on setting. A worship song intended for a church service is different from a personal faith-pop single.

Useful lyric frames:

- testimony: before/after transformation
- praise: direct address, declarative worship
- lament: honest need, unresolved pain, hope
- gratitude: remembered provision
- communal declaration: "we" voice
- personal prayer: intimate first-person address

Be respectful of doctrinal and community context. Avoid parodying sacred language or treating church idioms as mere decoration.

## Modulation and lift

Gospel and worship both use modulation, but differently.

| Device | Effect |
|---|---|
| Half-step final chorus | classic lift; can feel dated if automatic |
| Whole-step modulation | stronger, cleaner lift |
| Bridge build without modulation | modern worship intensity through density/range |
| Vamp rising sequence | choir/lead climax |
| Parallel minor/major shift | testimony contrast |

A key change works best when the melody, lyric, and arrangement are also intensifying.

## Quick recipes

### Simple worship chorus

```text
Key: D
| D | A/C♯ | Bm7 | Gadd9 |
Melody: mostly D-F♯-A-B, range under an octave.
Effect: singable, open, modern.
```

### Gospel tag

```text
Key: C
| C/E | F | F#°7 | Gsus G |
| C | Am7 | Dm7 G7 | C |
Effect: bass walk + passing diminished gives directed lift.
```

### Choir call-response vamp

```text
Lead: short improvised line
Choir: title phrase in harmony
Band: | 1 | 4 | 1 | 5 |
Build: unison → 3-part harmony → octave doubles → stop-time hits.
```

## Common pitfalls

- **Do not treat gospel as just "jazzier chords."** The vocal/community function matters.
- **Do not make a worship song too rangy for the congregation.** Singability is the point.
- **Do not add a key change as a substitute for real development.** Build melody, lyric, and density too.
- **Do not write choir parts as block chords without considering text clarity.** Words must land.
- **Do not assume every Christian song has the same theological vocabulary.** Context matters.
- **Do not imitate a church tradition casually for novelty.** Be specific and respectful.

## Quick decision matrix

| User wants | Try |
|---|---|
| More gospel | passing diminished chords, call-response, choir harmony, bass walkups |
| More worshipful | singable range, sus/add9 chords, repeated bridge/tag, gradual build |
| More roots gospel | acoustic instruments, shuffle/6/8, quartet harmony, plagal cadence |
| More modern CCM | pop-rock form, clean hook, bridge build, open guitar/keys texture |
| More soulful | gospel-R&B voicings, melisma space, organ/piano fills, slower groove |
| More congregational | lower key, narrower melody, simpler rhythm, clearer repeated text |

## Cross-references

- Choral writing → `../orchestration/choral-writing.md`
- Vocal idiom → `../instrument-idiom/vocals.md`
- Piano/keys idiom → `../instrument-idiom/piano-keyboards.md`
- Guitar idiom for worship/country gospel → `../instrument-idiom/guitar.md`
- Functional and chromatic harmony → `../harmony/functional-harmony.md`, `../harmony/chromatic-harmony.md`
- Jazz-influenced gospel harmony → `../harmony/jazz-harmony.md`
- Country gospel overlap → `country-americana.md`
- R&B/gospel overlap → `hip-hop-rnb.md`
- Lyric writing → `../songwriting/lyric-writing.md`
- Current scene research → `../research/web-music-trend-research.md`, `../research/reference-track-digging.md`

# K-Pop and J-Pop

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

K-pop and J-pop are the dominant popular music industries of South Korea and Japan respectively, each with its own deep history and distinctive conventions. Both are shaped by Western pop influence but have evolved into highly developed forms with characteristic structural, harmonic, and production approaches.

## When to consult this file

- The user is writing in K-pop, J-pop, or related East Asian pop styles
- Understanding the distinctive multi-section structures
- Production conventions for these markets
- Cross-genre work involving these traditions

## K-pop — overview

K-pop is the global term for South Korean popular music as exported internationally — primarily idol-group pop produced by entertainment companies (HYBE, SM, JYP, YG, and many smaller). The domestic Korean music industry is broader (ballads, indie, hip-hop, etc.); "K-pop" usually means the idol-group export product.

### History (briefly)

- **1990s**: Seo Taiji and Boys (1992) introduced rap and modern pop to Korean charts. H.O.T., Sechs Kies, S.E.S., Fin.K.L. defined first-generation K-pop idol groups.
- **2000s**: Second-generation groups emerge (TVXQ, BIGBANG, Wonder Girls, Girls' Generation, SHINee). International expansion begins (Japan first).
- **2010s**: Third generation (BTS, EXO, BLACKPINK, TWICE, Red Velvet) becomes globally successful. K-pop becomes a worldwide phenomenon.
- **2020s**: Fourth generation (NewJeans, IVE, LE SSERAFIM, ITZY, ENHYPEN, ATEEZ, aespa, Stray Kids) and continued global expansion. Genre boundaries blur with global pop.

### Structural conventions

K-pop has highly structured songs, often with more sections per minute than Western pop.

```
Intro (4-8 bars; often distinctive sound, vocal hook, or chant)
Verse 1 (8 bars; often rap-led, especially for boy groups)
Pre-Chorus (8 bars; melodic, building)
Chorus (8-16 bars; main hook, often danceable)
Post-Chorus / Drop (8 bars; second hook, often instrumental or vocal chant)
Verse 2 (8 bars; often features rap or different vocal arrangement)
Pre-Chorus (8 bars)
Chorus (8-16 bars)
Post-Chorus (8 bars)
Bridge (8-16 bars; often slower, modulates, vocal showcase)
Final Chorus (16-24 bars; modulated up; full intensity)
Outro (4-8 bars)
```

Total: often 3:00–3:30. Highly compressed compared to Western pop.

### K-pop musical signatures

**Multiple distinct hooks per song**

A typical K-pop song has 3+ memorable melodic ideas:
- The chorus melody (the "main hook")
- The post-chorus melody (often more memorable; often a chant, vocoded line, or instrumental hook)
- The bridge melody
- Sometimes the pre-chorus hook
- Sometimes a rap section's signature line

This contrasts with Western pop, where one chorus hook usually dominates.

**Section contrasts**

Each section often has a distinctly different feel:
- Rap verse with hip-hop production
- Melodic pre-chorus
- Anthemic chorus
- Drop-style post-chorus with EDM or trap influences
- Slow ballad-style bridge
- Final chorus with key change

This produces a "puzzle box" feel — the song crams diverse styles into 3 minutes.

**Genre fluidity**

K-pop freely mixes:
- Hip-hop (rap verses)
- EDM (drops, builds, electronic production)
- R&B (vocal melismas, harmonies)
- Pop (catchy melodies, chord progressions)
- Sometimes rock, jazz, Latin, Afrobeats — depending on the song's concept

The "concept" (often visual and thematic) drives the genre fusion.

**Modulation patterns**

- Bridge often modulates (often to relative or parallel mode)
- Final chorus often modulated up — half-step or whole-step (the "truck driver" modulation)
- Some K-pop songs use multiple key changes within a single song

**Chant-like hooks**

Many K-pop songs feature a chanted (often unison-shouted) phrase:
- The title (often)
- A short repeated phrase
- A nonsense vocal hook

This gives songs an instantly memorable element that translates across language barriers.

### Vocal arrangements

K-pop groups have multiple members; song arrangements distribute vocal lines among them.

- **Lead vocalist(s)**: handle main melody and high notes
- **Sub-vocalists**: harmonies and secondary lines
- **Rappers**: handle rap sections
- **Visual / dance focus members**: sometimes have less vocal time but vital to performance

Vocal arrangements often:

- Start a single voice in V1
- Add harmonies in pre-chorus
- Multiple voices on chorus (often unison + harmonies stacked)
- Solo or duo line in bridge
- Full ensemble (all members) on final chorus

### Production conventions

- **High production polish**: heavily produced, with polished mixes
- **Dance-oriented production**: most idol K-pop is designed to be danced to (choreographed)
- **EDM-influenced bass and drums**: 808s, sidechained synths, trap drums common
- **Vocal layering**: lots of doubled vocals, harmonies, ad-libs, vocal effects
- **Genre samples**: a song might use trap drums in V1, classical strings in pre-chorus, EDM drop in post-chorus

### Sub-categories within K-pop

- **Idol pop (girl groups, boy groups)**: the dominant export form
- **K-hip-hop**: separate scene; influenced by but distinct from US hip-hop (Epik High, Tablo, Beenzino, BewhY)
- **K-R&B**: Crush, Dean, Heize, Dynamic Duo, Zion.T
- **K-indie / K-rock**: smaller scene domestically (Hyukoh, Adoy, Jannabi, Silica Gel)
- **K-ballad**: the long-running domestic mainstay (sometimes performed by idol groups too)
- **Trot**: traditional Korean popular music (older audience; revived by younger artists like IU and Lim Young-woong)

## J-pop — overview

J-pop is the umbrella term for Japanese popular music. Like K-pop, it encompasses many sub-genres but has distinctive conventions.

### History (briefly)

- **1980s "City Pop"**: a sophisticated, jazz-fusion-influenced pop style (Tatsuro Yamashita, Mariya Takeuchi, Anri). Recently re-popularized globally via vaporwave / future funk online communities.
- **1990s J-pop boom**: massive domestic success (Mr. Children, Spitz, Hikaru Utada, Ayumi Hamasaki, Globe).
- **2000s J-pop**: Arashi, AKB48 (idol group system), continued mainstream success.
- **2010s onward**: Babymetal, anime music's global rise, idol groups, vocaloid (Hatsune Miku and others), etc.

### Structural conventions

J-pop generally follows verse-chorus form but often has more elaborate structures than Western pop. Songs are typically described in Japanese songwriting using these section terms (rather than English verse/chorus):

- **Intro** (イントロ) — instrumental opening, often 8–16 bars (longer than Western pop)
- **A melo** (Aメロ) — the "A melody", roughly equivalent to a verse
- **B melo** (Bメロ) — the "B melody", roughly equivalent to a pre-chorus / build
- **Sabi** (サビ) — the chorus / hook / emotional climax
- **C melo** (Cメロ) — a contrasting "C melody" that functions as a bridge, appearing before the final sabi
- **Ohsabi** (大サビ) or **last sabi** — a final, expanded chorus, often modulated and arrangement-thickened

A typical J-pop layout: `Intro – A melo – B melo – Sabi – A melo – B melo – Sabi – C melo – Ohsabi – Outro`. Compared to Western pop, J-pop often gives the B melo more compositional weight (it's not just a 4-bar pre-chorus but a substantial section with its own melodic identity), and the C melo / bridge tends to be more harmonically adventurous.

### Musical signatures

**Harmonic richness**

J-pop often uses sophisticated harmonic progressions inherited from city pop and jazz. Common features:
- Extended chords (9ths, 11ths, 13ths)
- Modal mixture (borrowed iv, ♭VII)
- Secondary dominants
- Cycle-of-fifths progressions
- Chromatic mediants (especially for emotional lift)

The "Japanese chord progression" (sometimes called *Royal Road* in Japan): IV-V-iii-vi (or sometimes stylized as IVΔ-V7-iii7-vi). Extremely common in J-pop and anime music. It cycles tension and release in a distinctive way.

**Melodic style**

- Wide melodic range (often spanning more than an octave within a single phrase)
- Lyrical, often modulating melodic lines
- Use of ♭5 or ♯4 as a passing tone for emotional inflection
- Pentatonic material (especially Japanese *yo* and *in* scale references) is sometimes used for Japanese traditional-adjacent color

**Anime music**

A subset of J-pop. Heavy in anime, video games, and related media.
- Dramatic, often complex (multi-section forms)
- Often combines orchestral/symphonic elements with band rock
- Vocaloid (synthesized vocal) sub-tradition is significant
- Major composers: Yoko Kanno (Cowboy Bebop), Hiroyuki Sawano (Attack on Titan), Yuki Kajiura (Madoka Magica), Kenji Kawai (Ghost in the Shell), Joe Hisaishi (Studio Ghibli — closer to film scoring).

### Production conventions

- **Polished production**: like K-pop, J-pop is highly produced
- **Often more "live band" feel** than K-pop's heavy electronic production
- **Idol music**: AKB48, Nogizaka46, etc. — idol groups with distinctive structures and conventions
- **Visual kei**: rock subgenre with elaborate visual aesthetic (X Japan, Dir En Grey)
- **Anime music**: theme songs and inserts; often by established J-pop artists or anisong specialists

## C-pop and other Asian pop

K-pop and J-pop are part of a wider regional pop ecosystem. Adjacent industries — Mandopop / C-pop, Cantopop, Thai pop, Vietnamese pop (V-pop), Filipino OPM, Indonesian pop — each have their own histories and conventions, with K-pop and Western pop influences layered on top of distinct regional foundations. They are *not* simply "K-pop with different language" — each has its own structural defaults, ballad traditions, and stylistic markers.

For working details on these industries, see `cpop-and-southeast-asian-pop.md`. For the technical issue of writing melodies in tonal languages (Mandarin, Cantonese, Vietnamese, Thai) — which K-pop and J-pop don't face — see `../fundamentals/prosody-and-language.md`.

## Composing K-pop / J-pop — practical advice

If the user wants to write in these styles:

1. **Listen extensively to recent material.** The styles evolve rapidly; what was K-pop in 2018 is different from K-pop in 2024.
2. **Study group dynamics.** K-pop especially is built around groups with multiple vocalists, rappers, and dancers. Plan vocal distribution.
3. **Plan multiple hooks.** Don't rely on one chorus hook. Plan post-chorus, bridge, and possibly verse hooks too.
4. **Plan the section contrasts.** Each section should feel sonically distinct — instrumentation, tempo feel, harmony, vocal treatment.
5. **Use the modulation conventions.** Final chorus modulation up is standard. Bridge modulations to relative/parallel are common.
6. **Don't fear genre switching.** K-pop especially welcomes mid-song genre changes. A trap V1 + EDM chorus + ballad bridge isn't unusual.
7. **Match production style to genre conventions.** K-pop has a distinctive sonic palette (heavy 808s, processed vocals, side-chained synths). J-pop has another (often live-feeling, sometimes more orchestral).
8. **Pay attention to the lyric**, even if writing in another language. Korean and Japanese lyric craft has its own conventions; collaborate with native lyricists or translators.

## Common pitfalls when advising

- **Don't generalize "Asian pop".** K-pop, J-pop, C-pop are distinct industries with different conventions. Be specific.
- **Don't apply Western pop song length to K-pop.** K-pop typically has more sections in 3 minutes than Western pop.
- **Don't ignore the visual / performance aspect.** K-pop especially is performed with choreography. Songs that don't have danceable sections may not work for idol groups.
- **Don't underestimate the harmonic sophistication.** J-pop especially uses jazz harmony liberally. Diatonic-only progressions may sound under-developed.
- **Don't apply 90s K-pop / J-pop conventions to current music.** Both have evolved significantly.
- **Don't assume English-language hooks are required.** While English phrases are common (especially in K-pop), the song should work in the primary language.

## Quick decision matrix

| Want | Try |
|------|-----|
| Modern K-pop idol track | Multi-hook structure, rap V1, melodic pre-chorus, anthemic chorus, post-chorus drop, bridge modulation |
| K-pop ballad | Slower tempo (70-90 BPM), piano + strings + vocal showcase, classical chord progressions |
| K-R&B | Mid-tempo, swung 16ths, jazzy chords, vocal layering and runs |
| 80s city pop revival | Major key, jazzy 7th chords, smooth bass, 100-120 BPM, fusion influences |
| Anime opening | Up-tempo, dramatic, often modulates, multi-section structure, instrumental hook |
| Anime ending | Slower, often more melodic and reflective, single dominant melody |
| Vocaloid | Synthesized vocals, often very fast, complex melodies, idiosyncratic harmony |
| Modern J-pop | Verse-pre-chorus-chorus structure with the *Royal Road* or similar distinctive J-pop progression |
| K-hip-hop | Separate from idol K-pop; closer to US hip-hop conventions but with Korean lyric |
| J-rock / visual kei | Distorted guitars, complex song forms, often dramatic vocals |

## Cross-references

- Pop song form (foundation for K-pop / J-pop) → `form/popular-song-forms.md`
- Modulation (key for K-pop final chorus, J-pop progressions) → `../harmony/modulation.md`
- Jazz harmony (key for J-pop city pop) → `../harmony/jazz-harmony.md`
- Pop production conventions → `../production-aware/arrangement-for-mix.md`
- Hip-hop conventions (relevant to K-pop rap sections) → `hip-hop-rnb.md`
- EDM conventions (relevant to K-pop drops) → `electronic-edm.md`
- R&B conventions (relevant to K-R&B and J-pop ballads) → `hip-hop-rnb.md`
- Songwriting craft (hook layering) → `../songwriting/hooks-and-memorability.md`
- Prosody and language (Korean / Japanese / multilingual hooks) → `../fundamentals/prosody-and-language.md`
- Korean traditional music and K-trad fusion → `korean-traditional.md`
- C-pop, Cantopop, T-pop, V-pop, OPM, Indo-pop (adjacent regional industries) → `cpop-and-southeast-asian-pop.md`
- Pop-rock conventions → `pop-rock.md`
- Game music (K-pop / J-pop are common reference points for game music; J-pop is foundational to anime / JRPG aesthetic) → `game-music.md`

# Jazz Styles

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

Jazz is not a single style but a tree of related traditions, each with its own harmonic, rhythmic, and improvisational conventions. This file covers the major styles and what makes each distinctive. The harmonic and chord-construction details are in `harmony/jazz-harmony.md` and `assets/jazz-voicings.md`; this file focuses on what differentiates the styles from each other.

## When to consult this file

- The user asks about a specific jazz subgenre (bebop, cool, modal, fusion, etc.)
- Composing or arranging in a jazz style
- Picking a jazz idiom for a project
- Understanding jazz historical/aesthetic differences

## Early jazz (c. 1900–1930)

Where jazz began. New Orleans roots, ragtime piano, blues vocal traditions, brass band heritage all converging.

**Key features**:
- 12-bar blues form pervasive
- Strophic song forms, simple harmonies (I, IV, V plus secondary dominants)
- Collective improvisation (multiple horns improvising simultaneously)
- 2-beat feel (emphasis on beats 1 and 3) in early styles, transitioning to 4-beat
- Brass-heavy ensembles (trumpet/cornet, trombone, clarinet)
- Strong polyphonic textures

**Sub-traditions**:
- **Dixieland / New Orleans jazz**: Louis Armstrong, King Oliver, Jelly Roll Morton, Sidney Bechet
- **Chicago style**: faster tempos, soloist-focused, 1920s development
- **Stride piano**: Fats Waller, James P. Johnson — left hand alternates bass note and chord; virtuosic right hand

## Swing era / Big band (c. 1930–1945)

The dominant popular music of its era. Large dance-band ensembles with section playing.

**Key features**:
- Big band: 3 trumpets, 3 trombones, 4 saxophones, rhythm section (piano, bass, drums, guitar)
- Tightly arranged section playing — harmonized brass and saxophone shouts
- 4/4 swing feel with the bass walking quarter notes
- Featured soloists (Duke Ellington's bandmembers, Benny Goodman's clarinet)
- 32-bar AABA standards as primary form
- Riff-based melodies
- Vocal features (Sinatra, Ella, Billie Holiday)

**Composers / arrangers**:
- Duke Ellington (the master arranger; integrated composer-bandleader)
- Count Basie (Kansas City swing)
- Benny Goodman, Glenn Miller, Tommy Dorsey (popular dance bands)
- Fletcher Henderson (foundational big band arranger)
- Mary Lou Williams (composer / pianist)

**Idiomatic features**:
- Block chord (homophonic) writing for sections
- "Drop 2" voicings standard
- Solos over functional harmony (II-V-I, blues changes)
- Plunger mutes and growl effects in brass
- Vocal scatting and "trading fours" (alternating 4-bar solos)

## Bebop (c. 1945–1955)

The first "art jazz" — small group, harmonically advanced, extremely fast and complex.

**Key features**:
- Small combos (typically quartet or quintet: piano, bass, drums + 1–2 horns)
- Fast tempos (often 200+ BPM)
- Intricate, fast-moving heads (themes) outlining complex harmony
- Improvisations using extended chord tones (9, 11, 13)
- Heavy use of secondary dominants, ii-V-I sequences
- Rhythmic complexity in the comping and solos
- Soloist-centric (the rhythm section "comps" — accompanies rhythmically)

**Composers / players**:
- Charlie Parker (alto sax — the central figure)
- Dizzy Gillespie (trumpet)
- Thelonious Monk (piano, composer with idiosyncratic style)
- Bud Powell (piano)
- Tadd Dameron, Kenny Clarke (drums)

**Idiomatic features**:
- "Bebop scales" (chromatic passing tones added to standard scales)
- Outlining ii-V-I changes constantly in solos
- Tritone substitutions
- Chromatic enclosure (surrounding a target note with chromatic neighbors)
- "Rhythm changes" (32-bar AABA based on Gershwin's "I Got Rhythm")
- "Bird blues" (Charlie Parker's reharmonization of standard blues)

## Cool jazz / West Coast (c. 1949–1955)

A reaction to bebop's intensity. Smoother, more arranged, more relaxed.

**Key features**:
- Slower to moderate tempos
- Thoughtful, lyrical solos rather than virtuosic explosions
- Arrangements with multiple voices (not just head + solos)
- More common chamber-jazz textures with non-traditional instruments (French horn, tuba, oboe)
- Emphasis on tone, dynamics, and space

**Players**:
- Miles Davis (*Birth of the Cool* nonet)
- Lennie Tristano, Lee Konitz, Warne Marsh (long contrapuntal lines)
- Chet Baker (lyrical trumpet, vocals)
- Stan Getz, Gerry Mulligan, Paul Desmond
- Chico Hamilton, Shorty Rogers (West Coast)

## Hard bop (c. 1953–1965)

A return to bebop's energy with added soul, blues, and gospel influences.

**Key features**:
- Bebop harmony but with stronger groove
- Blues and gospel motifs in melodies
- Heavier swing, more bluesy bass lines
- Often R&B-influenced, more "earthy" than cool jazz

**Players**:
- Art Blakey (Jazz Messengers — incubator for many great players)
- Horace Silver (composer of many hard bop standards)
- Cannonball Adderley, John Coltrane (early), Lee Morgan
- Wayne Shorter, Donald Byrd, Tina Brooks

## Modal jazz (c. 1959–onward)

A move away from chord-based improvisation toward modal frameworks. One scale per long passage instead of constant chord changes.

**Key features**:
- Static or slow-moving harmony
- Modal scales (Dorian, Mixolydian, Phrygian, Lydian) as the basis for solos
- Quartal and quintal voicings
- Long sustained passages on a single chord/mode
- Less reliance on functional ii-V-I motion

**Foundational works**:
- Miles Davis: *Kind of Blue* (1959) — the watershed; tracks like "So What" use only two modes
- John Coltrane: *My Favorite Things*, *A Love Supreme*
- Herbie Hancock, McCoy Tyner, Bill Evans (modal voicings on piano)

**Idiomatic features**:
- "So What" voicing (stacked 4ths)
- Long solos over single chords
- Modal melodies that emphasize characteristic notes (Dorian's natural 6, Phrygian's flat 2)
- Pedal points

## Free jazz / avant-garde (c. 1959–onward)

A radical break from harmonic and metrical conventions. Improvisation without preset chord changes or fixed meter.

**Key features**:
- Free of fixed harmonic structures
- Free of fixed meter (pulse may be present or absent)
- Group improvisation, often dense and chaotic
- Extended techniques (multiphonics, screams, noise)
- Spiritual or political ambitions in much of the music

**Players**:
- Ornette Coleman (the pioneer)
- John Coltrane (late period: *Ascension*, *Interstellar Space*)
- Albert Ayler, Pharoah Sanders
- Cecil Taylor (free piano, percussive style)
- Anthony Braxton, Sun Ra (composer-bandleader, Afrofuturism)
- AACM collective (Chicago)

## Fusion / Jazz-rock (c. 1969–onward)

Jazz combined with rock, funk, and electronic instruments. Electric jazz.

**Key features**:
- Electric instruments (electric piano, electric guitar, electric bass, synthesizers)
- Rock and funk rhythmic feels
- Often longer compositions with sectional structure
- Virtuosic instrumental playing
- Use of odd meters and complex rhythmic structures

**Foundational works / artists**:
- Miles Davis: *In a Silent Way* (1969), *Bitches Brew* (1970)
- Weather Report (Joe Zawinul, Wayne Shorter, Jaco Pastorius)
- Mahavishnu Orchestra (John McLaughlin)
- Return to Forever (Chick Corea)
- Herbie Hancock (*Head Hunters*)
- Frank Zappa (jazz-influenced rock)

## Smooth jazz / contemporary jazz (c. 1980–onward)

Commercial, radio-friendly jazz fusion. Clean production, emphasis on melody.

**Key features**:
- Polished production
- Lyrical, accessible melodies
- Pop and R&B influences
- Less harmonic complexity than mainstream jazz
- Saxophone often the lead voice
- Common in radio formats and adult contemporary

**Players**:
- Kenny G (the most commercially successful)
- George Benson, Earl Klugh
- Bob James, David Sanborn, Spyro Gyra
- The Crusaders

## Acid jazz / nu-jazz / electronic jazz (1990s–onward)

Jazz-influenced dance and electronic music.

**Players / projects**:
- Us3, A Tribe Called Quest (jazz samples in hip-hop)
- St Germain, Jazzanova (acid jazz / lounge)
- Brad Mehldau (modern acoustic jazz with rock influences)
- Esperanza Spalding, Robert Glasper (modern jazz with R&B and hip-hop)
- Snarky Puppy, Kamasi Washington (modern fusion-jazz revival)

## Latin jazz (c. 1947–onward)

Jazz fused with Cuban, Brazilian, and other Latin rhythms.

**Key features**:
- Afro-Cuban rhythms (clave, son montuno, mambo) for *Cubop* / Latin jazz
- Brazilian rhythms (samba, bossa nova) for *Brazilian jazz* / *bossa*
- Latin percussion (congas, bongos, claves, timbales)
- Jazz harmony over Latin grooves

**Players**:
- Dizzy Gillespie (*Cubop* pioneer with Chano Pozo)
- Stan Getz (*The Girl from Ipanema*, bossa nova)
- Antonio Carlos Jobim (composer of bossa nova standards)
- Tito Puente, Eddie Palmieri (salsa-jazz)
- Chick Corea (Latin influence in his music)
- Michel Camilo, Gonzalo Rubalcaba (modern Latin jazz)

## Modern jazz (post-2000)

Contemporary jazz draws on all the above and adds:
- Hip-hop influences (Robert Glasper, Esperanza Spalding)
- Electronic production (Mark Guiliana, James Francies)
- Metal and prog rock influences (Tigran Hamasyan)
- World music influences (Anouar Brahem, Avishai Cohen)
- Singer-songwriter sensibility (Norah Jones, Madeleine Peyroux)

## Idiomatic features across jazz styles

Some elements appear across most jazz styles:

| Feature | Detail |
|---------|--------|
| Swing eighth notes | Most jazz uses swung 8ths (60:40 to 67:33 ratio); fusion and Latin jazz often use straight 8ths |
| Walking bass | Quarter-note bass line through chord changes |
| Comping | Rhythmic chord accompaniment, especially by piano and guitar |
| Standard repertoire | Real Book / Great American Songbook tunes (Gershwin, Cole Porter, Rodgers, Kern, Berlin, Ellington, etc.) |
| Improvisation | Solo improvisation over the form is central |
| Trading | Soloists alternate (typically "trading 4s" — 4-bar exchanges) |
| Head-solos-head | Standard form: state melody, solos over changes, restate melody |

## Jazz harmony quick reference

(Full treatment in `harmony/jazz-harmony.md`.)

- ii-V-I is the central progression
- Tritone substitution (♭II7 for V7)
- Modal interchange (borrowing from parallel modes)
- Extended chord vocabulary (9, 11, 13, alterations)
- Chord-scale relationships (Dorian over m7, altered scale over V7alt, etc.)
- Coltrane changes (chromatic mediant cycle of major chords)

## Composing in a jazz style — practical guide

1. **Pick the substyle.** "Jazz" is too broad; bebop and smooth jazz are different worlds.
2. **Choose a form.** AABA (32-bar standard), 12-bar blues, 16-bar form, modal vamp. Match the substyle.
3. **Write a head.** Melodically interesting, harmonically connected to the changes.
4. **Voice the changes.** Use the appropriate jazz voicing style (drop 2 for guitar, rootless voicings for piano left hand, quartal for modal).
5. **Plan the arrangement.** Head – solo section (chord changes for soloists) – head out. For modal, indicate vamps and modes.
6. **Write idiomatically.** Bebop heads have specific contour profiles; modal heads have characteristic note emphasis.

## Common pitfalls when advising

- **Don't conflate substyles.** "Jazz harmony" varies enormously by substyle. Bebop changes are different from modal vamps.
- **Don't ignore the rhythm section's role.** Jazz isn't melody-and-chords; it's a band's collective performance. Comping, walking, and groove matter.
- **Don't write fully-determined music.** Jazz is about what's left to the players. Over-arranging removes the soul.
- **Don't ignore the swing feel.** Most jazz is swung; writing it in straight 8ths produces something else (fusion, Latin, or non-jazz).
- **Don't use "smooth jazz" pejoratively.** It's a legitimate style with its own audience and craft.

## Quick decision matrix

| Want | Try |
|------|-----|
| Sophisticated, fast, virtuosic | Bebop |
| Lush, lyrical, arranged | Cool jazz |
| Bluesy, soulful, grooving | Hard bop |
| Static, atmospheric, modal | Modal jazz |
| Free, abstract, exploratory | Free jazz |
| Electric, rock-influenced, virtuosic | Fusion |
| Polished, accessible, melodic | Smooth jazz |
| Hip-hop / electronic blend | Modern jazz / acid jazz |
| Latin-inflected | Latin jazz, bossa nova, Cubop |
| Standard repertoire feel | Swing era big band or AABA standard |
| Modern crossover | Glasper, Snarky Puppy, Kamasi Washington style |

## Cross-references

- Jazz harmony in detail → `harmony/jazz-harmony.md`
- Jazz voicings (drop 2, rootless, quartal) → `assets/jazz-voicings.md`
- Reharmonization techniques → `harmony/reharmonization.md`
- Modal harmony → `harmony/modal-harmony.md`
- Counterpoint in jazz → `counterpoint.md`
- Groove and swing → `rhythm-groove/groove-and-feel.md`
- Standard forms (12-bar blues, AABA) → `form/popular-song-forms.md`
- Latin rhythms → `genres/folk-roots-and-traditions.md`
- Improvisation as composition → `harmony/jazz-harmony.md`, `melody/motivic-development.md`

# Latin Pop and Reggaeton

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and regional scenes with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition-facing conventions in Latin pop, reggaeton, dembow-derived pop, bachata-pop, urbano crossover, and adjacent Latin mainstream styles. It is **not** a complete guide to Latin American or Afro-Caribbean music. For tradition-specific work, study deeper sources and work with practitioners.

## When to consult this file

- The user wants Latin pop, reggaeton, urbano, dembow, bachata-pop, or Spanish-language pop energy
- The user asks for a danceable Spanish/Latin crossover hook
- The user wants to blend pop harmony with dembow, Caribbean, or Latin rhythmic references
- The user asks for "current Latin" style; pair this file with active web research

## Core warning

Do not reduce Latin music to one rhythm. Reggaeton, bachata, salsa, cumbia, merengue, Latin trap, regional Mexican, Brazilian pop, and Latin rock are different traditions with different histories, grooves, instruments, and language conventions.

If the user says only "Latin feel," ask or infer a narrower frame:

- reggaeton / urbano
- Latin pop ballad
- bachata-pop
- salsa-influenced pop
- cumbia-influenced pop
- Latin trap
- regional Mexican / corridos tumbados influence
- Brazilian pop/funk/sertanejo axis, which deserves separate treatment

## Reggaeton / urbano composition priorities

Reggaeton is usually defined more by groove, vocal delivery, and production density than by complex chord progressions.

**Typical priorities**:

- 4/4 dance grid with dembow-derived groove
- repetitive, body-readable percussion pattern
- strong bass/kick relationship
- short, chantable hook phrases
- Spanish prosody and vowel clarity
- loop-based harmony, often two to four chords
- sparse verses and denser chorus/post-chorus
- call-response or crowd-response hook design

## Dembow-derived groove, composition view

Dembow is a family of related patterns, not a single immutable MIDI loop. A composition-facing shorthand:

```text
Feel: 4/4, tresillo / 3+3+2 energy inside a steady dance pulse
Core job: make the body know where the hook lands
```

Practical writing moves:

- Keep the vocal hook rhythm simple enough to sit on top of the groove.
- Let the bass reinforce the kick pattern; avoid busy harmonic bass motion.
- Use percussion fills to mark 4-bar and 8-bar phrase boundaries.
- Put hook words on open vowels and strong groove accents.
- If the verse is rhythmically dense, make the chorus more chant-like.

## Harmony

Latin pop and urbano harmony can be simple or lush. The groove often carries identity, so harmony should support rather than overcomplicate.

### Common loops

| Mood | Progression | Example in C / A minor |
|---|---|---|
| Bright pop | I-V-vi-IV | C-G-Am-F |
| Emotional minor-first | vi-IV-I-V | Am-F-C-G |
| Dark urbano | i-♭VI-♭III-♭VII | Am-F-C-G |
| Tense minor | i-♭VII-♭VI-V | Am-G-F-E |
| Romantic / bachata-pop | i-iv-♭VII-♭III | Am-Dm-G-C |
| Lift into chorus | iv-V-i or iv-VI-V | Dm-E-Am / Dm-F-E |

### Color tones

- add9 and sus chords keep loops modern without distracting from the groove
- minor 9 and maj7 can push toward R&B/Latin pop ballad
- dominant V in minor gives a more dramatic or classic pull
- guitar/keyboard montuno-like syncopation can imply harmony rhythmically

## Melody and vocals

Spanish-language hooks often benefit from vowel-forward writing and clear stress placement.

Useful moves:

- repeat a short hook cell with slight rhythmic variation
- leave space after the hook for percussion response
- alternate rap-like verse flow with sung chorus
- use melodic range sparingly in verse, then widen chorus or post-chorus
- build hooks around title repetition, call-response, or a memorable final vowel

Pair with `../fundamentals/prosody-and-language.md` for Spanish stress and syllable placement.

## Form

Common pop-urbano forms:

```text
Intro hook / tag → Verse → Pre or lift → Chorus → Post-chorus / dance hook → Verse 2 → Chorus → Outro tag
```

or:

```text
Intro → Chorus → Verse → Chorus → Feature / rap verse → Chorus → Outro
```

Many tracks reveal the groove and hook very early. If writing for streaming or short-form discovery, avoid long intros unless the texture itself is the hook.

## Arrangement

Common layers:

- drums: kick, snare/clap, dembow-derived percussion, hats/shakers
- bass: sub or electric/synth bass locked to kick
- harmonic bed: guitar, piano, synth pluck, pad, or sample-like loop
- lead vocal with doubles/ad-libs
- group chants or response vocals
- short ear-candy fills at phrase boundaries

The arrangement should make the dance pulse obvious even when the vocal is rhythmically dense.

## Bachata-pop crossover

Composition cues:

- guitar arpeggios or requinto-like figures
- romantic minor/major progressions
- syncopated percussion feel, often lighter than reggaeton
- vocal-forward topline with expressive ornaments
- duet or call-response format works well

Do not fake traditional bachata with a generic nylon guitar patch alone. The rhythmic guitar pattern and bass/percussion relationship are central.

## Salsa / cumbia / tropical pop influence

For salsa/cumbia-influenced pop, the rhythmic pattern and percussion architecture matter more than throwing in horns.

- Salsa influence: clave awareness, montuno-like piano/guitar patterns, horn responses, call-response vocals
- Cumbia influence: steady dance pulse, offbeat guitar/keyboard patterns, simple hook, accordion or synth-accordion color in some regions
- Merengue influence: fast duple energy, driving percussion, bright horn/synth hits

If the user wants authentic salsa/cumbia arrangement, this file is only a starting point.

## Latin trap / urbano crossover

Latin trap often blends trap drums and sub-bass with Spanish-language rap/sung flows.

Composition moves:

- half-time feel, 808 slides, dark minor loops
- sparse melody/chord sample or guitar loop
- melodic rap with Auto-Tune-like contour
- chorus may be more rhythmic than melodic
- leave room for ad-libs and vocal texture

## Quick recipes

### Reggaeton-pop chorus

```text
Key: A minor
Loop: | Am | F | C | G |
Topline: 3-note chant around ^5-^6-^5, answer phrase drops to ^3
Form: chorus hook by bar 1, post-chorus repeats title rhythm
Arrangement: dembow groove + sub + bright pluck + group vocal doubles
```

### Romantic Latin pop pre-chorus

```text
Key: C major / A minor
Loop: | Dm | E7 | Am | G |
Move: rising vocal sequence, then hold the last word over E7
Arrival: chorus starts on C or Am depending on brightness target
```

### Urbano bridge contrast

```text
Drop drums for 2 bars, keep filtered guitar/piano loop.
Use iv-V-i in minor, then bring dembow groove back on the chorus title.
```

## Common mistakes

- Calling any Spanish-language pop "reggaeton"
- Using dembow rhythm but writing a vocal that fights the groove
- Adding horns/guitar/accordion as surface markers without groove function
- Overcrowding the low end with bass, kick, and low piano
- Treating Latin America as one market or one rhythm
- Copying a current hit's hook rhythm too closely

## Cross-references

- Current trends and market checks → `../research/web-music-trend-research.md`
- Reference-track method → `../research/reference-track-digging.md`
- Style/copyright guardrails → `../research/style-reference-and-copyright.md`
- Groove and pocket → `../rhythm-groove/groove-and-feel.md`
- Pop forms → `../form/popular-song-forms.md`
- Prosody → `../fundamentals/prosody-and-language.md`
- Hip-hop/R&B crossover → `hip-hop-rnb.md`
- Brazilian pop / funk / samba-pop comparison → `brazilian-pop-and-funk.md`
- EDM crossover → `electronic-edm.md`

# MENA Pop and Maqam-Adjacent Fusion

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and regional scenes with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition-facing guidance for Middle Eastern and North African pop/fusion requests: Arabic pop, Turkish pop, Persian pop, Levantine/Gulf/North African influences, maqam-adjacent color, and Western pop/EDM/hip-hop crossovers. It is not a full guide to Arabic, Turkish, Persian, Kurdish, Amazigh, or other regional classical/folk traditions.

## When to consult this file

- The user wants Arabic pop, Turkish pop, Persian pop, MENA club/pop, or maqam-adjacent color
- The user says "Middle Eastern feel" and needs a safer, more specific translation
- The user wants microtonal/modal melody in a pop context
- The user asks for current regional pop references; pair with active web research

## Core warnings

- **Maqam is not just a scale.** It includes characteristic phrases, melodic direction, modulation behavior, ornament, tuning tendencies, and aesthetic conventions.
- **Quarter-tones are not a shortcut.** Some maqamat use intervals that differ from 12-TET, but equal 24-TET is only an approximation in many contexts.
- **MENA is not one sound.** Egyptian, Lebanese, Gulf, Turkish, Persian, Maghrebi, Kurdish, Amazigh, and diaspora pop scenes differ.
- Avoid vague cultural shorthand. Name the actual musical variable: augmented second, drone, ornament, iqa'/usul-inspired rhythm, oud-like pluck, maqam Hijaz-adjacent tetrachord, etc.

## Safer translation of vague requests

| User says | Translate to |
|---|---|
| "Middle Eastern" | Which region or reference? Arabic pop, Turkish pop, Persian ballad, Levantine dance, Gulf rhythm, Maghrebi chaabi/raï-adjacent, etc. |
| "Arabic scale" | Ask for maqam family or use "Hijaz-adjacent / Bayati-adjacent / Rast-adjacent color" cautiously |
| "Desert / mysterious" | Drone, sparse percussion, low register, ornamented melody — avoid stereotypes |
| "Belly dance feel" | Often a dance/rhythm/arrangement request; identify rhythm and instrumentation |
| "Maqam but pop" | Modal melody + pop form/harmony; do not claim full maqam authenticity |

## Melodic language

For pop-fusion, use maqam-adjacent constraints without overclaiming.

Useful abstractions:

| Idea | Pop adaptation |
|---|---|
| Jins / tetrachord color | Build hook around a 4–5 note cell with a clear characteristic interval |
| Sayr / melodic path | Decide whether melody opens low, rises gradually, or circles a focal pitch |
| Ornament | Slides, turns, mordents, grace notes, bends, melisma |
| Drone / tonic emphasis | Keep a stable center while melody supplies color |
| Modulation | Move to a related color in bridge or instrumental break, then return |

### Common 12-TET approximations

Use these as **approximations**, not authentic maqam definitions.

| Color | 12-TET pitch set in C | Caution |
|---|---|---|
| Hijaz-adjacent | C-D♭-E-F-G-A♭-B♭ | Strong augmented-second color; often overused as stereotype |
| Phrygian / ♭2 color | C-D♭-E♭-F-G-A♭-B♭ | Darker, less augmented-second bite |
| Harmonic minor | C-D-E♭-F-G-A♭-B | Western minor dominant color; not automatically Arabic/Turkish/Persian |
| Rast-adjacent approximation | C-D-E-ish-F-G-A-B♭-ish | Needs non-12-TET/tuning nuance in traditional contexts |
| Bayati-adjacent approximation | C-D-half-flat-ish-E♭-F-G-A♭-B♭ | 12-TET cannot capture the characteristic second degree well |

If microtonal nuance matters, consult `../techniques/microtonal.md` and current/practitioner sources.

## Harmony

Traditional maqam-based music is melody-centered rather than Western chord-progression-centered, but pop/fusion often uses chords.

Safer harmonic approaches:

| Goal | Approach |
|---|---|
| Preserve modal melody | Use drone/pedal + sparse chords |
| Pop accessibility | Use i-♭VII-♭VI-♭VII, i-iv-V, or I-♭VII-IV loops |
| Hijaz-adjacent drama | Use i-♭II or i-♭II-V-i carefully |
| Ballad color | Use minor add9, sus, maj7, and slow harmonic rhythm |
| Dance/club | Keep harmony simple; let rhythm, bass, and ornament carry identity |

Avoid forcing every melody note into a Western triad. Chords can flatten the modal identity if they move too often.

## Rhythm

MENA rhythms vary widely. Do not use one "Arab beat" for everything.

Composition-facing options:

| Rhythm idea | Use |
|---|---|
| Maqsum-like 4/4 | Versatile Arabic pop/dance frame |
| Saidi-like feel | Earthy, dance-oriented, often associated with upper Egyptian style |
| Malfuf-like 2/4 | Fast entrance, procession, dance energy |
| 6/8 or 12/8 | Gulf, North African, or pop-ballad contexts depending on style |
| Turkish usul-inspired cycles | Longer accent cycles, odd meters, asymmetry in some styles |
| Pop/EDM hybrid | Four-on-floor or trap drums with regional percussion layers |

If the user names a specific dance, region, or tradition, search and verify rather than guessing.

## Arrangement and instrumentation

Common color sources:

- oud-like plucked lines
- qanun-like tremolo/ornament figures
- ney/flute-like breathy lead
- violin/strings with slides and ornaments
- darbuka/doumbek/riq/frame-drum-inspired percussion
- synth lead with pitch bends
- call-response backing vocals
- drone or pedal bass

Do not rely on one sample-pack instrument as a cultural proxy. The melodic phrase, ornament, rhythm, and form matter more than the patch name.

## Vocal writing

Useful moves:

- allow melisma or ornament at phrase endings
- write space for breath and expressive turns
- place important words on sustained notes
- avoid over-harmonizing ornamented lines
- consider language stress, vowel length, and dialect-specific pronunciation
- use call-response or unison group vocals for dance hooks

## Form templates

### Arabic/Turkish/Persian pop ballad-inspired

```text
Instrumental motif → Verse → Pre-lift → Chorus → Instrumental answer → Verse 2 → Chorus → Bridge/modulation → Final chorus
```

### MENA club-pop fusion

```text
Intro percussion/motif → Vocal hook → Beat drop → Verse → Chorus/drop → Instrumental riff → Verse/rap → Final hook
```

### Maqam-adjacent instrumental cue

```text
Drone + free ornament → pulse enters → 4-bar motif → sequence/transposition → breakdown → return to tonic motif
```

## Quick recipes

### Hijaz-adjacent pop hook

```text
Center: C
Melody cell: C-D♭-E-F-E-D♭-C
Harmony: | Cm | D♭ | Cm | B♭ |
Arrangement: drone + plucked ostinato + hand percussion + modern sub
Caution: do not claim full maqam Hijaz authenticity without phrase/tuning study.
```

### Modal dance cue

```text
Rhythm: maqsum-like 4/4 frame
Bass: tonic pedal with octave movement
Melody: short call phrase, then group response
Form: 8-bar build, 8-bar full groove, 4-bar breakdown, return
```

### Ballad color

```text
Loop: | Am(add9) | G | Fmaj7 | E7sus E7 |
Topline: narrow verse, melismatic answer phrase, chorus peak on ^5 or ^♭6
Texture: soft strings + plucked motif + restrained percussion
```

## Common mistakes

- Treating "Phrygian dominant" as a complete MENA music toolkit
- Calling maqam a scale and ignoring melodic path/ornament
- Using quarter-tones randomly
- Over-chording a modal melody
- Using stereotyped "desert" language or threat coding
- Ignoring regional and language differences
- Copying a current pop hook because the surface scale matches

## Cross-references

- Current trend checks → `../research/web-music-trend-research.md`
- Reference-track method → `../research/reference-track-digging.md`
- Style/copyright guardrails → `../research/style-reference-and-copyright.md`
- Microtonal / maqam caveats → `../techniques/microtonal.md`
- Modal harmony → `../harmony/modal-harmony.md`
- Prosody → `../fundamentals/prosody-and-language.md`
- Folk/roots specificity caution → `folk-roots-and-traditions.md`
- Film scoring → `film-tv-scoring.md`

# Minor, Microgenre, and Hybrid Genre Handling

> **Snapshot note:** Last content review snapshot: 2026-04-27. Microgenre names, online-scene labels, and hybrid genre tags change quickly. Use this file as a handling framework, then verify current usage through active web research and reference listening.

This file is for genre names too small, too new, too local, too internet-native, or too hybrid to deserve a dedicated file yet. It helps the agent respond without pretending that a single paragraph can fully define a living scene.

Use this file when the user names a style such as:

- a microgenre: hyperpop, phonk, breakcore, witch house, vaporwave, future funk, blackgaze, slowcore, math rock, plugg, rage rap, etc.
- a platform-born or meme-driven tag
- a local scene label the skill does not yet cover (for regional popular scenes, also check `regional-scene-starters.md`)
- a hybrid label: "K-pop + Jersey club", "country-trap", "ambient gospel", "metalcore with J-pop chorus", etc.
- a genre term the agent does not recognize confidently

## Core principle

For minor and hybrid genres, the safest answer is:

```text
I will not over-define the scene. I will identify a few craft variables, verify current references if needed, and give a playable sketch that uses those variables without claiming to represent the whole genre.
```

## Microgenre triage workflow

### 1. Identify the genre type

| Type | Examples | First move |
|---|---|---|
| Internet microgenre | hyperpop, vaporwave, nightcore, drift phonk | check current reference cluster and platform context |
| Local club/dance form | Jersey club, footwork/juke, gqom, baile funk variant | verify tempo/groove grid and dance function |
| Guitar/band substyle | shoegaze, slowcore, math rock, post-rock, blackgaze | identify texture, riff language, form pacing |
| Rap/R&B substyle | plugg, rage, cloud rap, alt-R&B niche | identify drum pocket, synth palette, vocal/flow density |
| Metal/punk substyle | djent, doom, sludge, screamo, grindcore | identify tuning, riff rhythm, tempo/endurance, vocal role |
| Nostalgia/aesthetic genre | city-pop revival, synthwave, future funk, chillwave | identify era-signifiers vs modern form/low-end |
| Hybrid request | K-pop + amapiano, folk + hyperpop, gospel + drill | split into ingredients, then recombine intentionally |

### 2. Ask for one anchor if the label is unstable

```text
That tag can mean different things depending on platform and scene. Give me one reference track or artist that represents the exact lane, and I will extract craft variables rather than copying it.
```

If the user does not provide a reference, use active web research and say the result is approximate.

### 3. Build a craft-variable profile

```md
| Variable | What to check |
|---|---|
| Tempo / meter | BPM range, half-time/double-time feel, dance grid |
| Groove | drum pattern, swing, subdivision, loop length |
| Harmony | static vamp, modal center, functional changes, chromatic color |
| Melody/topline | chant, rap flow, maximalist leaps, narrow drone, noisy texture |
| Texture | distorted, glossy, lo-fi, dense, sparse, sidechained, wall-of-sound |
| Form | loop, verse/chorus, drop, build, breakdown, through-composed arc |
| Cultural/platform function | club, meme, playlist, live scene, fandom, video, underground label |
| Do-not-copy | signature sample, meme phrase, producer tag, artist vocal identity |
```

## Microgenre family notes

These are starting points, not definitions.

### Hyperpop / digicore / maximalist internet pop

Composition variables:

- compressed, high-contrast sections
- bright or distorted synth hooks
- pitch-shifted or heavily processed vocal identity
- abrupt modulations, tempo-feel shifts, or texture cuts
- short-form-friendly hook cells

Pitfall: copying a specific vocal-processing identity or one artist's melodic fingerprint.

Pair with `../genres/electronic-edm.md`, `../genres/pop-rock.md`, and `../research/style-reference-and-copyright.md`.

### Phonk / drift phonk / Memphis-rap-adjacent internet styles

Composition variables:

- looped minor motifs or dark modal centers
- hard, repetitive drum pattern; cowbell-like lead in some drift variants
- chopped vocal/sample aura
- car-edit/short-form or underground tape aesthetics depending on variant

Pitfall: treating all phonk as the same. Drift phonk, Brazilian phonk variants, and Memphis-rooted references can differ sharply.

Pair with `../genres/hip-hop-rnb.md`, `../genres/electronic-edm.md`, and `../research/web-music-trend-research.md`.

### Vaporwave / future funk / synthwave / chillwave

Composition variables:

- nostalgia-coded harmony, samples, timbres, or visual identity
- loop-based form and slow harmonic rotation
- sidechain or tape/lo-fi gestures depending on substyle
- disco/funk/city-pop-derived references in some branches

Pitfall: nostalgia aesthetics can become sample-dependent; avoid uncleared samples and rebuild ideas from scratch.

Pair with `../genres/electronic-edm.md`, `../genres/pop-rock.md`, and `../research/reference-track-digging.md`.

### Breakcore / jungle revival / footwork-juke-adjacent high-density rhythm

Composition variables:

- chopped breakbeats or rapid percussion edits
- high rhythmic density and abrupt form pivots
- bass repetition or sub pressure
- melodic material may be minimal, sentimental, noisy, or rave-derived

Pitfall: confusing breakbeat editing as random chaos. Strong tracks often have clear phrase logic.

Pair with `../rhythm-groove/rhythmic-devices.md`, `../genres/electronic-edm.md`, and `../form/narrative-and-transitions.md`.

### Jersey club / Baltimore club / regional club forms

Composition variables:

- dance-function first
- kick pattern and vocal chops as structural hooks
- call-and-response or chant-like fragments
- short loop cycles with strong drop points

Pitfall: borrowing the pattern without understanding dance/social function.

Pair with `../genres/electronic-edm.md`, `../genres/hip-hop-rnb.md`, and `../research/regional-trend-evolution-analysis.md`.

### Shoegaze / dream pop / slowcore / post-rock

Composition variables:

- texture as harmony: sustained guitars, blurred attacks, layered reverb/delay
- slower harmonic pacing or long crescendos
- vocal embedded in texture rather than always foregrounded
- simple motifs transformed by density

Pitfall: adding reverb to any guitar part and calling it shoegaze. The part must support the texture.

Pair with `pop-rock.md`, `../instrument-idiom/guitar.md`, and `../orchestration/arrangement-density.md`.

### Math rock / post-hardcore / emo-adjacent hybrids

Composition variables:

- irregular phrase lengths, odd meters, or additive accents
- clean interlocking guitar figures or angular riffs
- dynamic shifts between restraint and catharsis
- melodic hooks may be shouted, sung, or instrumental

Pitfall: using odd meter without a singable or memorable phrase identity.

Pair with `metal-punk-hardcore.md`, `pop-rock.md`, and `../rhythm-groove/odd-meters-polyrhythm.md`.

### Blackgaze / post-metal / doom / sludge-adjacent hybrids

Composition variables:

- heavy guitar texture plus long-form atmospheric build
- tremolo/riff cells, drones, pedal tones, modal centers
- extreme vocals may function as texture rather than lyric foreground
- slow or medium tempo can still feel intense through density

Pitfall: writing a normal metal riff and adding pads; the form and texture need to breathe.

Pair with `metal-punk-hardcore.md`, `../instrument-idiom/guitar.md`, and `../production-aware/energy-and-dynamics.md`.

### Plugg / rage / cloud-rap / alt-trap microstyles

Composition variables:

- synth/rhythm identity often more important than chord complexity
- loop economy, vocal pocket, ad-lib space
- beat may be minimal, glossy, distorted, or arcade-like depending on lane
- topline/flow density and vocal tone define much of the style

Pitfall: over-arranging. Many tracks need space for vocal identity and ad-libs.

Pair with `hip-hop-rnb.md`, `../rhythm-groove/groove-and-feel.md`, and `../production-aware/arrangement-for-mix.md`.

## Hybrid genre recipe

When the user requests "A + B":

### 1. Decide the anchor

Which ingredient controls the song's core identity?

| Anchor | Usually controls |
|---|---|
| Genre A as song form | verse/chorus/drop/bridge layout |
| Genre B as groove | drum pattern, bass, subdivision |
| Genre C as harmony | chord palette, cadences, voicings |
| Regional/traditional ingredient | rhythm cycle, mode, ornament, instrument idiom, language |
| Artist/reference ingredient | energy target only; avoid cloning |

### 2. Limit the number of signatures

For a first sketch, use **one primary signature element from each ingredient**.

Bad hybrid:

```text
K-pop section changes + hyperpop vocal processing + flamenco guitar + drill drums + gospel chords + orchestral strings + maqam melody.
```

Better hybrid:

```text
K-pop form anchor + Jersey-club-derived kick pattern + neo-soul color chords. Keep the topline simple and avoid adding more signatures until the hook works.
```

### 3. Use a merge table

```md
| Ingredient | Keep | Soften / avoid | Concrete move |
|---|---|---|---|
| K-pop | pre-chorus lift, post-chorus chant | overstuffed section count | 4-bar pre into 4-bar chant hook |
| Jersey club | kick pattern energy | exact sample-chop cliché | original vocal chop rhythm |
| Neo-soul | maj9/min9 voicings | too many reharm changes | 2-chord loop with chromatic passing chord |
```

## Unknown genre name protocol

If the agent does not recognize a genre term:

1. Do not guess confidently.
2. Use web research if the user needs accuracy.
3. Ask for one reference track if the answer can wait for user input.
4. If proceeding without clarification, say it is approximate.
5. Extract variables from available examples rather than defining the whole scene.

Response pattern:

```md
I don't want to over-define that tag from memory. Treating it as a current/scene-specific label, I'd check 3-5 references and look for shared tempo, groove, texture, form, and vocal behavior. For a first sketch, I would start with...
```

## Pitfalls

- **Do not turn a TikTok tag into a timeless genre rule.** Platform tags mutate quickly.
- **Do not assume a microgenre is only a sound palette.** Many are scenes with social, geographic, or platform contexts.
- **Do not over-hybridize.** Two strong ingredients usually beat six weak ones.
- **Do not use surface timbres as cultural shortcuts.** Name the specific tradition, scene, rhythm, or instrument idiom.
- **Do not copy memes, samples, or producer tags.** Make original material.

## Quick decision matrix

| User goal | Best response |
|---|---|
| "What is [microgenre]?" | Give cautious craft-variable summary + ask for reference or use web |
| "Make this more [microgenre]" | Change 1-3 variables: groove, texture, form, vocal approach |
| "Blend A and B" | Pick anchor, supporting ingredient, and one shared bridge variable |
| "Is this too close to reference?" | Load `../research/style-reference-and-copyright.md` and compare melody/lyrics/riff/sample/vocal identity |
| "Which references should I dig?" | Load `../research/reference-track-digging.md` + active web research |

## Cross-references

- Web trend research → `../research/web-music-trend-research.md`
- User listening context → `../research/user-listening-context-and-streaming-services.md`
- Regional trend evolution → `../research/regional-trend-evolution-analysis.md`
- Reference-track digging → `../research/reference-track-digging.md`
- Style/copyright guardrails → `../research/style-reference-and-copyright.md`
- Electronic / EDM → `electronic-edm.md`
- Hip-hop / R&B → `hip-hop-rnb.md`
- Pop / rock / indie → `pop-rock.md`
- Metal / punk / hardcore → `metal-punk-hardcore.md`
- Folk/roots or tradition-specific traditions → `folk-roots-and-traditions.md`
- Rhythm devices → `../rhythm-groove/rhythmic-devices.md`
- Instrument idiom overview → `../instrument-idiom/overview.md`

## Additional cross-references

- Regional scene starter map → `regional-scene-starters.md`
- Fast web-search cheatsheet → `../../assets/web-search-cheatsheet.md`

# Korean Traditional Music and K-Trad Fusion

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

Korean traditional music (*gugak*, 국악) is a 1500-year-old tradition with its own pitch systems, rhythmic frameworks, and aesthetic priorities — fundamentally distinct from Western art music. It splits into court / aristocratic music (*jeongak*, 정악) and folk / popular traditions (*minsogak*, 민속악), each with its own conventions and instruments. Modern Korean composers and producers increasingly fuse traditional elements with pop, rock, electronic, and film idioms — a movement loosely called K-trad or Korean traditional fusion.

This file covers the working orientation for composers wanting to engage with the tradition. It's not a substitute for study with practitioners or extended listening, but it's enough to make sensible choices and avoid the most common mistakes.

For broader Asian and folk traditions, see `folk-roots-and-traditions.md`. For modern Korean popular music, see `kpop-jpop.md`. For modal harmony (which has analogues in Korean modal systems), see `../harmony/modal-harmony.md`. For non-12-TET pitch considerations, see `../fundamentals/pitch-intervals-scales.md`.

## When to consult this file

- Composing for a project with Korean traditional elements (film, drama, game, theater, K-pop fusion)
- Writing for traditional Korean instruments
- Trying to evoke a Korean traditional feeling in pop, rock, or electronic contexts
- Analyzing or arranging a *gugak* piece
- Deciding when traditional fusion serves a piece and when it's tokenistic

## Two main streams

### Jeongak (정악) — court and aristocratic music

The refined, formal tradition associated with the royal court, Confucian rites, and the literati class.

**Repertoire**:
- *Jongmyo Jeryeak* (종묘제례악): ritual music for the royal ancestral shrine; UNESCO Intangible Cultural Heritage
- *Yeongsan Hoesang* (영산회상): a major suite-like work, central to chamber jeongak
- *Boheoja* (보허자), *Nakyangchun* (낙양춘): ceremonial pieces
- *Sujecheon* (수제천): ceremonial music for outdoor processions and royal events
- *Yeombuldodeuri* (염불도드리), *Cheonnyeonmanse* (천년만세): chamber pieces

**Aesthetic priorities**:
- Slow, deliberate tempo
- Restrained ornamentation
- Long sustained tones with subtle pitch inflection
- Strict adherence to mode (*jo*) and rhythmic cycle (*jangdan*)
- Ensemble unity over individual virtuosity

**Typical instrumentation**: court orchestra includes *pyeonjong* (편종, bronze bells), *pyeongyeong* (편경, stone chimes), *daegeum* (대금, large bamboo flute), *piri* (피리, cylindrical reed), *haegeum* (해금, two-string fiddle), *ajaeng* (아쟁, bowed zither), *gayageum* (가야금, plucked zither), *geomungo* (거문고, plucked zither with stick), *janggu* (장구, hourglass drum), *jwago* (좌고, large barrel drum).

### Minsogak (민속악) — folk and popular tradition

The vital, expressive tradition of the common people.

**Major forms**:
- *Pansori* (판소리): epic vocal storytelling with single drummer
- *Sanjo* (산조): instrumental virtuoso form, free-form solo with drum
- *Sinawi* (시나위): improvised ensemble music rooted in shamanic ritual
- *Nongak* / *Pungmul* (농악/풍물): outdoor percussion-and-dance ensemble
- *Japga* (잡가): popular vocal songs, semi-professional
- *Minyo* (민요): folk songs, regional varieties (gyeonggi, namdo, seodo, dongbu, jeju)

**Aesthetic priorities**:
- Expressive ornamentation (*sigimsae*, 시김새)
- Emotional intensity (*han* 한, *heung* 흥, *sinmyeong* 신명)
- Improvisation within strict frameworks
- Rhythmic drive and acceleration
- Individual virtuosity celebrated

## Pansori (판소리) — epic vocal storytelling

A solo singer (*sorikkun*, 소리꾼) performs a long narrative work accompanied only by a single drummer (*gosu*, 고수) on the *buk* (북, barrel drum). Performances historically lasted 4–8 hours.

**The five surviving works** (*pansori daseot madang*, 판소리 다섯 마당):
- *Chunhyangga* (춘향가) — love story
- *Simcheongga* (심청가) — filial piety
- *Heungboga* (흥보가) — fable of two brothers
- *Sugungga* (수궁가) — animal fable
- *Jeokbyeokga* (적벽가) — historical-military

**Two main schools**:
- *Dongpyeonje* (동편제): masculine, strong, declamatory; eastern Jeolla
- *Seopyeonje* (서편제): feminine, lyrical, emotionally elaborate; western Jeolla
- (*Junggoje* (중고제), a third school, has nearly died out)

**Vocal technique**: *sori* (소리) involves controlled rasp, dramatic pitch sweeps, and a piercing, projected tone built up over years of practice (the *deugum*, 득음, vocal training tradition involves shouting against waterfalls). The voice ranges from rough chest voice to clean falsetto, with elaborate ornamentation.

**Drummer's role**: the *gosu* punctuates the narrative with cries of *chuimsae* (추임새 — "eolssigu!", "jota!", "jihwaja!") and varies the *jangdan* (rhythmic cycle) according to dramatic context.

**Compositional structure**: alternation of *aniri* (아니리, spoken narration) and *chang* / *sori* (창/소리, sung passages) with *balim* (발림, gestures). Sung passages use specific *jangdan* matched to the dramatic mood.

## Sanjo (산조) — instrumental virtuosity

A solo instrumental form (1890s onward), typically *gayageum sanjo*, *geomungo sanjo*, *daegeum sanjo*, *haegeum sanjo*, *ajaeng sanjo*, or *piri sanjo*, accompanied by *janggu*. The performer plays through a sequence of *jangdan* sections that progressively accelerate.

**Standard sequence of jangdan in sanjo**:
- *Jinyangjo* (진양조): very slow, 6 large beats with common 24-beat phrase grouping — exposition, lyrical
- *Jungmori* (중모리): moderate, 12/4 — narrative
- *Jungjungmori* (중중모리): slightly faster, 4 large beats / 12 subdivisions — flowing
- *Jajinmori* (자진모리): fast, 4 large beats / 12 subdivisions — driving
- *Hwimori* (휘모리): very fast, 4/4 — climactic
- *Danmori* (단모리): extremely fast (in some traditions)

The piece is largely composed but allows soloist freedom in ornamentation. Major founders/schools: Kim Chang-jo (김창조, founder of gayageum sanjo), Baek Nak-jun (백낙준, geomungo), Han Beom-su (한범수, haegeum), and many others — each lineage has distinctive *seonyul* (선율, melodic content).

## Sinawi (시나위)

Improvised ensemble music, originally accompanying shamanic ritual (*gut*, 굿). Each instrument improvises its own line over the same *jangdan* and *jo* — heterophonic rather than harmonic. The result is dense, layered, and ecstatic.

A modern *sinawi* ensemble typically includes *daegeum*, *piri*, *haegeum*, *ajaeng*, *gayageum*, *geomungo*, and *janggu*. The piece is structured by *jangdan* changes (similar to *sanjo*) but each instrument's part is freshly improvised.

## Nongak / Pungmul (농악 / 풍물)

Outdoor percussion-and-dance ensemble, historically tied to agricultural communities. Four main instruments — collectively called *samul* (사물, "four things"):
- *Kkwaenggwari* (꽹과리, small gong)
- *Jing* (징, large gong)
- *Janggu* (장구, hourglass drum)
- *Buk* (북, barrel drum)

Plus *taepyeongso* (태평소, oboe-like wind), *sogo* (소고, small handheld drum), and dancers.

The 1978 *SamulNori* (사물놀이) ensemble led by Kim Duk-soo (김덕수) brought the four percussion instruments into a stage-concert format. *SamulNori* and its successors (Noreum Machi, Dulsori) became the dominant export form of Korean percussion music.

## Jangdan (장단) — rhythmic cycles

*Jangdan* is the cyclical rhythmic-metrical framework of Korean traditional music, played on the *janggu* or *buk*. Each *jangdan* has:
- A specific number of beats per cycle
- A characteristic accent pattern (*jeongganbo* notation visualizes this as squares)
- A typical tempo range
- An associated emotional or dramatic feel

Major *jangdan* in folk music:

| Jangdan | Beats / pattern | Tempo | Character |
|---------|-----------------|-------|-----------|
| *Jinyangjo* (진양조) | 6 large beats with triple subdivision; often grouped into 4-cycle / 24-beat phrases in pansori pedagogy | Very slow | Solemn, lyrical, exposition |
| *Jungmori* (중모리) | 12 beats, often felt as broad 12/4 | Moderate | Narrative, walking |
| *Jungjungmori* (중중모리) | 4 large beats / 12 subdivisions (12/8 feel) | Mod-fast | Flowing, dance-ready |
| *Jajinmori* (자진모리) | 4 large beats / 12 subdivisions (fast 12/8 feel) | Fast | Driving, energetic |
| *Hwimori* (휘모리) | 4 beats, 4/4 | Very fast (♩= 200+) | Climactic, frenzied |
| *Gutgeori* (굿거리) | 12 beats, 12/8 | Mod (♩.= 70) | Dance, festive |
| *Semachi* (세마치) | 9 beats, 9/8 (compound triple) | Moderate | Lyrical, often sung folk |
| *Eotmori* (엇모리) | 10 beats, asymmetric (3+2+3+2) | Moderate-fast | Limping, off-balance, supernatural |
| *Dongsalpuri* (동살풀이) | 4 beats, 4/4 | Mod-fast | Folk dance, lively |
| *Taryeong* (타령) | 12 beats, 12/8 | Mod-fast | Song-like, urban folk |

Different jangdan match different dramatic moments. In *pansori*, *jinyangjo* is associated with slow lyrical or weighty passages, while faster cycles such as *jajinmori* support action, motion, and heightened energy.

For Western-music-trained composers, the most counterintuitive *jangdan* is *eotmori* — its 10-beat asymmetric grouping (3+2+3+2) feels neither duple nor triple in a Western sense, and it's specifically associated with supernatural, otherworldly, or "off-balance" dramatic moments.

## Jo (조) — modal systems

Korean traditional music uses several modal systems (*jo*, 조). The two most important in folk music:

### Pyeongjo (평조)

Roughly analogous to a major-pentatonic color, but with characteristic ornamentation, intonation, and register practice. For composition-facing sketches, think ^1, ^2, ^3, ^5, ^6 while remembering that Korean pitch-name systems such as *hwangjong* (황종) are not simply fixed Western concert pitches.

Bright, balanced, often festive feel. Used in court music and brighter folk.

### Gyemyeonjo (계면조)

The minor-feeling mode. Roughly: ^1, ^♭3, ^4, ^5, ^♭7, with important tones shaped by *sigimsae* such as upward inflection (*chuseong*, 추성), downward release (*toeseong*, 퇴성), and vibrato-like shaking (*yoseong*, 요성). Its emotional core comes less from a fixed Western scale formula than from how those tones are bent, released, and timed.

Sad, lamenting, expressive. The mode of *pansori*'s most affecting passages, much *minyo* (folk song), and the *han* (한, deep sorrow / longing) aesthetic.

### Other modes

*Ujo* (우조) is associated with brighter, masculine character (often a transposed *pyeongjo* variant). Various regional folk traditions have their own modal preferences:
- *Gyeonggi minyo* (Seoul region): bright, often *pyeongjo*-derived
- *Namdo minyo* (southern, Jeolla): rich *gyemyeonjo*, melismatic, intense
- *Seodo minyo* (northwestern): plaintive, semi-Phrygian feel
- *Jeju minyo* (island): distinctive scales not shared with mainland traditions

When composing a Korean traditional-adjacent piece, identifying which regional mode or practice you are emulating matters as much as picking *gyemyeonjo* vs. *pyeongjo* in the abstract.

## Sigimsae (시김새) — ornamentation

The ornamental pitch inflections that animate Korean traditional melody. Specific named ornaments include:

- *Chuseong* (추성): a pitch-raising slide or upward inflection
- *Toeseong* (퇴성): a pitch-falling or downward-releasing inflection
- *Yoseong* (요성): a vibrating or wavering inflection; often wider and more structural than Western vibrato
- *Jeonseong* (전성): a turning/ornamental inflection
- *Goldoreum* (골도름): a fast trill-like figure

Different instruments have different idiomatic *sigimsae*. *Gayageum* uses left-hand string pulls; *daegeum* uses fingering and embouchure; *haegeum* uses bow pressure and finger position; the human voice has the widest palette. *Sigimsae* is the primary expressive vocabulary — playing the bare skeletal pitches without *sigimsae* sounds dead.

For Western-trained composers writing for Korean instruments: notating *sigimsae* explicitly is impractical. Indicate the desired ornaments in prose, or work with a player who knows the tradition. Modern *gugak* scores often use partial Western notation with annotations in *jeongganbo* (정간보, the traditional notation grid) or written prose for ornaments.

## Instruments — quick reference

### Strings

| Instrument | Description | Range (rough) | Idiom |
|------------|-------------|---------------|-------|
| *Gayageum* (가야금) | 12-string plucked zither (modern variants: 18, 21, 25 strings) | Two octaves around middle C | Lyrical, flowing; *sanjo* showpiece |
| *Geomungo* (거문고) | 6-string plucked zither, struck with stick | Lower than *gayageum* | Stately, masculine; literati instrument |
| *Ajaeng* (아쟁) | 7- or 8-string bowed zither (with rosined wooden stick or horsehair bow) | Low, cello-like | Mournful, deep; sometimes called "Korean cello" |
| *Haegeum* (해금) | 2-string fiddle held vertically | Mid to high, vocal-range | Speaking, lyrical, intense — closest to vocal expression among instruments |
| *Bipa* (비파) | Pear-shaped lute (rare in modern use) | Mid range | Historical / restored use |

### Winds

| Instrument | Description | Range | Idiom |
|------------|-------------|-------|-------|
| *Daegeum* (대금) | Large transverse bamboo flute with vibrating membrane | Wide, ~2 octaves from D4 | Spacious, breathy; principal flute of *jeongak* and *sanjo* |
| *Sogeum* (소금) | Small transverse bamboo flute | Higher than *daegeum* | Bright, used in ensemble |
| *Danso* (단소) | Small notched bamboo flute (5 holes) | Two octaves, mid-high | Gentle, often pedagogical |
| *Piri* (피리) | Cylindrical double-reed (3 main types: *hyangpiri*, *sepiri*, *dangpiri*) | Mid range, penetrating | The dominant melody instrument in court ensembles |
| *Taepyeongso* (태평소) | Conical double-reed, oboe-like, very loud | Mid-high, blasting | Outdoor music, *nongak*; cuts through anything |
| *Saenghwang* (생황) | Free-reed mouth organ (similar to Chinese *sheng*) | Mid range, harmonic | Court music; rare in folk |

### Percussion

| Instrument | Description | Role |
|------------|-------------|------|
| *Janggu* (장구) | Hourglass drum, played with stick (*chae*) on right and hand on left | The universal *jangdan*-keeper |
| *Buk* (북) | Single-headed barrel drum | *Pansori* drum; ensemble bass drum |
| *Jwago* (좌고) | Large barrel drum on a stand | Court ensembles |
| *Jing* (징) | Large bronze hand-held gong | *Jangdan* downbeats; outdoor music |
| *Kkwaenggwari* (꽹과리) | Small high-pitched gong | Lead instrument in *nongak* / *samulnori* |
| *Pyeonjong* (편종) | Set of 16 tuned bronze bells | Court ritual music |
| *Pyeongyeong* (편경) | Set of 16 tuned stone chimes | Court ritual music |
| *Bak* (박) | Wooden clapper of 6 boards | Court music section markers |

## Composing in Korean traditional style — practical advice

### Where to start

- **Pick the form first.** *Sanjo*-derived solo? *Sinawi*-derived ensemble? *Pansori*-style narrative? *Pungmul*-derived percussion piece? *Jeongak*-influenced slow chamber? Each has different conventions.
- **Pick the *jangdan*.** This determines tempo and feel more than any other choice. A *jinyangjo* piece is fundamentally different from a *jajinmori* piece, even with the same instruments and same melodic material.
- **Pick the mode (*jo*).** Bright (*pyeongjo*) or sorrowful (*gyemyeonjo*)? Regional influence?
- **Pick the instruments.** Solo *gayageum* with *janggu* is a different commitment from full *gugak* orchestra.

### Idiomatic gestures to learn

- *Jangdan* acceleration: a piece often progresses from slow → fast (*sanjo* archetype). Composing a static-tempo piece misses the form's central energy arc.
- *Sigimsae* inflections: build the melody knowing it will be ornamented. Skeletal Western melodies played on *gayageum* can sound non-idiomatic and bare.
- *Heterophony* over harmony: in *sinawi* and ensemble music, parts decorate the same melodic skeleton independently rather than supporting it harmonically. Don't force Western functional harmony into a *gugak* piece.
- Cyclical structure: *jangdan* repeats; the energy comes from variation within cycles, not from chord changes.

### Working with players

If you have access to *gugak* practitioners, collaborate. Notate the skeleton; let them ornament. Many contemporary *gugak* musicians can work from Western notation, *jeongganbo*, prose annotations, recordings, or hybrid sketches, but notation literacy and preferred workflow vary by performer and project. Be specific about which mode, *jangdan*, and emotional register you want; let the player help shape the *sigimsae*.

## K-trad fusion — modern hybrid music

Korean traditional fusion (*gugak* + popular music) is a live, rapidly evolving scene.

### Major artists / projects

- **이날치 (Leenalchi)** — Globally famous after the 2020 "Tiger is Coming" / Korean Tourism Organization videos. Combines *pansori* vocals with funk-rock band, retells *Sugungga* in modern dress.
- **잠비나이 (Jambinai)** — Post-rock / post-metal incorporating *piri*, *geomungo*, and *haegeum*. Internationally toured.
- **두번째달 (The Second Moon)** — Lyrical fusion of Korean traditional, Celtic, and Latin. *Arirang Fantasy* and the *Jeong-i* (정이) film soundtrack are signature.
- **추다혜차지스 (Choo Da-hye Charges)** — Traditional-vocal / shamanic-ritual-informed energy in a band context. Confrontational, post-traditional.
- **악단광칠 (ADG7)** — Crossover of *minsogak*, *jeongak*, and dance pop.
- **고래야 (Coreyah)** — World-music-fusion with emphasis on folk songs.
- **김덕수 사물놀이 (Kim Duk-soo SamulNori)** — Foundational *samulnori* group; many successors.
- **씽씽 (SsingSsing)** — Glam-rock-styled folk-song revival (now disbanded, members in solo work).
- **고요 (Goyo)** — Traditional vocal artist with electronic / chamber pop fusion.

### Common fusion patterns

| Approach | Example |
|----------|---------|
| Traditional vocal over modern band | Leenalchi, SsingSsing |
| Traditional instrument as featured solo color in pop | Many K-pop tracks (e.g., BTS "IDOL" uses traditional percussion samples) |
| Traditional rhythmic cycle (*jangdan*) under modern groove | Various jazz-fusion / electronic works |
| Modal melody over modern harmony | Common in K-drama OST and film score |
| Full traditional arrangement of pop song | Korean traditional covers of K-pop hits |
| Pop-arranged traditional piece | "Tiger is Coming", *arirang* fantasias |

### Modern uses

- **Film and drama OST**: *Seopyeonje* (1993, the film that sparked modern interest in *pansori*), *King and the Clown* (2005), *Sa-do* (2015), *The Throne*, *Hansan* (2022). Composers: Kim Su-cheol (김수철), Lee Byung-woo (이병우), Bang Jun-seok (방준석).
- **Game OST**: Korean games and global games with Korean settings increasingly use *gugak* elements (e.g., *Black Desert*, various MMORPGs).
- **Olympic / national event ceremonies**: 2018 PyeongChang Olympics opening / closing.
- **K-pop sampling**: producers sample *gayageum*, *daegeum*, *taepyeongso* for sonic color.
- **Theater and dance**: contemporary dance and theater frequently use *gugak* live or recorded.

## Common pitfalls when advising

- **Don't treat "Korean traditional" as one thing.** *Jeongak* and *minsogak* are as different as Bach chorale and Delta blues. Identify which you're working with.
- **Don't force Western harmony onto *gugak*.** Most traditional pieces are heterophonic or monophonic; chordal accompaniment is non-idiomatic in many traditional contexts and often makes the piece sound generic-fusion.
- **Don't ignore *jangdan*.** "Korean traditional rhythm" without specifying which *jangdan* is meaningless. *Jajinmori* and *jinyangjo* feel completely different.
- **Don't confuse *samulnori* with all Korean percussion.** *Samulnori* is one stage-formatted genre derived from *nongak*. Court percussion, *pansori* drumming, and *sanjo* drumming are different traditions with different conventions.
- **Don't reduce "Korean traditional" to a sound effect.** Adding *daegeum* sample to a Western pop song doesn't make it Korean traditional — it makes it pop with a *daegeum* sample. Genuine fusion engages with the traditional material's logic.
- **Don't forget *sigimsae*.** A melodic line on Korean instruments without ornament inflection sounds bare. Even sample libraries struggle to reproduce *sigimsae* well — when possible, work with players.
- **Don't assume tonal gravity.** *Gugak* modes don't have V→I cadences. Phrase endings work through other mechanisms: melodic descent to tonic, *jangdan* alignment, dynamic taper.
- **Don't romanticize.** "Mystical Eastern wisdom" framings are tokenistic and often inaccurate. Treat *gugak* as the highly-developed musical tradition it is, with as much specificity and rigor as you'd give classical Western music.

## Quick decision matrix

| Want | Try |
|------|-----|
| Lyrical, slow, sorrowful Korean traditional feel | *Jinyangjo* tempo with *gyemyeonjo* mode; *gayageum* or *haegeum* solo |
| Fast, driving traditional climax | *Jajinmori* or *hwimori* with full ensemble; *taepyeongso* lead |
| Festive Korean folk feel | *Gutgeori* tempo with *pyeongjo*; *samulnori*-style percussion |
| Solo virtuosic showpiece | *Sanjo* form: progressive *jangdan* sequence (*jinyangjo* → *jungmori* → ... → *hwimori*) |
| Court-music gravitas | *Jeongak*-style: slow tempo, full ensemble with *piri* lead, restrained ornamentation |
| Improvised ensemble feel | *Sinawi*-style: heterophonic ensemble over *jangdan* |
| Narrative / theatrical | *Pansori*-style: solo voice + *buk*, alternating *aniri* / *chang* or *sori* |
| Off-balance, supernatural, otherworldly | *Eotmori* (10-beat asymmetric) |
| Korean instruments in pop / electronic context | Pick 1–2 instruments for color (*daegeum*, *haegeum*, *gayageum*, *taepyeongso* are most adaptable); avoid trying to use all of them |
| K-trad fusion song | Reference Leenalchi, Jambinai, Coreyah, Second Moon for different fusion approaches |
| Film score with Korean traditional element | Study Kim Su-cheol, Lee Byung-woo, Bang Jun-seok for how to integrate traditional sound with cinematic scoring |

## Cross-references

- Modal systems and pentatonic frameworks → `../harmony/modal-harmony.md`, `../fundamentals/pitch-intervals-scales.md`
- Asymmetric meters (*eotmori* and similar) → `../rhythm-groove/odd-meters-polyrhythm.md`
- Other folk / roots / tradition-specific music practices → `folk-roots-and-traditions.md`
- Modern Korean popular music (K-pop and K-trad fusion overlap) → `kpop-jpop.md`
- Film scoring techniques (where K-trad fusion frequently lives) → `film-tv-scoring.md`
- Game scoring (Korean MMORPGs / mobile games often use *gugak* fusion) → `game-music.md`
- Microtonal pitch and *sigimsae*-style ornamentation → `../techniques/microtonal.md`, `../fundamentals/pitch-intervals-scales.md`
- Heterophony and texture (the dominant *sinawi* approach) → `../orchestration/voicing-and-texture.md`, `../counterpoint.md`
- Choral writing (relevant to *pansori* vocal traditions and Korean choral fusion) → `../orchestration/choral-writing.md`

# Metal, Punk, and Hardcore

> **Snapshot note:** Last content review snapshot: 2026-04-27. Use this file for durable composition conventions, but verify current subscene terms, production norms, artist references, and platform trends with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition-facing conventions in punk, hardcore, metal, pop-punk, metalcore, post-hardcore, thrash-adjacent writing, doom/sludge-adjacent writing, and heavy guitar music generally. It is **not** a complete guide to every metal or punk subgenre.

## When to consult this file

- The user wants punk, hardcore, metal, metalcore, pop-punk, post-hardcore, emo-adjacent, thrash, doom, sludge, or heavy guitar writing
- The user asks for riffs, breakdowns, chugs, tremolo picking, blast-like energy, mosh sections, or aggressive band texture
- The user wants to make a pop/rock song heavier
- The user wants to write guitar/bass/drum parts that feel playable rather than generic MIDI blocks
- The user wants intense energy without simply making everything louder

## Core idea

In metal, punk, and hardcore, **riff, rhythm, attack, and group energy** are often more structurally important than chord progressions.

A useful diagnostic:

```text
What is the physical action?
- head-nod / stomp
- sprint / pogo
- mosh / push-pull
- slow crush
- anthem singalong
- technical precision
```

Then write the rhythm, register, articulation, and form around that action.

## Substyle map

| Frame | Composition priorities | Common signs |
|---|---|---|
| Punk rock | speed, directness, short forms, shoutable hooks | downstrokes, power chords, fast backbeat |
| Pop-punk | punk energy + pop chorus | I-V-vi-IV, octave guitars, gang vocals |
| Hardcore punk | intensity, stops, short cells, group chants | fast tempos, breakdowns, shouted vocals, DIY directness |
| Post-hardcore / emo-adjacent | tension-release, dynamic contrast, melodic hooks | angular riffs, wide choruses, screamed/sung contrast |
| Heavy metal | riff architecture, distorted guitars, powerful vocals | power chords, modal minor, gallops, solos |
| Thrash-adjacent | speed, tight picking, palm-muted precision | E-string riffs, fast downpicking/alternate picking |
| Doom/sludge-adjacent | slow tempo, weight, repetition | low tuning, long sustain, chromatic descent |
| Metalcore | riff + breakdown + hook architecture | chugs, syncopation, half-time drops, clean chorus |
| Djent/prog-metal-adjacent | meter games, low strings, syncopated cells | odd groupings, tight gates, extended-range guitar |

## Harmony and pitch language

Heavy music often uses pitch as **riff contour** rather than functional harmony.

### Common pitch centers and modes

| Sound | Material |
|---|---|
| Straight punk/rock | major/minor pentatonic, I-IV-V, I-♭VII-IV |
| Dark metal | Aeolian, Phrygian, harmonic minor colors |
| Aggressive low riff | tonic pedal + ♭2/♭5/♭6 chromatic cells |
| Epic / power metal | natural minor, harmonic minor, modal mixture, sequence |
| Doom/sludge | chromatic descent, tritone, low pedal, slow modal loops |
| Pop-punk chorus | major key or relative minor with clear tonal pull |

### Progression families

| Use | Progression | Example in E |
|---|---|---|
| Punk anthem | I-♭VII-IV-I | E-D-A-E |
| Pop-punk chorus | I-V-vi-IV | E-B-C♯m-A |
| Dark minor | i-♭VI-♭VII-i | Em-C-D-Em |
| Phrygian bite | i-♭II-i | Em-F-Em |
| Metal sequence | i-♭VII-♭VI-V | Em-D-C-B7 |
| Doom loop | i-♭V-iv-i | Em-B♭-Am-Em |
| Hardcore chant | i / ♭VII / i stops | E5 / D5 / E5 hits |

Use full extended chords sparingly unless the substyle calls for prog, jazz-metal, post-rock, or emo color.

## Riff writing

A riff is not just a chord progression. It is a repeated physical pattern with identity.

### Riff variables

| Variable | Choices |
|---|---|
| Register | low-string pedal, mid-neck power chords, octave line, high tremolo line |
| Articulation | palm-muted, open, staccato, slide, bend, pinch harmonic, tremolo |
| Rhythm | straight 8ths, gallop, syncopated chug, triplet feel, odd grouping |
| Pitch | pedal + neighbor, chromatic cell, modal fragment, power-chord movement |
| Density | full band unison, guitar-only, bass counter-riff, drum-accented hits |

### Basic riff recipes

#### Punk downstroke riff

```text
Tempo: 176 BPM
Guitar: all downstrokes, power chords
| E5 E5 E5 E5 | A5 A5 B5 B5 |
Vocal: shoutable 3-5 note hook, little melisma.
```

#### Metal pedal riff

```text
Center: E
Pattern: E pedal on 8ths, accents on F and G
E E F E | E G F E |
Effect: Phrygian bite without needing complex harmony.
```

#### Metalcore breakdown seed

```text
Tempo: 150 BPM, half-time feel
0---0-0---0--- | 0-0---0--3-2-- |
Kick follows chugs; snare lands huge on 3.
```

#### Doom/sludge slow crush

```text
Tempo: 68 BPM
| E5 | B♭5 | A5 | E5 |
Let chords decay. Do not fill every gap.
```

## Rhythm and drums

| Device | Effect |
|---|---|
| Fast straight 8ths | punk urgency |
| D-beat-like drive | hardcore/punk propulsion |
| Double-time drums over mid-tempo riff | creates panic without changing guitar tempo |
| Half-time breakdown | makes the same tempo feel heavier |
| Gallop rhythm | classic metal propulsion |
| Blast-like texture | extreme intensity; use only when the style wants it |
| Stop-time hits | crowd impact, mosh cue, lyric punctuation |
| Odd grouping over 4/4 | prog/djent tension |

A heavy section often works because the drums simplify at the right moment. If every instrument plays maximum density, the impact can shrink.

## Guitar and bass idioms

### Guitar

- Power chords are default because distortion makes full triads muddy.
- Palm muting controls weight more than the chord symbol does.
- Open low strings make riffs feel idiomatic; transposition can break the part.
- Down-tuning changes both range and hand feel.
- Octave shapes cut through without adding harmonic complexity.
- Tremolo picking creates sustain-through-motion.

### Bass

- Bass can double the guitar riff for weight, but strategic divergence makes sections bigger.
- In punk/hardcore, eighth-note root motion can be the engine.
- In metal, bass may reinforce low guitar, add fills between vocal phrases, or lock with kick accents.
- In a breakdown, simpler bass often hits harder.

See `../instrument-idiom/guitar.md`, `../instrument-idiom/bass.md`, and `../instrument-idiom/drums-percussion.md` for playability.

## Vocals and hooks

Heavy vocals are highly style-specific. Composition guidance should avoid assuming that every vocalist can scream, belt, growl, or sustain high notes safely.

Useful hook strategies:

- gang vocal on the title
- call-response between lead and band
- screamed verse + sung chorus contrast
- rhythmic chant over one pitch
- melodic chorus after riff-heavy verse
- short repeated phrase before breakdown

### Lyric tone

| Frame | Lyric tendencies |
|---|---|
| Punk | direct, social, sarcastic, alienated, funny, blunt |
| Hardcore | communal, confrontational, ethical, scene-centered, cathartic |
| Metal | mythic, dark, existential, technical, political, horror, fantasy |
| Pop-punk | youth, relationships, irony, self-deprecation, nostalgia |
| Emo/post-hardcore | confession, tension, rupture, memory, interior conflict |

Do not force violent or grim content. Heavy music can be joyous, absurd, spiritual, political, technical, or cathartic.

## Form

Common forms:

```text
Intro riff → Verse → Chorus → Verse → Chorus → Bridge / breakdown → Final chorus
```

```text
Riff A → Riff B → Verse → Riff A → Breakdown → Riff C → Outro
```

```text
Fast section → stop-time chant → half-time breakdown → double-time return
```

Heavy forms often use **section contrast** more than smooth transition. Abruptness can be a feature if it is cued by a drum fill, scream, pick scrape, or stop.

## Arrangement and production-facing composition

Composition choices that make production easier:

- leave space between low guitar, bass, and kick
- avoid stacking full six-note guitar chords under dense cymbals
- place pick attacks and kick transients deliberately
- double rhythm guitars only when the part is tight enough
- use octaves or single-note hooks to make choruses cut
- reserve cymbal wash for climaxes; too much wash hides riff detail
- make breakdowns rhythmically memorable, not merely low and slow

## Quick recipes

### Make a pop song punkier

- raise tempo or imply double-time
- simplify chords to power chords
- shorten verse lines
- add gang vocal after the hook
- make drums more direct; fewer ghost-note details

### Make a rock riff heavier

- move the riff to a lower open-string center
- palm-mute selected notes, not everything
- remove full triads from distorted guitars
- let bass and kick lock on the main accents
- put a rest before the biggest hit

### Make a breakdown hit

- simplify pitch material
- widen the register: low chugs + high noise/lead/scream
- switch to half-time snare placement
- add one memorable rhythmic cell
- stop everything for one beat before the return

## Common pitfalls

- **Do not write full keyboard chords and expect them to translate to distorted guitar.** Guitar shapes and distortion physics matter.
- **Do not confuse heaviness with constant density.** Rests and contrast make impact.
- **Do not flatten punk, hardcore, and metal into one angry sound.** Their histories, tempos, scenes, and ethics differ.
- **Do not overcomplicate punk unless the user asks for prog/post-hardcore.** Directness is a craft value.
- **Do not ignore tuning.** A riff written in E standard may lose its identity when moved to drop C, and vice versa.
- **Do not assume every vocalist can do extreme techniques.** Ask or keep alternatives.

## Quick decision matrix

| User wants | Try |
|---|---|
| More punk | faster tempo, downstrokes, short phrases, gang hook |
| More hardcore | stop-time, shouted rhythm, half-time impact, fewer chords |
| More metal | lower pedal riffs, modal minor, palm muting, guitar/bass/kick lock |
| More doom | slower tempo, longer decay, fewer notes, chromatic weight |
| More pop-punk | bright chorus, octave guitars, I-V-vi-IV, big backing vocals |
| More metalcore | syncopated chugs, screamed/sung contrast, breakdown architecture |
| More prog/djent | odd groupings, extended-range register, metric displacement |

## Cross-references

- Pop and rock fundamentals → `pop-rock.md`
- Guitar idioms → `../instrument-idiom/guitar.md`
- Bass idioms → `../instrument-idiom/bass.md`
- Drum idioms → `../instrument-idiom/drums-percussion.md`
- Modal harmony → `../harmony/modal-harmony.md`
- Rhythm and displacement → `../rhythm-groove/rhythmic-devices.md`, `../rhythm-groove/odd-meters-polyrhythm.md`
- Arrangement density → `../orchestration/arrangement-density.md`
- Production-aware arrangement → `../production-aware/arrangement-for-mix.md`
- Current scene research → `../research/web-music-trend-research.md`, `../research/reference-track-digging.md`

# Musical Theatre

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

Musical theatre is a hybrid genre — drama set to music, where songs serve narrative and character functions as well as standalone musical pleasures. The conventions of musical theatre songwriting are distinct from pop, opera, and film scoring, with their own vocabulary of song types, structural devices, and idiomatic moves.

For pop song forms (which musical theatre borrows and modifies), see `form/popular-song-forms.md`. For classical operatic vocal writing, see `classical-periods.md`. For lyric writing specifically, see `songwriting/lyric-writing.md`. For vocal range and character, see `orchestration/instruments-ranges-character.md`.

## When to consult this file

- Writing for stage musicals (Broadway, West End, regional theatre)
- Songs that serve narrative functions
- "I Want" songs, "11 o'clock" numbers, list songs, patter songs
- Cabaret and revue material
- Understanding why a song "doesn't move the story forward" — usually a missing dramatic function
- Composing in styles ranging from Golden Age to contemporary

## What musical theatre songs do

Songs in a musical aren't just musical numbers — they're dramatic events. The classic distinction:

- **Spoken dialogue**: characters communicate information.
- **Song**: characters express something they can't express in speech alone — heightened emotion, internal feelings, decisions, transformation.

The transition from speech to song happens when emotion exceeds the capacity of speech. A character bursts into song because they have to.

Every song should accomplish at least one dramatic function:

1. **Reveal character** (who they are, what they want, what they hide)
2. **Advance plot** (a decision is made, a relationship changes, an obstacle is overcome)
3. **Establish setting / mood** (the world the characters inhabit)
4. **Comment on action** (often the chorus's role — observation, perspective)

A song that's purely entertaining without dramatic function may be cut in revisions. The ones that survive earn their place by serving the show.

## Standard song types

A vocabulary of song types developed over the 20th century. Most musicals contain several of these.

### "I Want" song

Early in the show. The protagonist articulates their central desire — what they want that they don't have. Drives the entire plot.

Examples:
- "Wouldn't It Be Loverly" (My Fair Lady, Eliza)
- "Part of Your World" (The Little Mermaid, Ariel)
- "Defying Gravity" (Wicked, Elphaba — late "I want" but functions as one)
- "Corner of the Sky" (Pippin, Pippin)
- "Waving Through a Window" (Dear Evan Hansen, Evan)

Structurally: usually mid-tempo or moderate, often building from quiet introspection to anthemic chorus. Lets the audience identify with the protagonist's longing.

### "I Am" song / Character song

Establishes who a character is. Often comedic or quirky for secondary characters.

Examples:
- "I Cain't Say No" (Oklahoma!, Ado Annie)
- "Popular" (Wicked, Glinda)
- "When You're Good to Mama" (Chicago, Matron "Mama" Morton)

### Love song / Duet

Two characters establish or develop their relationship in song. Often the romantic peak of the show.

Examples:
- "Tonight" (West Side Story)
- "If I Loved You" (Carousel)
- "All I Ask of You" (Phantom of the Opera)
- "A Million Years Ago" (Le Mis: "A Heart Full of Love" sequence)

Structurally: often AABA with the bridge providing emotional contrast.

### Comic song / Patter song

Rapid-fire lyrics, often with humorous content. Demands strong diction.

Examples:
- "I Am the Very Model of a Modern Major-General" (Pirates of Penzance)
- "The Internet Is for Porn" (Avenue Q)
- "Schadenfreude" (Avenue Q)
- "Wand'rin' Star" (Paint Your Wagon)

Structurally: high syllabic density, often verse-chorus with clearly punctuated chorus phrases.

### List song

A song organized as a list of items, examples, or instances. Common structural device.

Examples:
- "I've Got a Little List" (The Mikado, Ko-Ko)
- "Always Look on the Bright Side of Life" (Spamalot)
- "Modern Major-General" (also a patter song)
- "These Are a Few of My Favorite Things" (Sound of Music)

The "list" provides the structure; lyrics enumerate items with a refrain anchor.

### "11 o'clock" number

The big climactic song late in the show. Typically a star turn for the lead, often featuring high notes, big emotion, and a definitive moment of resolution or decision.

Examples:
- "Rose's Turn" (Gypsy)
- "Memory" (Cats)
- "I Dreamed a Dream" (Les Misérables — actually closer to mid-show but functionally similar)
- "Don't Cry for Me Argentina" (Evita)
- "Defying Gravity" (Wicked, Act 1 finale, but functions as 11 o'clock)

Named because in classic Broadway shows it would land around 11pm in an 8pm-curtain show.

### Reprise

A song that returned later in the show with new lyrics, modified context, or as climactic statement. Reprises connect emotional moments across the show.

Examples:
- "Memory" appears multiple times in Cats with different emotional weight
- "Do You Hear the People Sing" appears at the end of Les Mis Act 1 and at the very end of the show
- "I Dreamed a Dream" reprised at Fantine's death

Reprises rely on motivic recognition — when a song returns, the audience feels the resonance of the original.

### Group / ensemble number

The full company singing together. Often opens or closes acts. Provides spectacle and establishes the world.

Examples:
- "One" (A Chorus Line)
- "Tradition" (Fiddler on the Roof)
- "Belle" (Beauty and the Beast — full company)
- "Wells Fargo Wagon" (The Music Man — full town)

### Showstopper

A song so popular or impactful it stops the show — the audience interrupts with applause. Tends to be character-defining or emotionally peak moments. Many "I Want" songs and "11 o'clock" numbers are showstoppers.

### Comedy duet / Trio / Sextet

Multiple characters in dialogue through song. Often comedic. Each character has their own lyrical content.

Examples:
- "Anything You Can Do" (Annie Get Your Gun, Annie + Frank)
- "I Will Never Leave You" (Side Show, Daisy + Violet)
- The sextets in Mozart operas (precursor to musical theatre conventions)

## Structural conventions

### Act structure

Most musicals are 2 acts (with intermission). A few notable exceptions are 1-act shows or 3-act revues.

**Act 1**:

- Opens with a number that establishes the world (often the company singing).
- "I Want" song establishes protagonist's desire.
- Inciting incident (often songless) sets the plot in motion.
- Mid-Act 1: developing relationships, increasing complications.
- **Act 1 finale**: climactic number where the protagonist commits to the journey, often raising stakes for Act 2. Often the most ambitious musical statement so far.

**Act 2**:

- Opens with a song that reorients the audience (often a contrast in mood from Act 1 finale).
- Crisis and apparent disaster.
- "11 o'clock" number — the protagonist's transformative musical moment.
- Resolution.
- Final ensemble number bringing the company back.

### Underscore vs. song

In musical theatre, the orchestra plays continuously through some scenes — *underscore* (incidental music supporting dialogue and action). Songs are highlighted moments emerging from underscore.

The transition into a song is a craft moment: how does spoken dialogue lead to sung melody? Common techniques:

- **Speech-to-song through accompaniment**: the orchestra starts playing under dialogue; the dialogue gets more emotional; suddenly the character sings the next line.
- **Vamp introduction**: a short vamp under dialogue, building tension; the character launches into the song.
- **Song from silence**: dialogue stops; the character sings without orchestral preparation.

### Verse and refrain (older Broadway)

Older showtunes (1900s–1950s) often distinguish *verse* (a recitative-like introduction) from *refrain* (the famous melody). The "song" you know from a recording is usually the refrain; the verse is often omitted in cabaret performance.

Example: "Over the Rainbow" has a verse ("When all the world is a hopeless jumble") that's rarely performed; the refrain ("Somewhere over the rainbow") is what everyone knows.

This distinction is mostly historical; modern shows generally don't use it.

## Stylistic eras

### Golden Age (1940s–1960s)

Rodgers and Hammerstein, Lerner and Loewe, Frank Loesser, Jule Styne, Cole Porter (late). The classics: Oklahoma!, South Pacific, My Fair Lady, West Side Story, The Sound of Music, Guys and Dolls, Gypsy.

Characteristics:

- **Tunesmith aesthetic**: songs designed to be hummable and memorable.
- **Operetta and big-band influence**: orchestrations include strings, brass, woodwinds.
- **Verse-refrain structure** common.
- **AABA 32-bar form** dominant.
- **Strong character-driven storytelling**.
- **Dance numbers** as integral parts of the show.

### Concept musical era (1960s–1970s)

Sondheim's emergence (Company, Follies, A Little Night Music, Sweeney Todd, Sunday in the Park with George). Hair, Cabaret, Chicago, A Chorus Line.

Characteristics:

- **Concept** over linear story (Company is a series of vignettes about marriage; Follies is about memory and aging).
- **More complex harmony and form**: Sondheim borrowed from Ravel, Britten.
- **Ironic and ambivalent** rather than uplifting.
- **Sophisticated lyrics**: Sondheim's lyric craft set new standards.
- **Less reliance on hummable hits**; the song is for the show, not the radio.

### Megamusical era (1980s–1990s)

Andrew Lloyd Webber, Schönberg/Boublil. Cats, Les Misérables, Phantom of the Opera, Miss Saigon. Spectacle-driven; lyrical pop sensibilities; sung-through (no spoken dialogue) common.

Characteristics:

- **Sung-through** (or nearly so).
- **Big anthemic ballads** designed to chart on radio.
- **Visual spectacle**: helicopters (Miss Saigon), chandeliers (Phantom).
- **Pop-influenced harmony**.
- **Operatic-pop hybrid** vocal style.

### Contemporary (2000s–present)

A blend of styles. Multiple aesthetic streams active.

- **Pop musicals**: shows by pop songwriters (Spongebob, Mean Girls, Heathers, & Juliet). Pop song forms; pop production aesthetic.
- **Hip-hop musicals**: Hamilton, In the Heights (Lin-Manuel Miranda). Rap as primary musical form; pop melodies.
- **Continued Sondheim legacy**: Adam Guettel (The Light in the Piazza), Jason Robert Brown (Parade, The Last Five Years), Jeanine Tesori (Caroline, or Change; Fun Home).
- **Pasek and Paul** (Dear Evan Hansen, La La Land): pop-influenced contemporary musical theatre.
- **Disney stage adaptations**: Beauty and the Beast, The Lion King, Aladdin — children-friendly, score-driven.

## Vocal writing for musical theatre

Belt and mix voice are the dominant vocal modalities in modern musical theatre. Rare classical "head voice" is reserved for specific styles or older repertoire.

### Belt range

The belted "money note" — a loud, chest-voice (or chest-mix) high note — is a Broadway staple. Common belt notes:

- Female: D5–F5 belt; G5+ for big belters
- Male: F4–A4 belt; B4+ for big belters

A song's emotional climax often coincides with a belt to the highest note.

### Vocal "tags"

Final repeats with new high notes, ad-libbed riffs, or sustained climactic notes. The ending of "Defying Gravity" — held high note — is the prototype.

### Mixed voice for sustainability

Belted singing is harder to sustain than mixed voice. Long shows demand vocal stamina, so most singing is in mix; belts are saved for moments.

### Vocal style by character

Different characters often sing in different vocal styles:

- The ingenue: pure tone, lyrical mix.
- The leading man: full chest, baritone or tenor warmth.
- The comic: brassy, character voice; often less classical.
- The villain: dark, sometimes operatic; or sneering character voice.
- The chorus: blended ensemble voice.

## Lyric craft

Sondheim's principles (briefly):

- **Less is more**: clarity and concision.
- **Concrete language**: specific words, not abstract emotions.
- **Action verbs**: songs progress through action.
- **Internal rhyme**: enriches without forcing.
- **Subtext**: the lyric reveals what the character won't say directly.
- **Function over flowers**: the lyric serves the dramatic moment.

Bad theatrical lyrics: vague, abstract, disconnected from character or plot. Good lyrics: specific, character-driven, dramatically necessary.

For more on lyric writing, see `songwriting/lyric-writing.md`.

## Common pitfalls when advising

- **Don't recommend a song that doesn't serve a dramatic function.** "It would be a great song here" isn't enough — the song must do something for the story.
- **Don't ignore the character.** The same dramatic moment sung by different characters should sound different. Match style to character.
- **Don't write only for the recording.** A song works in context (with dialogue, action, staging). A song that's gorgeous on a cast album but doesn't land in the show isn't doing its job.
- **Don't use generic Broadway pastiche.** Modern audiences can hear Andrew Lloyd Webber pastiche from a mile away. If the show has a specific aesthetic (rock, R&B, hip-hop), commit to it.
- **Don't ignore the tempo of dramatic time.** Some moments need slow ballads; some need fast comic songs. The pacing of musical numbers across the show matters.
- **Don't write impossible vocal lines.** Demands like 8 belt notes per phrase or sustained high A's for 16 bars are not sustainable for an 8-show-week run. Write within practical vocal limits.

## Quick decision matrix

| Want | Try (song type) |
|------|-----------------|
| Audience identification with protagonist | "I Want" song early in Act 1 |
| Character introduction (secondary) | "I Am" song |
| Romantic peak | Duet, often AABA, building from intimate to anthemic |
| Comic relief | Patter song or comic duet |
| Itemized humor or content | List song |
| Climactic protagonist transformation | "11 o'clock" number |
| Connect emotional moments across show | Reprise of an earlier song |
| Establish the world / community | Group / ensemble number |
| Smooth transition from dialogue to song | Vamp accompaniment or speech-to-sung gradient |
| Big finale with company | Full ensemble number, often a reprise |
| Golden Age feel | AABA, verse-refrain, orchestral, big-band-derived |
| Modern Broadway feel | Pop verse-chorus with theatrical lyric craft |
| Hip-hop musical | Rap verses with sung choruses; *Hamilton* / *In the Heights* template |
| Sondheim-style sophistication | Through-composed, complex harmony, internal rhyme, ambivalent emotion |

## Cross-references

- Lyric writing → `songwriting/lyric-writing.md`
- Hooks and memorability → `songwriting/hooks-and-memorability.md`
- Topline (vocal melody) craft → `songwriting/topline-craft.md`
- Pop song forms (which musical theatre adapts) → `form/popular-song-forms.md`
- Vocal ranges and character → `orchestration/instruments-ranges-character.md`
- Energy arcs and transitions across the show → `form/narrative-and-transitions.md`
- Classical operatic forerunners → `classical-periods.md`
- Pop and rock styles that inform contemporary musical theatre → `pop-rock.md`
- Hip-hop conventions for hip-hop musicals → `hip-hop-rnb.md`

# Media and Commercial Music

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition for media outside film, TV, and games — primarily advertising, podcasts, branded content, corporate / explainer video, social media, trailers, and similar short-form commercial work. These are large industries with their own conventions, distinct from concert music or even film/TV scoring. The constraint is usually short duration, clear functional purpose, and tight client requirements.

This file complements `film-tv-scoring.md` (linear narrative scoring; closely related), `game-music.md` (interactive media), `../production-aware/pre-production-decisions.md` (decisions before composing). For specific musical genres a media piece might draw on (orchestral, electronic, folk, etc.), see the relevant `genres/*.md` file.

## When to consult this file

- Composing for advertising, commercials, brand identity
- Podcast theme and stinger composition
- Corporate / explainer / training video music
- Social media short-form content (TikTok hooks, Instagram reels, YouTube intros)
- Trailer music (film trailers, game trailers, app trailers)
- Production music libraries / stock music
- Promo music for radio, broadcast, streaming
- Branded content (sponsored video series, content marketing)

## What's distinctive about media music

Compared to film/TV scoring or concert music, media composition has these defining features:

| Feature | Implication |
|---------|-------------|
| Short duration (5–60 seconds typical) | The whole musical idea must land quickly; no long development |
| Functional / persuasive intent | Music serves a goal (sell, inform, brand, engage) — aesthetic value secondary |
| Tight client constraints | Briefs may specify mood, length, style, and required hits |
| Multiple deliverables | One piece often needs 60s, 30s, 15s, 6s versions; sometimes 5min full mix |
| Genre-fluent expectation | Composer must work in many idioms (orchestral, EDM, hip-hop, indie, folk) on demand |
| Quick turnaround | Days or hours, not weeks |
| Royalty / licensing structure | Different from film/TV; production music libraries are common |

## Advertising and commercial music

The largest category. Sub-types and their conventions:

### Brand identity / sonic logo

A 1–4 second musical signature for a brand. Examples: Intel "bong-bong-bong-bong" (5 notes), Netflix "ta-dum", McDonald's "ba-da-ba-ba-ba" (the "I'm Lovin' It" notes), THX deep note opener. These are intensively crafted and represent significant licensing budgets.

Compositional priorities:
- **Memorable from first listen** — the listener may hear this hundreds of times and must recognize it instantly each time
- **Distinct from competitors** — must not sound like another brand's signature
- **Adaptable** — the signature should work across contexts (TV, radio, web, in-store, on-hold, voice assistant)
- **Durable** — will be in use for years; must not feel "of its moment" too narrowly

Sonic logos are usually composed by sonic-branding specialists; budgets are substantial.

### TV / broadcast commercial (15s, 30s, 60s)

The classic format. Standard durations: 15 seconds, 30 seconds (most common), 60 seconds. Music is typically:

- **Underscore for VO**: bed under voice-over with controlled dynamic that doesn't overwhelm the words
- **Theme-driven**: a short identifiable theme that connects to the brand's emotional positioning
- **Genre-matched**: contemporary pop / hip-hop for younger demos; warmer / acoustic for family / lifestyle; orchestral for prestige (luxury, automotive, finance)
- **Climactic resolution**: builds to a moment of emotional payoff aligned with the brand reveal or call to action

Common structure for a 30-second spot:

| Time | Function |
|------|----------|
| 0:00–0:05 | Setup / attention grab; sparse music or sonic hook |
| 0:05–0:15 | Story development / VO; underscore (low dynamic) |
| 0:15–0:25 | Emotional turn / product reveal; music swells |
| 0:25–0:30 | Brand stamp / sonic logo; final hit / sustain |

### Online / digital advertising (varies, often 6s to 30s)

YouTube pre-roll, social ads, programmatic display. Increasing use of short-form formats:

- **6-second bumpers** (YouTube): the entire music idea has to land in 6 seconds. Often a single sonic gesture: a hook, a stinger, a sound-design swell.
- **15s and 30s variants**: similar to TV but optimized for muted-by-default playback (much social viewing has sound off until a click). Music plus visual must work together; music alone often can't carry the spot.
- **Auto-play with sound on tap**: music designed to "earn" the unmute — strong sonic hook in first second.

### Long-form branded content (1–10 minutes)

Brand documentaries, sponsored video series, "branded entertainment". Closer to film/TV scoring in length but with brand objectives in mind. Often original orchestral or hybrid scores.

## Podcast music

A growing area. Sub-types:

### Theme music (intro / outro)

The signature musical identity of a podcast. Typical: 15–60 seconds, distinctive instrument or sonic palette, identifiable melodic hook.

Compositional priorities:
- **Captures the show's tone** — comedic vs. serious, intimate vs. broadcast-style, niche vs. mass
- **Sets up the host's voice** — the music should leave a clean handoff to spoken word, often at a medium-low dynamic
- **Tolerates repetition** — listeners hear this every episode; it should reward repeated listening
- **Recognizable instantly** — listeners should know which podcast they're hearing within 2–3 seconds

### Bumpers and transitions

Short musical cues (2–10 seconds) used to mark section breaks within an episode. Often variations on the main theme; sometimes wholly different cues for specific recurring segments.

### Underscore (for narrative / interview podcasts)

Some podcasts (especially documentary-style) use underscore similarly to film. *Serial*, *This American Life*, *S-Town*, and similar narrative podcasts have substantial musical content.

Compositional priorities:
- **Subordinate to voice** — the listener is there for the words; music serves them
- **Atmospheric over melodic** — sustained pads, ambient textures, restrained motion
- **Emotionally specific** — music tells the listener how to feel about the moment without overstating

### Sponsored / advertising spots within podcasts

Mid-roll or pre-roll ads with their own music. Same conventions as broadcast advertising but compressed (often 30–90 seconds total ad with 10–20 seconds of musical content).

## Corporate / explainer / training video

A vast volume of content: product explainer videos, internal training, corporate communications, conference videos. Music typically:

- **Neutral-positive emotional palette** — uplifting without being saccharine, professional without being clinical
- **Genre-restrained** — usually contemporary pop-influenced or hybrid orchestral/electronic; rarely strongly genre-specific (no metal, no avant-garde)
- **VO-friendly** — designed to sit under voice-over comfortably
- **Long enough to cover full duration** — often 2–10 minutes; loops or extended composition

This category often uses production music libraries rather than custom composition. Custom composition for corporate work tends to be for higher-budget brand films, executive presentations, and major launch campaigns.

## Social media short-form

TikTok, Instagram Reels, YouTube Shorts, Twitter/X video. Distinctive features:

- **Music is often pre-existing** (TikTok's library of licensed tracks is the music; users select from it) rather than composed for the post
- **For branded social content**: short hooks (3–15 seconds) that work as standalone musical ideas
- **Trend-responsive**: musical styles on social media shift quickly; what was a hit hook 3 months ago may feel dated
- **Often sample-based or AI-assisted** in indie / creator contexts

For the composer, social media is mostly a distribution channel for licensed or library music. Custom composition for social media is mostly for brands with their own sonic identity to maintain.

## Trailer music

Trailer music is its own micro-industry, distinct from film scoring (even though trailers exist alongside films).

### Why trailer music is different

Film scores are written *for* the film, generally after picture lock; trailers are typically cut *to* pre-existing music. The trailer editor selects from a music library, finds tracks that match the desired emotional arc, and cuts picture to fit the music's rises and drops.

Trailer music libraries (Two Steps from Hell, Audiomachine, Position Music, etc.) are entire businesses producing tracks specifically for trailer use — designed with trailer cutting in mind.

### Trailer music conventions

Trailer cues are highly conventional. Typical structures:

- **Build-up**: slow, atmospheric opening (10–30s)
- **First drop / hit**: a major impact moment with thematic statement
- **Continued build with thematic development**
- **Climax / final drop**: maximum intensity, full orchestration
- **Tag / resolution**: short conclusion, often softer

Common trailer music elements:
- "Big" orchestral percussion (taikos, large drums, stadium-percussion)
- Choral pads (often Latin or vocalese, no specific text)
- Synth bass / sub bass for impact
- Risers, sweeps, sound-design transitions
- Brass / horn calls for hero / heroic feel
- String ostinatos for tension

### Trailer hits

Short percussive / harmonic accents (impacts, "whooshes", boom hits) used at picture cuts. These are essentially sound design and are usually part of trailer music tracks rather than separate elements.

### Studying trailer music

The "Inception trailer" (2010) — using "Mind Heist" by Zack Hemsey — is widely cited as the moment trailer music conventions consolidated. Listening to many recent trailers and identifying the musical conventions is the best way to learn the idiom. The "BWAAAH" sound is now a trailer-music cliché — part of the genre vocabulary.

## Production music / library music

Pre-composed music licensed for media use. Composers write tracks for libraries rather than for specific projects; clients (filmmakers, ad agencies, corporate video producers) license tracks from the library.

### Library music conventions

- **Each track is generic enough to fit many uses** but specific enough to have a recognizable mood
- **Tracks often come in versions**: full mix, no melody (so VO can replace the lead), no drums, 60s/30s/15s edits, stinger/loop/stem packages
- **Genre / mood metadata** is essential for discoverability: tags like "uplifting corporate piano", "epic action cinematic", "indie folk acoustic", "dark electronic tension"
- **Royalty structure**: composers earn writer's share + (sometimes) publisher's share; libraries typically own / co-own the publishing
- **Volume composition**: library composers often write hundreds of tracks per year

This is a viable composition career path with its own craft — writing efficiently, hitting briefs precisely, producing professional masters quickly.

### Library music vs. custom composition

| Aspect | Library | Custom |
|--------|---------|--------|
| Cost | Lower (licensing fee) | Higher (composer fee + recording) |
| Specificity | Generic / multi-use | Tailored to project |
| Speed | Instant (license and use) | Days to weeks |
| Identity | Shared with other users of the same library | Unique to project |
| Best for | Lower-budget productions, "background" needs | Hero spots, signature campaigns, branded content |

## Composing for media — practical workflow

1. **Read the brief carefully.** Briefs usually specify: duration, emotional / mood targets, genre / style, reference tracks, deliverables (multiple lengths), deadline, technical specs (file format, sample rate, loudness target).
2. **Identify the function.** What does the music need to *do*? Set mood? Build to a CTA? Underscore VO? Each role has different conventions.
3. **Match the duration target precisely.** If the brief says 30 seconds, the piece is exactly 30 seconds. Not 28, not 33.
4. **Plan the emotional arc within the duration.** A 30-second piece has roughly 3–4 emotional beats; a 6-second bumper has 1–2.
5. **Identify "sync points"** — moments in the visuals or VO where music must land specifically (logo reveal, product shot, key word in the script).
6. **Compose with the cut in mind.** Music for media is designed to support cuts; don't ignore where the visual editor needs the music to breathe or hit.
7. **Deliver multiple versions when required.** A standard ad music delivery often includes: 60s, 30s, 15s, 6s, and stems (drums-only, music-only, etc.).
8. **Master to broadcast / streaming standards.** Different platforms have different loudness targets (LUFS); the brief usually specifies. Common targets: -23 LUFS for European broadcast (EBU R128), -24 LKFS for US TV (CALM Act), -16 to -14 LUFS for streaming services and YouTube, varies for social media.

## Working with clients

Media composition is intensely client-driven. A few patterns:

### Brief interpretation

Briefs use evocative language ("uplifting", "modern", "human", "cinematic") that's open to interpretation. The first draft is often "wrong" — the client knows what they don't want but couldn't say what they did want until they hear something. Multiple revisions are normal.

### Reference tracks

Clients often provide reference tracks ("we want it to feel like X"). Treat these as goals, not as templates. The actual composition should be inspired by but not derivative of the reference.

Watch out: some clients want you to *exactly* match a reference, which can risk copyright issues. Push back diplomatically: "We can capture this energy / feel without being legally exposed." See `../production-aware/pre-production-decisions.md` for reference-track best practices.

### Revisions

Revisions are part of the work. Build them into the timeline and contract. Common revision types:

- **Tonal**: "more uplifting", "less dark"
- **Structural**: "build the chorus earlier", "we need to extend before the logo"
- **Genre**: "less orchestral, more electronic"
- **Mix**: "lower the bass", "drum hits should be punchier"
- **Sync**: "the swell needs to land 2 seconds later"

Each kind of revision has different implications for cost and time. A "tonal" revision might be a melodic adjustment; a "genre" revision might be a complete redo.

### Approval and ownership

Before final delivery, confirm:
- Who has final approval (creative director, brand manager, client lead)?
- What's the deliverable specification (file format, length, masters needed)?
- What rights are being purchased (broadcast rights only, perpetual rights, exclusive vs. non-exclusive)?
- Are there multiple usage windows (initial campaign + longer-term licensing)?

## Common pitfalls when advising

- **Don't ignore the brief.** Media composition is brief-driven. A piece that ignores the brief, no matter how musically excellent, is a failure.
- **Don't make the music more important than the message.** In ads, podcasts, corporate video, the music serves the communication. Distinctive music that overwhelms the script or visuals is a problem.
- **Don't ignore duration constraints.** A 30-second piece must be 30 seconds. Adjusting the visual to fit the music is rarely an option.
- **Don't compose without the sync points.** If a logo reveals at 28 seconds, the music must build to land at 28 seconds.
- **Don't underestimate revisions.** Plan for 2–5 revision rounds on most ad work.
- **Don't deliver poorly mastered files.** Broadcast and streaming have specific loudness targets; missing them creates extra work for the post-production team.
- **Don't ignore VO.** Most media music shares space with voice-over; the music must duck appropriately and leave frequency room for the voice.
- **Don't use copyrighted material as reference too literally.** "Make it sound exactly like Track X" risks legal issues. Capture the *feel* without copying the content.
- **Don't apply concert-music aesthetic priorities.** Commercial music is functional. The composer's pride is in the work serving the brief well, not in the abstract musical qualities.
- **Don't underprice your work.** Media composition has standard rates by category (broadcast TV ad, online video, corporate video, etc.). Know the going rates for your market.
- **Don't ignore loudness creep.** Mastering louder than the brief specifies to "make it sound bigger" causes broadcast / streaming compliance issues.

## Quick decision matrix

| Want | Try |
|------|-----|
| 30-second TV ad | Build through 3 emotional beats (setup / development / payoff); land brand stamp on final beat; underscore volume for VO |
| Podcast theme | 15–60s; distinctive sonic palette; clear handoff to spoken word; tolerates repetition |
| Sonic logo | 1–4 seconds; intensively memorable; distinct from competitors; works across delivery contexts |
| Corporate explainer | Neutral-positive palette; loops or extends to needed length; VO-friendly low dynamic |
| YouTube 6-second bumper | One sonic gesture / hook; lands instantly; "earns the unmute" |
| Trailer cue | Build-up + first drop + climactic drop + tag; cinematic orchestral or hybrid; lots of percussion / sound-design impact |
| Library track for "uplifting corporate" | Major key, mid-tempo (110–125 BPM), piano/strings/light electronic, lyric-free, easy-to-cut sections |
| Library track for "epic action cinematic" | Heavy percussion, brass, choral pads, building intensity, multi-section structure |
| Branded content longform | Closer to film scoring; full original score; brand emotional positioning shapes thematic palette |
| Social media custom | Short hook focused on instant recognition; plan for muted playback (visuals carry weight too) |
| Reference track says "feel like X" | Capture energy/mood; don't copy melodic content; check legal exposure |
| Multiple deliverables (60/30/15/6) | Compose the full piece first; create shorter versions by selecting strongest moments rather than time-compression |

## Cross-references

- Linear film and TV scoring (closely related, much technique transfers) → `film-tv-scoring.md`
- Game and interactive media scoring → `game-music.md`
- Algorithmic / AI-assisted composition (increasing role in library music and short-form) → `../techniques/algorithmic-and-AI-assisted.md`
- Pre-production decisions (key, tempo, length, references) → `../production-aware/pre-production-decisions.md`
- Energy and dynamics across short forms → `../production-aware/energy-and-dynamics.md`
- Form and narrative shaping (compressed for short-form) → `../form/narrative-and-transitions.md`
- Specific musical idioms a piece might draw on → various `genres/*.md`
- Mix-aware composition (mastering / loudness / VO interaction) → `../production-aware/arrangement-for-mix.md`
- Music for trailers specifically (where film and game trailers overlap) → `film-tv-scoring.md`, `game-music.md`

# Pop and Rock

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, charts, platform norms, and scene-specific references with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers the broad family of popular music styles centered on the verse-chorus song with band instrumentation: pop, rock, and their many subgenres. The harmonic and formal conventions are covered briefly here; cross-references point to detailed treatment elsewhere.

For form details, see `form/popular-song-forms.md`. For melodic conventions, see `melody/melodic-construction.md`. For songwriting technique, see `songwriting/*.md` files.

## When to consult this file

- The user is writing in a pop, rock, or related popular music style
- Identifying conventions of a specific subgenre (indie, punk, country, soft rock, etc.)
- Comparing aesthetic priorities across pop styles

## What's distinctive about pop/rock

These genres share:

- **Verse-chorus-based forms** — the chorus is the structural payoff (see `form/popular-song-forms.md`)
- **Hook-driven writing** — memorability of the chorus and lyrics is paramount
- **Band-based instrumentation** — drums, bass, guitar(s), vocals, often keys
- **3–5 minute song lengths** for radio and streaming
- **Simple harmony** typically — diatonic with some chromatic spice
- **Repetition** — choruses return identically; verses repeat melody with new lyrics

These are defaults, not requirements. Many great pop and rock songs break each of these.

## Early rock and roll (c. 1955–1963)

Roots in R&B, country, blues, and Western swing. The first commercial mass-market rock.

**Key features**:
- 12-bar blues form pervasive, often "rocked up"
- I-vi-IV-V "doo-wop" progression (50s progression)
- Rockabilly: country + R&B combination
- Strong backbeat (snare on 2 and 4)
- Triplet shuffle feel in many ballads
- Simple harmonies, prominent vocals
- Short songs (2–3 minutes)

**Artists**: Chuck Berry, Little Richard, Elvis Presley, Buddy Holly, Fats Domino, Bo Diddley, Roy Orbison, The Everly Brothers.

## British Invasion / 60s pop and rock (c. 1963–1969)

The Beatles, Stones, and other British bands transform popular music. American responses include Motown, Beach Boys, and the rise of psychedelia.

**Key features**:
- Verse-chorus form solidifies; AABA still common
- Adventurous harmony emerging (modal mixture, modulation)
- Studio production becomes integral to the song
- Concept albums (*Sgt. Pepper's*, *Pet Sounds*)
- 4/4 with strong backbeats, occasionally folk-influenced 6/8

**Artists**: The Beatles, The Rolling Stones, The Who, The Kinks, The Beach Boys, Bob Dylan (folk-rock crossover), Motown artists (R&B-pop), girl groups, surf rock.

## Soul and R&B in pop (c. 1955–onward)

Crossover R&B that became mainstream pop. African-American gospel, blues, and jazz traditions in popular form.

**Key features** (more detail in `genres/hip-hop-rnb.md`):
- Gospel-derived melodies (call and response, melisma)
- Heavy backbeat
- Horn sections, organ, and tight rhythm sections
- Vocal-centered arrangements
- ii-V-I and bluesy harmony

**Artists**: Ray Charles, Sam Cooke, Otis Redding, Aretha Franklin, James Brown, Stevie Wonder, Marvin Gaye, the Motown family.

## Folk-rock and singer-songwriter (c. 1965–onward)

Folk traditions meet rock instrumentation. Singer-songwriter as auteur emerges.

**Key features**:
- Lyric-centric writing
- Acoustic guitar or piano as foundation
- Often strophic or simple verse-chorus forms
- Storytelling lyrics, personal/political themes
- Vocal melodies with conversational range

**Artists**: Bob Dylan, Joni Mitchell, Crosby Stills & Nash, Simon & Garfunkel, James Taylor, Carole King, Cat Stevens, Neil Young, Jackson Browne.

## Hard rock and heavy metal (c. 1968–onward)

Louder, heavier, more distorted. Often longer and more complex song structures.

**Key features**:
- Distorted electric guitars central
- Power chords (root + 5, no 3rd)
- Heavy drums, often with complex fills
- Vocal styles ranging from melodic to aggressive
- Riffs as compositional foundation
- Often longer songs with extended instrumental sections

**Sub-traditions**:
- **Hard rock**: Led Zeppelin, Deep Purple, AC/DC, Aerosmith, Guns N' Roses
- **Classic metal (NWOBHM)**: Black Sabbath, Iron Maiden, Judas Priest
- **Thrash metal**: Metallica, Megadeth, Slayer, Anthrax
- **Death/black metal**: aggressive, extreme styles
- **Power metal, prog metal, doom, sludge**: diverse subgenres

**Idiomatic harmony**: power chords, ♭VI-♭VII-i progressions, modal (Aeolian, Phrygian, Phrygian dominant for darker styles), chromatic riffs.

## Progressive rock (c. 1968–onward)

Rock musicians borrowing from classical, jazz, and avant-garde. Long compositions, complex structures, virtuosic playing.

**Key features**:
- Long songs (5+ minutes; concept albums with 20+ minute sides)
- Odd meters (5/4, 7/8, etc. — see `rhythm-groove/odd-meters-polyrhythm.md`)
- Complex harmonies and modulation
- Multi-section structures
- Synthesizers and elaborate keyboard work
- Concept-driven albums

**Artists**: Yes, Genesis, King Crimson, Emerson Lake & Palmer, Pink Floyd (prog leanings), Rush, Jethro Tull, Dream Theater (prog metal), Tool (modern prog metal), Tame Impala (modern psych-prog).

## Punk and post-punk (c. 1976–onward)

A reaction against prog and corporate rock. Short, aggressive, simple.

**Key features**:
- Fast tempos (160+ BPM common)
- Simple harmonies (often I-IV-V or just I-V)
- Power chords and aggressive vocals
- Short songs (often under 2 minutes)
- DIY ethic; lo-fi production
- Lyrics political or personal

**Punk sub-traditions**:
- **Original punk**: Sex Pistols, The Clash, Ramones, The Damned
- **Hardcore**: Black Flag, Bad Brains, Dead Kennedys
- **Pop punk**: Green Day, Blink-182, The Offspring (more melodic)
- **Post-punk**: Joy Division, Wire, Public Image Ltd. (artier, atmospheric)
- **Indie / alternative descended from post-punk**: Sonic Youth, Pavement, etc.

## New wave / synth-pop (c. 1978–1988)

Punk-inspired pop with synthesizers. Often danceable, often quirky.

**Key features**:
- Synthesizers and drum machines prominent
- Catchy hooks
- Often visual/fashion-driven aesthetic
- Production-focused (Trevor Horn, Tony Visconti)

**Artists**: Talking Heads, Blondie, Devo, Kraftwerk (foundational electronic), Depeche Mode, Duran Duran, Eurythmics, Tears for Fears, Erasure, Pet Shop Boys.

## Indie and alternative rock (c. 1985–onward)

Bands outside major-label corporate rock. Diverse aesthetics held together by independent ethos.

**Sub-traditions**:
- **College rock / early indie**: REM, Replacements, Hüsker Dü, Pixies
- **Grunge**: Nirvana, Pearl Jam, Soundgarden, Alice in Chains (early 90s)
- **Britpop**: Oasis, Blur, Pulp, Suede (mid-90s UK)
- **Lo-fi indie**: Pavement, Built to Spill, Sebadoh
- **Post-rock**: Slint, Mogwai, Godspeed You! Black Emperor (instrumental, atmospheric)
- **Indie-pop / twee**: Belle & Sebastian, The Magnetic Fields
- **Math rock**: Don Caballero, Battles, Hella (complex meters)
- **Dream pop / shoegaze**: My Bloody Valentine, Slowdive, Ride, Cocteau Twins
- **Modern indie**: Arcade Fire, The National, Vampire Weekend, Bon Iver, Phoebe Bridgers, Mitski

## Country (c. 1920s–onward)

American working-class music with deep roots in folk, blues, gospel.

**Sub-traditions**:
- **Hillbilly / old-time / bluegrass**: acoustic, fiddle and banjo, rural harmonies (Bill Monroe, the Carter Family)
- **Honky-tonk**: barroom country, steel guitar, Hank Williams
- **Nashville sound**: orchestrated, polished country (Patsy Cline, Jim Reeves, 1950s-60s)
- **Outlaw country**: anti-Nashville, raw (Willie Nelson, Waylon Jennings, Merle Haggard)
- **Country pop / contemporary**: pop production with country instrumentation (Garth Brooks, Shania Twain, Taylor Swift's early career)
- **Bro country**: 2010s mainstream country with rock and rap influences
- **Americana / alt-country**: indie-influenced country (Wilco, Old 97's, Lucinda Williams)

**Idiomatic features**:
- Straightforward verse-chorus form
- Storytelling lyrics
- I-IV-V harmony predominant; modal mixture occasional
- Pedal steel guitar, fiddle, banjo (traditional)
- Vocal harmonies (often parallel 3rds and 6ths)
- 12-bar blues forms occasional

## Pop (c. 1980–onward)

The mainstream radio format. Constantly evolving, eclectic.

**Sub-eras**:
- **80s pop**: Michael Jackson, Madonna, Whitney Houston, Prince
- **90s pop**: Mariah Carey, boy bands (Backstreet Boys, NSYNC), girl groups (Spice Girls)
- **2000s pop**: Britney Spears, Beyoncé, Justin Timberlake; rise of Auto-Tune (Cher's "Believe", T-Pain)
- **2010s pop**: Lady Gaga, Katy Perry, Bruno Mars, EDM-influenced production; Lana Del Rey (alt-pop)
- **Modern pop (2020s)**: Billie Eilish, Olivia Rodrigo, Dua Lipa, The Weeknd; SoundCloud and bedroom pop influences; Latin pop crossover

**Modern pop characteristics**:
- Verses 8 bars, choruses 8 bars, post-choruses common
- Tempos 100–130 BPM range typical
- Production-driven (the producer often shapes the sound more than the songwriter)
- Vocal hooks layered with backing vocals
- Side-chained pads and synths
- Trap influences in beats
- Diatonic harmony with occasional Mixolydian (♭VII) or Aeolian (i-♭VI-♭VII)

## Songwriting in pop/rock — practical advice

For writing in this idiom:

1. **Hook first.** Most pop songs begin with the hook (chorus melody and/or title). Build the rest around it.
2. **Think rhythmically.** The rhythm of the chorus melody is often more important than the pitches. A great hook has a distinctive rhythmic profile.
3. **Match section to function.** Verses tell, choruses pay off, bridges contrast. See `form/popular-song-forms.md`.
4. **Keep harmony simple.** I-V-vi-IV (and its rotations) accounts for thousands of hit songs. Complexity isn't the point.
5. **Make the chorus sing higher.** Tessitura lift between verse and chorus is a primary contrast tool.
6. **Land choruses on the title.** The title often appears at the start (or end) of the chorus, on a memorable note.
7. **Watch the song length.** Modern pop trends shorter (3 minutes); rock can go longer (4–5 minutes for radio). Albums can have longer tracks.

## Production conventions across the family

Different sub-genres have distinctive production:

| Style | Production |
|-------|------------|
| Classic rock | Live band feel; analog warmth; minimal correction |
| Punk | Lo-fi, raw, fast; minimal overdubs |
| 80s pop | Reverb-heavy, gated drums, synthesizer pads |
| 90s grunge | Loud guitars, dry vocals, "garage" feel |
| Modern pop | Heavily processed; auto-tuned vocals; layered production |
| Country | Live-tracked drums and acoustic instruments; vocal up-front |
| Modern country | Pop production with country instrumentation |
| Indie | Often deliberately lo-fi; idiosyncratic production choices |
| Dream pop / shoegaze | Heavy reverb and delay; vocal buried in the mix |

## Common pitfalls when advising

- **Don't generalize across the family.** "Pop" and "metal" and "country" are different genres with different priorities. Specify which.
- **Don't apply prog/jazz complexity to pop.** A simple I-V-vi-IV progression executed well beats a clever modulation in a pop chorus.
- **Don't ignore production as part of composition.** In modern pop especially, the song and the production are inseparable. The same chord progression sounds completely different across production styles.
- **Don't dismiss "simple" songs.** A great hook is harder to write than a complex one; both have legitimate craft.
- **Don't write rock without thinking about live performance.** Many rock songs are designed to be played live by a band; complex arrangements that can't be performed face issues.

## Quick decision matrix

| Want | Try |
|------|-----|
| Catchy modern pop | Verse-pre-chorus-chorus-post-chorus, diatonic harmony, hooky vocal |
| 80s nostalgia | Synth pads, gated drums, big chorus with chorus-effect on guitars |
| Stadium rock | Powerful backbeat, simple progression, anthemic chorus |
| Classic singer-songwriter | Voice + acoustic guitar (or piano), 4-5 minutes, narrative lyrics |
| Punk energy | Fast tempo, simple chords, short song, raw production |
| Heavy metal | Power chords, distorted riffs, modal (Aeolian / Phrygian), aggressive vocals |
| Progressive rock | Long form, odd meters, virtuosic playing, multiple sections |
| Indie / alternative | Lo-fi production, idiosyncratic structure, conversational vocals |
| Modern country | Verse-chorus with story; pedal steel; pop production polish |
| Dance-pop crossover | EDM-influenced beat, big hook, half-time feel sections |

## Cross-references

- Pop song form → `form/popular-song-forms.md`
- Songwriting hooks and craft → `songwriting/hooks-and-memorability.md`, `songwriting/lyric-writing.md`, `songwriting/topline-craft.md`
- Pop harmony patterns → `assets/progressions-catalog.md`
- Production and arrangement → `production-aware/arrangement-for-mix.md`, `orchestration/arrangement-density.md`
- Backbeat and rock groove → `rhythm-groove/groove-and-feel.md`
- Hip-hop and R&B (related family) → `genres/hip-hop-rnb.md`
- Country/Americana and roots-rock crossover → `country-americana.md`
- Metal, punk, hardcore, and heavy guitar idiom → `metal-punk-hardcore.md`
- Gospel/CCM crossover and worship-pop lift → `gospel-and-ccm.md`
- Brazilian pop/funk crossover → `brazilian-pop-and-funk.md`
- Folk traditions → `genres/folk-roots-and-traditions.md`
- EDM (related modern dance music) → `genres/electronic-edm.md`

# Regional Scene Starter Map

> **Snapshot note:** Last content review snapshot: 2026-04-27. Regional scenes and genre labels change quickly. Use this file as a dispatch map, then verify current references with active web research and local-language sources when possible.

This file collects regional popular-music scenes that are important enough to recognize but not yet covered by a full dedicated guide in this skill. It is not a history. It is a composition-facing map: what to listen for, which craft variables matter, and where the agent should be careful.

Use this file when the user asks about:

- regional Mexican, corridos tumbados, norteño/banda-adjacent pop
- sertanejo, pagode, arrocha, piseiro, Brazilian regional pop beyond the main Brazilian guide
- raï, Maghreb pop, Arabic club-pop subscenes
- Turkish arabesk-pop, Anatolian pop-adjacent modern styles
- Punjabi pop, bhangra-pop, filmi/non-filmi crossovers
- gqom, singeli, gengetone, coupé-décalé, kuduro, or other local dance forms
- any regional scene that is too specific for a general genre file but too important to ignore

For microgenres and internet tags, use `minor-and-hybrid-genres.md`. For broader trend analysis, use `../research/regional-trend-evolution-analysis.md`.

## Core principle

Do not reduce a regional scene to one rhythm, one instrument, or one scale. Start by asking:

```text
Is the user asking for the current scene, a traditional root, a pop crossover, a dance function, or a reference-track lane?
```

Then extract craft variables:

| Variable | What to check |
|---|---|
| Function | dance, radio, wedding, devotional, film, club, live band, social video |
| Rhythm | groove cell, subdivision, tempo-feel, percussion role |
| Harmony | loop, modal center, functional pop changes, dominant pull, drone |
| Melody | vocal ornament, call-response, chant, melisma, range, hook length |
| Language/prosody | syllable density, code-switching, rhyme placement, vowel shape |
| Ensemble | live-band idiom, electronic production, accordion/brass/strings/synths/percussion |
| Form | strophic, verse/chorus, drop, call-response, dance break, long vamp |
| Currentness | chart signal, local playlist, short-form trend, practitioner interview |

## Regional Mexican / corridos tumbados / norteño-banda-pop lane

Composition variables:

- story-forward vocal phrasing; lyric narrative often carries identity
- guitar or bajo sexto-like rhythmic identity in many lanes
- accordion/brass/banda or sierreño-type timbral associations depending on reference
- minor-key and modal color can appear, but do not assume every track is dark
- modern crossovers may borrow trap/hip-hop low-end, vocal phrasing, or release strategy

Agent caution:

- Ask whether the user means traditional corrido, modern corridos tumbados, banda-pop, sierreño, norteño, or a crossover.
- Do not invent Spanish lyrics unless the user wants draft text and language confidence is adequate.
- For current references, check regional charts, Latin charts, local press, and actual recordings.

Pair with `latin-pop-and-reggaeton.md`, `country-americana.md`, `instrument-idiom/guitar.md`, and `../research/web-music-trend-research.md`.

## Brazilian regional pop beyond the main guide

This skill already has `brazilian-pop-and-funk.md`. Use this starter map when the user names specific scenes such as sertanejo, pagode, arrocha, piseiro, forró-pop, brega-funk, tecnobrega, or axé-adjacent pop.

Composition variables:

- Portuguese prosody is central; do not map Spanish-language Latin pop assumptions onto Brazilian music.
- Many scenes are dance/social-function driven, so groove and vocal participation matter as much as harmony.
- Live-band instrumentation and electronic production can coexist.
- Repetition, call-response, and crowd-singable hooks may be more important than harmonic novelty.

Agent caution:

- Ask for a reference lane because Brazilian scenes vary sharply by region and function.
- Separate funk carioca from samba/pagode/sertanejo/forró-related idioms.
- Use local Portuguese searches and current playlists for fresh references.

Pair with `brazilian-pop-and-funk.md`, `../fundamentals/prosody-and-language.md`, and `../research/regional-trend-evolution-analysis.md`.

## Raï / Maghreb pop / Arabic club-pop subscenes

Composition variables:

- vocal ornament, phrase shape, and language delivery are central
- electronic club grids can sit under region-specific melodic behavior
- maqam-adjacent colors may appear, but the scene should not be reduced to a scale
- call-response, percussion, and synth timbre often define the modern lane

Agent caution:

- Ask whether the user wants raï, Algerian/Moroccan/Tunisian pop, Arabic club-pop, or diaspora crossover.
- Do not claim maqam accuracy without practitioner-level verification.
- Treat references as necessary; this area is too broad for generic advice.

Pair with `mena-pop.md`, `../techniques/microtonal.md`, and `../research/reference-track-digging.md`.

## Turkish arabesk-pop / Anatolian-pop-adjacent modern styles

Composition variables:

- melodic contour and ornament often carry emotional identity
- minor/modal color, chromatic pull, and strong vocal delivery can be more important than Western chord novelty
- modern productions may combine electronic drums, strings, bağlama-like colors, and pop form
- long emotional build or intense refrain can be central

Agent caution:

- Ask whether the user means classic arabesk influence, modern Turkish pop, rock/Anatolian influence, or club crossover.
- Do not flatten Turkish makam-informed practice into equal-tempered minor scale vocabulary.

Pair with `mena-pop.md`, `folk-roots-and-traditions.md` as a legacy route, `../techniques/microtonal.md`, and current research files.

## Punjabi pop / bhangra-pop / South Asian crossover

Composition variables:

- dhol/dholak-like rhythmic identity and strong dance function in many bhangra-derived lanes
- vocal hooks may rely on short, high-energy call phrases
- Punjabi pop can cross with hip-hop, EDM, filmi, and diaspora club production
- lyric language, pronunciation, and social function matter heavily

Agent caution:

- Separate Punjabi pop/bhangra-pop from Hindi film-pop and from classical raga-derived writing.
- Ask for a reference if the user wants modern diaspora production.

Pair with `south-asian-film-pop.md`, `hip-hop-rnb.md`, `electronic-edm.md`, and `../fundamentals/prosody-and-language.md`.

## Local African dance scenes beyond Afrobeats/amapiano

Examples: gqom, singeli, gengetone, kuduro, coupé-décalé, alte-adjacent scenes, local house/club hybrids.

Composition variables:

- groove, tempo-feel, and dance function are usually primary
- loop architecture may matter more than chord progression
- percussion sound choice and rhythmic density can define the style
- vocal/chant placement often responds to dance and social function

Agent caution:

- Do not route everything through "Afrobeats" or amapiano.
- Use local/regional references and current scene names.
- Ask whether the user wants a respectful influence, a club-functional track, or a pop crossover.

Pair with `afrobeats-and-amapiano.md`, `electronic-edm.md`, `rhythm-groove/groove-and-feel.md`, and `../research/web-music-trend-research.md`.

## How to answer a regional-scene request

Use this template:

```md
**I will treat this as [specific scene/lane], not as a generic regional color.**

**Reference check needed**
- Current scene? yes/no
- User's target platform/region:
- One or two anchor tracks:

**Craft variables**
- Groove:
- Vocal/lyric behavior:
- Harmony/melody:
- Instrumentation:
- Form/function:

**Starter sketch**
- Key/tempo:
- 4-bar groove/chord seed:
- Hook/topline idea:
- What to avoid:
```

## Cross-references

- Microgenres and hybrid labels → `minor-and-hybrid-genres.md`
- Broader regional analysis → `../research/regional-trend-evolution-analysis.md`
- Web research workflow → `../research/web-music-trend-research.md`
- Reference digging → `../research/reference-track-digging.md`
- User streaming context → `../research/user-listening-context-and-streaming-services.md`
- Latin pop/reggaeton → `latin-pop-and-reggaeton.md`
- Brazilian pop/funk → `brazilian-pop-and-funk.md`
- Afrobeats/amapiano → `afrobeats-and-amapiano.md`
- South Asian film-pop → `south-asian-film-pop.md`
- MENA pop → `mena-pop.md`

# Jazz Harmony

Jazz harmony is functional harmony with extensions as default, substitutions as basic literacy, and voicings as a defining language. The tonal logic is mostly the same as classical functional harmony, but the surface is denser and the substitution toolkit is wider. This file covers what makes jazz harmony jazz, not the broader system.

## When to consult this file

- The user asks about ii-V-I, jazz changes, "comping"
- Tritone substitution, backdoor ii-V, Coltrane changes
- Upper-structure triads, rootless voicings, drop-2/drop-3
- Chord-scale relationships ("what scale do I play over G7♭9")
- Bebop chromaticism, bop language
- Modal jazz harmonic approach (also see `modal-harmony.md`)
- Reharmonization in a jazz context (also see `reharmonization.md`)

For chord spelling and quality, see `chord-construction.md`. For functional logic, see `functional-harmony.md`. For voicings as visual shapes, see `assets/jazz-voicings.md`. For modal jazz specifically, see `modal-harmony.md`.

## What makes jazz harmony jazz

Three axes:

1. **Extensions are the default, not the exception.** A jazz `Cmaj7` is rarely just C-E-G-B; it usually has 9, sometimes 13. A `G7` is rarely just G-B-D-F; it has 9 (or ♭9), and often 13 (or ♭13). The 7th is the *minimum*.
2. **Functional motion is dense.** Every diatonic chord can be approached by its own ii-V. A 32-bar standard might have 30 chord changes, where a comparable pop song might have 8.
3. **Substitution is part of the literacy.** Tritone subs, backdoor ii-Vs, secondary dominants — these aren't "advanced" in jazz; they're starting vocabulary.

A jazz musician who reads `Cmaj7 – Am7 – Dm7 – G7` understands that as a default starting point. The actual played version will have substitutions, alterations, and voicings that the chart doesn't specify.

## Extensions and alterations — jazz defaults

Every chord type has a default extension treatment in jazz. Use these as the starting assumption; deviate for effect.

| Chord type | Default extensions | Common alterations |
|------------|-------------------|---------------------|
| Major 7 | 9, ♯11, 13 (Lydian flavor) | Sometimes natural 11 implied via sus, but ♯11 is more common |
| Dominant 7 (resolving) | 9, 13 | ♭9, ♯9, ♭13 (or ♯5) for tension; commonly all four "altered" tensions |
| Dominant 7 (non-resolving / static) | 9, ♯11, 13 (Lydian dominant) | Less alteration; the chord doesn't need to resolve |
| Minor 7 | 9, 11, occasionally 13 | (rarely altered) |
| Minor 7♭5 (half-diminished) | 11, ♭13; sometimes natural 9 (Locrian ♮2) | (extensions stay where they are) |
| Diminished 7 | 9, 11, ♭13 (each tone of the chord can serve as a "root") | Symmetric — many tensions work |
| Minor-major 7 | 9, 11, 13 | (melodic minor flavor) |

The key idea: when a player sees `G7` resolving to `Cmaj7`, they may add ♭9 or ♭13 (or both) to enrich the dominant pull. When they see `G7♯11` or a dom7 chord that's *not* resolving, they leave the natural extensions or use Lydian dominant flavor.

## ii-V-I — the central jazz cadence

The ii-V-I (or ii-V-i in minor) is the cellular unit of jazz harmonic motion. Every standard contains them; most progressions can be analyzed as chains of ii-Vs.

### Major key ii-V-I

In C major: `Dm7 – G7 – Cmaj7`.

Voice-led smoothly:
- 3rd of `Dm7` (F) is the 7th of `G7`, then resolves down to E (the 3rd of `Cmaj7`)
- 7th of `Dm7` (C) is the 11th-or-implied of `G7`, then can hold or move to B (the 7th of `Cmaj7`)
- The two "guide tones" (3rd and 7th of each chord) are F & C in Dm7, F & B in G7, E & B in Cmaj7. Half-step motion drives the resolution.

### Minor key ii-V-i

In C minor: `Dm7♭5 – G7♭9 – Cm(maj7)` (or Cm6, or Cm7 — the choice depends on style).

Voice-led smoothly:
- The `iiø7` (Dm7♭5) has the same 3rd and 7th as Dm7 in major (F and C), but the ♭5 (A♭) confirms minor key.
- The `V7♭9` borrows the ♭9 from harmonic minor (A♭ becomes the ♭9 of G).
- Resolution to a tonic that's typically `m6`, `m(maj7)`, or `m7`.

### ii-V-I chains

Chains of ii-Vs that target multiple keys are the bread and butter of jazz changes.

`Dm7 – G7 – Cmaj7 | Cm7 – F7 – B♭maj7 | B♭m7 – E♭7 – A♭maj7`

This is a sequence of ii-V-Is descending by whole step. "All the Things You Are" famously uses ii-V-I chains, modulating constantly through the form.

ii-Vs can also be used without the resolution — a "ii-V to nowhere" sets up a key, then redirects.

## The blues form (briefly — see `../genres/jazz-styles.md` for genre depth)

12-bar blues is a foundational form in jazz. Standard form:

| Bars | Chords |
|------|--------|
| 1 | I7 |
| 2 | IV7 (quick change) or I7 |
| 3 | I7 |
| 4 | I7 |
| 5 | IV7 |
| 6 | IV7 (or ♯iv°7 to V) |
| 7 | I7 |
| 8 | VI7 (V/ii) or I7 |
| 9 | ii7 (or V7) |
| 10 | V7 |
| 11 | I7 (or ii-V turnaround) |
| 12 | V7 (or turnaround) |

The "jazz blues" extends this with more substitutions:

| Bars | Chords |
|------|--------|
| 1 | I7 |
| 2 | IV7 |
| 3 | I7 |
| 4 | I7 (or v7-I7) |
| 5 | IV7 |
| 6 | ♯iv°7 |
| 7 | I7 |
| 8 | III7 (V/vi) |
| 9 | VI7 (V/ii) |
| 10 | ii7 |
| 11 | V7 |
| 12 | I7 (turnaround: I-VI-ii-V or I-♭III-II-♭II) |

Bird blues and minor blues add more substitution layers. The blues form is where many jazz substitution tricks are first learned.

## Substitutions

Substitution is replacing a chord with a different chord that performs the same function (or a related function). The five most important:

### 1. Tritone substitution

Replace any dominant 7 chord with the dominant 7 a tritone away. The two share the same tritone (the chord's most defining interval), so they have the same essential function.

`G7` (G B D F) and `D♭7` (D♭ F A♭ C♭) share the tritone B-F (= C♭-F). So `D♭7` can substitute for `G7` in any context, including resolving to `Cmaj7`.

`Dm7 – D♭7 – Cmaj7` is a tritone-sub ii-V-I. The bass moves chromatically (D – D♭ – C), which is exactly why tritone subs feel so smooth.

The notation: `D♭7` can be written as `subV7/I` (substitute V of I) when functioning as a tritone sub.

Tritone subs work on secondary dominants too: `V7/V` (D7 in C major) substituted by `A♭7`. In a long chain, tritone-substituting some V7s and not others creates chromatic descending bass lines.

### 2. Backdoor ii-V

A `iv–♭VII7–I` motion (from the parallel minor's ii-V). In C major: `Fm7 – B♭7 – Cmaj7`.

The `B♭7` looks like it should resolve to E♭ (its conventional target), but it resolves *up* a whole step to C, with smooth voice leading. The 3rd of B♭7 (D) and the 7th of B♭7 (A♭) resolve to the 9th of Cmaj7 (D) and the 3rd of Cmaj7 (E) respectively.

Backdoor ii-Vs are common in standards. They give a momentary minor flavor (the iv chord) before resolving to major.

### 3. Diminished substitution

Any dominant 7♭9 chord shares three notes with a diminished 7 chord built on its 3rd, 5th, or ♭7. So `°7` chords can substitute for dominant chords (and vice versa).

`G7♭9 = G B D F A♭`. The °7 chord built on B (B D F A♭) is the same except for the missing root.

Practical use: replace `V7` with a `°7` for a smoother voice-leading effect, often with a chromatic bass approach.

### 4. ii-V insertion

Insert a ii-V before any chord. Don't have a `Dm7 – G7` before `Cmaj7`? Add one. Don't have a `Em7 – A7` before `Dm7`? Add one. Every chord can be approached by its own ii-V.

A 32-bar tune with 16 chord changes can become one with 32 chord changes by inserting ii-Vs in front of every chord.

This is how players add density and forward motion to simple progressions, and how reharmonization commonly works.

### 5. Coltrane substitution (Coltrane changes)

Replace a single ii-V-I with a sequence of three key centers a major third apart, each with its own ii-V-I.

A compact C-centered Coltrane-cycle illustration: `Cmaj7 – E♭7 – A♭maj7 – B7 – Emaj7 – G7 – Cmaj7`. The Maj7 chords are a M3 apart (C, A♭, E), each connected by a dominant 7. Treat this as a transposed/simplified pattern, not a literal full lead-sheet excerpt from "Giant Steps".

Practical use: replace a long static major-key passage with a Coltrane sequence to add motion. Famously hard to improvise over but conceptually elegant.

## Modal interchange in jazz

Jazz uses modal interchange (borrowing from parallel minor) more aggressively than classical or pop. Examples:

- The minor IV (`iv`) appears constantly in major-key tunes.
- The `♭III` and `♭VII` show up as borrowed chords or modal flavor.
- Minor-key tunes borrow the major-key V (with leading tone) for cadences.

The line between "modal interchange" and "constant chromatic harmony" gets blurry in jazz. Most tunes are happy to borrow whatever's needed.

## Voicing approaches

Jazz harmony lives or dies by voicing. The same chord symbol can be voiced dozens of ways, each with a different feel.

### Shell voicings

Just root + 3 + 7 (omitting the 5). Common in jazz piano left hand. Three-note voicings are clear and don't fight the bass.

`Cmaj7` shell: C (root), E (3), B (7).

### Rootless voicings

Omit the root entirely (the bass player handles it). Replace the root slot with the 9th. Frees the chord for tensions.

`Cmaj7` rootless: E G B D (3 5 7 9). `G7` rootless: F A B D (♭7 9 3 5) — common left-hand voicing.

### Drop voicings

Take a close-position 4-note voicing and drop the 2nd or 3rd voice from the top down an octave.

Starting from `Cmaj7` close position (low → high): `C – E – G – B` (root, 3, 5, 7).

- **Drop 2**: drop the 2nd-from-top voice (`G`) an octave down → `G – C – E – B`. The 5th sits in the bass; the rest stays compact above.
- **Drop 3**: drop the 3rd-from-top voice (`E`) an octave down → `E – C – G – B`. Wider gap between the bass note and the upper triad-shape; useful when you want a clear "bass + chord" separation.

Drop voicings open up the chord for stepwise voice leading on guitar and piano. Drop 2 in particular is the standard jazz guitar voicing because it lays well on the top four strings (or middle four strings). For full inversion tables, see `assets/jazz-voicings.md`.

### Quartal voicings

Stacked 4ths. Modal, McCoy Tyner / "So What" / Herbie Hancock. Common for Dorian and other modal contexts.

`So What` voicing for D Dorian: D G C F A (D, G, C, F, A — a stack of 4ths up to a 3rd at the top).

### Spread voicings

Wide intervals across multiple octaves. Common in piano left-hand-low / right-hand-high arrangements.

### Upper structure triads (UST)

Already covered in `chord-construction.md`. A jazz favorite for getting tensions on top: play a shell or rootless voicing in the left hand and a triad in the right hand whose notes are upper extensions of the chord.

## Chord-scale theory — a brief

For each chord, there's a "default" scale that contains its chord tones plus the available extensions. Improvisers and composers use these as starting points (not as rigid rules — bebop chromaticism violates them constantly).

| Chord | Default scale |
|-------|---------------|
| Imaj7 | Ionian (or Lydian for color) |
| Imaj7♯11 | Lydian |
| ii7 | Dorian |
| iiø7 | Locrian (or Locrian ♮2 if the natural 9 is in the chord) |
| iii7 | Phrygian (or Aeolian if borrowed) |
| IV7 (V/♭VII or blues IV) | Mixolydian or Lydian dominant |
| V7 (resolving) | Mixolydian, with alterations available: altered scale, half-whole diminished, Phrygian dominant |
| V7sus4 | Mixolydian with no 3rd; or "Phrygian over V" for darker color |
| vi7 | Aeolian (or Dorian) |
| vii°7 | Half-whole diminished, or harmonic minor of the next chord |
| Im(maj7) | Melodic minor |
| Im7 (modal context) | Dorian |

This isn't a substitute for ear and taste — it's a starting framework. Many great soloists violate chord-scale defaults constantly.

## Common reharm patterns (brief)

For full treatment see `reharmonization.md`. Quick patterns specific to jazz:

- **Replace V7 with subV7**: tritone sub.
- **Use tritone subs for dominant-function chords, not literally any chord**: `G7 → D♭7` works because both chords contain the same guide-tone tritone. For a non-dominant chord like `Cmaj7`, first reinterpret its function or use another reharm device (chromatic mediant, approach ii–V, pedal, or dominantization); don't call the result a tritone substitution unless it behaves like a dominant substitute.
- **Replace a single chord with a ii-V**: turn `Cmaj7` (lasting 4 bars) into `Cmaj7 – Dm7 – G7 – Cmaj7` or `Cmaj7 – Em7 – A7 – Dm7 – G7 – Cmaj7`.
- **Pedal the V**: replace several beats of various chords with a static V7 (often with extensions changing).
- **Add a turnaround**: the standard `Cmaj7 – Am7 – Dm7 – G7` or its variants (`Cmaj7 – A7 – D7 – G7`; `Cmaj7 – Bm7♭5 – E7 – Am7 – D7 – G7`).

## Genre notes within jazz

- **Swing**: simpler harmony, often diatonic with secondary dominants. Standard ii-V-I motion.
- **Bebop**: dense ii-V chains, tritone subs, fast-moving changes. "Anthropology", "Cherokee".
- **Cool jazz**: smoother, modal-flavored, less dense. "Kind of Blue".
- **Hard bop**: gospel-flavored, bluesy, with rich extended chords. Horace Silver, Art Blakey.
- **Modal jazz**: see `modal-harmony.md`.
- **Free jazz**: harmony is largely abandoned; pitch organization happens in other ways. See `techniques/20th-century-techniques.md`.
- **Fusion**: jazz harmony with rock/funk rhythm; often more modal, often with electric voicings.
- **Contemporary jazz / post-bop**: extended harmonic vocabulary, slash chords, polychords, chromatic mediants. Wayne Shorter, Brad Mehldau.

## Common pitfalls when advising

- **Don't suggest extensions without considering the bass.** A `Cmaj9♯11` sounds great with C in the bass; the same notes with E in the bass become an `Em7 add ♯5` (dissonant nightmare).
- **Don't pile alterations on a non-resolving dominant.** `7♯11` flavor is for static or modal dominants; `7♭9♯9` and `7alt` are for resolving dominants.
- **Don't tritone-sub everything.** Indiscriminate tritone subs lose the resolution feel. Use them where chromatic descent is wanted.
- **Don't ignore the chord-scale default for context.** A `D7` in the middle of a C major context (V/V) wants Mixolydian or altered. The same `D7` in a D blues wants D Mixolydian with blue notes.
- **Don't chase substitution without ear.** Substitutions are tools; the goal is sound, not complexity.

## Quick decision matrix

| Want | Use |
|------|-----|
| Smooth, textbook jazz cadence | ii-V-I with extensions |
| Chromatic bass descent into tonic | Tritone-sub the V (D♭7 → C) |
| Momentary minor flavor before major tonic | Backdoor ii-V (iv-♭VII7-I) |
| Add motion to static harmony | Insert ii-V before each chord |
| Modal vamp instead of changes | Static modal chord with quartal voicings |
| Maximum tension on dominant | V7♭9♯9♭13 (altered) |
| Bright dominant that doesn't resolve | V7♯11 (Lydian dominant) |
| Suspended sound | sus4 voicings, often with 9 |
| Full vertical color | Upper structure triads |
| Lyrical, vocal-friendly extensions | 9 and 13, mostly natural; avoid stacked alterations |

## Cross-references

- Chord spelling and qualities → `chord-construction.md`
- Functional logic underlying ii-V-I → `functional-harmony.md`
- Voice leading guide tones → `voice-leading.md`
- Voicing shapes (drop 2/3, rootless, quartal) → `assets/jazz-voicings.md`
- Modal jazz specifically → `modal-harmony.md`
- Reharmonization techniques in depth → `reharmonization.md`
- Modulation in jazz → `modulation.md`
- Jazz styles by sub-genre → `../genres/jazz-styles.md`
- Bebop melodic language → `../melody/melodic-construction.md`, `../genres/jazz-styles.md`

# South Asian Film-Pop and Raga-Derived Pop

> **Snapshot note:** Last content review snapshot: 2026-04-26. Use this file for stable compositional conventions, but verify current artists, film industries, charts, platform norms, and regional scenes with active web research. See `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers composition-facing conventions for South Asian film-pop and related pop-fusion writing: Hindi film music/Bollywood, Tamil/Telugu/Malayalam/Kannada film-pop, Pakistani pop, Punjabi pop/bhangra influence, indie-fusion, and diaspora pop. It is not a full guide to Hindustani, Carnatic, folk, devotional, qawwali, ghazal, or regional traditions.

## When to consult this file

- The user wants Bollywood/Hindi-film, South Indian film-pop, Punjabi pop, or South Asian fusion energy
- The user asks for raga-inspired melody or tala/rhythm-cycle influence in a pop context
- The user wants cinematic song-sequence writing, dance hooks, or playback-style vocal drama
- The user wants current South Asian references; pair with active web research

## Core warnings

- **Raga is not just a scale.** A raga includes characteristic phrases, important tones, ascent/descent behavior, ornament, register treatment, and affective associations.
- **Tala is not just a time signature.** It is a cyclical rhythmic framework with accents, claps/waves, and performance conventions.
- **Film-pop is hybrid by design.** It may borrow from classical, folk, rock, EDM, hip-hop, qawwali, ghazal, devotional, and Western orchestral/film-score language.
- Do not treat "Indian" or "South Asian" as one sound. Region, language, film industry, and era matter.

## Film-pop composition priorities

South Asian film songs often serve both musical and dramatic functions.

**Common priorities**:

- strong melodic identity and vocal expressivity
- section contrast tied to scene/narrative energy
- danceability or cinematic sweep, depending on context
- call-response between vocal, percussion, and instrumental hooks
- memorable instrumental prelude/interlude motifs
- flexible genre fusion
- language-specific prosody and vowel placement

## Song-sequence mindset

A film-pop song may need to function as:

- romantic confession
- dance number
- devotional or quasi-devotional moment
- comic relief
- montage
- item number / club sequence
- grief or longing song
- heroic entrance
- wedding/festival song

Composition should support that dramatic function. A "good hook" is not enough if it does not match the scene's emotional role.

## Melody: raga-derived without faking raga

For a pop composer, safer raga-inspired writing means borrowing **constraints and gestures** rather than claiming full raga authenticity.

Useful abstractions:

| Raga-derived idea | Pop-safe adaptation |
|---|---|
| Characteristic phrase | Build hook around a short recurring turn, not a generic scale run |
| Important resting tone | Land the chorus on a stable focal pitch repeatedly |
| Ornament / gamaka | Use slides, turns, grace notes, or bends in a style-aware way |
| Ascent/descent difference | Let verse rise differently from how it descends |
| Drone center | Keep tonic/dominant drone or pedal under harmonic motion |
| Affective color | Choose register, ornament, and rhythm to match mood |

Avoid saying "this is Raga X" unless you know the grammar. Say "raga-inspired" or "uses a raga-adjacent melodic constraint" when appropriate.

## Harmony

Traditional classical raga performance is not chord-progression-based in the Western pop sense, but film-pop often uses Western harmony freely.

Common pop-fusion options:

| Goal | Progression / approach |
|---|---|
| Drone/modal | i or I pedal with changing upper melodies |
| Romantic minor | i-♭VI-♭VII-i or i-iv-V-i |
| Cinematic lift | ♭VI-♭VII-i or IV-V-vi |
| Western pop crossover | I-V-vi-IV or vi-IV-I-V |
| Qawwali/Sufi-pop color | I-♭VII-IV-I or i-♭VII-♭VI-♭VII with call-response |
| Ghazal/ballad color | slower harmonic rhythm, m7/maj7/add9 colors, ornamented melody |

If the melody is strongly raga-derived, avoid thick chord changes that contradict the characteristic tones.

## Rhythm and groove

### Tala-inspired thinking

Use cyclical phrasing and recurring accents, not just odd meters.

Practical moves:

- create a repeated percussion cycle that marks phrase identity
- use claps/fills at cycle returns
- write melody phrases that resolve at the cycle point
- let the bass/percussion announce section changes

### Common pop-facing grooves

| Groove family | Composition use |
|---|---|
| Dhol / bhangra influence | energetic dance hooks, Punjabi pop, wedding/festival energy |
| Tabla-inspired ballad | romantic or devotional softness; avoid fake over-programming |
| Keherwa-like 8-beat feel | versatile pop/film groove frame |
| Dadra-like 6-beat feel | romantic, folk, light-classical feeling |
| EDM/hip-hop hybrid | modern film-pop, club scenes, youth-market tracks |
| Folk/regional percussion | strong regional identity; requires deeper reference study |

## Vocal writing and prosody

South Asian film-pop is often vocal-forward. Melody should leave room for expressive slides, turns, and breath.

Useful moves:

- write longer vowels on emotional words
- allow pickup ornaments into important syllables
- reserve the highest note for a lyric/dramatic payoff
- use call-response between lead and backing/group vocals
- keep tongue-twister rhythmic hooks singable at tempo
- adapt syllable density to the language: Hindi/Urdu, Tamil, Telugu, Punjabi, Bengali, Malayalam, etc. differ

## Arrangement

Common layers:

- drone/pedal: tanpura-like pad, synth drone, low string pedal
- percussion: tabla/dhol/dholak/mridangam-inspired or programmed hybrid
- harmonic bed: piano, guitar, strings, synths
- melodic hooks: flute, shehnai-like reed, violin, sitar/sarod-like pluck, electric guitar, synth lead
- vocal: lead, doubles, group response, ad-libs
- cinematic layers: strings, brass, choir, risers, impacts

Avoid surface-token orchestration. A sitar-like timbre over a generic pop loop does not automatically create idiomatic South Asian writing.

## Form templates

### Film-pop romantic song

```text
Instrumental motif → Verse / antara-like section → Hook / mukhda-like return → Interlude → Verse 2 → Hook → Bridge or high-energy reprise
```

### Dance / wedding / festival track

```text
Intro chant or instrumental riff → Groove drop → Verse → Pre-lift → Hook → Dance break → Verse/rap → Hook → Final chant
```

### Raga-inspired pop ballad

```text
Drone intro → Ornamented verse → Wider chorus/hook → Instrumental response → Verse 2 variation → Final chorus with ad-libs
```

## Quick recipes

### Raga-inspired pop hook without naming a raga

```text
Center: D drone
Melody rule: emphasize D-F-G-A-C; avoid busy chord changes
Hook contour: grace-note turn into A, descend G-F-D
Harmony: | Dm(add9) | Dm/C | B♭maj7 | C |
Arrangement: drone + soft percussion + flute/voice response
```

### Punjabi-pop dance hook

```text
Tempo feel: energetic 4/4
Loop: | Am | G | F | G |
Percussion: dhol-like high-energy accents + modern kick/sub
Topline: short chantable title phrase, group response after each line
Form: hook early, dance break after first chorus
```

### Cinematic South Asian fusion bridge

```text
Drop to drone + solo vocal pickup.
Add strings in rising sequence.
Re-enter chorus with percussion fill at cycle return.
```

## Common mistakes

- Treating raga as a Western scale preset
- Writing thick Western chord loops under every raga-like phrase
- Using tabla/sitar/flute as surface decoration only
- Ignoring language prosody and vowel placement
- Confusing Hindi-film conventions with all South Asian music
- Assuming Bollywood = one modern sound
- Over-orchestrating so the vocal ornament has no room

## Cross-references

- Current trend checks → `../research/web-music-trend-research.md`
- Reference-track method → `../research/reference-track-digging.md`
- Style/copyright guardrails → `../research/style-reference-and-copyright.md`
- Microtonal / raga / non-12-TET caveats → `../techniques/microtonal.md`
- Prosody → `../fundamentals/prosody-and-language.md`
- Film scoring → `film-tv-scoring.md`
- Pop forms → `../form/popular-song-forms.md`
- Arrangement density → `../orchestration/arrangement-density.md`

# Chord Construction

This file covers how chords are built — their qualities, extensions, alterations, and notation. It's the vocabulary file. For *function* (what chords do in a key), see `functional-harmony.md`. For *connection* (how chords move between each other), see `voice-leading.md`. For *voicing* (how a chord is laid out across registers and instruments), see `orchestration/voicing-and-texture.md` and `assets/jazz-voicings.md`.

## When to consult this file

- The user asks "what is a Cmaj7♯11" or any chord-spelling question
- The user wants to add color (extensions, alterations) but doesn't know which to use
- Notation/symbol questions ("how do I write half-diminished")
- Choosing between chord qualities for a given mood
- Understanding what a chord *contains* (its component notes/intervals)

## Triads

A triad is three notes stacked in thirds: root, third, fifth. The four qualities differ in the size of the two thirds.

| Quality | Symbol | Stacking | Example (root C) | Sound |
|---------|--------|----------|------------------|-------|
| Major | `C`, `CM`, `Cmaj` | M3 + m3 | C E G | Bright, stable |
| Minor | `Cm`, `Cmin`, `C-` | m3 + M3 | C E♭ G | Dark, stable |
| Diminished | `C°`, `Cdim` | m3 + m3 | C E♭ G♭ | Tense, unstable (tritone between root and 5th) |
| Augmented | `C+`, `Caug` | M3 + M3 | C E G♯ | Tense, ambiguous (symmetric — divides octave equally) |

Major and minor are stable (consonant fifth); diminished and augmented are unstable (dissonant or symmetrically ambiguous fifth) and almost always resolve.

### Inversions

A triad's bass note determines its inversion:

- **Root position** — root in bass (`C` = C-E-G with C below)
- **First inversion** — third in bass; figured `⁶` in classical, `C/E` in lead-sheet
- **Second inversion** — fifth in bass; figured `⁶/₄`, written `C/G`

Second inversion (the "6/4 chord") is unstable in classical practice and used only in specific contexts (cadential, passing, neighboring — see `voice-leading.md`). In pop and rock, slash chords with the fifth in the bass are common and unproblematic.

## Seventh chords

Add a seventh on top of a triad. Now there are seven common qualities, distinguished by the quality of both the underlying triad and the seventh interval (M7 or m7) above the root.

| Quality | Symbol | Notes (root C) | Built from | Common context |
|---------|--------|----------------|------------|----------------|
| Major 7 | `Cmaj7`, `CΔ`, `CM7` | C E G B | major triad + M7 | Tonic chord in jazz/pop; lush, restful |
| Dominant 7 | `C7` | C E G B♭ | major triad + m7 | The functional dominant; tense, unresolved |
| Minor 7 | `Cm7`, `C-7` | C E♭ G B♭ | minor triad + m7 | All-purpose minor; jazz, R&B, neo-soul |
| Half-diminished | `Cm7♭5`, `Cø7`, `Cø` | C E♭ G♭ B♭ | dim triad + m7 | The `iiø7` of minor ii-V-i; melancholy |
| Fully diminished | `C°7`, `Cdim7` | C E♭ G♭ B𝄫 (=A) | dim triad + d7 | Symmetric (stacked m3s); tense passing chord |
| Minor-major 7 | `CmMaj7`, `Cm(maj7)`, `C-Δ` | C E♭ G B | minor triad + M7 | "James Bond" sound; line cliché destination |
| Augmented major 7 | `Cmaj7♯5`, `C+Δ` | C E G♯ B | aug triad + M7 | Ambiguous, glassy/bright-tension; jazz and impressionist |

Two more, less common:
- **Augmented 7** (`C7♯5` / `C+7`): C E G♯ B♭ — altered dominant; appears in altered V chords
- **Diminished major 7** (`C°Maj7`): C E♭ G♭ B — rare, very dissonant and coloristic; not symmetric like a fully diminished 7th chord

### Inversions of 7th chords

| Inversion | Bass | Lead-sheet | Figured |
|-----------|------|------------|---------|
| Root position | root | `C7` | `7` |
| First | third | `C7/E` | `⁶/₅` |
| Second | fifth | `C7/G` | `⁴/₃` |
| Third | seventh | `C7/B♭` | `²` or `⁴/₂` |

## Extensions: 9, 11, 13

Stack thirds beyond the seventh: 9 (root + M9 or compound 2nd), 11 (compound 4th), 13 (compound 6th). These add color without changing the chord's underlying function.

### Default (unaltered) extensions per chord type

| Chord | 9 | 11 | 13 | Notes |
|-------|---|----|----|-------|
| Major 7 | natural 9 | ♯11 (avoid natural 11 — clashes with major 3rd) | natural 13 | Lydian color preferred |
| Dominant 7 | natural 9 | natural 11 (rare; usually as sus4 or omit) | natural 13 | The 11 is uncomfortable above the M3 |
| Minor 7 | natural 9 | natural 11 | natural 13 (rarely used) | Dorian-flavored full extensions |
| Half-diminished | natural 9 or ♭9 | natural 11 | ♭13 | `Locrian ♮2` is common in jazz minor ii–V–i; plain Locrian/♭9 is also possible by context |
| Fully diminished | natural 9 (sometimes ♭9) | natural 11 | ♭13 | Symmetric; many extensions sound OK |
| Minor-major 7 | natural 9 | natural 11 | natural 13 | Melodic minor flavor |

The "avoid note" concept: an extension that clashes too harshly with a chord tone. The most famous is the natural 11 over a major 7th (clashes with the M3 by a m9 — the harshest interval in jazz). The fix is to either omit, alter to ♯11, or treat as sus4 (which omits the 3rd).

### Altered extensions

For dominant 7 chords specifically — the chord that *wants* tension — extensions are routinely altered to increase the pull toward resolution.

| Alteration | What it means | Effect |
|------------|---------------|--------|
| ♭9 | flatten the 9th (so D♭ over C7) | Phrygian-dominant flavor; common in V→i in minor |
| ♯9 | sharpen the 9th (so D♯/E♭ over C7) | "Hendrix chord" with M3 below; bluesy bite |
| ♭5 / ♯11 | flatten the 5th or sharpen the 11th (G♭/F♯ over C7) | Lydian dominant flavor; same note, different naming |
| ♯5 / ♭13 | sharpen the 5th or flatten the 13th (G♯/A♭ over C7) | Whole-tone or altered scale flavor |

An altered dominant draws from the altered-scale color (7th mode of melodic minor): ♭9, ♯9, ♭5/♯11, and ♭13/♯5 are available, usually as a selected subset rather than all crammed into one voicing. A common notation: `C7alt` or a more specific label such as `C7♯9♭13`.

### Sus chords

Replace the 3rd with a 4th (`sus4`) or 2nd (`sus2`). The chord has no 3rd, so it's neither major nor minor — quality is suspended.

| Symbol | Notes (root C) | Notes |
|--------|----------------|-------|
| `Csus4` (or `Csus`) | C F G | Most common; "wanting to resolve down to 3" |
| `Csus2` | C D G | More open, modern; less directional |
| `C7sus4` (or `C7sus`) | C F G B♭ | Dominant 7 with sus 4; common in pop, gospel, jazz |
| `C9sus4` | C F G B♭ D | Adds the 9th; very common in jazz/pop ("Maiden Voyage" chord) |

`sus4` traditionally resolves down to a 3 (Csus4 → C). In modern pop and jazz it often doesn't — it's used for its own flavor.

### Add chords (no 7th)

Add an extension to a triad without the 7th in between:

| Symbol | Notes (root C) | Notes |
|--------|----------------|-------|
| `Cadd9` | C E G D | Major triad + 9; common in pop ballads, John Mayer territory |
| `Cmadd9` | C E♭ G D | Minor add9; melancholy and lush |
| `C6` | C E G A | Major 6 — "old-fashioned" or jazz-flavored alternative to Cmaj7 |
| `Cm6` | C E♭ G A | Minor 6 — Dorian flavor (note: A is the natural 6 of C minor in Dorian) |
| `C6/9` | C E G A D | "Six-nine"; standard final-chord alternative to Cmaj7 in jazz |

`add9` differs from `9` in that `9` implies a 7th is present, while `add9` does not.

## Slash chords

`X/Y` means chord X with bass note Y. Two flavors:

1. **Inversions** — Y is a chord tone of X. `C/E` is C major in first inversion. Used to control bass motion (often stepwise).
2. **Polychord-flavored** — Y is *not* a chord tone of X. `C/D` is C major over a D bass — functionally close to `D9sus4` but voiced differently. `F/G` is a common dominant substitute (`G9sus4`-like).

Pop and gospel use slash chords heavily for stepwise bass lines: `C – G/B – Am – Em/G – F – C/E – Dm – G`.

## Polychords and upper-structure triads

A polychord stacks two triads, written as `top chord / bottom chord` (e.g., `D/C` = D major triad over C major triad).

Upper-structure triads (UST) are a jazz technique: pick a triad whose notes are tensions over the underlying dominant 7. For `C7`:

| UST | Resulting chord tones | Tensions |
|-----|----------------------|----------|
| D/C7 | D F♯ A | 9, ♯11, 13 (Lydian dominant) |
| D♭/C7 | D♭ F A♭ | ♭9, 11, ♭13; useful color, but the natural 11 clashes with the 3rd if the shell includes E |
| E♭/C7 | E♭ G B♭ | ♯9, 5, ♭7 (Hendrix flavor) |
| F♯/C7 | F♯ A♯ C♯ | ♯11, enharmonic ♭7, ♭9; respell A♯ as B♭ when explaining function |
| A♭/C7 | A♭ C E♭ | ♭13, root, ♯9 |

Players use USTs as voicings — left hand plays the C7 shell (3 + ♭7), right hand plays the upper triad. Quick way to get rich, alterations-loaded voicings without thinking note-by-note.

## Quartal and quintal voicings

Stack 4ths or 5ths instead of 3rds. The result has no clear root and no clear quality (major/minor); it's chord-as-color.

- **Quartal** (stacked 4ths): C–F–B♭–E♭ — standard McCoy Tyner / modern jazz / "So What" voicing.
- **Quintal** (stacked 5ths): C–G–D–A — open, ambient, common in film and ECM-style jazz.

Use for modal contexts, ambiguous tonality, or when you want a less "vertical" sound. See `assets/jazz-voicings.md` for shapes.

## Clusters

Stacked 2nds. Chromatic clusters (C–C♯–D) are extremely dissonant; diatonic clusters (C–D–E) are dissonant but more usable. Used for color, texture, or shock — Bartók, Lutosławski, modern film, certain jazz.

## Symbol notation conventions

Different traditions write the same chord differently. The agent should recognize all and use the one that matches the user's vocabulary.

| Chord | Real Book / jazz | Pop / lead-sheet | Roman numeral | Nashville |
|-------|------------------|------------------|---------------|-----------|
| C major | `C` | `C` | `I` (in C major) | `1` |
| C minor | `C-` | `Cm` | `i` | `1m` |
| C major 7 | `CΔ` or `Cmaj7` | `Cmaj7` | `IΔ` | `1Δ` |
| C dominant 7 | `C7` | `C7` | `V7` (where applicable) | `1 7` |
| C minor 7 | `C-7` | `Cm7` | `i7` | `1m7` |
| C half-dim | `Cø7` | `Cm7♭5` | `viiø7` (in minor) | varies |
| C diminished | `C°7` | `Cdim7` | `vii°7` | `1°7` |

Idiom hints:
- The Real Book uses `Δ` (delta) for major 7.
- Pop and rock often write `Cmaj` or `Cmaj7`; rarely use `Δ`.
- Some jazz scores use `m` and others use `-` for minor.
- Roman numerals use case (capital = major, lowercase = minor); always specify the key.

## Voicing vs. spelling

A chord's *spelling* (its abstract pitch-class content) is fixed. Its *voicing* (how those pitch classes are distributed across registers and instruments) is open.

Common voicing approaches (full treatment in `orchestration/voicing-and-texture.md` and `assets/jazz-voicings.md`):

| Approach | Notes |
|----------|-------|
| Close position | All chord tones within one octave, no gaps |
| Open position | Wider spacing, often with chord tones an octave apart |
| Drop 2 | Take the 2nd voice from top and drop it an octave (jazz guitar, piano LH) |
| Drop 3 | Take the 3rd from top and drop it an octave |
| Rootless | Omit the root (rhythm section bass plays it); typical of jazz piano comping |
| Shell | Just root + 3 + 7 (dropping the 5; jazz piano LH) |
| Quartal | Stacked 4ths (modal, McCoy Tyner, modern) |
| Spread | Wide voicing, 2-octave+ spread; cinematic |

A chord's mood depends as much on voicing as on spelling. `Cmaj7` close voiced (C-E-G-B) is restful. The same chord spread (low C, then E G B in upper register) is open and cinematic. With ♯11 in the top voice (C-E-G-B-F♯) it becomes Lydian-flavored. The user's mood request often translates more to voicing than to chord choice.

## Six quick voicing rules

1. **Don't pile up close intervals in low register.** Below middle C, m2/M2 intervals turn muddy. Spread the voicing in the bass.
2. **The 3rd defines the chord's quality.** If you're going to omit something, omit the 5th (the most expendable note), not the 3rd.
3. **The 7th defines the chord's flavor in jazz.** It's even more important than the 3rd in many jazz voicings.
4. **Tensions (9, 11, 13) want to be on top.** Putting tensions in the middle creates muddy clusters.
5. **The bass is half the chord.** Slash chords and inversions change the chord's affect more than altering an inner voice.
6. **For piano, your LH should rarely play more than 4 notes.** Three is often plenty. Save space.

## Cross-references

- Function (what chords do in a key) → `functional-harmony.md`
- How to connect chords smoothly → `voice-leading.md`
- Non-diatonic chords (secondary dominants, borrowed, Neapolitan, +6) → `chromatic-harmony.md`
- Jazz extensions, ii-V-I, alterations in context → `jazz-harmony.md`
- Voicing distribution across instruments → `orchestration/voicing-and-texture.md`
- Specific voicing shapes (rootless, drops, quartal) → `assets/jazz-voicings.md`
- Modes and which extensions belong to which mode → `modal-harmony.md`, `assets/modes-cheatsheet.md`
- Chord symbol notation across systems (jazz, pop, RN, Nashville) → `assets/chord-symbol-conventions.md`

# Chromatic Harmony

Chromatic harmony introduces notes outside the prevailing key without leaving the key. The diatonic system gets enriched with borrowed colors, secondary tonal centers, and chromatic voice-leading chords. This is the layer that turns a workmanlike progression into one that feels searching, lush, or surprising.

## When to consult this file

- The user wants more harmonic color without changing keys
- "What's a secondary dominant", "what's modal mixture", "what's a Neapolitan"
- A specific chromatic chord appears in analysis
- The user asks how to sneak surprise into an otherwise diatonic progression
- Songwriting goals like "wistful", "yearning", "dramatic", "old-Hollywood" — chromatic harmony is the most common toolkit

This file does not cover full key change. For that, see `modulation.md`. For chromatic harmony in jazz idiom (with extensions and substitution), see `jazz-harmony.md`. For "outside" without functional logic, see `modal-harmony.md` and `techniques/20th-century-techniques.md`.

## The four chromatic toolkits

Almost every chromatic chord falls into one of four families. They overlap, but it's a useful map:

1. **Tonicization** — apply functional dominant motion to a chord that isn't tonic. Secondary dominants, secondary leading-tone chords.
2. **Modal mixture (borrowing)** — borrow chords from the parallel mode (most often parallel minor in major keys).
3. **Special chromatic chords** — Neapolitan, augmented 6ths, common-tone diminished, common-tone augmented.
4. **Chromatic mediants** — chords whose roots are a third away and which share a chromatic third relationship.

## Tonicization

A *tonicization* briefly treats a non-tonic diatonic chord as a temporary tonic by approaching it via its own dominant (or leading-tone chord). The "key" of the music doesn't change — the moment passes — but the local sense of arrival on the tonicized chord adds weight.

### Secondary dominants

A secondary dominant is a `V` (or `V7`) of a non-tonic chord. Notation: `V/x` reads "V of x". Built by taking the chord that's a perfect 5th *above* the target.

In C major, the diatonic chords (other than `vii°`, which is itself dissonant) can each be tonicized:

| Target | Secondary dominant | Notes (C major) | Chord |
|--------|--------------------|-----------------| ------|
| ii (Dm) | V/ii | A C♯ E (G) | A or A7 |
| iii (Em) | V/iii | B D♯ F♯ (A) | B or B7 |
| IV (F) | V/IV | C E G B♭ | C7 (the only secondary that doesn't introduce a new note for the triad — the ♭7 makes it dominant) |
| V (G) | V/V | D F♯ A (C) | D or D7 |
| vi (Am) | V/vi | E G♯ B (D) | E or E7 |

`V/vii°` doesn't exist functionally (you don't tonicize an unstable chord).

The diagnostic for a secondary dominant: a major or dominant 7 chord whose root is a P5 above another diatonic chord, immediately followed by that chord.

In progression: `C – A7 – Dm – G7 – C` is `I – V7/ii – ii – V7 – I`. The A7 tonicizes Dm by introducing C♯ (the leading tone to D), creating a momentary ii-becomes-tonic feeling.

### Secondary leading-tone chords

Same idea, but using a `vii°` instead of `V`. Notation: `vii°/x`.

`vii°/V` in C major = F♯° (F♯ A C). It contains the leading tone of G (F♯) and resolves to G with the same dominant function as `V/V`, but with a softer color (no real root motion of a fifth — the bass moves by half-step).

Secondary leading-tone chords are common when you want chromatic voice leading without the strong root motion of `V/x`.

### Where secondary dominants and leading-tone chords go

Almost always to their target. A `V/V` that doesn't go to V is just a chromatic chord — the listener will hear it as a deception or modulation. The functional power of these chords *is* their resolution to the target.

Common patterns in progressions:

- **`I – V/V – V – I`** — the most common chromatic insertion. Adds drive to the cadence.
- **`I – V/vi – vi – IV – V – I`** — tonicizes the relative minor briefly. Common in ballads, film.
- **`I – V/IV – IV – V – I`** — the V/IV (which is a I7 chord; e.g., C7 in C major) is a blues-flavored move.
- **`vi – V/V – V`** — using the V/V as a predominant.

### Tonicizing in the cycle of fifths

A long cycle-of-fifths sequence can be reinterpreted with secondary dominants at every step:
`I – V/vi – vi – V/ii – ii – V – I` becomes a chromatic sequence where every chord is briefly tonicized before passing on. This is a Bach-flavored move; a slowed-down version is also common in jazz ("ii-V chains").

## Modal mixture (borrowing)

Modal mixture borrows chords from the parallel mode. In a major key, that's the parallel minor (C major borrows from C minor). In a minor key, it's borrowing from parallel major (less common, but real — most often the Picardy third).

### Borrowed chords from parallel minor (in major keys)

| Borrowed chord | What it is in C major | Source in C minor | Effect |
|----------------|----------------------|-------------------|--------|
| `iv` | F minor (F A♭ C) | iv of C minor | Wistful, "smiling-through-tears"; the most common borrowed chord |
| `♭III` | E♭ major (E♭ G B♭) | ♭III of C minor | Bittersweet, distant, sometimes heroic |
| `♭VI` | A♭ major (A♭ C E♭) | ♭VI of C minor | Dark, cinematic, "Hollywood" |
| `♭VII` | B♭ major (B♭ D F) | ♭VII of C minor | Modal, rock-flavored, plagal-direction motion |
| `iiø7` | Dm7♭5 (D F A♭ C) | iiø7 of C minor | Approach to V with darker color |
| `vii°7` (with ♭6) | B D F A♭ | vii°7 of harmonic minor | Stronger leading-tone chord |

Borrowed chord progressions:

- **`I – ♭VI – ♭VII – I`** — the "lift" progression. Common in cinematic pop, anthemic rock.
- **`I – iv – I`** — minor plagal cadence. The Beatles, R&B.
- **`I – V – ♭VI – V`** — the deceptive cadence taken further; very dramatic.
- **`I – ♭III – IV – I`** — modal-flavored, hovering between major and minor.

### Borrowed dominants

Sometimes a borrowed `V` from parallel minor is used. `Vsus♭9` or `V7♭9` borrows ♭6 from minor. Adds darkness to the dominant.

### Picardy third (and the reverse)

The Picardy third is a major tonic chord at the end of a minor-key piece. It's a borrowing from parallel major used at the cadence — purely for effect. Common in Baroque and modern adaptations.

The reverse (minor tonic ending a major piece) is rarer but used — a sudden plunge into minor at the end can be devastating.

## Special chromatic chords

### The Neapolitan (♭II)

A major chord built on the lowered second scale degree. In C major: D♭ major (D♭ F A♭). Almost always in first inversion (`♭II⁶`), with F in the bass.

Functions as a *predominant* — substitutes for `ii` or `iv`. Resolves to V (or V7), then to I.

`I – ♭II⁶ – V – I` is the classic Neapolitan progression. The bass typically walks: C → F → G → C, all stepwise except the last fifth.

In minor, the Neapolitan is even more at home — `♭II⁶` is closely related to the iv chord in minor, and the chromatic descent ♭2 → 1 echoes the Phrygian flavor.

Effect: dramatic, slightly archaic, intensifies the predominant before V. Beethoven uses it constantly.

### Augmented 6th chords

Three chords that are essentially predominants with a chromatic kick — they're built specifically to drive into V via a half-step pull from both sides.

The defining interval is the **augmented 6th** between ♭6 (in the bass) and ♯4 (in an upper voice). When this interval expands outward by half-step in each voice, it lands on the octave at scale degree 5 — i.e., the root of `V`. So the chord tightly pulls into `V`.

The three flavors differ in what's between the ♭6 and ♯4:

| Type | Notes (in C major resolving to V=G) | Other notes |
|------|-------------------------------------|-------------|
| **Italian (It⁶)** | A♭ (♭6) – C – F♯ (♯4) | Just the 3rd added |
| **French (Fr⁶)** | A♭ – C – D – F♯ | Add a 4th above the bass (D, the ^2) — has a whole-tone quality |
| **German (Ger⁶)** | A♭ – C – E♭ – F♯ | Add a m3 above the 3rd (E♭, the ♭3) — sounds almost identical to a dominant 7 chord on A♭, hence enharmonic potential for modulation |

All three resolve to `V`, with the ♭6 → ^5 down by half-step and the ♯4 → ^5 up by half-step (so they collide on the octave at ^5). The German 6th creates parallel 5ths going to V, traditionally avoided by routing through a cadential 6/4 (`Ger⁶ → I⁶/₄ → V`).

Effect: highly dramatic, pre-cadential, nineteenth-century operatic. Wagner uses augmented 6ths heavily.

### Common-tone diminished and common-tone augmented

A diminished 7 chord that shares a *common tone* with the chord that follows. Often written `°7` with the implicit understanding that it's a passing/embellishing function.

In C major, `C°7` (C E♭ G♭ B𝄫) — sharing the common tone C — embellishes a Cmaj7 chord. Used as a passing chord to delay or color the resolution.

Similarly, an augmented chord can share a common tone and provide a chromatic step-up to a target chord.

These are decorative chromatic harmonies, not functional — they're voice-leading flourishes.

### Chromatic mediants

Two chords whose roots are a third apart and share a chromatic relationship. Examples in C:

- C major → E major (chromatic third up)
- C major → A♭ major (chromatic third down)
- C major → E♭ major (also chromatic — borrowed ♭III)
- C major → A major (chromatic — V/ii context, or just chromatic mediant)

Chromatic mediants don't have strong functional drive; they're used for *color*. Schubert, Liszt, late Romantic, and film composers love them. The harmony "swerves sideways" rather than going through dominant logic.

`C → A♭ → C` (I – ♭VI – I) is bittersweet.
`C → E → C` (I – III – I) is uncanny, magical.

These are also the root motions used in neo-Riemannian theory (P, L, R, and chained transformations) — see `techniques/20th-century-techniques.md`.

## Chromatic voice-leading patterns

Several common chromatic moves work as patterns the agent should recognize:

### Line clichés

A static chord with a chromatically descending (or ascending) inner line. Iconic example: a C minor chord with a line moving C → B → B♭ → A → A♭ → G in an inner voice.

This produces the chord sequence `Cm – Cm(maj7) – Cm7 – Cm6 – Fm/A♭ – Cm` (the line cliché). The inner voice walks `C → B → B♭ → A → A♭ → G` while the bass usually pedals on `C` (or moves to `A♭` for the fifth chord). Used in James Bond themes, "Stairway to Heaven", "Michelle", "My Funny Valentine". Adds drama through static harmony with internal motion.

### Descending chromatic bass

A bass line that walks down by half-step under any chord progression. Common pattern in C: `C – G/B – Am – Am7/G – F♯ø7 – F – C/E – D7 – G7 – C`. The bass: C–B–A–G–F♯–F–E–D–G–C.

### Chromatic ascending bass (less common but striking)

Bach uses this. Also common in modern rock and film.

### Plagal chromaticism (the IV-iv-I move)

`IV – iv – I`. The major IV with chromatic descent in an inner voice (often A → A♭ → G in the example C major). One of the most affecting moves in pop. The Beatles' "In My Life" and many others.

## Genre notes

- **Classical (Baroque)**: secondary dominants, Neapolitan, line clichés.
- **Classical (Romantic)**: aggressive chromatic harmony — augmented 6ths, chromatic mediants, modal mixture, deceptive cadences. Wagner, Liszt, Brahms, Mahler.
- **Jazz**: chromatic harmony is built into the language — secondary dominants are constant, modal mixture pervasive. See `jazz-harmony.md`.
- **Pop/rock**: modal mixture (especially `iv` and `♭VI/♭VII`) is everywhere; secondary dominants are less common but appear in 50s/60s songwriting and modern singer-songwriter.
- **R&B/neo-soul**: rich chromatic harmony, often with extensions stacked on borrowed chords. Stevie Wonder, D'Angelo.
- **Film/TV**: chromatic mediants and modal mixture are the language of suspense, wonder, and emotional depth. The "hero's journey" minor-to-major modal mixture is a cliché for a reason.
- **K-pop/J-pop**: heavy modal mixture, especially in bridges and pre-choruses. Borrowed `iv` is ubiquitous in ballads.

## Common pitfalls when advising

- **Don't suggest chromatic chords without justifying their resolution.** A Neapolitan that doesn't go to V is just a confusing chord.
- **Modal mixture is most effective when sparing.** A song that borrows `iv`, `♭VI`, `♭VII`, *and* `♭III` in the same chorus loses the contrast that makes any individual borrowing feel special.
- **Augmented 6ths are register-sensitive.** Voiced in low register they're muddy. Keep the ♯4 high enough to be heard.
- **Secondary dominants need rhythmic space.** Cramming `V/ii – ii – V/V – V – I` into one bar feels rushed; give each tonicization at least a beat to register.
- **Don't reduce all chromatic chords to "tension".** Chromatic chords have specific colors. `iv` is wistful; `♭VI` is heavy; `♯iv°` is brittle. Choose for the affect, not generic "more interesting".

## Quick decision matrix — chromatic chord by mood

| Want | Try |
|------|-----|
| Wistful, "smiling through tears" | Borrowed `iv` in major |
| Lifted, anthemic dark | `♭VI` and `♭VII` (in major) |
| Old Hollywood, lush | Augmented 6ths approaching V |
| Magical, uncanny | Chromatic mediant (e.g., I → III in major) |
| Driving, urgent cadence | Secondary dominant `V/V` |
| Bittersweet, archaic | Neapolitan `♭II⁶ – V – I` |
| Bluesy in a major key | `V/IV` (which is the I7) |
| Suspenseful, hovering | Common-tone diminished |
| Cinematic surprise | Deceptive cadence to `♭VI` instead of `vi` |

## Cross-references

- Functional context (why these chords resolve where they do) → `functional-harmony.md`
- Voice-leading specifics (especially augmented 6th resolution) → `voice-leading.md`
- Modulation (when the chromaticism stays — full key change) → `modulation.md`
- Modal music (when there's no functional motion to chromatically embellish) → `modal-harmony.md`
- Jazz extensions and substitutions (which use secondary dominants and tritone subs heavily) → `jazz-harmony.md`
- Reharmonization techniques that draw on chromatic harmony → `reharmonization.md`
- Chromatic mediants and neo-Riemannian transformations → `techniques/20th-century-techniques.md`

# Functional Harmony

Functional harmony organizes chords by the role they play in establishing and resolving tension toward a tonic. It is the dominant framework for tonal Western music — most classical music from ~1650 to ~1900, the bulk of pop, rock, country, R&B, and gospel, and conservative jazz harmony all operate inside it. Modal music and post-tonal music use different organizing principles (see `modal-harmony.md` and `techniques/20th-century-techniques.md`).

## When to consult this file

- The user asks about chord progressions, "what chord goes here", "why does X→Y sound good"
- Cadences and how to end phrases
- Diatonic motion in major or minor keys
- The bedrock harmonic logic of pop, rock, country, gospel, and most film/TV
- Explaining *why* a progression "works"

Pair with `chromatic-harmony.md` whenever non-diatonic chords (secondary dominants, borrowed chords, Neapolitan, augmented 6ths) appear. Pair with `jazz-harmony.md` for extended chords, substitutions, and ii-V-I in a jazz context. Pair with `modal-harmony.md` when the user describes music that "doesn't seem to land on a regular tonic" or uses a modal framework (Dorian rock, Mixolydian folk, Phrygian flamenco-adjacent pop).

## The three functions

Every diatonic triad in a major or minor key belongs to one of three functional groups, defined by what the chord *does* — where the ear expects it to go.

### Tonic (T) — `I`, `vi`, `iii`

The home. Tonic chords sound stable; the ear treats them as an arrival, not a question. `I` is the prototype. `vi` and `iii` share two of `I`'s three notes, so they can substitute for tonic with a softer, more bittersweet character — `vi` more clearly than `iii`. `iii` is the rarest of the three: it straddles tonic and dominant, sharing two notes with `V` as well.

In minor: `i`, `♭VI`, `♭III` play the corresponding roles.

### Predominant / Subdominant (PD or S) — `ii`, `IV`, `vi`

The chords that prepare a dominant. They generate gentle motion away from tonic. `IV` is bright and stepwise from `V`; `ii` is rougher and pulls more strongly toward `V` (its root is a fifth above `V`'s root, the strongest root motion in tonal music). `vi` is functionally ambiguous: tonic-ish in one context, predominant in another (especially when followed by `ii` or `IV`).

In minor: `iiø` (half-diminished) and `iv` are the standard predominants. `♭VI` can also serve.

### Dominant (D) — `V`, `vii°`

The strongest tension chord. The leading tone (^7) and the dominant 7th's ♭7 both want to resolve — the former up to ^1, the latter down to ^3. `vii°` is `V7` with the root removed, so it serves the same dominant function with a softer color and a leading-tone bass.

In natural minor, the V chord is minor (`v`), which has no leading tone and feels modal. To get functional dominant motion in minor, you raise ^7, producing the `V` chord and (incidentally) the harmonic-minor scale.

A typical phrase moves **T → PD → D → T**. Skipping PD is fine and common (`I → V → I`). Skipping D is rarer in classical practice but normal in pop (`I → IV → I` plagal motion). Going *backwards* (D → PD) is felt as retrograde motion and used deliberately for a stalled or static effect.

## Why functional harmony works

Three perceptual mechanisms drive the system:

1. **Leading-tone resolution.** The half-step from ^7 to ^1 is felt as gravity — the ear strongly anticipates the resolution. `V` contains ^7; `I` contains ^1.
2. **Tritone resolution.** `V7` contains the tritone between ^4 and ^7. Tritones are perceptually unstable and resolve inward (^4→^3 and ^7→^1) or outward, but most strongly inward by half-step contraction. This is why dominant 7th chords are *the* tension chord of the system.
3. **Root motion by descending fifth (or ascending fourth).** The strongest root motion in tonal music is by descending fifth (`V → I`, `ii → V`, `vi → ii`...). This generates the cycle: `iii → vi → ii → V → I`. Each step has the same kinematic feel — partly because each new chord's root was the previous chord's fifth, so there's a felt continuity.

If the user remembers nothing else: ^7 wants ^1, the tritone wants to collapse inward, and roots like to drop by a fifth.

## Cadences (essential)

A cadence is a punctuation mark — how a phrase ends. The choice of cadence determines whether the phrase feels concluded, paused, surprised, or unresolved. See `assets/cadence-reference.md` for the full table; the four most important:

- **Perfect Authentic Cadence (PAC)**: `V → I` with both chords in root position and ^1 in the soprano of the `I`. Maximum closure. The musical equivalent of a period.
- **Imperfect Authentic Cadence (IAC)**: `V → I` with weaker conditions (inverted, or ^3/^5 in the soprano). Closure, but lighter — like a comma at the end of a clause.
- **Half Cadence (HC)**: phrase ends on `V`. Sounds like a question. Often used as the antecedent in a period.
- **Deceptive Cadence**: `V → vi` (instead of expected `V → I`). The ear is surprised; tonal weight delays. Common in pop bridges and in classical dramatic moments.

Plagal (`IV → I`) is gentle and "amen"-flavored; common in rock and gospel as a soft ending. Phrygian half cadence (`iv⁶ → V` in minor) is bittersweet and archaic, common in Baroque slow movements.

## Common diatonic progressions

These are the workhorses. Memorize them, but understand they're not laws — they're high-probability patterns whose ubiquity is itself a fact about how trained ears expect to hear music.

### Major key

- **`I – IV – V – I`** — the rock/folk/country backbone. Energetic and clear.
- **`I – V – vi – IV`** ("axis progression") — the most common pop progression of the last 30 years. Bright, anthemic. Can be rotated to start on any chord (`vi – IV – I – V`, `IV – I – V – vi`, etc.); each rotation has its own emotional shading.
- **`I – vi – IV – V`** ("'50s progression" / doo-wop) — nostalgic, rolls easily.
- **`ii – V – I`** — the basic jazz cadence, but also pervasive in classical and pop. Smooth, propulsive.
- **`I – iii – IV – V`** — more lyrical/lifted than `I – IV – V – I` because the `iii` has a tonic-ish softness.
- **`I – V/vi – vi – V – I`** — tonicizes the relative minor for color. Common in film and ballads.
- **`vi – ii – V – I`** — a deferred opening; starts in the relative-minor mood and lands in major.
- **`I – V⁶ – vi – iii⁶ – IV – I⁶ – ii⁶ – V`** ("Pachelbel canon" sequence) — descending stepwise bass in long form.

### Minor key

- **`i – iv – V – i`** (with raised ^7 on `V`) — tense, classical. Note that `V` requires harmonic minor's raised ^7; `v` (natural minor) sounds modal/unfinished.
- **`i – ♭VI – ♭III – ♭VII`** (or any rotation) — the natural-minor / Aeolian rock progression. Doesn't need a leading tone; epic, modal-flavored.
- **`i – ♭VII – ♭VI – V`** ("Andalusian cadence") — descending tetrachord bass; flamenco-/Phrygian-adjacent.
- **`i – iv – ♭VII – ♭III – ♭VI – iiø – V – i`** (full minor cycle of fifths) — Bach territory.
- **`i – V – i – ♭VII – ♭VI – V`** (lament bass) — descending tetrachord again, but with the V holding the tension. Used in baroque laments and modern ballads.

## Functional harmony in pop vs. classical

Same toolkit, different priorities:

- **Classical** values voice leading, careful preparation of dissonance, and strong cadences at structural points. Functional motion is felt as the primary engine; melody emerges from the harmony.
- **Pop / rock** often inverts this — melody and groove come first, and harmony serves them. Pop tolerates static harmony (one chord for 8 bars), parallel motion, and "wrong" voice leading because the rhythm and timbre carry the energy. Pop also blurs the major/minor distinction more freely and uses modal-flavored progressions (especially Mixolydian and Aeolian) that aren't strictly functional.
- **R&B / gospel** treats functional harmony as a starting point, then enriches it with extensions, secondary dominants, modal mixture, and chromatic passing chords (see `chromatic-harmony.md` and `jazz-harmony.md`).
- **Country** is conservative — `I IV V` plus `vi` and `ii` covers most of the genre. Pedal steel and vocal harmony do the emotional work; the chords themselves stay simple.
- **Film/TV underscore** uses functional harmony for clarity in clear emotional moments (love themes, victory cues), but reaches for modal and post-tonal language for ambiguity, fear, and otherworldliness.

## Tonicization (briefly)

A *tonicization* is a moment where another diatonic chord is briefly treated as a temporary tonic, usually by inserting its dominant before it. `V/V` (the dominant of the dominant) is the most common. In C major: `D7` → `G` is `V/V → V`. The ear momentarily treats G as a local tonic, which makes the eventual return to `C` feel earned.

Full treatment is in `chromatic-harmony.md`. For functional purposes: tonicization extends the system without breaking it. Modulation (which goes further and stays in the new key) is in `modulation.md`.

## Voice leading at the cadence

Voice leading isn't strictly part of "functional harmony", but the dominant→tonic resolution depends on it. The two crucial moves at the V→I:

- **^7 → ^1** (leading tone resolves up). In `V → I` in C major: B → C. Don't double the leading tone in `V` — you'll get parallel octaves on resolution.
- **^4 → ^3** (the tritone resolves inward). In `V7 → I` in C major: F → E.

Both moves are handled automatically by sensible voice leading. Full treatment in `voice-leading.md`.

## Functional tendencies you can rely on

These are not hard rules, but they're robust enough to give the user as defaults — and to break deliberately for effect.

- **`V` strongly wants `I`**. The strongest expectation in tonal music.
- **`vii°` resolves to `I`** like `V` does, but more softly.
- **`V/x` strongly wants `x`** (any tonicization).
- **Predominants prefer `V` next**. `IV → V` and `ii → V` are extremely common; `IV → I` is plagal (also common, but feels different).
- **Chords with the leading tone (V, vii°) don't comfortably go to predominants.** Going `V → IV` in classical is a "regression" and feels stalled; in rock it's ubiquitous and fine.
- **`vi` after `V`** is the standard deception — the V→vi deceptive cadence.
- **Root motion by 2nd, 3rd, and 5th** are all common, ranked by strength: descending fifth > descending third > ascending second > the rest. Ascending fifth (e.g., I→V) is also very common as the *expected* setup move, before the descent home.

## Common pitfalls when advising

- **Don't impose classical voice-leading rules on pop.** A user writing a guitar-strummed indie song is not going to thank you for warning about parallel fifths.
- **Don't reduce every progression to function names.** Sometimes a progression *is* `vi – IV – I – V`; sometimes it's `Em – C – G – D` and the names tell the player what to do. Use whichever vocabulary the user is using.
- **Predominant doesn't have to be `IV` or `ii`.** `vi` can act predominantly. `iii` rarely does, but can. Function is contextual — a chord's role is set by what's around it.
- **Modal progressions ≠ broken functional progressions.** A song in Mixolydian (`I – ♭VII – I`) isn't a "wrong" major-key song; it's modal. Don't try to "fix" it by inserting a leading tone unless the user asks for that sound.
- **A progression isn't strong or weak in isolation.** `I – IV – I – V` sounds strong as a verse and weak as the climax. Strength is contextual.
- **Beware the phrase "this progression is sad/happy/etc."** Mode (major/minor) and tempo and instrumentation usually matter more than progression. `I – V – vi – IV` at 60 BPM with strings is "sad"; the same chords at 140 BPM with distorted guitars is "anthemic".

## Quick decision matrix

When the user wants a feeling, not a specific technique:

| User wants | Try |
|------------|-----|
| Open, anthemic, bright | `I – V – vi – IV` or `I – IV – V – I`; major key |
| Nostalgic, '50s-flavored | `I – vi – IV – V` |
| Pensive, walking | `I – iii – IV – V` or `I – V/vi – vi – IV` |
| Sad but lifted | `vi – IV – I – V` (axis rotation) |
| Mournful, classical | `i – iv – V – i` in minor |
| Epic, modal-rock | `i – ♭VI – ♭III – ♭VII` |
| Flamenco-/Phrygian-adjacent, dramatic | `i – ♭VII – ♭VI – V` (Andalusian) |
| Suspended, unresolved | end the phrase on V (HC) or on iii or vi |
| Delayed resolution | `V → vi` deceptive, then back to `V → I` next phrase |
| Static, hypnotic | one chord for many bars, or a 2-chord oscillation (`I ↔ vi`, `i ↔ ♭VII`) |

Always pair with rhythm, tempo, and timbre suggestions — chord choice alone rarely conveys the user's full intent.

## Cross-references

- Cadence catalog → `assets/cadence-reference.md`
- Standard progression catalog by mood/genre → `assets/progressions-catalog.md`
- Chromatic chords (secondary, borrowed, Neapolitan, +6) → `harmony/chromatic-harmony.md`
- Jazz extensions, ii-V-I as a jazz construct, substitutions → `harmony/jazz-harmony.md`
- Voice-leading rules → `harmony/voice-leading.md`
- Modulation (going to a new key for real) → `harmony/modulation.md`
- Modal harmony (when there's no functional motion at all) → `harmony/modal-harmony.md`
- Reharmonization (replacing functional chords with substitutes) → `harmony/reharmonization.md`

# Modulation

Modulation is a real change of key — not a tonicization that comes back, but a passage that establishes a new tonic and stays there long enough for the listener to accept it. This file covers the techniques for getting from one key to another, and how to choose between them based on the kind of move you want.

## When to consult this file

- The user asks about key changes, modulating, "going up a step at the last chorus"
- Pivot chords, common tones, enharmonic moves
- Distant key relationships and how to bridge them
- The "truck driver modulation" and other pop-specific patterns
- Form-level modulation (development sections, bridges)

For chromatic chords that don't change key (secondary dominants, borrowed chords), see `chromatic-harmony.md`. For key relationships in formal context (sonata form, AABA bridges), see `form/classical-forms.md` and `form/popular-song-forms.md`.

## Tonicization vs. modulation — the boundary

There's no sharp line. The difference is one of duration, structural weight, and whether the new key gets a cadence.

| Feature | Tonicization | Modulation |
|---------|--------------|------------|
| Duration | A beat to a few bars | Several bars to entire sections |
| Cadence in new key? | No | Yes (typically PAC or strong V→I) |
| Returns to original key? | Yes, immediately | May or may not |
| Listener perception | Color in the original key | A new key has been established |

A `V/V – V – I` is tonicization. A passage that goes to the dominant key, cadences there, sits there for 8 bars, then returns is modulation. Some passages (Romantic music especially) live in the gray zone — you've gone "to" the new key but you're never quite settled.

## Key relationships

Keys differ in *closeness* by how many sharps or flats away they are on the cycle of fifths.

### Closely related keys

For a major key, the closely related keys (sharing six of seven scale notes — differing by one accidental) are:

- The dominant (V) — e.g., C major → G major
- The subdominant (IV) — e.g., C major → F major
- The relative minor (vi) — e.g., C major → A minor
- The supertonic minor (ii) — e.g., C major → D minor
- The mediant minor (iii) — e.g., C major → E minor
- The relative minor of the subdominant (vi of IV / iv of I) — e.g., C major → D minor or A minor (overlap with above)

In practice: V, IV, and the relative minor are the "easy" modulations. They share the most chords with the original key.

For a minor key, the closely related keys are: the relative major (♭III), the dominant minor (v), the subdominant minor (iv), the submediant major (♭VI), the supertonic diminished (rarely used as a key center), and the leading-tone subtonic major (♭VII).

### Distant keys

Anything more than one accidental away. Most striking distant relationships:

- **Up a half-step** (e.g., C → D♭): the "truck driver" modulation in pop
- **Up a whole step** (e.g., C → D): also common pop modulation, slightly less abrupt
- **Tritone** (e.g., C → F♯): far apart but enharmonically reachable through tritone sub
- **Chromatic mediants** (e.g., C → E or C → A♭): no shared diatonic chords, very colorful

Distant keys require more setup — you can't just slide in. The longer the distance, the more the modulation needs preparation or boldness to feel earned.

## Modulation techniques

### 1. Pivot chord (common chord) modulation

The most stable, classical technique. Find a chord that exists in both the original key and the destination key. Approach it as a chord of the original key, leave it as a chord of the destination key.

Example: modulating from C major to G major.

- The chord `Am` is `vi` in C major and `ii` in G major.
- Set up: `C – F – Am` (functioning as `I – IV – vi` in C)
- Reinterpret: `Am – D7 – G` (now functioning as `ii – V – I` in G)

The pivot chord isn't heard as a key change yet. The change is felt only when the destination key's dominant arrives (D7 in this example) and resolves to G.

Common pivot chord choices for moving to a closely related key:

| From → To | Common pivots |
|-----------|---------------|
| C major → G major | Am (vi=ii), Em (iii=vi), C (I=IV) |
| C major → F major | Dm (ii=vi), Am (vi=iii), F (IV=I) |
| C major → A minor | C (I=♭III), F (IV=♭VI), G (V=♭VII), Am (vi=i) |
| C major → D minor | F (IV=♭III), Am (vi=v), Dm (ii=i) |

Pivot chord modulation is the workhorse of classical and many pop modulations.

### 2. Direct (phrase) modulation

Just *change keys* with no preparation. The new key's tonic appears suddenly. Surprising and abrupt; usually at a clear formal boundary (start of a verse, chorus, or new section).

Effect: dramatic, especially when the new key is distant. A Beatles trademark — many of their songs modulate by direct juxtaposition between sections.

Pop "truck driver" modulations are usually direct: end the chorus on the original key, repeat the chorus a half-step or whole-step higher. Done correctly, the listener feels the lift; done lazily, it feels gimmicky.

### 3. Common-tone modulation

Hold a single note (or a small set of notes) across the change, and reinterpret it in the new key. The ear latches onto the held tone as continuity while everything else shifts.

Example: a C major chord (C-E-G) → reinterpret E as the tonic of E major. Hold E, change G→G♯ and C→B. Suddenly you're in E major, with E as the common tone.

Common-tone modulation is favorite of Schubert and Romantic composers. Distant keys can be reached almost effortlessly by spotting a shared tone.

### 4. Sequential modulation

A short passage repeats at successively transposed pitch levels until it lands in a new key. Common in Baroque and Classical music; used in Romantic for build-ups.

Example: a 2-bar phrase repeated up a step, then up another step, then resolved in the new key.

Effect: feels like a journey, accumulating energy. Movie trailer music uses this constantly.

### 5. Chromatic modulation

Chromatic voice leading drives the change. The chords don't share diatonic membership, but stepwise voice motion makes the transition feel inevitable.

Most often involves a fully diminished 7 chord (which is a pivot in disguise — `°7` chords belong to multiple keys at once because of their symmetry) or a German 6 reinterpreted as a dominant 7.

### 6. Enharmonic modulation

Reinterpret a chord by spelling its notes differently. Two famous cases:

**German 6 ↔ dominant 7**: A `Ger⁶` in one key has the same pitch content as a dominant 7th in a key a half-step above. In C major, the German 6 (A♭ C E♭ F♯) is enharmonically A♭7 (A♭ C E♭ G♭) — `V7` of D♭. Reinterpret and you're in D♭. An `It⁶` has only three distinct notes, so it can lead into this move by voice expansion, but it is not itself enharmonically identical to a full `V7`.

**Diminished 7 chord**: A `°7` chord is symmetric (stacked m3s), so enharmonically it has four possible roots and can resolve into four different keys. `B°7 = D°7 = F°7 = A♭°7` (with appropriate spelling). It can be the `vii°7` of C, E♭, G♭, or A. Pick which interpretation you want and resolve there.

Enharmonic modulation is the magician's trick — it lets you teleport between keys that look unrelated on paper.

### 7. Modal pivot

Reinterpret the current tonic as a different scale degree of a destination key. This works because the same chord can serve as different functions in different keys.

E.g., the C major chord can be the `I` of C, `IV` of G, `V` of F, or `♭VI` of E minor (if borrowed). Land in C with the previous key's logic, then approach C as if it were `IV` of G — and you've modulated to G.

## Pop-specific modulation patterns

### "Truck driver" modulation (up a half-step or whole-step)

The defining pop modulation: at the last chorus, the entire song shifts up a half-step or whole-step. No preparation, no pivot, just a direct lift.

Examples:
- "Man in the Mirror" (Michael Jackson) — half-step up at the final chorus
- "Living on a Prayer" (Bon Jovi) — half-step up at the final chorus
- "Love on Top" (Beyoncé) — multiple consecutive half-step modulations

Effect: a sudden injection of energy, often paired with a vocal climax. Cliché but reliable.

How to set it up: the bar before the modulation is often a `V7` of the new key (so `V7` of D♭ before modulating up a half-step from C to D♭). This makes it feel earned rather than abrupt.

### Modulation to the relative minor (or back)

A song in major modulates briefly to the relative minor for the bridge, then returns. Works because they share the same key signature — the modulation is more of a shift in tonal center than a change of pitch material.

### Modulation by chorus-shift in K-pop / J-pop

K-pop ballads often modulate within a section — the chorus might start in one key, shift to a new key for the second half, and the song never returns. This produces an ascending emotional contour across the song.

### Modal modulation

Stay in the same set of pitches but shift the tonal center. C major → A minor is technically modulation but uses no new pitches. Common in folk, modal jazz, and some film.

## Setting up a modulation that feels earned

Beyond technique, the *placement* and *preparation* matter:

- **Modulate at structural points.** End of verse → bridge in new key. End of bridge → final chorus a step higher. Don't modulate mid-phrase unless you want it to feel jarring.
- **Pre-domesticate the new key.** A bar or two of the new key's V or V7 before landing on the new tonic helps the ear adjust.
- **Match modulation distance to emotional weight.** A half-step up is energetic. A tritone away is uncanny. A modulation to the relative minor is reflective. Pick the distance that fits the moment.
- **Repetition reinforces the new key.** A modulation that immediately leaves doesn't feel like a real modulation — it's a tonicization.
- **Cadence in the new key.** A V→I (or ii–V–I) in the new key seals the modulation. Without it, the listener is suspicious.

## Returning to the original key

Modulating away is half the work; coming back is the other half. Options:

1. **Cadential return**: end the modulated section with a cadence that resolves into the original key's tonic. Often this means ending on the dominant of the original key.
2. **Pivot back**: same technique as the outbound modulation, in reverse.
3. **Direct return**: just put the original key's tonic at the next formal boundary. Fine if there's been enough time in the new key to make the original feel "fresh" again.
4. **Don't return**: stay in the new key. Common in modulating songs that build energy by climbing keys.

## Genre notes

- **Classical**: pivot chord modulation is standard. Sonata form requires modulation from tonic to dominant (or relative major in minor-key sonatas) in the exposition; development modulates further; recapitulation returns. Modulation is the structural engine.
- **Jazz**: ii-V-I sequences let you modulate freely — chain ii-Vs that target any key. Jazz often doesn't "modulate" in the classical sense; instead, the harmonic ground shifts continuously.
- **Pop**: most songs don't modulate at all (constraints of memorability and verse-chorus structure). When they do, it's usually truck-driver style or a brief bridge-key excursion.
- **R&B/neo-soul**: more modulation than pop average; key shifts at bridges and outros are common.
- **K-pop**: heavy modulation, especially in ballads and dramatic concept songs. Multiple key changes within a single song are common.
- **Film/TV**: modulation is constant — themes are presented in different keys to match scene contexts. Cinematic modulations often use chromatic mediants and enharmonic moves for emotional effect.
- **Musical theatre**: songs often modulate at the bridge to "open up" or at the final chorus to enhance the climax. Used self-consciously and often for narrative reasons.

## Common pitfalls when advising

- **Don't suggest a modulation without thinking about how to come back (or commit to staying).** A modulation that hangs in the new key without resolution feels incomplete.
- **Don't modulate to a distant key without setup.** A surprise modulation can be exciting, but it has to be motivated — usually by a strong sectional break or a specific emotional event.
- **Don't modulate too often.** A song that changes key every 8 bars feels homeless. The listener needs time to settle into each key for the modulation to mean anything.
- **Don't confuse mode shift with modulation.** Going from C major to C minor is modal mixture (or change of mode), not modulation in the technical sense — the tonal center stays.
- **Don't ignore the melody during modulation.** Your vocal line has to land on stable tones in the new key. A modulation that strands the singer on a chromatic note feels unprepared.

## Quick decision matrix

| Want | Use |
|------|-----|
| Smooth, stable modulation to a related key | Pivot chord |
| Sudden energy lift (pop chorus) | Direct (truck driver) modulation up a half/whole step |
| Magical, distant key transition | Common-tone modulation (Schubert flavor) |
| Cinematic build-up | Sequential modulation through ascending steps |
| Surprise reveal of a far-away key | Enharmonic modulation (German 6 → dom 7, or °7 reinterpretation) |
| Reflective shift to "the same place but minor" | Modulate to relative minor |
| Bridge-key excursion in a pop song | Modulate to ♭VI, ♭III, or IV for the bridge |

## Cross-references

- Chromatic chords used as pivots → `chromatic-harmony.md`
- Functional cadences in the destination key → `functional-harmony.md`
- Voice-leading at the modulation point → `voice-leading.md`
- Sonata-form modulation → `form/classical-forms.md`
- Pop-form modulation patterns → `form/popular-song-forms.md`
- K-pop modulation conventions → `genres/kpop-jpop.md`
- Jazz key-change techniques → `jazz-harmony.md`
- Film-score modulation usage → `genres/film-tv-scoring.md`

# Modal Harmony

Modal harmony is what happens when music is organized around a tonal center but doesn't use functional dominant→tonic logic. Instead of leading-tone resolution and tritone gravity, modal music establishes its center through repetition, drone, characteristic notes, and a handful of mode-specific chord moves. It's a fundamentally different system from functional harmony — not a degraded version of it.

## When to consult this file

- The user describes music that "doesn't seem to land" or "hovers"
- Mixolydian rock, Dorian jazz, Phrygian flamenco, Lydian film music
- Any music in a church mode, with a drone, or with a static modal vamp
- Modal jazz (Miles Davis, Coltrane modal era)
- Folk traditions that don't use Western tonal harmony
- The difference between modal music and modal mixture

For functional harmony (and modal mixture as a coloring of functional harmony), see `functional-harmony.md` and `chromatic-harmony.md`. For the modes themselves as scales, see `fundamentals/pitch-intervals-scales.md` and `assets/modes-cheatsheet.md`.

## Modal vs. functional vs. modal mixture — three different things

It's easy to conflate these. They're distinct:

| Concept | What it means |
|---------|---------------|
| **Modal harmony** | Music organized around a tonal center without relying on functional dominant→tonic motion. It often uses non-Ionian modes, but Aeolian and some major-like modal contexts can also be modal when the center is established by repetition, drone, characteristic notes, or vamps rather than leading-tone cadence. |
| **Modal jazz** | A specific mid-20th-century jazz movement where each section uses a single mode for extended periods. Subset of modal harmony. |
| **Modal mixture** | Borrowing chords from a parallel mode while staying in functional harmony. Lives inside functional logic. |

A song can use modal mixture without being modal. A modal song doesn't necessarily use modal mixture (it doesn't borrow from anywhere — it's wholly in one mode).

## The seven diatonic modes

All built from the same set of pitches (the white keys of the piano) but starting on different notes. Each has a distinct character.

| Mode | Starts on | Formula (whole/half steps from tonic) | Character |
|------|-----------|---------------------------------------|-----------|
| Ionian | C | W W H W W W H | Major; functional default |
| Dorian | D | W H W W W H W | Minor with raised ^6; cool, jazzy |
| Phrygian | E | H W W W H W W | Minor with lowered ^2; dark, half-step-above-tonic color |
| Lydian | F | W W W H W W H | Major with raised ^4; floating, otherworldly |
| Mixolydian | G | W W H W W H W | Major with lowered ^7; rock, blues, Celtic |
| Aeolian | A | W H W W H W W | Natural minor; reflective |
| Locrian | B | H W W H W W W | Diminished tonic; rarely used as tonal center |

For any reference key (e.g., C as the tonic of every mode), the mode formulas in scale-degree terms:

| Mode (root C) | Notes | Distinguishing alterations |
|---------------|-------|----------------------------|
| C Ionian | C D E F G A B | (default major) |
| C Dorian | C D E♭ F G A B♭ | ♭3, ♭7, natural 6 |
| C Phrygian | C D♭ E♭ F G A♭ B♭ | ♭2, ♭3, ♭6, ♭7 |
| C Lydian | C D E F♯ G A B | ♯4 |
| C Mixolydian | C D E F G A B♭ | ♭7 |
| C Aeolian | C D E♭ F G A♭ B♭ | ♭3, ♭6, ♭7 (natural minor) |
| C Locrian | C D♭ E♭ F G♭ A♭ B♭ | ♭2, ♭3, ♭5, ♭6, ♭7 |

## Characteristic notes

The "characteristic note" of a mode is the scale degree that distinguishes it from the more familiar major or minor scale. Establishing the mode means *featuring* this note — placing it on strong beats, in the melody's contour, in the bass, or in characteristic chords.

| Mode | Characteristic note (vs. major/minor) | What it does |
|------|---------------------------------------|--------------|
| Dorian | natural ^6 (vs. minor's ♭6) | Brightens the minor scale |
| Phrygian | ♭2 | Flamenco-/Andalusian-adjacent half-step color |
| Lydian | ♯4 | Otherworldly lift |
| Mixolydian | ♭7 (vs. major's ^7) | Rock/blues/Celtic flavor; no leading tone |
| Aeolian | ♭6 + ♭7 (no leading tone) | Pure natural minor; modal/folk |

Without the characteristic note, the music will collapse perceptually into the closest major or minor mode. A Dorian piece without the natural 6 sounds like minor. A Mixolydian piece without the ♭7 sounds like major.

## How modal music establishes its tonal center (without a leading tone)

Functional harmony gives you the leading tone and the V→I cadence as your closure mechanism. Modal music has none of that — Mixolydian's ^7 is ♭7, Dorian's ^7 is ♭7, etc. So how does the listener know what the tonic is?

Several mechanisms, used singly or in combination:

1. **Drone or pedal tone.** A held bass note on the tonic (or tonic + fifth) anchors the mode. Common in folk, raga, certain rock (hurdy-gurdy traditions, Sabbath riffs).
2. **Repetition and metric placement.** The tonic falls on downbeats, opens phrases, ends them. Repetition builds expectation of return.
3. **Characteristic chord move.** Each mode has a signature chord move that confirms the modal center (see below).
4. **Modal vamps.** A short repeating chord progression that loops between the tonic and a characteristic chord. The tonic's primacy is established by repetition.
5. **Static harmony on the tonic chord.** Long stretches of unchanging harmony allow the melody to explore the mode without competing pulls.

## Characteristic chord moves per mode

Each mode has signature chord progressions that establish modal identity without functional drive.

### Dorian (natural ^6 in a minor frame)

The defining chord move: **`i – IV`** (because IV is *major* in Dorian — it contains the natural ^6).

In D Dorian: `Dm – G – Dm – G` is the prototypical Dorian vamp. The G chord (containing B, the natural 6 of D) confirms Dorian.

Other Dorian moves:
- `i – IV – ♭VII – i` (Dm – G – C – Dm)
- `i – ♭III – IV – ♭VI`
- "So What" by Miles Davis: pure D Dorian with `Dm7` as the home chord.

### Phrygian (♭2 in a minor frame)

The defining move: **`i – ♭II`** (because the ♭II is so distinctive). The bass moves down by half-step from the tonic.

In E Phrygian: `Em – F – Em – F` is the basic vamp. The F chord (containing the ♭2 of E) sets it up.

Phrygian dominant (♭3 raised — see below) is more common than pure Phrygian in Western music. The Andalusian cadence `i – ♭VII – ♭VI – V` lives at the boundary of Phrygian and Phrygian dominant.

### Lydian (♯4 in a major frame)

The defining move: **`I – II`** (because II is *major* in Lydian — it contains the ♯4).

In F Lydian: `F – G – F – G`. The G chord (containing B, the ♯4 of F) is what makes it Lydian instead of major.

Lydian gives a floating, "looking up" quality. Common in film (Williams' "E.T."), jazz (`Maj7♯11` chord), and some prog.

### Mixolydian (♭7 in a major frame)

The defining move: **`I – ♭VII`**.

In G Mixolydian: `G – F – G – F`. The F chord (containing the ♭7 of G) confirms Mixolydian.

This is the bedrock of rock — `D – C – G – D` (Mixolydian in D), Beatles' "Norwegian Wood" (Mixolydian-flavored), Stones' "Sympathy for the Devil", AC/DC, much of Led Zeppelin. Also Celtic music.

Other Mixolydian moves:
- `I – ♭VII – IV – I`
- `I – v – ♭VII – IV`

### Aeolian (natural minor — ♭3, ♭6, ♭7, no leading tone)

The defining move: **`i – ♭VII`** (similar to Mixolydian's I-♭VII but in a minor frame). Without a leading tone, Aeolian feels "modal minor" rather than functional minor.

In A Aeolian: `Am – G – Am – G`. Or the famous rock progression `Am – F – C – G` (i – ♭VI – ♭III – ♭VII).

Most "minor key rock" is Aeolian rather than functional minor. The progression `i – ♭VII – ♭VI – ♭VII` is ubiquitous.

### Locrian (diminished tonic)

Rarely used as a true tonal center. The tonic chord is diminished and unstable, so it's hard to feel as "home". When Locrian is used, it is usually a momentary color, a metal sonority, or a highly dissonant film/avant-garde device.

### Ionian (major) and "modal major"

Strictly, Ionian is just major and operates with full functional logic. But "modal major" is sometimes invoked to describe major-key music that avoids strong V→I cadences and leans on plagal motion (`I – IV – I – IV`) and parallel motion. Some folk and pop fits this description.

## Modal mixture vs. modal music

Modal mixture borrows chords from parallel modes while staying in functional harmony. It uses borrowed chords for color and resolves them functionally. The home key is unambiguous; the borrowing is local.

Modal music doesn't have a home key in the functional sense — it has a *modal center*. The mode is the system, not a borrowing from one.

A song that opens with `Cmaj7 – B♭maj7 – Cmaj7 – B♭maj7` (I and ♭VII alternating) is Mixolydian if it stays there. If the same chords show up in a song that's clearly in C major and resolves through `G7 – C` at cadences, the `B♭` is borrowed (modal mixture) and the song is functional.

## Modal jazz

A specific genre, distinct from "jazz with modal flavor". Modal jazz typically:

- Uses a single mode for extended sections (often 8 or 16 bars per mode)
- Replaces ii-V-I motion with static modal vamps
- Frees the soloist to explore one mode at length, building intensity through dynamic and rhythmic development rather than through chord-change navigation
- Was pioneered around 1959 with Miles Davis's *Kind of Blue*, John Coltrane's modal period (*Impressions*, *India*), and continued by McCoy Tyner, Herbie Hancock, and many others.

The Mode-of-the-moment in modal jazz is *the* harmony — there's no chord-scale mismatch because there's no chord change against which to be wrong.

Common modal jazz forms:

- **AABA with modal sections**: e.g., 16 bars D Dorian, 8 bars E♭ Dorian, 8 bars D Dorian.
- **Two-mode swap**: e.g., 8 bars D Dorian, 8 bars E♭ Dorian, repeat.
- **Single mode for the whole tune**: rare in jazz; common in raga-influenced jazz fusion.

Voicings in modal jazz are typically quartal (stacked 4ths), pentatonic-derived, or open with extensions emphasizing the characteristic note.

## Modal harmony in pop

Pop uses modes constantly without naming them as such:

- **Mixolydian**: blues-influenced rock and folk-rock — Beatles, Rolling Stones, much of 70s rock.
- **Dorian**: cool jazz-pop, certain Beatles songs ("Eleanor Rigby"), darker indie.
- **Aeolian**: most "minor key" pop and rock, especially since the 80s. The whole indie-rock/alt-rock world.
- **Phrygian**: occasional flamenco-adjacent or metal-adjacent moments. Less common in pop.
- **Lydian**: rare in pop verses; common as a brief uplift, often via a Lydian IV moment.

The dominant pop tonalities are Aeolian and Mixolydian-flavored major. Most pop today is closer to modal than to functional.

## Modal harmony in film and TV

Film composers reach for modes constantly:

- **Lydian**: wonder, magic, awe (Spielberg/Williams territory)
- **Mixolydian**: heroism, adventure
- **Dorian**: cool, urban, ambiguous
- **Phrygian**: severity, starkness, claustrophobic tension; avoid using it as shorthand for cultural otherness
- **Aeolian**: melancholy, loss
- **Locrian**: alien, deeply unsettling

Combined with characteristic instrumentation (high strings + woodwinds for Lydian, low strings + horns for Phrygian, etc.), modes give film composers an almost color-palette use of harmony.

## Special modal scales

Beyond the seven diatonic modes, several non-diatonic modes are common:

### Phrygian dominant

Phrygian with a *raised* 3rd. In E: E F G♯ A B C D. The 5th mode of harmonic minor.

Sounds: flamenco-adjacent, klezmer-adjacent, maqam-adjacent in equal temperament, and certain metal idioms. The combination of the ♭2 and the major 3rd creates a strongly marked augmented second (between F and G♯ here). Use the label as a pitch-color shortcut only, not as a claim that the music is authentically flamenco, klezmer, or maqam-based.

### Lydian dominant (Lydian ♭7)

Lydian with a lowered 7th. In F: F G A B C D E♭. The 4th mode of melodic minor.

Sounds: jazzy, "the bright dominant". Common over `V7♯11` chords. Steely Dan, fusion.

### Altered scale (super Locrian)

The 7th mode of melodic minor. All the alterations stacked: ♭2, ♭3, ♭4 (= M3 enharmonically), ♭5, ♭6, ♭7. Used over altered dominant chords (`7alt`).

### Hungarian / gypsy minor

Harmonic minor with a raised 4th. Strong augmented-2nd color; highly marked and dramatic.

### Melodic minor (jazz minor, ascending only)

Minor with raised 6 and 7. Often used in jazz over minor-major 7 chords. Each of its modes has a name and a use: Dorian ♭2, Lydian augmented, Lydian dominant, Mixolydian ♭6, Locrian ♮2, altered.

## Common pitfalls when advising

- **Don't impose functional logic on modal music.** A Mixolydian tune doesn't need a leading tone "for resolution" — it resolves by other means. Suggesting one will destroy the modality.
- **Don't confuse modal mixture with modal music.** A pop song with a borrowed `iv` is functional with mixture. A pop song built on `i – ♭VII – ♭VI – ♭VII` is Aeolian.
- **Establish the mode early.** A modal piece needs the characteristic note in the first phrase; otherwise the listener defaults to major or minor and feels confused when the mode reasserts itself.
- **Use the characteristic chord moves.** Telling the user "write in Dorian" is vague. Telling them "alternate i and IV — your IV will be major because of the natural 6th" is actionable.
- **Don't overload on alterations.** A piece in pure Phrygian is rare and difficult; most "Phrygian" tracks use Phrygian color on top of an underlying major or minor system. Pick the level of modal commitment that fits.

## Quick decision matrix — mode by feeling

| Want | Try |
|------|-----|
| Cool, ambiguous, mature | Dorian |
| Dark half-step-above-tonic color, flamenco-adjacent or metal edge | Phrygian or Phrygian dominant |
| Magical, floating, wonder | Lydian |
| Rock, blues, anthemic without leading tone | Mixolydian |
| Reflective minor, folky | Aeolian |
| Alien, dread | Locrian |
| Klezmer-adjacent, flamenco-adjacent, or maqam-adjacent pitch color | Phrygian dominant, plus idiom-specific rhythm, ornament, phrasing, and listening study |
| Jazzy bright dominant | Lydian dominant |
| Tense, fully altered dominant | Altered scale |

## Cross-references

- The modes as scales (with examples and theory) → `fundamentals/pitch-intervals-scales.md`
- Modes cheat sheet (formulas, characteristic chords) → `assets/modes-cheatsheet.md`
- Functional harmony (the alternative system) → `functional-harmony.md`
- Modal mixture (borrowing across modes within functional logic) → `chromatic-harmony.md`
- Modal cadences → `assets/cadence-reference.md`
- Modal jazz in genre context → `genres/jazz-styles.md`
- Modal usage in film → `genres/film-tv-scoring.md`
- Modal melodies (which scale degrees to feature in lines) → `melody/melodic-construction.md`

# Reharmonization

Reharmonization is the act of replacing the chords of an existing piece (or section, or phrase) while keeping its melody, form, or character largely intact. It's a synthesis discipline — every reharm technique pulls from chromatic harmony, jazz substitution, modal interchange, or voice-leading reasoning. This file pulls those techniques together and frames them as decisions about *what role* the new chord plays.

## When to consult this file

- The user has a chord progression they want to rework
- "How do I make this progression more interesting"
- "Reharmonize this melody"
- Replacing chords for variety, mood change, sectional contrast (e.g., reharmonizing the bridge of an AABA tune)
- Setting up a richer arrangement of a simple song
- Jazz-style chart writing

For the underlying tools, see: `chord-construction.md` (extensions, alterations), `functional-harmony.md` (functional substitutes), `chromatic-harmony.md` (secondary dominants, modal mixture, special chords), `jazz-harmony.md` (tritone subs, ii-V chains, Coltrane). This file is the application layer.

## Why reharmonize

Three motivations cover almost every case:

1. **The melody is good but the harmony is plain or generic.** The melody has already proven itself memorable; the chords are doing the bare minimum. Reharm adds depth.
2. **The progression has been heard too many times.** Pop-songbook reharms (e.g., axis progression all the way through) feel familiar; substitution adds personality.
3. **Sectional contrast.** AABA forms benefit from reharmonizing the second A or the bridge; final choruses can feature reharm to elevate the energy or color.

## The melody is sacred

When reharmonizing under a fixed melody, every new chord must contain the melody note (or accommodate it as an extension or non-chord tone). This is the central constraint.

If the melody plays C on beat 1, your reharm chord must:
- Contain C as a chord tone (root, 3, 5, 7, etc.), OR
- Contain C as an extension (9 of B♭maj7, ♯11 of F♯ aren't the same thing — read carefully — 11 of G7sus4, etc.), OR
- Treat C as a non-chord tone that resolves quickly (passing tone, suspension, appoggiatura)

Most reharms keep the melody as a chord tone or a clear extension. NCT-treated reharm is a more sophisticated move and works best when the chord changes faster than the melody.

## Reharm by chord function — substitution within roles

The simplest reharm: replace a chord with another that performs the same function.

### Tonic substitutes

`I` can be substituted by `vi`, `iii`, or `Imaj7` (with extensions for color).

`I – IV – V – I` becomes `vi – IV – V – I` or `iii – IV – V – I`. The function is the same; the color shifts.

### Predominant substitutes

`IV` and `ii` are interchangeable as predominants. `vi` can serve too. So:

`C – F – G – C` (I-IV-V-I) becomes `C – Dm – G – C` (I-ii-V-I) or `C – Am – G – C` (I-vi-V-I, though this is heard differently because vi is also tonic-flavored).

### Dominant substitutes

`V` and `vii°` share function. `V7` for full power; `vii°` for softer dominant. In jazz, `subV7` (tritone sub) is added to the dominant family.

`G – C` becomes `Bm7♭5 – C` (vii°-derived) or `D♭7 – C` (tritone sub).

## Reharm by adding chromaticism

Take a diatonic progression and chromatically extend it.

### 1. Insert secondary dominants

Before any diatonic chord (other than `vii°`), insert its `V` or `V7`.

`C – Dm – G – C` (I-ii-V-I) → `C – A7 – Dm – G – C` (I-V/ii-ii-V-I).

The A7 tonicizes Dm. Adds drive without changing the underlying progression's logic.

Stack secondary dominants for more density: `C – E7 – A7 – Dm – G – C` (the E7 is V/vi → which would resolve to Am, but here it goes to A7 instead, which is V/ii → Dm — so you've made a chain of secondary dominants down the cycle).

### 2. Insert ii-Vs

Even more substantial than secondary dominants: insert a full `ii-V` before each chord.

`C – F – Dm – G – C` (I-IV-ii-V-I) → `C – Gm7 – C7 – F – Em7 – A7 – Dm – G – C` (I, then ii-V to F, F, then ii-V to Dm, Dm, V, I).

Now the progression has 9 chords where it had 5. Used heavily in jazz reharms of pop songs.

### 3. Borrowed chords

Replace a diatonic chord with its parallel-mode equivalent. The most common moves:

- Replace `IV` with `iv` (the "Beatles move"): `C – F – G – C` becomes `C – Fm – G – C`.
- Insert `♭VI` before `V` for cinematic color: `C – F – G` becomes `C – F – A♭ – G`.
- Use `♭VII` instead of `V` for modal flavor: `C – F – G – C` becomes `C – F – B♭ – C`.

Combine with secondary dominants for richer reharm: `C – Fm – B♭ – Cmaj7` (I – iv – ♭VII – I, with smooth voice leading).

### 4. Tritone substitution

Replace any `V7` with the dom7 a tritone away.

`C – Dm – G7 – C` → `C – Dm – D♭7 – C`. The chromatic descent (D – D♭ – C) is the gain.

In a chain: `C – A7 – Dm – G7 – C` → `C – E♭7 – Dm – D♭7 – C` (both dominants tritone-substituted; bass: C – E♭ – D – D♭ – C, mostly chromatic).

## Reharm by changing the bass line

Bass motion drives much of the perceived progression. Reharm by designing a bass line, then choosing chords that fit it.

### Stepwise descending bass

A bass that walks down stepwise from tonic to dominant (or further). Fit chords above:

C major, descending bass: `C – B – A – G – F – E – D – G`.

| Bass | Possible chord |
|------|----------------|
| C | C |
| B | G/B (V in first inversion) or Em/B |
| A | Am |
| G | C/G (I in second inversion) or G |
| F | F or Dm/F |
| E | C/E (I in first inversion) or Em |
| D | Dm or G/D (V in second inversion) |
| G | G |

The progression becomes melodic in the bass while supporting the original melody above. This is the Pachelbel-canon territory and the foundation of many ballads.

### Pedal bass

Hold a single bass note and let chords change above it. The pedal note is usually the tonic or the dominant.

Tonic pedal: `C – F/C – G/C – C – Am/C – Dm/C – G/C – C`. The bass C never moves. Used for tense static-harmony moments.

Dominant pedal: bass on G (or 5 of any tonic), various chords above. Builds tension before the resolution.

### Chromatic descending bass

Each bass note is a half-step below the previous. Find chords that work above it.

`C – G/B – Am – Am7/G – F♯m7♭5 – F – C/E – D7 – G7 – C`. Bass: C–B–A–G–F♯–F–E–D–G–C. Chromatic from C down to D.

This is a classic jazz reharm move and shows up in dozens of standards.

## Reharm by recontextualizing function

Take a chord and reinterpret it as having a different function in a different (briefly tonicized) key.

Example: a `Cmaj7` in the original progression. Reharm by treating the melody's C as the 3rd of `A♭maj7` instead — now you're momentarily in A♭ major, and the melody note serves as the chord's 3rd.

`Cmaj7 – Dm7 – G7 – Cmaj7` reharmed: `A♭maj7 – Dm7 – G7 – Cmaj7`. The first chord is no longer tonic but a chromatic mediant; melody note C is now the 3rd of A♭maj7. Suddenly the progression sounds Schubertian.

This kind of reharm is harder — you have to verify the melody fits the new chord — but yields the most striking results.

## Reharm by sequence

If the original progression repeats a phrase, reharm one of the repetitions to a different key center, then sequence it.

`Cmaj7 – Am7 – Dm7 – G7 || Cmaj7 – Am7 – Dm7 – G7` (repeating phrase) becomes `Cmaj7 – Am7 – Dm7 – G7 || E♭maj7 – Cm7 – Fm7 – B♭7`. The second half is the same progression a m3 up.

Sequential reharm is structurally satisfying — the listener perceives the parallel and feels the lift of the new key.

## Reharm by line — counterpoint reharm

Treat the upper voice (or an inner voice) as a counterpoint line and rebuild the harmony to support it. The most striking version is the **pedal-melody reharm**: a single sustained note in the melody, against shifting chords that each interpret the held note as a different chord tone or extension.

Example: the melody sustains a long `C` for four bars. Pick chords whose chord tones (or available tensions) include `C`, and design a bass line that creates motion underneath:

| Bar | Chord | `C` functions as |
|-----|-------|------------------|
| 1 | `Cmaj7` | root |
| 2 | `A♭maj7` | major 3rd |
| 3 | `Fmaj7` | 5th |
| 4 | `D♭maj7` | major 7th |

The melody `C` never moves; the harmony around it shifts through chromatic-mediant relationships. Each chord recasts the same pitch as a different chord tone, producing a constantly-changing emotional reading of one note.

Other common targets for sustained `C`:
- `Am7` — `C` is the m3 (chord tone)
- `Dm7` — `C` is the ♭7 (chord tone, very lush)
- `F♯ø7` — `C` is the ♭5 (chord tone, dark)
- `B♭maj7` — `C` is the 9 (extension)
- `E♭maj7` — `C` is the 13 (extension)

When designing a reharm under a held note, the test is strict: every chord's pitch content must contain the melody note as a chord tone, or accept it as a clearly available extension (9, 11, 13, or altered tension where idiomatic). Notes that fall on "avoid" extensions — like a natural 11 over a major 7 — produce sour clashes and break the line.

Counterpoint reharm under a *moving* upper line works the same way but with more constraints: every beat the melody changes, every chord must contain (or accept) that beat's note. Bill Evans, Brad Mehldau, and the late-Romantic harmonic tradition (Wolf, Strauss) all reward close study here.

## Reharm by texture — same chords, different feel

Sometimes "reharm" doesn't mean changing chords at all — it means changing the *voicing* and *bass position* of the existing chords.

Original: `C – Am – F – G` (close-position triads).

Reharm by texture: same chords, but voiced as `C9 – Am11 – Fmaj9 – G13sus4 – G13`, with rootless voicings and a busier bass line.

The ear hears this as significantly more interesting, even though the harmonic skeleton is identical.

## Reharm priorities — what to focus on

When reharmonizing a piece, decide what you want to *change* and what you want to *preserve*:

| Preserve | Change | Result |
|----------|--------|--------|
| Melody, function | Surface chord choices | Subtle reharm; same emotional weight, more colors |
| Melody, color | Function | Bold reharm; same notes, but a new emotional reading |
| Melody, texture | Bass line | Walking-bass reharm; jazzy |
| Melody, key | Sectional treatment | Different sections sound like different songs |
| Melody | Everything else | Full reharm; could be Brad Mehldau territory |

Most pop-style reharms preserve function and change surface. Most jazz reharms change both function and surface. Avant-garde reharms can preserve only the melodic shape.

## Less is more

The hardest discipline in reharm is restraint. A heavily reharmonized verse can feel exhausting; the listener loses track of the song. The most affecting reharms often:

- Leave most of the original progression alone
- Change the harmony at one or two key moments (the title hook, the second-A in AABA, the bridge)
- Use just enough new color to make those moments feel special

A reharm that swaps every chord can feel like exhibitionism. A reharm that swaps three chords at the perfect moments feels like genius.

## Common reharm patterns in pop and jazz

### Bridge reharm

The bridge is the most reharmonized section in popular forms. AABA bridges typically modulate to a related key and use richer harmony than the A sections.

Common moves: bridge in the relative minor, bridge in the parallel minor, bridge starting on `IV` and modulating, bridge as a ii-V-I chain.

### Final chorus reharm

The last chorus (or last A) can reharm the original progression with extensions, secondary dominants, or modulation. This signals "we've arrived" musically.

Pop final-chorus reharm: original is `I – V – vi – IV`; final chorus is `I – V/vi – vi – V/V – V – IV – I` (more chord changes, more chromatic motion) — or just modulate up a half-step.

### Outro reharm

The outro can reharm by *simplifying* — strip the extensions, repeat a single chord with variation, or land on an unresolved dominant or modal vamp. Effective when contrasted with a fully harmonized song body.

### Intro reharm

A song in C major might intro on a related key (the relative minor, a chromatic mediant), creating uncertainty before the verse's clear arrival on `C`.

## Genre notes

- **Classical (esp. Romantic)**: reharms operate at the level of variation movements and through-composed development. Late Romantic music constantly reharmonizes its own themes.
- **Jazz**: reharm is the genre. Standards exist as much for reharmonization as for melody. Bill Evans' reharms of standards are landmark.
- **Pop**: reharm is more conservative; usually applied at the bridge or final chorus. Acoustic singer-songwriters often reharm covers as a defining stylistic choice.
- **R&B / neo-soul**: heavy reharm with rich extensions and modal interchange. Stevie Wonder, D'Angelo, Robert Glasper.
- **Film/TV**: reharm of themes is the language of leitmotif. The same theme appears in different keys and harmonic contexts to reflect narrative.
- **K-pop / J-pop**: reharm at the bridge and second chorus. Often modulates and adds extensions. Sometimes reharm becomes a hook unto itself.

## Common pitfalls when advising

- **Don't reharm just to reharm.** "Make it more interesting" without a target feeling produces busy, unmotivated chords.
- **Don't lose the song.** A reharm that obscures the original progression so completely that the song loses its identity has gone too far. (Unless that's the goal.)
- **Don't put dissonant extensions in the bass.** The lowest note is the most defining; reharms that move the bass should keep the bass intervals consonant.
- **Don't ignore tempo.** A reharm that adds chord changes on every beat will work at 60 BPM and not at 140 BPM. Density is tempo-dependent.
- **Don't ignore the melody's strong notes.** Reharms that put a melody's structural note on a chord where it's a weird tension will sound wrong, even if the chord itself is fine.

## Quick decision matrix

| Want | Try |
|------|-----|
| Smoother bass motion | Slash chords / inversions / pedal bass |
| More tension before resolution | Insert secondary dominants or tritone subs |
| Wistful, "Beatles" minor color in major | Borrowed iv |
| Modal flavor | Replace V with ♭VII |
| Cinematic darkness | ♭VI inserted before V; modal mixture chords |
| Jazzy density | Insert ii-Vs before each chord |
| Sectional contrast (e.g., bridge) | Modulate to a chromatic mediant or parallel mode |
| Old Hollywood lush | Augmented 6ths approaching V |
| Schubertian magic | Reinterpret a chord via common-tone modulation |
| Gospel/R&B richness | Replace simple chords with extended (9, 11, 13), use slash chords for stepwise bass |
| Final-chorus lift | Modulate up a step or half-step |

## Cross-references

- The toolkit (functional substitutions) → `functional-harmony.md`
- Chromatic toolkit (secondary dominants, modal mixture, Neapolitan, +6) → `chromatic-harmony.md`
- Jazz substitution toolkit (tritone sub, backdoor ii-V, Coltrane) → `jazz-harmony.md`
- Modulation as reharm at the section level → `modulation.md`
- Voice leading constraints during reharm → `voice-leading.md`
- Voicings that change feel without changing chords → `chord-construction.md`, `assets/jazz-voicings.md`
- Form-level reharm (where to apply, when) → `form/popular-song-forms.md`, `form/classical-forms.md`
- Reharm in genre context → `genres/jazz-styles.md`, `genres/film-tv-scoring.md`, `genres/kpop-jpop.md`

# Voice Leading

Voice leading is the discipline of moving each individual voice (line) smoothly from one chord to the next. It's the engine of tonal music. A progression that looks ordinary on paper can sound exquisite or clumsy depending on voice leading; a progression that looks adventurous can sound tame if voice-led predictably.

## When to consult this file

- The user asks how to "connect" chords smoothly
- 4-part writing, chorale style, SATB
- The user has a chord progression that "sounds clunky" or "doesn't flow"
- Questions about parallel 5ths, parallel octaves, "rules" of harmony
- Suspensions, anticipations, passing tones, neighbor tones (NCTs)
- Cadential 6/4 chords
- How jazz/pop voice leading differs from classical

For the function of chords, see `functional-harmony.md`. For chord spelling and qualities, see `chord-construction.md`. For texture/orchestration considerations, see `orchestration/voicing-and-texture.md`.

## The four voices (SATB)

Classical voice leading uses four voices, named by vocal part. Their typical ranges:

| Voice | Range (typical) | Range (extended) |
|-------|-----------------|------------------|
| Soprano | C4 – G5 | A3 – C6 |
| Alto | G3 – D5 | E3 – F5 |
| Tenor | C3 – G4 | B2 – A4 |
| Bass | E2 – C4 | C2 – E4 |

Adjacent voices should not exceed an octave apart, except bass-tenor (which may exceed an octave because the bass is registrally distinct).

These ranges are templates. Pop, jazz, and instrumental writing can be looser, but the underlying principle — keep voices in their comfort zone, avoid extremes — applies everywhere.

## Doubling rules

When writing a 4-voice texture for a 3-note triad, one note must be doubled.

| Chord type | Default double | Avoid doubling |
|------------|----------------|----------------|
| Major triad in root position | Root | Leading tone (creates parallel 8ves on resolution) |
| Major triad in first inversion | Root, soprano note, or 5th | Leading tone |
| Major triad in second inversion | Bass (the 5th) | Other notes |
| Minor triad | Root or 3rd | (no major issue) |
| Diminished triad (rare in root pos) | 3rd (the soft note) | Root or 5th (creates tritone exposure) |

In strict common-practice four-part writing, avoid doubling the leading tone (^7) in a `V` or `vii°` chord because both copies tend to resolve up to ^1, often producing exposed or parallel octaves. In pop, jazz, guitar voicings, and non-resolving textures, this is a style-aware caution rather than a universal ban.

For 7th chords, all four voices can be different, so doubling isn't required. If you do double in strict tonal writing, the root or 5th is usually safer than the 7th; doubling the 7th often creates awkward resolution because both copies tend to resolve downward.

## Motion types

Between any two voices, there are four kinds of motion:

| Motion | Definition | Example (from C-E to D-F) |
|--------|------------|---------------------------|
| **Parallel** | Both voices move in the same direction by the same interval | C-E → D-F (both up a M2; interval stays a M3) |
| **Similar** | Both voices move in the same direction by different intervals | C-E → D-G (both up, but one M2 and one P4) |
| **Contrary** | Voices move in opposite directions | C-E → B-F (one down, one up) |
| **Oblique** | One voice moves, the other stays | C-E → C-F |

Contrary motion is the most "independent" — it's the gold standard for keeping each voice melodically distinct. Similar motion is fine. Parallel motion is fine for most intervals but produces specific problems for two of them.

## The parallel-fifth and parallel-octave prohibition

In common-practice classical (Baroque through late Romantic), two prohibitions:

- **Parallel perfect fifths**: two voices that are a P5 apart moving to another P5 (e.g., C-G → D-A).
- **Parallel perfect octaves**: two voices that are a P8 apart moving to another P8 (e.g., C-C → D-D).

Why? P5 and P8 are the strongest consonances after the unison — so strong that in parallel they collapse the perception of two voices into one. The whole point of polyphony is voice independence; parallel P5s/P8s undermine that.

This is a context-specific aesthetic, not a universal rule. Where it applies and where it doesn't:

| Context | Status |
|---------|--------|
| Classical vocal counterpoint, chorale, fugue | Strictly forbidden |
| Romantic symphonic writing | Mostly avoided in 4-voice texture; tolerated in doublings (e.g., octaves between violin sections) |
| Impressionist (Debussy, Ravel) | Used deliberately for color |
| Jazz piano voicing | Tolerated and ignored |
| Rock guitar (power chords are parallel 5ths) | The whole genre runs on it |
| EDM / synth lead doubling | Routine |
| Hip-hop / lo-fi sample chops | Whatever the sample does |
| Choral pop arrangement | Avoid for SATB realism |

When advising, identify the genre frame first. Tell a metal songwriter to avoid parallel 5ths and they'll laugh.

## Hidden (or "direct") fifths and octaves

Two voices arriving at a P5 or P8 by similar motion (rather than parallel). Less serious than parallel 5ths/8ves, but in strict counterpoint they're avoided when the outer voices are involved and the soprano leaps. The detection is:

- Both voices move in the same direction
- They arrive at a P5 or P8
- The soprano didn't approach by step

Beginners don't need to memorize this — it's a refinement layer for chorale-style writing. For pop, jazz, and most film, ignore.

## Tendency tones

Some scale degrees have strong directional pull. Voice leading respects these tendencies by default and breaks them deliberately for effect.

| Tendency tone | Wants to go to | Why |
|---------------|----------------|-----|
| Leading tone (^7) in V or vii° | ^1 (up by m2) | Half-step gravity to tonic |
| 7th of dom7 (♭^7 of V7, which is ^4 of the key) | ^3 (down by m2 in major; half-step) | Tritone resolution inward |
| 7th of any 7th chord | down by step (preferably half-step, the 7th wants to "fall") | Continues the falling-7th sequence |
| ♭^6 (in minor or borrowed) | ^5 (down by m2) | Half-step gravity below |
| Chromatic neighbors (e.g., a passing ♯^4) | back to whichever scale degree they came from / are heading to | Local chromatic resolution |

In `V7 → I` (C major):
- B (^7) → C (^1) — leading tone up
- F (^4, the 7th of G7) → E (^3) — 7th down
- G (^5, common tone) — stays
- D (^2, the 9 of G7 if present) → C (^1) or E (^3) — usually steps

If the user writes a `V7 → I` and it sounds wrong, check that ^7 went up and the 7th went down. 95% of "this cadence sounds bad" problems are tendency-tone mismanagement.

## A practical 5-step workflow for voice-leading a progression

Given a chord progression, here's a systematic way to voice it.

1. **Decide the soprano line first.** What's the melody (or the top of the chordal accompaniment) doing? This is the most exposed voice and the one the listener tracks.
2. **Decide the bass line.** Often dictated by the chord progression itself (root motion). Adjust by using inversions when stepwise bass is wanted.
3. **Fill in the inner voices** (alto, tenor) by minimum motion — keep common tones, move other voices by step when possible.
4. **Resolve tendency tones.** Leading tones up, 7ths down, ♭6 down. Override only if you need a specific effect.
5. **Check for parallel 5ths/8ves and bad doublings.** Fix by re-voicing — usually by inverting one chord, or by reassigning a voice to a different chord tone.

### Worked example: G7 → C in C major, four voices

Starting voicing for `G7` (low → high): `G – F – B – D` (bass root, tenor ♭7, alto 3rd, soprano 5th — using one common arrangement). Resolving each voice:

| Voice | G7 | C | Motion |
|-------|----|----|--------|
| Soprano | D (5 of G7) | E (3 of C) | step up |
| Alto | B (3, leading tone) | C (root) | step up — leading tone resolves |
| Tenor | F (♭7) | E (3) | step down — 7th resolves down |
| Bass | G (root) | C (root) | leap up P4 |

Result on `C`: bass `C`, tenor `E`, alto `C`, soprano `E`. The 5th is omitted (a normal outcome when the leading tone and 7th both resolve correctly without doubling). All tendency tones resolved correctly. No parallel 5ths or 8ves.

If you'd prefer a `C` chord with all three triad members present, use a `G7` voicing that doubles the root and omits the 5th: `G – F – B – G` (bass root, tenor ♭7, alto 3rd, soprano root doubled). Then the resolution is:

| Voice | G7 | C |
|-------|----|----|
| Soprano | G | G (common tone, the 5th of C) |
| Alto | B | C |
| Tenor | F | E |
| Bass | G | C |

Result: complete `C` triad with doubled root. This is the textbook resolution shown in most harmony classes — tendency tones resolved, no parallels, every triad member present.

## Cadential, passing, and neighboring 6/4 chords

Second-inversion triads (6/4 chords) are unstable in classical practice. They're used in three specific contexts:

- **Cadential 6/4**: a `I⁶/₄` over a dominant bass, immediately resolving to `V` (or `V7`). Functionally, the chord is *not* tonic; it's a dominant decorated by suspensions. Notation: sometimes shown as `V⁶/⁴–⁵/³` to clarify the 6/4 is a delay of the V.
- **Passing 6/4**: connects two chords via stepwise bass motion. E.g., `I` → `V⁶/₄` → `I⁶`. The bass walks up or down stepwise, and the 6/4 is just a passing harmony.
- **Neighboring 6/4** (sometimes "pedal 6/4"): bass stays, upper voices oscillate. E.g., `I` → `IV⁶/₄` → `I` over a held tonic bass. Embellishment, not real motion.

In pop and rock, slash chords with the 5th in bass are routine and don't need this categorization — the listener doesn't expect classical voice-leading expectations.

## Suspensions, anticipations, and other non-chord tones

Non-chord tones (NCTs) are notes that don't belong to the chord at the moment they sound. They create melodic interest and small dissonances that resolve to consonance.

### Suspensions

A note is held over from the previous chord into the new one, becomes dissonant, then resolves down by step.

Three steps: **preparation** (consonance in the previous chord) — **suspension** (dissonance with the new chord) — **resolution** (down by step to consonance).

Common suspension types (named by the interval above the bass at the suspension and the resolution):

| Type | Example | Notes |
|------|---------|-------|
| 4–3 | Bass C, voice held on F resolves to E | Most common; "suspended 4" |
| 7–6 | Bass C, voice held on B resolves to A | Common in renaissance |
| 9–8 | Bass C, voice held on D resolves to C | Less stable; 9-8 over a root is sometimes weak |
| 2–3 | Suspension in the bass (rare) | Bass D held into chord wanting C; resolves up to E (if upper voices are spelled to make the dissonance an interval of 2 above bass) |

Pop/rock: the `sus4` chord (`Csus4` → `C`) is functionally a 4-3 suspension, and resolving it is the same operation. Pop often *doesn't* resolve `sus4`, treating it as its own color — that's a stylistic choice, not a voice-leading mistake.

### Anticipation

The next chord's note arrives early. Common in pop and gospel:

`G7 → C` where the soprano sings B → (C arrives on the "and" of beat 4 before the actual chord change on beat 1) → C continues.

Anticipations release tension early and create a forward-leaning feel. Almost always at cadences.

### Passing and neighbor tones

- **Passing tone** — fills in stepwise between two chord tones. C → D → E (D is passing).
- **Neighbor tone** — steps to a non-chord tone and returns. C → D → C (D is upper neighbor) or C → B → C (lower neighbor).

These are weak NCTs (barely felt as dissonance). They're the connective tissue of melodic lines.

### Appoggiatura

A leap to a non-chord tone that resolves by step (usually down). The dissonance is *emphasized* — landed on, lingered on, then released. Appoggiaturas are emotional. The opening of "Yesterday" is appoggiatura-heavy.

### Escape tone (échappée)

A stepwise motion away from a chord tone, then a leap in the opposite direction to the next chord tone. C → D → A. The D is the escape tone. Less common than the others; classical filler.

## A 9-item voice-leading checklist for problem progressions

When a user's progression sounds clumsy, walk through:

1. Are tendency tones resolving? (Leading tones up, 7ths down.)
2. Any parallel 5ths or 8ves? (In styles that care.)
3. Any doubled leading tones?
4. Inner voices moving by leap when they could be moving by step?
5. Voice crossing (alto going below tenor, etc.)? Avoid in chorale style; OK in some textures.
6. Voice overlap (one voice moving past where the next-higher voice was)? Same — avoid in chorale, fine in some textures.
7. Bass line leaps too much? Stepwise bass usually feels smoother; consider inversions to walk the bass.
8. Soprano line directionless? Even chordal accompaniments benefit from a top voice that has shape.
9. Common tones not held? Holding common tones between chords reduces unnecessary motion and clarifies the harmony.

## Jazz voice leading

Jazz uses voice leading as the glue between extended chords, but with different priorities:

- **Guide tones** are the most important voices: the 3rd and the 7th of each chord. They define the chord's quality. In a ii–V–I progression, the guide tones move chromatically: in C major, Dm7 has F and C, G7 has F and B, Cmaj7 has E and B. So one guide tone holds (B in G7 → B in Cmaj7), the other moves by half-step (F in Dm7 → F in G7 → E in Cmaj7).
- **Rootless voicings** put guide tones on the bottom and tensions on top. The bass player handles roots.
- **Drop voicings** (drop 2, drop 3) are derived by taking a close-position chord and dropping a voice an octave to add air.
- **Parallel motion is fine.** Parallel 9th chords moving in parallel are a McCoy Tyner specialty.

See `jazz-harmony.md` and `assets/jazz-voicings.md`.

## Pop and rock voice leading

Even more relaxed:

- Power chords are parallel P5s by design.
- Chord shapes (e.g., open chords on guitar) determine voicing more than abstract voice-leading concerns.
- The vocal melody is the soprano and is composed independently of the chord voicings.
- Pad and synth layers often move in parallel for thickening, not voice independence.

Pop voice-leading wisdom is largely about: keep the bass musical; avoid muddy low intervals; keep one or two common tones between chords if possible to reduce "jumpy" feel.

## Modal and contemporary voice leading

In modal music, the leading-tone resolution is gone (no leading tone in most modes), so voice-leading focuses on the characteristic notes of the mode. Parallel motion is more accepted because the music isn't governed by tendency tones.

In post-tonal music (12-tone, atonal), voice leading is about register, contour, and pitch-class proximity, not tonal function. Composers often think in terms of voice-leading distance (how many semitones each voice moves) — a "smooth" voice leading minimizes total motion.

## Cross-references

- Functional context (why tendency tones tend) → `functional-harmony.md`
- Specific chord shapes and their voicings → `chord-construction.md`, `assets/jazz-voicings.md`
- Voicing across instruments → `orchestration/voicing-and-texture.md`
- Chromatic voice-leading patterns (line clichés, descending chromatic basses) → `chromatic-harmony.md`
- Jazz-specific voice leading → `jazz-harmony.md`
- 20th-century voice leading (parsimonious, neo-Riemannian) → `techniques/20th-century-techniques.md`

# Brass

Use this file when the user asks for trumpet, horn, trombone, tuba, brass section writing, fanfares, horn lines, pop horn stabs, mutes, or brass playability.

## Core idea

Brass is powerful, directional, and endurance-limited. It can sound heroic, warm, sacred, funky, brutal, or comic depending on register, articulation, and voicing.

```text
Brass writing = register + breath + endurance + voicing + articulation
```

## Family roles

| Instrument | Strengths | Cautions |
|---|---|---|
| Trumpet | brilliance, lead lines, fanfares, pop stabs | high/loud writing fatigues quickly |
| Horn | warmth, noble midrange, blend with winds/strings | exposed attacks and high writing need care |
| Trombone | power, glissando, strong bass/mid lines, section chords | fast high agility is harder |
| Tuba | low foundation, weight, comic/agile potential | dense low writing can blur |
| Flugelhorn | soft trumpet-family warmth | less cutting than trumpet |

## Register and endurance

Brass can play high and loud, but not forever.

- Save the highest notes for climaxes.
- Give rests after loud high passages.
- Repeated staccato high notes are tiring.
- Low brass needs space for clarity.
- Horns blend well but can be buried if orchestrated too softly against dense textures.

## Voicing brass

### Four-part section

Common orchestral brass spread:

```text
Trumpet / Horn / Trombone / Tuba
Top voice clear, low voice grounded, inner voices not too close in low register.
```

### Pop horn section

Common smaller section:

```text
Trumpet + tenor sax + trombone
or trumpet + alto/tenor sax + trombone
```

Write rhythmic stabs and riffs like a section, not like block keyboard chords. Leave rests between hits.

### Horn pads

Horns are excellent at warm sustained harmony. Keep motion smooth and avoid constant high writing.

## Articulation and effects

| Technique | Use |
|---|---|
| Marcato accents | fanfare, impact, trailer/media hits |
| Legato | warm chorale, noble line |
| Staccato | rhythmic stabs, comedy, lightness |
| Fall / doit | jazz/pop horn idiom |
| Glissando | trombone slide effect, comic or expressive |
| Mutes | color change; trumpet/trombone especially |
| Rip | dramatic ascent into a note |
| Swell | cinematic build, brass pad motion |

## Brass by style

| Style | Brass writing priority |
|---|---|
| Classical/film | chorales, fanfares, counterlines, climactic support |
| Jazz/big band | section voicings, articulations, lead trumpet range, sax/brass balance |
| Funk/soul/pop | short stabs, riffs, call-response, tight articulations |
| Gospel | shout support, hits, organ/choir response |
| Metal/cinematic hybrid | low brass weight, unison hits, octave reinforcement |
| Musical theatre | clear rhythmic punctuation, character-specific color |

## Quick examples

### Simple fanfare

```text
Trumpets/horns: 1 - 5 - 1 octave shape
Trombones/tuba: root/fifth support
Rhythm: dotted or triplet pickup into downbeat.
```

### Pop horn stab

```text
Chord: C9
Voicing: trumpet E, sax B♭/D, trombone G or C
Rhythm: short syncopated hits after vocal line.
```

### Warm horn pad

```text
4 horns in close/mid spacing, soft dynamics.
Move mostly stepwise; let strings carry motion if needed.
```

## Common pitfalls

- **Do not write brass at maximum volume and height for long stretches.** Endurance matters.
- **Do not close-voice low brass too densely.** It muddies.
- **Do not write pop horns as sustained orchestral pads unless that is the intended sound.** Stabs/riffs often work better.
- **Do not forget breathing.** Brass phrases need air.
- **Do not overuse mutes/effects without a clear color reason.**

## Cross-references

- Instrument ranges → `../orchestration/instruments-ranges-character.md`
- Winds contrast and sax writing → `winds.md`
- Voicing and texture → `../orchestration/voicing-and-texture.md`
- Jazz styles → `../genres/jazz-styles.md`
- Gospel and CCM → `../genres/gospel-and-ccm.md`
- Film/TV scoring → `../genres/film-tv-scoring.md`

# Bass

Use this file when the user asks for basslines, electric bass, upright bass, synth bass, walking bass, root movement, low-end arrangement, or bass/drum lock.

## Core idea

Bass defines harmonic ground and physical pulse. A good bassline is not just the chord root; it tells the listener where the body and harmony are going.

```text
Bass part = root logic + rhythm + register + relationship to kick
```

## Range and register

Standard 4-string electric bass:

```text
E1 A1 D2 G2
```

5-string bass often adds low B0. Upright bass is commonly written sounding an octave lower than notated.

| Register | Use | Caution |
|---|---|---|
| B0-E1 | sub weight, modern low-end impact | pitch can become indistinct |
| E1-C2 | main bass foundation | conflicts with kick if too busy |
| C2-G2 | melodic fills, walking clarity | may thin out the low end if overused |
| G2+ | solos, fills, counterline | can fight vocal/guitar/piano midrange |

## Bassline roles

| Role | Description |
|---|---|
| Root anchor | supports harmony clearly; common in pop/rock/country |
| Root-fifth motion | stable folk/country/rock motion |
| Walking bass | stepwise/chromatic connection, jazz/blues/swing |
| Riff bass | memorable repeated figure, funk/rock/metal/synth-pop |
| Counter-melody | independent line against vocal/harmony |
| Sub bass | felt more than heard; EDM/hip-hop/pop |
| Pedal point | one bass note under changing chords |

## Writing by genre

| Style | Bass priority |
|---|---|
| Pop | simple hook-supporting motion, tight with kick |
| Rock | root/5th/octave drive, lock with drums/guitars |
| Punk | relentless 8ths, chord-root clarity, occasional pickups |
| Metal | double guitar riff, kick accents, low pedal precision |
| Country | root/fifth, walkups into chord changes, supportive tone |
| Jazz | walking quarter notes, guide-tone direction, swing feel |
| Funk/R&B | syncopated riff, ghost-note feel, space as groove |
| Gospel | walkups, passing tones, shout/vamp energy |
| EDM/hip-hop | sub rhythm, kick relationship, repetition and sound envelope |

## Connection devices

### Diatonic walk

```text
C → D → E → F under C → F
Effect: smooth lift into IV.
```

### Chromatic approach

```text
C → C# → D
Effect: pushes into Dm or D.
```

### Fifth motion

```text
C → G → C or C → F
Effect: stable, country/folk/rock-friendly.
```

### Enclosure

```text
Approach target D from E and C# before landing.
Effect: jazz/gospel pull.
```

### Pedal

```text
Bass holds C while chords change above: C - F/C - G/C - C
Effect: grounded, cinematic, worship/pop build.
```

## Kick relationship

Common choices:

- **Unison with kick**: tight, modern, punchy.
- **Bass fills kick gaps**: groovier, funk/R&B feel.
- **Bass long notes under busy kick**: stable but can feel static.
- **Kick marks downbeats, bass syncopates**: danceable push.

If the track feels muddy, simplify bass rhythm before changing the chords.

## Quick examples

### Country walkup

```text
Key: G, progression G → C
Bass: G - A - B - C
Use: line into IV.
```

### Pop chorus bass

```text
Progression: C - G - Am - F
Bass: C C G G | G G D G | A A E A | F F C F
Effect: clear roots with mild motion.
```

### Funk/R&B seed

```text
Center: E minor
E - rest - E G - A - G - E
Leave space; let drums answer.
```

### Metal pedal

```text
E pedal eighths, accents with guitar chugs.
Only leave pedal for section changes or fills.
```

## Common pitfalls

- **Do not make the bass constantly melodic under a busy vocal.** Support first.
- **Do not ignore the kick drum.** Bass rhythm without kick context is incomplete.
- **Do not put every chord tone in the bassline.** Root direction and groove matter more.
- **Do not overuse the lowest notes.** Very low notes lose pitch clarity if not arranged carefully.
- **Do not write upright bass like electric bass without considering sustain and attack.**

## Cross-references

- Groove and feel → `../rhythm-groove/groove-and-feel.md`
- Drums/percussion → `drums-percussion.md`
- Country bass → `../genres/country-americana.md`
- Metal/punk bass → `../genres/metal-punk-hardcore.md`
- Gospel bass → `../genres/gospel-and-ccm.md`
- Production-aware low end → `../production-aware/arrangement-for-mix.md`

# Instrument Idiom Overview

This folder gives composition-facing guidance for writing parts that feel natural on real instruments. It complements `../orchestration/instruments-ranges-character.md`, which gives range and character summaries. Use these files when the user asks for a part, riff, accompaniment, section writing, or playability check.

## What "idiomatic" means

An idiomatic part respects the way an instrument naturally produces sound:

- physical layout: strings, keys, valves, sticks, breath, hand span
- natural resonances: open strings, registers, pedal tones, vowels
- gesture: strum, bow, tongue, roll, bend, comp, walk, breathe
- fatigue: brass endurance, vocal tessitura, repeated leaps, hand stretches
- ensemble role: bass foundation, drum timekeeping, guitar comping, string pad, vocal lead

Idiomatic does **not** mean predictable. It means the part gives the performer a believable path to the sound.

## When to consult instrument-idiom files

| User asks | Start here |
|---|---|
| Piano voicing, left-hand pattern, keyboard accompaniment | `piano-keyboards.md` |
| Guitar chords, riffs, strumming, capo, tuning, playable voicing | `guitar.md` |
| Bassline, walking bass, synth/electric bass role | `bass.md` |
| Drum groove, fill, kit writing, percussion layer | `drums-percussion.md` |
| String quartet/section writing, bowing, divisi, double stops | `strings.md` |
| Flute/clarinet/oboe/bassoon/sax line or section | `winds.md` |
| Brass fanfare, horn line, trumpet/trombone/tuba section | `brass.md` |
| Solo vocal, backing vocals, choir-adjacent pop writing | `vocals.md` |

## Fast idiom diagnostic

When a part sounds fake or awkward, check these in order:

1. **Range** — is it possible?
2. **Tessitura** — is it comfortable for more than a moment?
3. **Gesture** — does the pattern match how the instrument moves?
4. **Breath / decay** — does the phrase allow sound to be renewed?
5. **Register color** — is the chosen register giving the intended timbre?
6. **Density** — is the part competing with too many other parts?
7. **Notation / communication** — could a player understand the intent quickly?

## Idiom vs. orchestration

- **Orchestration** asks: which instruments should play this musical idea?
- **Instrument idiom** asks: once chosen, how should the idea be shaped so the instrument speaks naturally?

Example:

```text
Idea: fast repeated notes
Orchestration answer: strings, flute, marimba, guitar tremolo, repeated synth, snare roll
Idiom answer: each instrument repeats differently; do not copy-paste the same MIDI pattern.
```

## Real performers vs. sample libraries

A sample library can make impossible writing sound plausible for a few seconds, but that does not make it performer-friendly. Conversely, real performers can solve awkward writing if the musical goal is clear.

Use this rule:

```text
If the user is writing for a real session, prioritize playability and notation.
If the user is producing in a DAW, prioritize believable gesture and sonic clarity.
If unknown, give both versions.
```

## Common part-writing modes

| Mode | What matters most |
|---|---|
| Lead line | range, breath/bow/picking, phrase shape, register color |
| Accompaniment | rhythmic pocket, voicing, density, non-conflict with vocal |
| Riff | physical repeatability, accent pattern, timbre, memorability |
| Pad / sustain | register spacing, re-articulation, blend, dynamic contour |
| Counterline | independence, register separation, answer timing |
| Section writing | voicing, balance, fatigue, blend, orchestration role |

## Common pitfalls

- **Do not write only from piano-roll logic.** Instruments have bodies.
- **Do not equate possible with good.** A note may exist but sound weak or strained.
- **Do not forget recovery.** Brass, winds, singers, and drummers need endurance planning.
- **Do not use the same voicing across instruments.** A piano chord, guitar chord, string voicing, and horn voicing behave differently.
- **Do not over-specify technique when a musical description is enough.** "Warm, connected, not too loud" can be more useful than too many markings.

## Cross-references

- Instrument ranges and characters → `../orchestration/instruments-ranges-character.md`
- Voicing and texture → `../orchestration/voicing-and-texture.md`
- Arrangement density → `../orchestration/arrangement-density.md`
- Notation and conventions → `../fundamentals/notation-and-conventions.md`
- Rhythm and groove → `../rhythm-groove/groove-and-feel.md`
- Production-aware arrangement → `../production-aware/arrangement-for-mix.md`

# Strings

Use this file when the user asks for violin, viola, cello, double bass, string quartet, string section, bowed texture, pizzicato, divisi, or string playability.

## Core idea

Strings are built for sustain, phrasing, resonance, and register color. They can also be percussive, but their strongest idiom is shaped motion over time.

```text
String writing = register + bowing + fingering/open strings + texture + divisi plan
```

## Family reference

| Instrument | Lowest open string | Common role |
|---|---|---|
| Violin | G3 | high melody, agile line, shimmer |
| Viola | C3 | inner warmth, dark melody, texture glue |
| Cello | C2 | lyrical tenor/bass, rich melody, foundation |
| Double bass | E1 sounding | low foundation, weight, pizz/arco bass |

## Register color

| Register | Effect |
|---|---|
| Low strings | weight, darkness, intimacy, grit |
| Middle register | warmth and blend |
| High register | intensity, shine, tension, fragility |
| Open strings | resonance and folk/classical clarity |
| Sul ponticello / near bridge | glassy, tense color |
| Sul tasto / over fingerboard | soft, veiled color |

## Bowing and articulation

| Technique | Composition use |
|---|---|
| Legato | long lyrical lines, pads |
| Detaché | clear separated notes without harshness |
| Staccato / spiccato | light rhythmic drive |
| Tremolo | suspense, shimmer, crescendo support |
| Pizzicato | short plucked rhythm; not true sustain |
| Marcato | strong accents, cinematic rhythm |
| Glissando | expressive slide, transition, effect |
| Harmonics | fragile high color |
| Double stop | two-note sonority; check playability |

## Section writing

### Pad / sustain

- spread voices across the section
- avoid low close voicings
- move inner voices slowly
- use dynamics and bow changes to create life

### Rhythmic ostinato

- keep bowing comfortable
- use repeated patterns with slight contour
- leave room for accents and phrase breaths
- consider spiccato or tremolo depending on tempo

### Lyrical melody

- violin/cello are natural melodic carriers
- viola can be beautiful for darker, intimate melodies
- mark phrase shape more than every tiny nuance
- allow bow direction changes at musical phrase points

### Divisi

Divisi splits section players among multiple notes. It makes richer chords possible but reduces strength per note.

```text
Violins divisi into 3 notes = softer per note than all violins on one line.
```

Use divisi for lush pads; avoid it when you need maximum unison power.

## Double-stop caution

Double stops are idiomatic only when the interval and fingering work on adjacent strings. Good starting choices:

- open-string drones
- 3rds/6ths in moderate tempo
- perfect 5ths across adjacent strings
- simple two-note folk/country/fiddle shapes

Avoid fast chromatic multi-stop writing unless you know the instrument well.

## Pizzicato writing

- Pizzicato decays quickly; it does not sustain like harp or guitar.
- Fast pizzicato can be awkward depending on tempo and string crossing.
- Low pizzicato bass/cello is punchy but needs space.
- Bartók pizzicato is a special snap effect; use sparingly.

## String textures

| Texture | Use |
|---|---|
| Unison line | power, clarity, cinematic sweep |
| Octaves | broad romantic/film sound |
| Divisi pad | lush harmony |
| Tremolo chord | suspense/build |
| Pizzicato ostinato | light pulse, comedy, chamber pop |
| Counterline | emotional answer to vocal |
| Drone | folk, modal, ancient/ritual association when used specifically |

## Quick examples

### Warm pop string pad

```text
Cello: root + fifth, long notes
Viola: 3rd or 7th in middle register
Violins: 9th/top melody note, soft legato
Move one voice at a time.
```

### Cinematic lift

```text
Low strings: ostinato in 8ths
Violins: long rising line
Violas/celli: inner counter-motion
Add tremolo only near the climax.
```

### Country/folk fiddle response

```text
Short 2-bar answer after vocal line.
Use open strings, slides, and double-stop hints.
Do not overfill the vocal.
```

## Common pitfalls

- **Do not write every string pad as a keyboard chord.** Voice-leading and bowing matter.
- **Do not overuse divisi if you need power.** Divisi thins each note.
- **Do not expect pizzicato to sustain.** It decays.
- **Do not write impossible double stops casually.** Check strings/fingering.
- **Do not keep strings in one dynamic forever.** Bowed sustain needs shaping.

## Cross-references

- Instrument ranges → `../orchestration/instruments-ranges-character.md`
- Voicing and texture → `../orchestration/voicing-and-texture.md`
- Arrangement density → `../orchestration/arrangement-density.md`
- Notation and articulation → `../fundamentals/notation-and-conventions.md`
- Film/TV scoring → `../genres/film-tv-scoring.md`
- Country/Americana fiddle contexts → `../genres/country-americana.md`

# Guitar

Use this file when the user asks for guitar chords, riffs, voicings, strumming, fingerpicking, capo use, tuning, distorted guitar, acoustic accompaniment, or guitar playability.

## Core idea

Guitar is shape-based. A progression that looks simple on piano may be awkward on guitar, and a simple guitar shape can sound huge because of open strings, resonance, and strumming motion.

```text
Guitar part = tuning + position + string set + right-hand pattern
```

## Standard tuning reference

```text
E2 A2 D3 G3 B3 E4
```

Open strings matter. Keys like E, A, D, G, C, and their relative minors often feel natural because open strings support common voicings. Capos let singer-friendly keys keep guitar-friendly shapes.

## Chord-shape thinking

| Need | Guitar-friendly move |
|---|---|
| Bright pop openness | add9/sus shapes, open strings, capo |
| Rock weight | power chords, octaves, palm-muted roots |
| Folk movement | alternating bass, hammer-ons/pull-offs inside chord shapes |
| Country sparkle | double-stops, open strings, hybrid picking |
| R&B/gospel comp | 3-note shell voicings, 6ths, slides, syncopation |
| Metal precision | low-string pedal, palm muting, drop tuning, tight rhythmic cells |

## Acoustic guitar patterns

### Basic strum

```text
Down on strong beats, up on offbeats.
Good for pop, folk, country, worship.
```

### Muted sixteenth strum

```text
Right hand keeps continuous motion; left hand controls chord/stops.
Good for modern pop, funk-pop, country-pop.
```

### Alternating bass

```text
Bass note alternates root/5th or root/3rd while upper strings strum.
Good for folk, country, singer-songwriter.
```

### Fingerpicking

```text
Thumb: bass strings
Fingers: inner/top strings
Good for intimate ballads and arpeggiated patterns.
```

## Electric guitar patterns

### Power chord riff

```text
Root + fifth, sometimes octave.
Use distortion-friendly voicing; avoid full low triads.
```

### Octave line

```text
Root + octave with muted middle string.
Cuts through rock/pop without adding chord density.
```

### Double-stops

```text
Two-note shapes, often 3rds/6ths or bluesy bends.
Good for country, soul, rock fills.
```

### Palm-muted pulse

```text
Right hand mutes near bridge; pitch stays tight and percussive.
Good for punk, metal, pop-rock drive.
```

### Swells / ambient guitar

```text
Volume pedal or knob + delay/reverb.
Good for worship, film, post-rock, pads.
```

## Capo guidance

Use a capo when:

- the singer needs a different key but the guitar part depends on open shapes
- the song needs brighter voicings
- the arrangement wants multiple guitars in different registers
- a simple part needs shimmer without new harmony

Example:

```text
Sounding key: B♭
Capo: 3
Played shapes: G-D-Em-C
Effect: B♭ major pop progression with open-string guitar resonance.
```

## Tuning guidance

| Tuning | Use |
|---|---|
| Standard | general pop/rock/jazz/country |
| Drop D | low D power, folk/rock drones, easier power chords |
| DADGAD | modal folk, drones, suspended color |
| Open G / Open D | slide, blues, roots, resonant major shapes |
| Down-tuned standard | metal/doom weight, lower vocal keys |
| Extended-range | modern metal/djent low register |

Do not transpose guitar riffs casually without checking tuning. A riff may depend on an open string or one-finger shape.

## Voicing tips

- Low full triads under distortion get muddy; use power chords or spread voicings.
- Open-string add9/sus voicings sound modern and guitar-like.
- Jazz chords on guitar often omit roots or fifths.
- A single high-string counterline can be more effective than a new chord.
- For two guitars, separate roles: one rhythm bed, one hook/counterline, or different capo positions.

## Quick examples

### Open pop progression

```text
Capo 2, shapes: G - D/F# - Em7 - Cadd9
Sounding: A - E/G# - F#m7 - Dadd9
Use: pop, worship, country-pop.
```

### Punk / rock power loop

```text
E5 - A5 - B5 - A5
All downstrokes or tight alternate picking.
```

### Country double-stop fill

```text
Over G: play B-D → C-E → B-D on top strings, with a small slide into C-E.
Use between vocal phrases, not over every lyric.
```

## Common pitfalls

- **Do not write six-note piano chords for distorted guitar.** Distortion wants simpler voicings.
- **Do not ignore open strings.** They are part of the instrument's sound.
- **Do not assume every chord inversion is easy.** Guitar voicings are shape-dependent.
- **Do not put busy guitar fills under every vocal line.** Answer the singer.
- **Do not forget right-hand rhythm.** Strum/pick pattern can define the style more than the chord names.

## Cross-references

- Metal/punk/hardcore guitar → `../genres/metal-punk-hardcore.md`
- Country guitar → `../genres/country-americana.md`
- Brazilian guitar contexts → `../genres/brazilian-pop-and-funk.md`
- Voicing and texture → `../orchestration/voicing-and-texture.md`
- Chord symbols → `../../assets/chord-symbol-conventions.md`

# Drums and Percussion

Use this file when the user asks for drum grooves, fills, kit writing, percussion layers, groove energy, playable drum parts, or rhythm-section arrangement.

## Core idea

Drums organize time, energy, and section identity. A drum part is not just a beat; it tells the listener how to move and when the form changes.

```text
Drum part = pulse + backbeat/anchor + subdivision + fills + density arc
```

## Drum kit roles

| Element | Typical role |
|---|---|
| Kick | low pulse, dance/body accent, bass relationship |
| Snare | backbeat, impact, phrase punctuation |
| Hi-hat | subdivision, groove texture, openness/closure |
| Ride | sustained time feel, jazz/rock chorus lift |
| Toms | fills, tribal/arena energy, section contrast |
| Crash | section marker, chorus entry, climax |
| Ghost notes | groove detail, funk/R&B/soul feel |

## Groove building blocks

### Backbeat

```text
Snare on 2 and 4.
Use: pop, rock, country, R&B, many dance styles.
```

### Four-on-floor

```text
Kick on every quarter note.
Use: dance-pop, EDM, disco, worship/arena builds.
```

### Half-time

```text
Snare on 3 instead of 2 and 4.
Use: heavy chorus, breakdown, trap/R&B weight.
```

### Double-time feel

```text
Perceived subdivision speeds up while tempo stays same.
Use: punk/hardcore lift, chorus acceleration.
```

### Shuffle / swing

```text
Triplet-based subdivision.
Use: blues, country, roots gospel, jazz, some rock.
```

### Train beat

```text
Fast snare/brush pattern suggesting locomotive motion.
Use: country, folk-rock, Americana.
```

## Fill design

A fill should answer a formal question:

| Need | Fill type |
|---|---|
| mark section change | short tom/snare fill into downbeat |
| create suspense | rest before downbeat or rising fill |
| intensify chorus | busier fill + crash entry |
| cue breakdown | stop-time hit, silence, or single snare flam |
| avoid clutter | tiny pickup fill or no fill |

Good fills often start simple and become more active only where the form needs it.

## Density arc

A common arrangement arc:

```text
Verse: kick + hat, light snare
Pre: add ghost notes or toms
Chorus: full snare/crash/ride or open hats
Bridge: strip down or change surface
Final chorus: full kit + extra percussion
```

Do not use full cymbal wash everywhere. Cymbals can hide vocals, guitars, and consonants.

## Percussion layers

| Layer | Use |
|---|---|
| Shaker | continuous subdivision, brightness |
| Tambourine | chorus lift, gospel/pop brightness |
| Claps | crowd/community feel, pop backbeat |
| Conga/bongo | hand-percussion groove, Latin/pop contexts |
| Pandeiro/tamborim/surdo-like roles | Brazilian-informed contexts; use with specificity |
| Timpani / orchestral bass drum | cinematic impact |
| Snare roll / cymbal swell | transition build |
| Taiko-like ensemble drums | epic media scoring; avoid vague cultural shorthand |

Name the rhythmic function if you are not sure of the specific tradition. For serious tradition-based percussion, consult practitioners and current references.

## Human playability

A drummer has four limbs, but coordination is style-specific. MIDI can create patterns that no drummer would choose.

Check:

- Are kick, snare, hat/ride, and toms playable by two hands/two feet?
- Does the drummer have time to move from hat to tom to crash?
- Are ghost notes realistic at the tempo?
- Does the part require constant maximum velocity?
- Are fills placed where a player would naturally breathe with the band?

## Groove by style

| Style | Starting drum logic |
|---|---|
| Pop | clear kick/snare pattern, controlled hats, section lifts |
| Rock | strong backbeat, fills into sections, crash support |
| Punk | fast hats/ride, direct snare, few ornamental details |
| Metal | kick precision, cymbal control, riff accents, breakdown contrast |
| Country | train beat, shuffle, simple backbeat, supportive fills |
| Gospel | responsive fills, shout energy, 6/8 or shuffle options |
| R&B/funk | ghost notes, syncopated kick, pocket over density |
| Brazilian pop/funk | groove-specific percussion loop; verify substyle |
| Film/game | dynamic cueing, orchestral percussion, hybrid hits |

## Quick examples

### Pop verse to chorus

```text
Verse: kick on 1/3, snare on 2/4, closed hats 8ths
Pre: add 16th hat lift
Chorus: open hats or ride, crash on downbeat, stronger kick pattern
```

### Metal breakdown

```text
Half-time snare on 3.
Kick follows guitar chugs.
Use rests before major hits.
```

### Country train beat

```text
Snare/brush subdivision suggests forward motion.
Kick supports 1 and 3 or root/fifth bass motion.
```

## Common pitfalls

- **Do not write drums as constant fills.** Groove first, fills second.
- **Do not make every chorus louder by adding cymbals only.** Change rhythm, register, or arrangement too.
- **Do not ignore the bassline.** Kick and bass are a pair.
- **Do not use culturally specific percussion as wallpaper.** Know the function or keep it generic.
- **Do not overhumanize by randomizing everything.** Groove has intentional consistency.

## Cross-references

- Groove and feel → `../rhythm-groove/groove-and-feel.md`
- Rhythmic devices → `../rhythm-groove/rhythmic-devices.md`
- Bass relationship → `bass.md`
- Metal/punk/hardcore drums → `../genres/metal-punk-hardcore.md`
- Brazilian pop/funk percussion → `../genres/brazilian-pop-and-funk.md`
- Arrangement density → `../orchestration/arrangement-density.md`

# Piano and Keyboards

Use this file when the user asks for piano voicings, keyboard accompaniment, left-hand patterns, gospel/jazz comping, synth/keys arrangement roles, or piano playability.

## Core idea

Piano can cover melody, harmony, rhythm, and bass at once, but idiomatic writing still needs hand spacing, register control, and pedal/decay awareness.

```text
Piano part = register + hand shape + rhythm + sustain plan
```

## Playability basics

- Standard piano range: A0-C8.
- Comfortable hand span varies; write octaves as normal, tenths as optional unless the player is advanced.
- Fast repeated wide leaps are harder than they look in MIDI.
- Low closed chords get muddy; use open fifths, octaves, or tenths in the bass.
- Pedal can connect harmony, but too much pedal blurs rhythm and chromatic motion.

## Register guide

| Register | Best use | Caution |
|---|---|---|
| A0-C2 | power, sub-like bass, dramatic hits | dense chords become mud |
| C2-C4 | left-hand bass, stride, arpeggios, comp foundation | avoid constant conflict with bass guitar |
| C4-C6 | melody, chord voicing, accompaniment | crowded vocal range if too busy |
| C6-C8 | sparkle, filigree, doubled melody | can become brittle or piercing |

## Common accompaniment patterns

### Block-chord comp

```text
RH: 3- or 4-note chord stabs
LH: bass note or octave
Use: pop, gospel, theatre, rock ballad
```

### Broken-chord arpeggio

```text
LH: root-5th-10th shape
RH: chord tones or melody support
Use: ballad, cinematic, singer-songwriter
```

### Alberti-like pattern

```text
low-high-middle-high
Use: classical/light pop reference, gentle motion
```

### Stride / country-gospel left hand

```text
LH: bass note on beat 1/3, chord or octave answer on 2/4
Use: old-time, gospel, ragtime-adjacent, roots color
```

### Gospel / R&B comp

```text
LH: bass walk or rootless support
RH: 3rds/7ths + extensions, rhythmic fills between vocals
Use: gospel, R&B, neo-soul
```

### Synth/key pad

```text
Sustained voicing, slow filter/dynamic motion, minimal rhythm
Use: pop, worship, ambient, film/game underscore
```

## Voicing rules of thumb

- Put 3rd and 7th in the middle register for harmonic clarity.
- Put melody/top note in a singable register if the piano is carrying the hook.
- Spread low voicings; close high voicings can sparkle.
- For solo piano, include bass, inner rhythm, and melody; for ensemble piano, leave space.
- In dense band arrangements, two- or three-note voicings often beat full chords.

## Genre-specific moves

| Style | Useful piano/keys move |
|---|---|
| Pop ballad | broken octaves, soft RH triads/add9, pedal-controlled sustain |
| Gospel | passing diminished chords, octave melody, response fills |
| Jazz | shell voicings, rootless 3-7-9-13 shapes, comp syncopation |
| Country | simple triads, walkups, honky-tonk octaves, piano/organ support |
| Rock | octave riffs, repeated chord stabs, organ pads |
| EDM/pop | sidechain-friendly pads, plucks, simple voicings, hook ostinato |
| Film/game | ostinato, register layering, pedal point, sparse motif |

## Quick examples

### Pop ballad in C

```text
LH: C2-G2-E3-G2 | A1-E2-C3-E2 | F1-C2-A2-C2 | G1-D2-B2-D2
RH: melody or soft triads above C4
Effect: emotional but playable.
```

### Gospel tag in C

```text
| C/E | F | F#°7 | Gsus G |
RH: voice-lead E-A-C → F-A-C → F#-A-C-E♭ → G-C-D-F / G-B-D-F
```

### Sparse pop comp

```text
LH: root only on bar 1
RH: 3rd + 7th + 9th on offbeats
Leave space for vocal.
```

## Common pitfalls

- **Do not stack full triads below C3.** Spread them.
- **Do not write nonstop RH fills under a vocal.** Fill between phrases.
- **Do not assume two hands can cover every MIDI layer.** Divide into piano + pad if needed.
- **Do not use sustain pedal to hide unclear harmony.** Pedal is color, not glue for bad writing.
- **Do not make left hand and bass guitar fight.** In a band, piano LH can thin out.

## Cross-references

- Chord voicings → `../harmony/voice-leading.md`, `../harmony/jazz-harmony.md`
- Gospel and CCM → `../genres/gospel-and-ccm.md`
- Country and Americana → `../genres/country-americana.md`
- Arrangement density → `../orchestration/arrangement-density.md`
- Production-aware frequency spacing → `../production-aware/arrangement-for-mix.md`

# Vocals

Use this file when the user asks for vocal melody, singer range, tessitura, backing vocals, ad-libs, choir-adjacent pop writing, belting, rap/sing-rap, or lyric-melody fit.

## Core idea

The voice is an instrument with words attached. A vocal line must fit range, tessitura, vowel, breath, language, and emotional delivery.

```text
Vocal writing = range + tessitura + vowel + breath + text stress + persona
```

## Range vs. tessitura

- **Range**: the total notes a singer can produce.
- **Tessitura**: where the line sits most of the time.

A melody can be technically within range but still exhausting if it sits too high, too loud, or on hard vowels.

## Practical range guidance

Use these as starting points only. Individual singers vary a lot.

| Voice context | Comfortable writing idea |
|---|---|
| Pop lead | tailor to singer; choruses often sit higher than verses |
| Congregational/worship | keep range narrower and avoid sustained extremes |
| Musical theatre | range and belt/mix choices are style-specific |
| Rock/metal | high belt or harsh vocals require singer-specific knowledge |
| R&B/gospel | allow melisma/ad-lib space; do not overfill the written melody |
| Rap/sing-rap | rhythm, consonants, breath, and vowel placement are central |
| Choir/SATB | use choral ranges and balance; see `../orchestration/choral-writing.md` |

## Breath and phrase length

A singer needs breath before long phrases and after dense consonants.

Good defaults:

- 2-bar phrases are usually safe.
- 4-bar phrases need planned breathing or simpler rhythm.
- Fast lyrics need shorter phrases or rests.
- Sustained high notes need setup and recovery.
- Ad-libs should not cover the main lyric unless intentionally chaotic.

## Vowels and high notes

High notes are easier on some vowels than others, and vowel modification is common in singing.

Composition-facing guidance:

- Place big high notes on open or singable vowels when possible.
- Avoid cramming many consonants at the top of the range.
- If the title word has a difficult vowel, consider placing the emotional high point before or after it.
- For non-English lyrics, check vowel and stress with a native speaker when possible.

## Melody and lyric stress

A vocal melody should respect natural word stress unless the distortion is intentional.

Check:

- important syllables land on strong beats or longer notes
- weak syllables are not consistently over-accented
- melisma does not obscure key words
- repeated hook text still sounds natural
- consonant clusters have enough time

Pair with `../fundamentals/prosody-and-language.md`.

## Lead vocal writing by section

| Section | Vocal strategy |
|---|---|
| Verse | conversational range, lyric detail, rhythmic identity |
| Pre-chorus | rising energy, more repetition, harmonic tension |
| Chorus | title/hook clarity, wider range or stronger rhythm |
| Post-chorus | chant, vowel hook, repeated tag, simpler text |
| Bridge | new viewpoint, register shift, stripped rhythm, or climax |
| Final chorus | ad-libs, octave doubles, harmony stacks, tag repetition |

## Backing vocals

### Common backing-vocal roles

| Role | Use |
|---|---|
| Unison double | strengthen hook without changing harmony |
| Octave double | widen chorus, gender/texture contrast |
| 3rds/6ths harmony | pop/country/folk sweetness |
| Block harmony | gospel, musical theatre, pop stack |
| Call-response | R&B, gospel, country, funk, pop |
| Pad vowels | cinematic/pop support |
| Gang vocals | punk, pop-punk, stadium chorus, group identity |
| Ad-libs | final chorus lift, R&B/gospel release |

### Harmony stack strategy

- Start with the melody.
- Add one harmony below or above for clarity.
- Add more notes only where the lyric repeats or the chorus needs width.
- Keep consonants together if text clarity matters.
- Thin backing vocals during verse if the lead lyric is detailed.

## Rap and rhythmic vocal writing

Rap/topline rhythm is composition.

Check:

- breath points
- internal rhyme placement
- bar-line tension/release
- consonant density
- relationship to kick/snare
- when the flow changes across sections
- whether the hook simplifies after dense verse flow

For sing-rap, pitch may be minimal but contour still matters.

## Extreme or high-intensity vocals

For screams, growls, belting, and very high rock/metal vocals:

- write alternatives if the singer is unknown
- avoid long strings of maximum intensity
- plan rests after high or harsh phrases
- do not prescribe unsafe technique
- write the emotional contour, not just "scream here"

## Quick examples

### Pop chorus lift

```text
Verse range: A3-E4
Chorus range: C4-G4
Hook title lands on G4 once, then resolves downward.
Effect: lift without exhausting the singer.
```

### Backing vocal stack

```text
Lead: title phrase
BGV 1: third below on final two words
BGV 2: octave above only on repeat
Ad-lib: after the title, not over it.
```

### Congregational melody

```text
Range: about an octave
Rhythm: mostly quarter/eighth notes
Hook: repeated phrase with clear vowels
Avoid: long high melisma or unpredictable syncopation.
```

## Common pitfalls

- **Do not write for an abstract voice type when the song has a real singer.** Tailor to the person.
- **Do not confuse range with comfort.** Tessitura is the daily problem.
- **Do not put the hardest lyric syllable on the highest sustained note by accident.**
- **Do not let backing vocals cover the lead lyric.** Support the hook.
- **Do not prescribe extreme vocal techniques casually.** Keep it safe and singer-specific.
- **Do not ignore language.** Prosody errors make even good melodies feel wrong.

## Cross-references

- Prosody and language → `../fundamentals/prosody-and-language.md`
- Choral writing → `../orchestration/choral-writing.md`
- Topline craft → `../songwriting/topline-craft.md`
- Hooks and memorability → `../songwriting/hooks-and-memorability.md`
- Gospel and CCM vocals → `../genres/gospel-and-ccm.md`
- Metal/punk/hardcore vocals → `../genres/metal-punk-hardcore.md`

# Melodic Construction

A melody is a sequence of pitches in time perceived as a single coherent line. "Coherent" is doing a lot of work in that sentence — it's the engineering problem of melodic writing. This file covers the parameters you adjust to make a line cohere, breathe, and stay memorable.

## When to consult this file

- The user asks how to write a melody, fix a melody, or evaluate one
- Range, contour, climax, melodic peaks
- Why a melody "feels right" or "feels wrong"
- The interaction between melody and harmony at the line-by-line level

For motif development across a piece, see `motivic-development.md`. For phrase grammar (period, sentence), see `phrase-structure.md`. For lyric–melody interaction (prosody), see `songwriting/lyric-writing.md` and `songwriting/topline-craft.md`. For melodic conventions in jazz/improvisation, see `harmony/jazz-harmony.md` and `genres/jazz-styles.md`.

## The parameters of a melody

A melody is the joint trajectory of these dimensions:

1. **Pitch content** — which scale degrees are used, how chromatic, how often the line touches consonance vs. tension.
2. **Contour** — the shape over time: up, down, undulating, arched, terraced.
3. **Range** — the distance from lowest to highest note. Affects perceived size and which voice/instrument can sing it.
4. **Tessitura** — which part of the range is used most. A melody can have wide range but live mostly in one zone.
5. **Interval profile** — proportion of steps vs. leaps; the size and direction of the leaps; whether leaps are "filled in" by following stepwise motion in the opposite direction.
6. **Rhythm** — durations, density, syncopation, repetition. Rhythm is half of melody. Probably more.
7. **Climax** — where the highest pitch (or longest note, or loudest moment) lands relative to the line's length.
8. **Relationship to harmony** — which notes are chord tones, which are non-chord tones (passing, neighbor, suspension, anticipation, appoggiatura, escape tone).

When a melody isn't working, it's almost always one or two of these dimensions misbehaving.

## Contour shapes (the macro shape)

Most strong melodies fall into one of a few archetypal contour shapes. None is "best" — they suit different feelings.

- **Arch**: rises to a peak somewhere past the middle, then falls. The most common shape in folk and classical. Feels balanced and sentence-like.
- **Inverted arch**: dips to a low point, returns. Less common. Feels reflective or recovering.
- **Ascending**: climbs throughout, ending high. Feels open, aspirational, unresolved (especially if it ends on ^2, ^4, or ^7).
- **Descending**: falls throughout. Feels conclusive, melancholy, or resigned.
- **Undulating / wave**: oscillates across a roughly horizontal axis. Common in chant, rap, and conversational melodies.
- **Terraced / step-up**: each phrase is a register higher than the last, like a staircase. Builds energy across multiple phrases. Common in pop pre-choruses and gospel build-ups.
- **Pendulum**: alternates between two notes or two zones. Hypnotic, often used for tension or trance.

A common stronger structure is to nest one shape inside another — a wave shape on the small scale, riding an arch on the large scale.

## Range and tessitura

- **Speech range** is roughly an octave. Anything within the speech range feels conversational — natural for verses, raps, narrative ballads.
- **Singable range** for an untrained voice: about 1.5 octaves. Skilled singers comfortable with 2 octaves; pros can push to 2.5 or 3.
- **Tessitura matters more than range.** A melody with a 2-octave range that lives in the middle 5th feels modest. A melody with a 5th of range that sits at the top of the singer's range feels intense.
- **Chorus tessitura is usually higher than verse tessitura.** This is a primary mechanism by which choruses "lift" — see `form/narrative-and-transitions.md`.
- **For instrumental melodies, the same idea applies.** Using the comfortable middle of an instrument feels easy; venturing into the extremes makes the line feel strained or virtuosic. See `orchestration/instruments-ranges-character.md`.

### Vocal range zones

Most vocal melodies sit in one of three zones:

| Zone | Position in voice | Effect |
|------|-------------------|--------|
| Low chest | Bottom 4th of singer's range | Intimate, conversational, sometimes "spoken" feel |
| Mid chest / mix | Middle of range | Comfortable, sustainable, default |
| Belt / head voice | Upper 4th to top | Intense, emotional, can't sustain forever |

A song that lives entirely in mid chest feels even-keeled. A song that builds from low chest in the verse to belt in the chorus has the most natural "lift" — and is the structural default of most pop ballads.

## Steps vs. leaps

The basic guideline: stepwise motion is the connective tissue, leaps are the punctuation.

- **Mostly stepwise melodies** (folk, chant, lullabies) feel smooth, vocal, predictable.
- **Leap-heavy melodies** (some jazz heads, fanfares, "Somewhere Over the Rainbow") feel athletic and dramatic. They demand more from the singer/player.
- **A leap larger than a 5th creates expectation.** The classical convention is to fill it in: leap up, then walk back down stepwise (or leap down, walk back up). This isn't a rule but a strong tendency that the ear notices when broken.
- **Multiple consecutive leaps in the same direction** (e.g., outlining a triad: C–E–G) feel arpeggiated and melodically thin. Useful when you want the line to feel like a chord, not a melody.
- **A leap to a non-chord tone is usually disruptive.** Leaps tend to land on chord tones; the surprise of leaping to an extension (a 9th, 11th, 13th) is striking and useful in jazz, neo-soul, and some film writing.

## The climax

Most strong melodies have a single highest moment — the climax. It's usually:

- **Past the midpoint.** Around the 60–75% mark of the line is a common location. (The "golden ratio" gets cited; it's a reasonable rule of thumb.)
- **The highest pitch in the line.** Or the longest. Or the most rhythmically charged.
- **Approached by stepwise ascent or a single dramatic leap.** The approach is what makes the climax feel earned.
- **Followed by descent.** The release. Without descent the climax doesn't feel like a peak — it feels like a plateau.

If a melody feels flat, check whether there's a clear climax. If it feels exhausting, check whether the climax comes too early or repeats too often.

A song's overall climax (the highest note in the entire piece) is usually saved for the final chorus or the bridge's payoff. Spending the highest note early leaves nowhere to go.

## Relationship to harmony — chord tones and non-chord tones

Every note in a melody is either a chord tone (at the moment it sounds, given the underlying harmony) or a non-chord tone. Chord tones are stable; non-chord tones are colored or tense. The richness of a melody comes from the dance between them.

The standard non-chord tones (NCTs):

- **Passing tone (PT)** — connects two chord tones by step. Approached and left by step in the same direction.
- **Neighbor tone (NT)** — moves up (or down) a step from a chord tone and returns to it. Approached and left by step in opposite directions.
- **Suspension** — a chord tone is held over from the previous chord, becomes a dissonance against the new harmony, and resolves down by step. The classic 4-3, 7-6, 9-8 suspensions.
- **Anticipation** — the next chord's note arrives early. Common in pop on the "and" of beat 4 anticipating beat 1.
- **Appoggiatura** — a leap to a non-chord tone, which then resolves by step (usually down). Lyrical and emotive.
- **Escape tone (échappée)** — moves stepwise from a chord tone, then leaps in the opposite direction.

These categories aren't pedantic — they correspond to perceptually different kinds of "tension". A passing tone is barely felt; an appoggiatura is a heart-tug.

For melodies to sound rich without sounding "wrong", a useful pattern: aim notes on strong beats at chord tones, decorate with NCTs on weak beats. This is the default. Inverting it — putting NCTs on strong beats — produces yearning, suspended-feeling melodies common in ballads.

### Strong-beat note choice and emotional weight

The pitch on a strong beat over a given chord carries that beat's emotional weight. Quick reference for what each chord tone sounds like in the melody:

| Chord tone | Stability | Affect when on a strong beat |
|------------|-----------|------------------------------|
| Root | Most stable | Grounded, declarative |
| 3rd | Stable; defines major/minor | Brightness or darkness depending on major/minor |
| 5th | Stable, relatively neutral | Open, "out there" |
| 7th (in 7th chord) | Tense | Yearning, unresolved |
| 9th | Tense, sweet | Lush, modern |
| ♯11 | Tense, ambiguous | Floating, Lydian |
| 13th | Tense, warm | Gospel, soul |
| Altered tensions (♭9, ♯9, ♭13) | Highly tense | Edgy, bluesy, jazzy |

Songwriters often choose melody notes for their *vertical* feeling against the chord, not just their horizontal flow. A chorus melody that lands on the 9th of the I chord (D over Cmaj9) feels modern and lifted; one that lands on the root feels grounded.

## Rhythm in melody

Pitch is half of melody. Rhythm is the other half — and the half that often does more work.

### Rhythmic identity

A motif's rhythm is usually more recognizable than its pitch contour. The opening four notes of Beethoven's 5th — three short notes and a long one — would be recognized as "the theme" even with completely different pitches. Rhythmic identity is how the listener locks onto a motif.

When writing a memorable melody, give it a distinctive rhythm — not necessarily complex, but characterful. Repetitive rhythms with subtle variations are often more memorable than rhythms that change every bar.

### Syllabic density

In vocal melodies, the number of syllables per beat affects the song's emotional pacing. Three modes:

- **Sparse (1–2 syllables per bar)**: open, ballad-like, gives space for sustained notes.
- **Moderate (3–6 syllables per bar)**: the default for pop verses.
- **Dense (7+ syllables per bar)**: rap-adjacent, conversational, stuffed with information.

Mixing densities across sections is a primary contrast tool. Verses often dense; choruses often sparse (or vice versa).

### Anacrusis and phrase rhythm

Many memorable melodies start before the downbeat — the anacrusis (pickup). It creates forward motion into the strong beat. "Happy Birthday" starts on the upbeat. Most pop choruses start on or just before beat 1 of the bar after the chord change.

Where phrases land also matters. A phrase that ends on a downbeat feels conclusive; one that ends on the off-beat feels suspended, encouraging continuation.

## Repetition, variation, and surprise

Memorable melodies repeat. Pure variation is forgettable; pure repetition is monotonous. The art is the ratio.

- **Exact repetition** — fine for short motifs (1–2 bars) but tedious over long stretches.
- **Sequenced repetition** — same shape, different starting pitch. Up a step, down a step, up a third. The default development tool.
- **Rhythmic repetition with melodic variation** — same rhythm, different pitches. Holds together rhythmically while exploring melodically.
- **Melodic repetition with rhythmic variation** — same pitches, different rhythm. Less common; works for hooks where the pitches are the identity.
- **Fragment-and-extend** — repeat a short motif; the third or fourth time, extend it into something new. The "rule of three" then break.

See `motivic-development.md` for systematic development of motifs across a piece.

## Genre-specific notes (high level — see genre files for depth)

- **Classical period melodies** are often clear-arched, balanced phrases with cadential closure. Mostly stepwise; leaps usually filled in.
- **Romantic melodies** are wider in range, more chromatic, more emotionally extreme. Wagner, Tchaikovsky.
- **Bebop melody** is leap-driven, chord-tone-heavy, fast, and rhythmically irregular. Outlines harmony aggressively. Charlie Parker.
- **Pop / rock melodies** are usually narrow-range, repetition-heavy, and rhythmic. Verse melodies often live within a 5th; choruses open up.
- **Hip-hop / R&B vocal melodies** have heavy rhythmic syncopation and often use blue notes (♭3, ♭5, ♭7 in major). Melisma is common in R&B.
- **EDM "melodies"** are often short, repetitive synth motifs whose interest is timbral and rhythmic, not melodic in the classical sense.
- **K-pop melodies** often combine multiple distinct hooks within a single song — each section has its own memorable line. The "post-chorus" or "drop chorus" introduces a third melodic identity.
- **Film/TV melody** ranges from full-blown themes (Williams, Goldsmith, Zimmer) to fragmentary motivic gestures. Theme writing is a specific discipline — see `genres/film-tv-scoring.md`.
- **Folk melodies** prioritize singability — narrow range, mostly stepwise, predictable phrase lengths.

## Common pitfalls when advising

- **Don't write a melody by stacking chord tones.** Strict chord-tone-only melodies sound mechanical. The non-chord tones are where the soul lives.
- **Don't write melodies note-by-note without thinking about contour.** A line that goes ↑↓↑↓↑↓ with no shape is forgettable, even if every note is "right".
- **Don't aim for too high a peak too early.** If bar 2 is the highest note, the rest of the melody is descent; the listener doesn't have a destination.
- **Don't ignore rhythm.** A melody that's pitch-only is half a melody. Always consider syllabic density, syncopation, and the rhythm of the contour.
- **Don't forget the tessitura.** Especially for vocal melodies, where a line lives matters more than where it occasionally goes. Match it to the singer.
- **Don't fight the underlying harmony.** A melody that consistently lands on the 7th over every chord can sound deliberately strange (good if intended) or accidentally sour (bad if not). Know which you're doing.

## Diagnostic checklist for "this melody isn't working"

When the user has a melody that feels off, walk through:

1. **Is there a clear contour shape?** Or is it noodly?
2. **Where's the climax?** Is there one? Is it placed well?
3. **Is the range appropriate** for the voice/instrument and the section's role (verse vs. chorus)?
4. **Is the step/leap balance right** for the genre?
5. **Are the strong-beat notes mostly chord tones?** If they're all NCTs, the melody might feel suspended.
6. **Is there enough repetition** for memorability? Too much for interest?
7. **Does the rhythm have a clear identity** or is it shapeless?
8. **Does the line answer itself** at phrase boundaries (antecedent–consequent), or is every phrase just trailing off?
9. **Does the climax or peak moment fall on a memorable lyric or harmonic event?** Or is the highest note arbitrary?
10. **Is the melody distinguishable from the bass line and the chord voicings?** Sometimes a melody fails because it's overlapping too closely with another part.

Diagnose, then prescribe specific fixes.

## Cross-references

- Motif and theme development → `motivic-development.md`
- Phrase grammar (period, sentence) → `phrase-structure.md`
- Counterpoint (when there's more than one melodic line) → `counterpoint.md`
- Lyric–melody interaction → `songwriting/lyric-writing.md`, `songwriting/topline-craft.md`
- Hooks specifically → `songwriting/hooks-and-memorability.md`
- Melody over chord changes (chord-scale) → `harmony/jazz-harmony.md`, `fundamentals/pitch-intervals-scales.md`
- Modal melodic writing (which scale degrees to feature) → `harmony/modal-harmony.md`
- Genre-specific melodic conventions → relevant `genres/*.md`

# Woodwinds and Saxophones

Use this file when the user asks for flute, oboe, clarinet, bassoon, saxophone, woodwind section writing, wind solos, breath-aware phrasing, or agile color lines.

## Core idea

Wind writing is breath-shaped. A wind player can phrase beautifully, but they need air, register-aware lines, and idiomatic articulation.

```text
Wind part = breath + register color + articulation + line shape
```

## Family roles

| Instrument | Characteristic strengths | Cautions |
|---|---|---|
| Flute | agility, shimmer, airy lyricism, high sparkle | low register is soft; high can pierce |
| Oboe | penetrating lyric line, plaintive color | exposed intonation/tone; needs breathing room |
| Clarinet | wide range, smooth legato, agile arpeggios | register break matters; low/high colors differ sharply |
| Bassoon | dry bass color, comic/agile lines, warm tenor lyricism | very low fast writing can blur |
| Saxophone | vocal, jazz/pop/R&B/rock flexibility, strong projection | orchestral vs pop/jazz idiom differs |

## Register color

| Register concept | Effect |
|---|---|
| Low flute | intimate, breathy, easily covered |
| High flute | bright, brilliant, can dominate |
| Oboe middle | clear, expressive, pastoral/poignant |
| Clarinet chalumeau | dark, warm low register |
| Clarinet clarion | clear, singing middle-high register |
| Bassoon tenor | lyrical and human |
| Sax middle | vocal and flexible |
| Sax high | intense, soloistic, sometimes raw depending on style |

## Breath and phrasing

- Leave breathing points at phrase boundaries.
- Long lines need either rests or staggered breathing in ensemble writing.
- Fast passages still need air planning.
- Avoid writing nonstop high loud passages unless the dramatic need is strong.
- Repeated staccato can be fatiguing at fast tempos.

## Articulation vocabulary

| Marking / idea | Use |
|---|---|
| Legato slurs | smooth lyrical phrases |
| Tongued staccato | clarity, lightness, rhythmic definition |
| Accents | dance/riff emphasis |
| Flutter tongue | special color, tension, effect |
| Grace notes | folk/jazz/classical ornament depending on style |
| Falls/scoops | sax/jazz/pop idiom; style-specific |

## Writing by function

### Solo melody

- choose register for emotional color
- keep phrase lengths singable
- use rests as expressive breath
- support with sparse accompaniment

### Section pad

- use smooth voice-leading
- avoid too many exposed dissonances in weak registers
- blend flute/clarinet/oboe carefully; each projects differently

### Agile figure

- winds can be very agile, but key and interval pattern matter
- stepwise/scalar lines are safer than constant wide leaps
- clarinet and flute handle arpeggios well; oboe/bassoon need more care

### Pop/R&B sax line

- write like a vocal hook or response line
- leave space for bends/falls if style-appropriate
- do not over-notate every expressive inflection unless needed

## Doubling and blend

| Combination | Effect |
|---|---|
| Flute + violin | bright, agile shimmer |
| Oboe + clarinet | focused reed blend |
| Clarinet + viola/cello | warm inner color |
| Bassoon + cello | dry/warm bass support |
| Sax section in harmony | pop/jazz horn-section power |
| Flute octave above melody | sparkle, lift |

## Quick examples

### Oboe lyrical line

```text
Range: middle register
Phrase: 4 bars with a rest or held note near bar 3
Support: soft strings/piano, not dense brass.
```

### Clarinet color shift

```text
Start in low chalumeau register for warmth.
Leap or sequence into clarion register for lift.
```

### Pop sax hook

```text
2-bar riff after vocal title.
Use repeated rhythmic cell + one fall at phrase end.
```

## Common pitfalls

- **Do not write wind parts without breathing points.** Air is part of the phrase.
- **Do not ignore register color.** The same note pattern changes identity across registers.
- **Do not bury low flute under dense accompaniment.** It will disappear.
- **Do not write saxophone as if it were just orchestral woodwind.** Pop/jazz phrasing is idiomatic.
- **Do not overuse special effects.** Flutter tongue and falls are expressive, not default.

## Cross-references

- Instrument ranges → `../orchestration/instruments-ranges-character.md`
- Brass contrast → `brass.md`
- Voicing and texture → `../orchestration/voicing-and-texture.md`
- Jazz styles → `../genres/jazz-styles.md`
- Film/TV scoring → `../genres/film-tv-scoring.md`
- Notation and articulation → `../fundamentals/notation-and-conventions.md`

# Phrase Structure

A phrase is the basic unit of musical "thought" — a self-contained melodic-harmonic gesture that ends with some kind of cadential closure. It's the musical equivalent of a sentence in language. Phrase structure is how phrases combine into larger units (periods, sentences, hybrid forms) and how their grammar shapes the listener's expectation.

## When to consult this file

- The user asks about "phrasing" or how phrases work
- Antecedent / consequent, periods, sentences
- "Why does this melody feel incomplete" — often a phrase-structure issue
- Writing balanced melodies in classical or pop styles
- Understanding how a 16-bar verse "works" as a unit
- Asymmetric or expanded phrase structures

For melodic content within a phrase, see `melodic-construction.md`. For motivic development across phrases, see `motivic-development.md`. For cadences as phrase endings, see `harmony/functional-harmony.md` and `assets/cadence-reference.md`. For phrase structure as it interacts with form (verse, chorus), see `form/popular-song-forms.md` and `form/classical-forms.md`.

## What is a phrase

A phrase is a complete musical thought — a unit that begins, develops, and ends with a cadence (or cadential gesture). It corresponds roughly to a singer's breath or an instrumentalist's bow stroke, though it can be longer than a single breath.

Three diagnostic features:

1. **Internal coherence** — the notes belong together; the listener perceives them as a single gesture.
2. **Cadential closure** — the phrase ends on a cadence (or cadence-like resolution), giving the listener a clear "ending point" before the next phrase begins.
3. **Pacing** — phrases have a sense of breath; a phrase that doesn't pause is one extended gesture, not two phrases.

A phrase can be as short as 2 bars or as long as 16+ bars, depending on tempo and style. The "default" classical phrase is 4 bars; the default pop phrase is also typically 4 or 8 bars, depending on tempo.

## Phrase length conventions

| Genre / context | Default phrase length |
|-----------------|----------------------|
| Classical period (Mozart, Haydn) | 4 bars (with frequent 4+4 = 8-bar periods) |
| Classical late-Romantic (Wagner, Mahler) | Variable, often 6, 8, or asymmetric |
| Pop / rock verse | 4 bars (with 4+4 forming an 8-bar half-section) |
| Pop / rock chorus | 4 or 8 bars |
| Blues | 4 bars (12-bar form is 3 × 4-bar phrases) |
| Jazz standards | 8-bar phrases (32-bar AABA = four 8-bar phrases; 16-bar A) |
| Folk songs (anglo-american) | 4 or 8 bars |
| Hip-hop verse | 16 bars (a "16"); often subdivided into 4 × 4 |
| EDM | Long phrases, often 16 or 32 bars, defined by structural changes |
| K-pop | Mixed: pre-chorus often 4+4, chorus often 8 bars, post-chorus often 4 bars |

These are defaults. Phrase length is a creative choice; deviation from the default creates expectation and surprise.

## The 4-bar default

Why 4 bars? Several converging reasons:

- **Speech rhythm**: a typical sung lyric phrase fits in roughly 4 bars at moderate tempo.
- **Breath length**: 4 bars at moderate tempo (~120 BPM) is roughly 8 seconds — comfortable for a singer.
- **Memory**: a 4-bar phrase is short enough for the listener to hold in working memory and feel as a single unit.
- **Cultural inheritance**: Western European folk and dance music established 4-bar phrasing as default.

Phrases that are 3 bars or 5 bars are perceived as asymmetric — striking in classical contexts, more accepted in jazz and contemporary pop. Phrases of 2 bars are often felt as half-phrases (or as "antecedents" of a 4-bar period). Phrases of 6+ bars feel expanded.

## Periodic structure (antecedent + consequent)

The classical period is two phrases that complement each other:

- **Antecedent**: the first phrase, ending with weak cadence (typically half cadence on V, or imperfect authentic cadence). Sets up an expectation.
- **Consequent**: the second phrase, ending with stronger cadence (typically PAC). Resolves the expectation.

Together they form a balanced, sentence-like statement. The grammar:

```
Antecedent (4 bars, ends HC or IAC)
Consequent (4 bars, ends PAC)
———
Period (8 bars total)
```

### Parallel period

Antecedent and consequent share the same opening (often identical for the first 1–2 bars), with the consequent diverging to provide stronger closure.

`a a' b a` is a typical small-scale parallel period: motif `a` opens both halves, `a'` is its variation in the antecedent, `b` is contrasting material, and a return to `a` (slightly altered for cadence) closes the consequent.

### Contrasting period

Antecedent and consequent have different opening material. The connection between them is the question-answer relationship of their cadences (HC → PAC), not melodic similarity.

Less common than parallel period in classical music, but still standard.

### Modulating period

The consequent ends in a different key (often the dominant) instead of the original tonic. Common in sonata-form expositions for transitions.

## Sentence structure

The sentence (a term from Schoenberg, formalized by William Caplin) is a different grammatical structure: a phrase that grows from itself rather than answering itself.

Three parts:

1. **Presentation** (typically 4 bars): a basic idea (~2 bars) followed by its repetition (~2 bars). The repetition can be exact, sequenced, or varied.
2. **Continuation** (typically 4 bars): increasing fragmentation, accelerated harmonic rhythm, "liquidation" of the basic idea.
3. **Cadence**: the conclusion, embedded in the continuation or as its own brief gesture.

```
Presentation (basic idea + repetition, 4 bars)
Continuation (fragmentation, accelerated motion, 4 bars including cadence)
———
Sentence (8 bars total)
```

A famous example: the opening of Beethoven's Piano Sonata No. 1 in F minor. The first 4 bars present a motif and its repetition; the next 4 fragment and accelerate, ending in a cadence.

### Sentence vs. period — the difference

- **Period**: question + answer. Antecedent ends weakly, consequent ends strongly. Two equal phrases in dialogue.
- **Sentence**: statement + development + conclusion. One growing phrase, internally articulated.

Both are 8 bars in classical practice. The grammar is different. A sentence has a more directional, growing feel; a period has a more balanced, dialogic feel.

## Hybrid structures

Real music often blends sentence and period grammar.

- **Compound period**: a period where each phrase is itself a sentence. Total: 16 bars.
- **Compound sentence**: a sentence where the presentation is itself a small period. Total: 16 bars.
- **Sentence with consequent**: a sentence completes, then a parallel sentence (or period) follows as a "consequent" at the larger scale.

These hybrids are classical conventions but the principles apply broadly. A pop verse-chorus pair can be analyzed as a hybrid period (verse = antecedent, chorus = consequent) or as a sentence (verse-chorus-bridge-chorus structure).

## Asymmetric and expanded phrases

Most phrases are 4 or 8 bars. Asymmetry is striking precisely because it deviates from the default.

### 5-bar phrases

A 4-bar phrase plus a one-bar extension. The extension often elaborates the cadence ("delaying" the resolution). Common in folk traditions (Irish reels, Scottish strathspeys) and in 20th-century classical.

### 6-bar phrases

Two 3-bar groups, or 4+2 (a 4-bar phrase with a 2-bar tag), or 2+2+2 (three short groups). Mahler uses these often.

### 7-bar phrases

Often 4+3 or 3+4 — a regular 4 with one added or one taken away. Brahms is the master.

### Expanded phrases (extension techniques)

A phrase of "default" length expanded by:

- **Repetition of an internal motif** before completing the phrase
- **Extension of the cadence** through deceptive cadence, evaded cadence, or pedal point
- **Internal sequence** that lengthens the development before the cadence
- **Insertion of a contrasting interpolation**

The result: a 4-bar phrase can become 6, 7, or 9 bars without losing its essential grammar.

### Truncated phrases

Phrases shorter than expected. A 3-bar phrase where 4 was expected. Used for breathlessness, urgency, or specific effects (e.g., "cutting off" a phrase to suggest interruption).

## Phrase elision

When the cadence of one phrase coincides with the start of the next, eliding them into a single beat. Effect: forward propulsion; the listener doesn't get the breath they expected.

Common in Bach, Brahms, and modern jazz. In pop, less frequently — pop phrases tend to have clear gaps between them for vocal breath.

Notation: an elision often shows up as a strong-beat downbeat that's both the resolution of the previous phrase and the opening of the next. Sometimes marked with a "phrase elision" indication in the score.

## Phrases in popular music

Pop song phrasing follows similar grammar but with different defaults.

### Verse phrases

Pop verses are typically 4-bar phrases, with two or four phrases per verse (8 or 16 bars total). Within a verse, phrases often follow:

- **Parallel period** structure (each pair of phrases is parallel)
- **Repetitive** structure (each phrase is a slight variation of the same idea)
- **Building** structure (phrases get more rhythmically dense as the verse progresses, leading into the pre-chorus)

### Chorus phrases

Pop choruses often use:

- **Same-melody phrasing**: each 4-bar phrase has the same melody (the chorus hook)
- **Bookending**: the chorus opens and closes on the title hook, with development in the middle 4 bars

The chorus melody is *less* grammatically complex than verses because memorability is the priority. A chorus where every phrase ends differently is harder to remember than one where the phrases are nearly identical.

### Pre-chorus phrasing

Pre-choruses often use phrases shorter than the verse (2-bar phrases instead of 4) to accelerate the harmonic and melodic rhythm into the chorus.

### Bridge phrasing

Bridges are the most "developed" section in pop. They often introduce a new phrase length (e.g., a 4-bar phrase among 8-bar verses), modulate, or use sentence-like grammar with continuous development.

## Phrasing in jazz

Jazz phrases are often longer (8 bars) and looser (driven by swing feel and chord changes rather than strict 4-bar grammar).

A 32-bar AABA standard has eight phrases of 4 bars each (or four 8-bar units of two paired phrases). The bridge (B) often departs from the A's phrase structure.

Jazz solos typically build phrases that span 4, 8, or sometimes 16 bars, with breath points at chord-change boundaries. A great improviser thinks in long phrases; a beginner often plays in 1–2 bar fragments.

## Phrasing in hip-hop

Hip-hop verses are usually 16 bars, often divided into 4 × 4-bar phrases or sometimes 8 × 2-bar phrases. The phrase structure is heavily influenced by rhyme scheme — the end of a couplet (or quatrain) often coincides with phrase end.

Rappers vary phrase length deliberately: shorter phrases (2 bars) for impact, longer phrases (4+) for storytelling, 8-bar runs for breathless intensity.

## Phrasing in EDM

EDM works at the section level rather than the phrase level. A "drop" is typically 16 or 32 bars, structured by build-ups, drops, breakdowns, and returns rather than phrase grammar in the classical sense. But within a section, basslines and lead synths still follow phrasing principles — short repeated motifs that build over many bars.

## Phrasing in film/TV

Film scoring varies wildly. A short cue may be a single phrase. A longer cue can use through-composed phrasing (each phrase contrasts) or motivically tight phrasing (every phrase derived from the leitmotif).

Trailer music often uses extended phrases that build over 30+ seconds, with the "drop" or climax functioning as a structural cadence.

## Common pitfalls when advising

- **Don't impose 4-bar default on all genres.** Hip-hop is built on 16-bar phrases; some folk traditions have asymmetric phrasing as default. The 4-bar phrase is a Western European convention.
- **Don't ignore phrase boundaries when composing.** Melodies that don't breathe (no clear cadential moments) feel exhausting. The listener needs a place to land before the next idea begins.
- **Don't make every phrase the same length.** A song with eight 4-bar phrases lined up identically is rhythmically rigid. Mix in variation — a 6-bar phrase among the 4-bars, or an extended cadence — to add interest.
- **Don't confuse phrase length with section length.** A "16-bar verse" is the section; the phrases inside it might be 4 bars each. Always specify which scale you're discussing.
- **Don't forget to land cadences.** A phrase that doesn't end somewhere harmonically is not really a phrase — it's an incomplete thought. Even pop phrases need at least an implied cadential resolution.
- **Don't impose classical period structure on music that wants something else.** Sentence structure, hybrid forms, and asymmetric phrases all have their place.

## Quick decision matrix

| Want | Use |
|------|-----|
| Balanced, conversational feel | Parallel period (antecedent + consequent) |
| Building, growing feel | Sentence (presentation + continuation) |
| Striking, surprising | Asymmetric phrase (3, 5, 7 bars) |
| Hypnotic, meditative | Repetitive phrasing with little variation |
| Driving, propulsive | Phrase elision (no breath between phrases) |
| Conclusive ending | Strong PAC at the end of a clear period |
| Open-ended | End on HC or deceptive cadence; leave the period unfinished |
| Pop chorus memorability | Identical phrases bookending the chorus |
| Pop verse build | Phrases that get rhythmically denser as the verse progresses |
| Jazz solo coherence | Longer phrases (8 bars) that develop a single motif |

## Cross-references

- Melodic content within a phrase → `melodic-construction.md`
- Motivic development across phrases → `motivic-development.md`
- Cadences (the harmonic punctuation that ends a phrase) → `harmony/functional-harmony.md`, `assets/cadence-reference.md`
- Counterpoint (multiple phrases happening simultaneously) → `counterpoint.md`
- Pop song structure (how phrases combine into verse/chorus) → `form/popular-song-forms.md`
- Classical forms (periods/sentences in larger structures) → `form/classical-forms.md`
- Lyric phrasing and prosody → `songwriting/lyric-writing.md`
- Genre-specific phrasing conventions → relevant `genres/*.md` files

# Arrangement for the Mix

Modern recorded music is a hybrid art: composition + arrangement + production + mixing. This file is about the *arrangement* decisions that affect how a track will mix — frequency distribution, instrumental ranges, parts that may compete or complement each other in the mix.

This skill doesn't cover mixing itself (compression, EQ, reverb, stereo placement) — that's a different specialty. But the composer/arranger makes choices that determine whether the mix will be easy or hard, clear or muddy. This file is about those choices.

For density and layering generally, see `orchestration/arrangement-density.md`. For voicing across instruments, see `orchestration/voicing-and-texture.md`. For energy and dynamics across sections, see `energy-and-dynamics.md`.

## When to consult this file

- The user has an arrangement that "sounds muddy" or "doesn't feel full"
- Mix-aware composition decisions
- Multiple instruments competing in the same range
- Frequency-conscious arrangement choices
- Kick + bass relationship (the perennial challenge)
- "Why does my song sound smaller than [reference track]?"
- Arrangement choices that aid mixing later

## The frequency spectrum

Audio occupies a frequency range from roughly 20 Hz to 20 kHz. The spectrum is divided (informally) into zones:

| Zone | Frequency | What lives here |
|------|-----------|-----------------|
| Sub-bass | 20–60 Hz | The lowest fundamentals — sub-kick, 808 sub bass, deep synth bass |
| Bass | 60–200 Hz | Bass guitar fundamentals, kick drum body, low piano notes |
| Low-mids | 200–500 Hz | Piano LH, low strings, lower vocal range, "warmth" |
| Mids | 500 Hz–2 kHz | Most instruments' fundamentals — vocals, guitars, lower brass, snare body |
| High-mids / presence | 2–6 kHz | Vocal clarity, guitar bite, snare crack, cymbal sticks |
| Highs | 6–12 kHz | Hi-hats, cymbals, vocal "air", string brilliance |
| Air | 12–20 kHz | Cymbal shimmer, reverb tails, "openness" |

Each zone can support some instruments. When two instruments occupy the same zone with similar timbres, they "fight" in the mix — frequency masking.

A well-arranged track has each zone "filled" by something distinct, with limited overlap. A poorly arranged track has multiple instruments competing in the same zone.

## The mid-range problem

Most instruments live in the 200 Hz – 2 kHz range. This is where:

- Vocals sit (the bulk of vocal energy)
- Guitars (most chords, strums, leads)
- Pianos (mid-range chords)
- Synths (most pads and leads)
- Brass (most parts)
- Most strings (violin and viola sit here)

Result: mid-range gets crowded fast. The mix engineer has to carve space with EQ, but the arranger can help by:

- **Limiting the number of mid-range instruments** at any given moment.
- **Distributing instruments to different mid-range zones** (one in 300 Hz, another in 600 Hz, another in 1 kHz).
- **Using extreme registers when possible** (high piano, low cellos) to free the middle.
- **Designing parts that don't all play at once** (one mid-range instrument plays during the verse, another during the chorus).

## Kick and bass — the perennial challenge

Kick drum and bass guitar (or sub bass) both occupy the 60–200 Hz range. They constantly compete.

Solutions:

- **Frequency separation**: kick has more energy at 60–80 Hz; bass at 100–200 Hz. Or vice versa. The key is choosing complementary spectral profiles.
- **Sidechaining**: the bass ducks (briefly drops in volume) every time the kick hits. Common in EDM. Creates "pumping" effect.
- **Rhythmic complementarity**: kick on 1 and 3, bass on 2 and 4 (or vice versa). They don't hit at the same time.
- **Note choice**: when kick hits, bass plays the root (matches the kick's fundamental); when kick rests, bass can play other notes more freely.
- **Layered kick**: a kick drum that's tuned to the song's key (or has tonal center) so it harmonizes with the bass.

A well-arranged track has a clear kick-and-bass relationship that's been thought through. A poorly arranged track has them stepping on each other.

## Stereo space

Mono is the center of the stereo field. Stereo allows panning — placing instruments left or right.

### Standard panning conventions

| Element | Position |
|---------|----------|
| Lead vocal | Center (always) |
| Kick drum | Center |
| Snare | Center or slightly off-center |
| Bass | Center (low frequencies don't pan well in headphones / club systems) |
| Hi-hats | Slightly to one side (stereotypically right for pop) |
| Cymbals | Spread across stereo field |
| Rhythm guitars (doubles) | Hard left + hard right (the "double-tracked" effect) |
| Lead guitar | Often center, sometimes off-center |
| Synth pads | Spread wide |
| Backing vocals | Spread to fill stereo field around lead |
| Strings | Spread (high strings to one side, low strings to other) |

### Mono compatibility

Many listeners hear music in mono — phone speakers, club PA systems, voice-call systems. The arrangement should still work in mono. This means:

- Don't put critical elements only on one side.
- Avoid out-of-phase stereo effects that will collapse in mono.
- Test mixes by listening in mono.

The composer/arranger contributes to mono compatibility by choosing panning that works in both formats.

## Headroom and dynamics

A track that's "loud" throughout has no dynamic range — every section sounds the same volume. A track with dynamics has loud and quiet moments.

For arrangement:

- **Verses should be quieter than choruses** in terms of average volume.
- **The pre-chorus build creates dynamic motion** — getting louder.
- **The bridge often has a dynamic dip** (quieter moment) that lets the final chorus hit harder.

Modern mastering typically compresses everything to high loudness, but the *arrangement* should still have dynamic shape; otherwise the mastering produces a flat, fatiguing result.

## The mid-side perspective

Audio engineers think in "mid" (mono content) and "side" (stereo content). The arranger can help by:

- **Putting the most important elements in the mid** (vocal, kick, bass, snare).
- **Putting decorative elements in the sides** (atmospheric synths, doubled guitars, ambience, reverb tails).
- **Keeping bass mono** — sub frequencies don't have stereo information naturally.

## Layering — frequency-aware

When adding layers, think frequency, not just "another instrument".

- **Adding a 2nd guitar in the same range as the 1st**: doubling — works if intentional, but doesn't add density.
- **Adding a high-pitched instrument** (piccolo, glockenspiel): adds top end, complements existing parts.
- **Adding a low instrument** (sub bass, low string): adds bottom, fills out the spectrum.
- **Adding mid-range instrument when mids are full**: clutter; will fight existing parts.

The skill is layering complementarily.

## Common arrangement-for-mix mistakes

### Too many things in the mid

Lead vocal + 3 mid-range synths + acoustic guitar + electric guitar all in 500 Hz – 2 kHz. The mix engineer has to carve enormously to make space; the result is thinned.

Fix: redistribute or remove some elements.

### Bass guitar in the same register as the kick fundamentals

Bass playing low E, F, G can muddle with kick fundamentals. The bass becomes inaudible, the kick becomes flabby.

Fix: have the bass play in a higher register, or shape kick to leave bass-friendly space.

### Vocal doubled at the same range

Lead vocal + harmony vocal an octave below in the same range — they fight.

Fix: pan the harmony off-center, or put the harmony in a different range (typically a 3rd or 5th above).

### High-frequency clutter

Multiple cymbals, hi-hats, and high synths all hitting at the same time. The high end becomes a wash of noise.

Fix: rhythmic complementarity (one element at a time on the high frequencies) or remove some elements.

### No sub-bass

Modern productions need sub-bass content (60 Hz and below) for clubs, headphones, and modern speakers. Without it, the track sounds "thin" on professional systems.

Fix: add a sub-bass synth or layered kick with sub content.

### Stereo collapse

Critical content (e.g., the lead riff) panned hard right. In mono, it's perfectly there; in stereo on headphones, the listener tilting their head can lose it.

Fix: keep critical content in or near center.

### Frequency masking between parts

Snare and lead guitar both occupy 200–500 Hz. They conflict.

Fix: arrange them not to play simultaneously, or put one in a different range.

## Genre-specific considerations

### Pop

- Vocal is everything. All other parts must clear space for the vocal in the mid-range.
- Kick and bass tightly produced; sidechained.
- Stereo wide for choruses (more layers spread across the stereo field) and tighter for verses.
- Reverb and delay used selectively — too much washes the vocal.

### Rock

- Guitar-centric. The arrangement is usually about guitars + drums + bass + vocals.
- Doubled rhythm guitars (hard panned) is the standard rock production trick.
- The kick and bass guitar relationship is critical; modern rock production tightens both.

### Hip-hop

- 808 sub-bass below the kick; the two together create the foundation.
- Drums are highly compressed and processed.
- Vocal is up-front, often with reverb but tightly mixed.
- Sample-based tracks have to work around the original recording's frequency content.

### EDM

- Sub-bass is foundational; modern productions have heavy 60 Hz – 100 Hz content.
- Kick and bass sidechained explicitly.
- Wide stereo on synths and pads; mono on the bass.
- Drops are designed to sound massive — frequency content from sub to air.

### R&B / neo-soul

- Vocal often sits forward with subtle layering.
- Bass is heavy; deep low end common.
- Drums often sound vintage (sampled or analog-emulated).
- Pads and atmospheric synths fill the mid-side space.

### Country

- Vocal is paramount; everything serves it.
- Acoustic guitar in mid; electric guitar (or pedal steel) often slightly off-center.
- Bass and kick more prominent than in many pop genres.
- Less aggressive frequency processing than pop or EDM.

### Classical / orchestral

- Natural acoustic recording approach often preferred.
- Sections of the orchestra placed in stereo according to seating.
- Less frequency-conscious arrangement; the room's acoustics matter more.
- But scoring choices affect how the orchestra balances on its own.

## A workflow for arrangement-aware composition

If the user is composing with the eventual mix in mind:

1. **Plan the frequency budget.** Decide which zones each instrument will occupy. Aim for at least one element per zone (sub, bass, low-mid, mid, high-mid, high) without crowding any.
2. **Set up kick-and-bass relationship.** Decide which has more 60 Hz content, which has 100–200 Hz. Plan their rhythmic relationship.
3. **Identify the lead element** (usually vocal). Reserve clear space for it in mid-range; arrange other parts to avoid its register.
4. **Design layers in pairs.** Each new element you add should complement an existing one, not duplicate it.
5. **Plan dynamic arc.** Identify which sections are quieter/louder; arrange parts accordingly.
6. **Consider mono compatibility.** Test (or imagine) the mix in mono; ensure critical elements are center.
7. **Think about transitions.** What instruments enter and exit at section breaks? What's added at the chorus, what's removed at the bridge?
8. **Reference checking.** Compare your arrangement to professional tracks in the same genre. Where are their elements in the spectrum? What's the kick-bass relationship?

## Common pitfalls when advising

- **Don't ignore the mix.** Composition decisions become mix problems. Arrange thoughtfully now to save mix work later.
- **Don't assume more layers = better.** Modern pop is highly produced but selectively layered. Each layer must earn its place.
- **Don't write all the parts in the mid-range.** The mid is where the vocal lives; other parts need to make space.
- **Don't ignore the kick-bass relationship.** It's the foundation; if it's muddy, everything sounds muddy.
- **Don't forget about mono.** Many listeners will hear in mono. Test arrangements in mono.
- **Don't over-process at the arrangement stage.** Arrangement is about parts; processing is about mixing. They're related but distinct.
- **Don't skip referencing.** Always compare to professional tracks in the same genre. Arrangement standards differ enormously by genre.

## Quick decision matrix

| Want | Try |
|------|-----|
| Bass that doesn't fight kick | Sidechain bass to kick; or have bass play in different register from kick fundamentals |
| Crowded mid-range fix | Remove one element; or push one to a different zone |
| Punchy vocal | Carve mids around vocal range; minimal compression-fighting elements there |
| Big stereo image | Doubled guitars panned wide; reverb tails spread wide; lead in center |
| Mono-compatible mix | Critical elements in center; avoid out-of-phase stereo tricks |
| Modern pop sound | Sub bass + sidechained bass; kick sample with character; tight vocals; wide pads |
| Fuller chorus than verse | Add mid-frequency layer (rhythm guitar, pad); add backing vocals; widen stereo |
| Thinner verse | Strip to vocal + minimal accompaniment; let chorus add layers |
| Rock production | Hard-panned rhythm guitars; tight kick + bass; vocal in center |
| Hip-hop low end | 808 sub + tuned kick + sidechained bass content; tight 100–200 Hz |
| EDM drop massive | Sub bass + main bass + lead synth + chord stab + drums all hitting; full frequency spectrum |

## Cross-references

- Density and layering → `orchestration/arrangement-density.md`
- Voicing and texture → `orchestration/voicing-and-texture.md`
- Instrument ranges (informs frequency planning) → `orchestration/instruments-ranges-character.md`
- Energy and dynamics across sections → `energy-and-dynamics.md`
- Pre-production decisions (key, tempo, length — set before arrangement) → `pre-production-decisions.md`
- Genre-specific production conventions → various `genres/*.md` files (especially pop-rock, hip-hop-rnb, electronic-edm, kpop-jpop)
- Form and arrangement (how density changes across sections) → `form/popular-song-forms.md`, `form/narrative-and-transitions.md`

# Voicing and Texture

This file is about how to distribute musical material across instruments and registers — *voicing* (the vertical arrangement of notes in a chord) and *texture* (the overall fabric created by all the parts together). These are the decisions that turn a chord-and-melody sketch into actual orchestration.

For instrument ranges and character, see `instruments-ranges-character.md`. For decisions about how many things should be playing at once, see `arrangement-density.md`. For chord spelling itself, see `harmony/chord-construction.md`. For voicing in voice-leading terms (how voices move chord-to-chord), see `harmony/voice-leading.md`.

## When to consult this file

- The user has a chord progression and asks how to voice it
- A piece sounds "muddy", "thin", or "harsh" — usually a voicing/texture issue
- Decisions about close vs. open voicing
- Doubling decisions (which notes to double across instruments)
- Texture types (homophony, polyphony, heterophony)
- Layering pads, strings, or synths

## Texture types

Texture is the overall fabric of a piece — how many independent musical lines are happening, and how they relate to each other.

### Monophony

A single melodic line with no accompaniment. Just one voice (one player or many in unison/octave).

Where it appears:

- Plainchant
- Solo vocal/instrumental lines (introduction of a theme; cadenzas)
- Unison playing in folk and traditional music
- The opening of a piece (often a single line establishes the theme)

Effect: focused, intimate, exposed. Every note is highlighted.

### Homophony

A melody with chord accompaniment. The melody is the "voice" the listener tracks; the chords support it.

This is the dominant texture in:

- Most pop, rock, country, R&B
- Hymns and chorales
- Most film scoring (melody + harmony)
- Romantic art songs
- Solo voice plus piano or guitar

Subtypes:

- **Melody-dominated homophony**: a clear lead melody, with subordinated chordal accompaniment.
- **Chordal homophony**: all parts move together rhythmically (block chords); typified by hymns and traditional choral writing.

### Polyphony / Counterpoint

Multiple independent melodic lines happening simultaneously. Each voice has its own contour and identity. See `counterpoint.md` for full treatment.

Where it appears:

- Bach fugues, inventions
- Renaissance motets and madrigals
- Some contemporary classical and jazz
- Limited use in pop (countermelodies, harmonized backing vocals)

### Heterophony

Multiple voices playing variants of the same melody simultaneously. One voice plays the basic line; others play decorated versions. The lines aren't independent (as in polyphony) but they're not unison either.

Common in:

- Many tradition-specific musics (Gamelan, Indian classical, some African traditions)
- Some folk music
- Modern classical pieces explicitly using heterophony

### Mixed texture

Most music uses combinations. A pop song might be primarily homophonic but include moments of monophony (vocal-only intro), brief counterpoint (a countermelody), and momentary heterophony (backing singers harmonizing in different rhythms).

## Voicing — vertical arrangement of notes

Voicing is how you distribute the notes of a chord across registers. The same chord can be voiced dozens of ways, each producing a different feel.

### Close vs. open voicing

**Close voicing**: all chord tones within a single octave, tightly packed. The voices are as close together as possible.

Example: Cmaj7 close: C E G B (all within an octave).

**Open voicing**: chord tones spread across multiple octaves with gaps.

Example: Cmaj7 open: C (low) – E (middle) – G (middle) – B (high), with gaps between voices.

| Quality | Close | Open |
|---------|-------|------|
| Sound | Compact, "blocky" | Spacious, "spread" |
| Texture | Dense, tight | Airy, full |
| Common use | Choral singing, jazz piano | Orchestral writing, film scoring |
| Risk | Muddy in low register | Loses chord identity if too spread |

### Drop voicings (jazz)

Take a close voicing of a chord and "drop" one voice down an octave. Standard jazz technique.

- **Drop 2**: take the 2nd voice from the top, drop it an octave. The most common jazz guitar voicing.
- **Drop 3**: take the 3rd voice from the top, drop it an octave.
- **Drop 2+4**: drop both 2nd and 4th from top.

Drop voicings open up close-position chords without losing their identity. They voice well on guitar and piano.

For chord-by-chord shapes, see `assets/jazz-voicings.md`.

### Spread voicings

Wide intervals across multiple octaves, often with chord tones an octave apart. Common in cinematic and film scoring.

A "spread" voicing of Cmaj7 might be:
- Bass: C2
- Pad mid: G3
- Strings: C5 E5
- Top: B5

The chord covers nearly four octaves. Big, lush, "movie sound".

### Quartal and quintal voicings

Stack 4ths or 5ths instead of 3rds. Less rooted, more modern.

Example: D3 G3 C4 F4 (quartal) — the famous "So What" voicing.

Used in modal jazz (McCoy Tyner, Herbie Hancock), modern film, and contemporary classical.

For more, see `harmony/chord-construction.md` and `assets/jazz-voicings.md`.

### Cluster voicings

Stack 2nds (chromatic or diatonic). Highly dissonant; used for color, texture, or specific effect.

Example: C D E F (diatonic cluster). Or C C♯ D D♯ (chromatic cluster).

Common in: 20th-century classical (Bartók, Ligeti, Penderecki), avant-garde jazz, some film scoring (especially horror).

## Voicing principles

Six rules of thumb (with style-dependent exceptions):

### 1. Don't pile close intervals in low register

Below middle C, intervals smaller than a 3rd or 4th turn muddy. The lower the register, the wider the spacing should be.

A bass note plus a 3rd above it (e.g., C2 + E2) is muddy. A bass note plus a 5th or 6th or octave (C2 + G2 or C2 + C3) is clear.

Rule: give the bass space. Don't crowd it.

### 2. The 3rd defines the chord's quality

If you must omit something, omit the 5th. The 3rd tells the listener whether it's major or minor; without it, the chord is ambiguous.

The exception: power chords (root + 5th, no 3rd) are deliberately ambiguous. That's their character.

### 3. The 7th defines the chord's flavor in jazz

Even more important than the 3rd in many jazz voicings. A `Cmaj7` with no 7th is just a C major triad. The 7th is the difference.

Many jazz piano left-hand voicings include just root (or no root) + 3 + 7. The 5 is omitted; the 3 and 7 carry the chord.

### 4. Tensions want to be on top

Extensions (9, 11, 13) and altered tones (♯9, ♭9, ♯11, ♭13) sound best in the upper register. Stuffing them in the middle creates muddy clusters.

A `Cmaj9` voiced C E G B D (all close) is muddy. C E G B with D up an octave on top is clear.

### 5. The bass is half the chord

Slash chords and inversions change the chord's affect more than altering an inner voice. A `Cmaj7/E` (E in bass) feels lighter and brighter than a root-position `Cmaj7`. A `Cmaj7/G` feels suspended.

When voicing for emotional effect, consider the bass first.

### 6. Avoid more than 4 simultaneously sounding parts in any register

Beyond 4 voices in a single register, distinguishing parts becomes difficult. Spread voices across registers; don't pile 6 voices into a single octave.

Exception: choral SATB writing (4 voices in two octaves works because of vocal timbres) and clusters (where the dissonance is the point).

## Doubling decisions

Doubling = giving two or more instruments the same line. Common doublings and their effects:

### Octave doublings

Same line, two octaves apart. Effect: rich, hymnic, "big".

| Doubling | Effect |
|----------|--------|
| Soprano + tenor (octave apart) | Strong vocal melody |
| Violin I + cello (octave apart) | Lush string melody |
| Flute + clarinet (octave apart) | Bright, slightly thinner than strings |
| Trumpet + horn (octave apart) | Powerful brass melody |

### Unison doublings

Same line, same octave. Effect: combined timbre; thicker than either instrument alone.

| Doubling | Effect |
|----------|--------|
| Violin + cello in tenor register | Warm, glowing |
| Flute + violin | Sweet, ethereal |
| Trumpet + clarinet (mid register) | Bright, full |

Unison can be warmer than either instrument alone, but if the timbres clash, the doubling sounds smudged. Test combinations.

### Doubling at the third (or other intervals)

Two instruments harmonize the line at a 3rd, 6th, 4th, or other interval. Effect: parallel motion, harmonized line.

This is *not* counterpoint — it's still one line, harmonized. Common in pop backing vocals, brass section writing, country music.

### Doubling for orchestral support

A sustained melodic line can have multiple instruments doubling at unison, octave above, and octave below for full orchestral richness. Common in late Romantic and film scoring.

### When NOT to double

- When the line has too much detail (rapid runs, articulations) — doubling loses precision.
- When you want a solo to feel "alone" — doubling negates that.
- When two instruments don't blend (e.g., trumpet + electric guitar usually sounds messy).
- When the doubled line dominates over the rest of the texture.

## Voicing for specific ensembles

### Piano (solo)

- Left hand handles bass and inner harmony; right hand handles melody.
- Bass typically a single note or octave; LH chord either rootless or with root in bass.
- Avoid more than 4 notes per hand in most cases.
- Use pedal sparingly to sustain harmonies; clear pedal at chord changes.

### String quartet

- 4 voices: violin I, violin II, viola, cello.
- Standard practice: melody in V1 (or sometimes cello, viola); inner voices in V2 and viola; bass in cello.
- Avoid voice crossings in chorale-style writing; allow them in fugal/contrapuntal writing.
- Use the cello's high register for solo passages; the viola's high register for warmer solos.

### Symphony orchestra

- Strings provide harmonic foundation and melodic backbone.
- Woodwinds add color, often doubling strings or playing solo.
- Brass adds power for climaxes and fanfares; often plays sustained chords behind active strings.
- Percussion accents and provides rhythmic punctuation.
- Orchestrate from the bass up: bass line first, then chord, then melody.

### Choir (SATB)

- Soprano: melody (usually).
- Alto: inner harmony, often doubled with tenor an octave above or below the soprano.
- Tenor: middle voice, often the harmonic anchor.
- Bass: foundation.
- Avoid leaps larger than an octave in any voice; aim for stepwise motion in inner voices.
- Doubling rules (`harmony/voice-leading.md`) matter: don't double the leading tone, etc.

### Pop band (drums, bass, guitar, keys, vocals)

- Bass: root (mostly) or melodic line; foundation.
- Drums: rhythm.
- Guitar(s): chordal accompaniment, riff, lead, or color.
- Keys (piano/synth): chord pad, secondary melody, color.
- Vocals: melody (lead) + backing harmonies.
- Voicings often kept simple for groove; complexity in production rather than voicings.

### Pop production / DAW

- Bass: kick + bass guitar / synth bass, often sidechained for groove.
- Mid layer: rhythm guitars, electric piano, pad synth.
- High layer: lead vocal (with double, harmonies), arpeggios, plucks, top synths.
- Effects layer: reverbs, delays, risers, atmospherics.

A common approach: voice the rhythm guitar in close voicing in mid register (E3–E5 zone), the synth pad in spread voicing across the full range, the lead vocal as the highest sustained melodic voice.

## Texture techniques in arrangement

### Melody + bass + chord (most common pop texture)

Bass on the bottom (single note line), chord in the middle (close voicing), melody on top.

Simple but effective. The sweet spot for pop, country, folk, rock.

### Melody + countermelody + bass

Add a second independent melodic line in a middle register. The melody and countermelody can be heard as dialogue.

Example: a violin melody on top, a cello countermelody in middle register, a bass below. Three independent voices.

### Pad + lead

In synth-based music: a sustained chordal pad fills the harmonic space, while a lead synth or vocal carries the melody. The pad is often spread across multiple octaves; the lead is in a clear single register.

### Layered texture

Multiple parts at different rhythmic densities:

- Sustained pad (very slow, sometimes one chord per bar)
- Mid-rhythm part (rhythm guitar at quarter notes or eighth notes)
- Active part (16th-note hi-hat or arpeggio)
- Lead (melody at variable density)

Each layer occupies a different rhythmic and registral space. Combined, they fill the texture without competing.

### Call and response

Two voices (or groups) alternate. Common in gospel, blues, Latin, African music. The two voices share the texture across time rather than simultaneously.

### Hocket

A medieval technique (and modern revival): a single melodic line is split between two voices, with each voice playing every other note. Creates a stuttering, distributed melodic line.

Modern uses: Steve Reich, prog rock, certain electronic music.

## Diagnosing texture problems

When the user says "it sounds wrong" texturally, common diagnoses:

| Symptom | Likely cause | Fix |
|---------|--------------|-----|
| "Muddy" | Too many low notes; close voicing in bass register | Spread the bass; remove inner voices below middle C |
| "Thin" | Not enough mid-range; too much top, too much bottom, gap in middle | Add a mid-register part (alto-range instrument or chord pad in mid) |
| "Cluttered" | Too many active parts at once | Thin out: have some parts hold while others move |
| "Lifeless" | All parts moving together rhythmically | Add rhythmic independence; let parts hold against moving parts |
| "Lopsided" | One side of the spectrum (high or low) dominates | Balance: if all parts are high, add bass; if all low, add top |
| "Doubling sounds smudged" | Two instruments doubling don't blend | Try a different doubling combination, or keep them at the octave instead of unison |
| "Melody can't be heard" | Melody in same range as accompaniment | Push the accompaniment to a different register, or thin its density |
| "Static / boring" | Texture doesn't change across sections | Vary texture: solo verse, full chorus, contrasting bridge |

## Texture by genre — quick reference

| Genre | Typical texture |
|-------|-----------------|
| Solo classical piano | Two-hand homophony; LH bass + chords, RH melody |
| Romantic symphony | Layered orchestral textures; melody + counter + harmony + bass |
| Choral SATB | Homophonic chordal or polyphonic |
| Bach fugue | Strict polyphony |
| Pop ballad | Melody + bass + chord; piano or guitar accompaniment |
| Modern pop production | Layered: kick/bass + rhythm guitar + pad + lead + atmosphere |
| Hip-hop | Beat (drums + bass) + sample/loop + vocal |
| Trap | Drum + 808 bass + melodic synth + vocal |
| EDM | Layered drums + bass + lead + pad + atmosphere |
| Jazz combo | Walking bass + drums + piano comping + horn melody |
| Big band | Rhythm section + horn section harmonized in 3-5 voices |
| Country | Vocal + acoustic guitar + bass + drums + (often) pedal steel/fiddle |
| Folk | Vocal + acoustic guitar (or piano), minimal additions |
| Indian classical | Drone + sitar/sarod (melodic) + tabla (rhythmic) |
| Gamelan | Heterophonic layers of metallophones at different rhythmic densities |

## Common pitfalls when advising

- **Don't recommend a "lush" voicing without specifying.** "Add lushness" could mean: spread voicing, octave doublings, more layers, modal interchange chords, etc. Be specific.
- **Don't ignore the genre's texture conventions.** A spread orchestral voicing in a country song would feel strange. A close jazz voicing in a film cue might feel cramped.
- **Don't pile up doublings without thinking.** Every additional doubling reduces clarity. Add doublings sparingly and with purpose.
- **Don't ignore the dynamic balance.** A muted trumpet doubling a forte string section will be inaudible. Match dynamics to the doubling intent.
- **Don't forget that texture changes across the song.** A song with the same texture throughout is monotonous. Plan texture variation as part of the form.
- **Don't conflate texture with arrangement density.** A song can have a thin texture (just two voices) with high density (very busy rhythm) or thick texture (many voices) with low density (sustained chords). Both are independent dimensions.

## Quick decision matrix

| Want | Try |
|------|-----|
| Intimate, exposed feeling | Monophony or sparse two-voice texture |
| Pop default | Melody + bass + chord (homophonic) |
| Big movie sound | Spread orchestral voicing across full range |
| Lush ballad | Close voicing in mid register + bass + sustained pad above |
| Modal jazz feel | Quartal voicings, McCoy-style stacked 4ths |
| Renaissance feel | Polyphonic SATB, equal voice independence |
| Cinematic chord | Spread voicing with octave doublings (low cello, mid winds, high strings, top piccolo) |
| Funk pocket | Tight close voicing on rhythm guitar/keys; bass independent; drums prominent |
| Synth pop | Layered: kick + sub bass + mid synth + arpeggio + lead + pad |
| K-pop chorus | Big spread voicing, vocal layering with harmonies, full instrumentation |
| Ambient | Sustained pad voicing, very slow harmonic rhythm, no rhythmic activity |

## Cross-references

- Instrument ranges and characters → `instruments-ranges-character.md`
- Density and how many parts at once → `arrangement-density.md`
- Choral writing (4-part vocal voicing, doubling, spacing) → `choral-writing.md`
- Chord spelling and qualities → `../harmony/chord-construction.md`
- Voice leading (chord-to-chord motion of voices) → `../harmony/voice-leading.md`
- Specific voicing shapes (drop 2, drop 3, rootless, etc.) → `../assets/jazz-voicings.md`
- Counterpoint → `../counterpoint.md`
- Production-side voicing decisions → `../production-aware/arrangement-for-mix.md`
- Genre-specific texture → various `../genres/*.md` files

# Reference Track Digging

> **Snapshot note:** Last content review snapshot: 2026-04-26. This is a method for turning reference tracks into compositional decisions without copying protected expression or flattening genre differences.

A reference track is not a command to imitate. It is evidence: a compact example of tempo, form, groove, harmony, melody, arrangement, production density, vocal delivery, and cultural context.

## When to use this file

Use this file when the user says:

- "Make it feel like [song/artist/genre]."
- "What should I reference for this style?"
- "I want current K-pop/J-pop/EDM/hip-hop/ballad energy."
- "Find me the sound of [scene/region/year]."
- "How do I study references without just copying?"

For current references, pair this with `web-music-trend-research.md`. If the user has a streaming service, public playlist, Replay/Wrapped summary, or Last.fm/scrobbling history, pair with `user-listening-context-and-streaming-services.md`. For living-artist, copyright, AI-assisted, sample, interpolation, or too-close imitation issues, pair with `style-reference-and-copyright.md`.

## Ethical and creative guardrails

- Extract **craft variables**, not melodies, lyrics, signature riffs, or distinctive recordings.
- Avoid imitating a living artist's exact style or vocal identity.
- If a reference is highly recognizable, use multiple references so the result is a blend, not a clone.
- For traditional or regional music, do not treat one track as the whole tradition. Study several examples and, when possible, work with practitioners.
- When the user asks for copyrighted lyrics, melody, or a too-close rewrite, redirect toward analysis and original alternatives.

## The three-layer reference stack

A strong reference set usually has three layers:

| Layer | Purpose | Example use |
|---|---|---|
| Anchor references | The closest sonic/emotional target | "This is the energy and form center." |
| Boundary references | Show what *not* to drift into | "This is too clubby / too rock / too sentimental." |
| Root references | Older or deeper idiom sources | "This explains where the groove or harmonic language comes from." |

For a hybrid style, build one stack per ingredient. Example: "K-pop with K-trad percussion and neo-soul harmony" needs K-pop anchors, K-trad rhythm/instrument roots, and neo-soul harmony references.

## Personal-library first pass

When the user's own taste matters, ask for lightweight streaming evidence before digging generic references:

- 5 anchor tracks they already love
- 2 boundary tracks that are close but wrong
- a public playlist link, if available
- a Replay/Wrapped/Last.fm summary, if they choose to share it

Then compare those with current platform or chart signals. User taste is the target; charts are context.

## Minimum viable reference set

For a fast composition answer:

1. **3 anchor tracks**
2. **2 contrast tracks**
3. **1–2 root/context tracks**

For a deeper style study:

1. 5–8 current tracks
2. 3 historical/idiom roots
3. 2 adjacent-market examples
4. 1 "wrong direction" example

## Listening pass template

For each reference, fill this out:

| Category | What to capture |
|---|---|
| Tempo / meter | BPM, swing/straight, half-time/double-time feel |
| Form | intro length, verse/pre/chorus/bridge layout, hook timing |
| Groove | kick/snare placement, hi-hat density, percussion pattern, bass rhythm |
| Harmony | key center, loop length, harmonic rhythm, modal/functional language |
| Melody / topline | range, contour, repeated cells, leaps, climax placement |
| Lyrics / prosody | syllable density, rhyme pattern, language-tone constraints, hook words |
| Arrangement | layer entry/exit, density arc, register spread, contrast between sections |
| Signature element | one element that makes it memorable |
| Transferable move | what can be borrowed abstractly |
| Do-not-copy element | exact melody, lyric phrase, riff, sample, vocal identity, or signature production tag |

## Reference matrix

Use a matrix to prevent shallow imitation:

```md
| Track | Role | Tempo | Form clue | Groove clue | Harmony clue | Topline clue | Arrangement clue | Transferable idea |
|---|---|---:|---|---|---|---|---|---|
| Track A | Anchor | 100 | hook by bar 9 | dembow-like | 4-chord loop | chant hook | sparse verse → dense chorus | early group-vocal hook |
| Track B | Boundary | 128 | drop-centered | four-on-floor | static | short phrases | club build | avoid this much EDM pressure |
| Track C | Root | 92 | call-response | traditional cycle | drone/modal | ornamented | live percussion | adapt call-response, not surface sample |
```

## How to dig by style

### Pop / K-pop / J-pop / C-pop

Look for:

- hook placement and section contrast
- pre-chorus lift strategy
- chorus density and group-vocal treatment
- dance-break or post-chorus role
- language-specific prosody and syllable density
- current chart/platform context

Pair with:
- `../genres/kpop-jpop.md`
- `../genres/cpop-and-southeast-asian-pop.md`
- `../fundamentals/prosody-and-language.md`
- `web-music-trend-research.md`

### Hip-hop / R&B / neo-soul

Look for:

- drum pocket and swing/microtiming feel
- bass/sub relationship
- sample/chop logic or interpolation risk
- harmonic color: minor 9, maj9, sus, planing, gospel moves
- vocal/rap flow density and rhyme placement

Pair with:
- `../genres/hip-hop-rnb.md`
- `../rhythm-groove/groove-and-feel.md`
- `../harmony/jazz-harmony.md`

### EDM / electronic

Look for:

- loop length and phrase grid
- build/drop/breakdown timing
- kick/bass relation as composition, not just mixing
- hook type: vocal chop, lead motif, bass riff, texture
- subgenre tempo and drum pattern

Pair with:
- `../genres/electronic-edm.md`
- `../production-aware/energy-and-dynamics.md`
- `../form/popular-song-forms.md`

### Film / game / media

Look for:

- dramatic function, not just genre
- theme transformability
- loopability / editability
- density under dialogue or gameplay
- stinger and transition language

Pair with:
- `../genres/film-tv-scoring.md`
- `../genres/game-music.md`
- `../genres/media-and-commercial-music.md`

### Traditional / regional fusion

Look for:

- the actual tradition being referenced
- rhythmic cycle, mode, ornament, language, and instrument idiom
- whether the source is concert, ritual, pop-fusion, film-score, or sample-pack mediated
- what requires a practitioner rather than a text description

Pair with:
- `../genres/korean-traditional.md`
- `../genres/folk-roots-and-traditions.md`
- `../techniques/microtonal.md`
- `../fundamentals/prosody-and-language.md`

## Turning references into original composition prompts

Bad prompt:

```text
Write a song like Track X.
```

Better prompt:

```text
Use Track X only as a craft reference:
- 96 BPM, minor-key dance-pop
- hook enters before 0:40
- verse is sparse and low-register
- pre-chorus rises by sequence
- chorus uses chantable 3-note hook
- avoid copying melody, lyric, or signature synth riff
Create a new 8-bar chorus progression and topline contour with similar energy.
```

## Quick anti-copy checklist

Before giving the user a generated musical idea, check:

- Is the melody a new contour, not the reference melody with small changes?
- Are lyrics original and not a paraphrase of a recognizable hook?
- Is the chord loop generic enough or transformed enough?
- Are signature riffs/samples/production tags avoided?
- Are references blended rather than copied one-to-one?
- Did you explain the transferable craft principle?
- If the request names a living artist/current hit, did you apply `style-reference-and-copyright.md`?

## Final advice

A reference track is most useful after it has been decomposed. The agent should answer with playable, concrete moves — chords, rhythm cells, contour, form plan — not vague adjectives like "make it more modern."

## Quick asset

- Fast web-search cheatsheet → `../../assets/web-search-cheatsheet.md`

# Motivic Development

A motif is the smallest melodic-rhythmic unit that carries musical identity — short enough that the ear can recognize it on return, distinctive enough that it doesn't blend into the background. Motivic development is the technique of growing a piece by transforming its motifs. It's how Beethoven builds movements from four-note cells, and how a film composer builds a 90-minute score from a single hummable theme.

## When to consult this file

- The user wants to know how to "develop" a melodic idea instead of just repeating it
- Composing a longer piece (sonata movement, film score, concept piece) where unity matters
- Writing variations on a theme
- Analyzing how a famous theme was developed
- Bridging from "I have a great hook" to "I have a coherent piece"

For melodic construction (creating a memorable line from scratch), see `melodic-construction.md`. For phrase grammar (how motifs combine into phrases), see `phrase-structure.md`. For variation as a formal type, see `techniques/theme-and-variation.md`.

## What is a motif

A motif (also "motive", or in Wagner-influenced contexts "leitmotif") is a short, recognizable unit. Three properties define motivic identity:

1. **Pitch shape (intervallic content and contour)** — the up-down profile and the size of intervals between notes.
2. **Rhythm** — the pattern of long and short, often more recognizable than the pitches.
3. **Harmonic context** — what chord(s) the motif sits over.

Some motifs are defined primarily by rhythm (Beethoven's "fate" motif: three short, one long; the pitches change but the rhythm is the identity). Others are primarily intervallic (the descending m3 of the "Star Wars" "Force theme"). Most are a combination.

A motif is *short* — typically 2–8 notes, 1–2 bars. Anything longer becomes a phrase or a theme.

A *theme* is a complete melodic statement, usually 4–16 bars, often built from motifs as building blocks. A theme can be developed by modifying it as a whole, or by extracting motifs from within it and developing those.

## Why develop motifs (and when not to)

Development creates *unity through variation* — the listener feels coherence (the motif keeps returning) and motion (the motif keeps changing). Without development, you have either:

- **Pure repetition** (the same motif over and over): can be hypnotic but quickly fatiguing
- **Pure variety** (constantly new motifs): incoherent; the listener has no anchor

Both have legitimate uses. Static dance music (techno, ambient) uses pure repetition for trance. Aleatoric and free-improv contexts use pure variety. Most other music sits between the extremes.

When **not** to develop:

- **Pop song verses and choruses** that are designed to be remembered exactly. A pop hook should return identical (or nearly so) to be memorable. Development happens at the *form* level (different sections), not within a single section.
- **Short pieces** where there's no room. A 30-second cue doesn't need development — just a clear statement.
- **Hypnotic / trance pieces** where stability is the point.

When **to** develop:

- **Longer instrumental forms** (sonata, fugue, theme-and-variations, film cue, prog rock).
- **Bridges and pre-choruses** in pop, where transformation of the verse motif creates lift.
- **Solos** in jazz — the soloist develops short ideas across the form.
- **Film scores** — leitmotifs change with narrative.

## Classical transformation operations

The classical toolkit for transforming motifs. These are formal operations; many are intuitive and have analogues in pop and jazz.

### 1. Repetition (exact)

The motif is restated unchanged. Anchors the listener. Usually the first transformation: a motif is presented, then immediately repeated. The repetition confirms motivic identity.

Beethoven's 5th: bars 1–4 are the motif; bars 5–8 are nearly exact repetition (a step lower).

### 2. Sequence (transposed repetition)

The motif is repeated at a different pitch level. This is the most common development tool — sequences appear constantly.

Sequence types:

- **Real sequence**: exact intervallic content preserved (chromatic transposition required to match exactly).
- **Tonal sequence**: stays within the diatonic key; intervals adjust as needed.
- **Modulating sequence**: the sequence itself causes a key change (each step lands in a new key).

A famous sequence: Pachelbel's Canon ground bass — the bass figure repeats while the upper voices sequence above it.

### 3. Inversion

The motif's intervals are reversed in direction. Where the original goes up by a 3rd, the inversion goes down by a 3rd.

Two types:

- **Strict inversion**: invert each interval exactly (a M3 up becomes a M3 down).
- **Tonal inversion**: invert within the key (a 3rd up becomes a 3rd down, but the quality may change).

Inversion is heavily used in Bach fugues and 20th-century serial music.

### 4. Retrograde

The motif's pitches are played in reverse order. Less common; usually felt by the analyst more than the listener (the ear doesn't easily recognize a sequence played backward).

### 5. Retrograde inversion

Both retrograde and inversion combined. A staple of 12-tone composition (the four forms of a tone row are: prime, retrograde, inversion, retrograde inversion).

### 6. Augmentation and diminution

**Augmentation**: each note's duration is multiplied (often doubled). The motif sounds slower and more grand. Common in fugue stretto, Romantic codas.

**Diminution**: each note's duration is divided (often halved). The motif sounds faster and more agitated. Common in development sections.

These rhythmic transformations preserve pitch content while changing temporal scale.

### 7. Fragmentation

Take a fragment of the motif (a few notes, a half) and develop it independently. Common in development sections of sonata-form movements — Beethoven's specialty. The 4-note "fate" motif is constantly fragmented down to 2 or 3 notes.

### 8. Extension and liquidation

**Extension**: the motif is extended at either end with new material. A 4-note motif becomes a 6-note phrase by adding two notes.

**Liquidation**: the motif is gradually stripped down to its bare essence (or even less). Used at endings to dissolve identity. The opposite of extension.

### 9. Contrast and combination

Two motifs in dialogue. Statement → response. Or two motifs combined in counterpoint (one in soprano, the other in bass simultaneously).

This is the engine of fugue (subject + countersubject) and of duets in opera and musical theatre.

### 10. Embellishment

Decorate the motif with passing tones, neighbor tones, ornaments, or rhythmic subdivisions. The motif is recognizable but more elaborate.

The opposite, **simplification**, strips ornaments to expose the bare motif. Common in classical recapitulation: the theme returns "purer".

### 11. Modal transformation

Recast the motif in a different mode. A major-key motif appears in parallel minor (or vice versa) at a key moment. Common in classical (recapitulation in parallel minor for tragic effect), film (motif darkens to indicate threat), and certain pop contexts.

### 12. Harmonic recontextualization

The motif's pitches stay the same, but the underlying harmony changes. Same notes, different chord — different feeling. A pop bridge often does this.

## Case study: Beethoven's 5th, opening movement

The motif: 𝅘𝅥𝅮𝅘𝅥𝅮𝅘𝅥𝅮𝅗𝅥 (three eighth notes plus a long tied note) — the "fate" motif. About 1 second long.

How it's developed:

- **Repeated** immediately at a step lower (sequence)
- **Repeated** through the whole orchestra in different registers
- **Sequenced** through different harmonies
- **Fragmented** to just three eighth notes when the motif's long note is dropped
- **Augmented** at climactic moments — three quarters and a half-note
- **Inverted** in some passages
- **Embellished** in the recapitulation
- **Combined** with secondary themes in counterpoint
- Overall: the motif appears in nearly every bar of the movement, in dozens of transformations. The unity is overwhelming, the variety inexhaustible.

This is the textbook example because Beethoven made the motif so concentrated and so transformable. Almost every classical-music textbook uses the 5th to teach motivic development.

## Pop and rock motivic development

Pop usually develops motifs at the form level (different sections) more than within a section. But within sections, the principles still apply:

### Verse-chorus contrast

The verse motif and the chorus motif are usually different in:

- Range (chorus is higher)
- Rhythm (chorus is often more sustained or anthemic)
- Density (chorus often has fewer syllables, longer notes)

These contrasts function as motivic contrast at the form level.

### Pre-chorus as motif transformation

The pre-chorus often takes the verse motif and transforms it — typically by:

- Raising the tessitura
- Sequencing upward
- Changing the harmonic rhythm (faster chord changes)
- Compressing the rhythm

This creates the feeling of "lift" before the chorus arrives.

### Bridge as motif development

The bridge often takes a fragment of the chorus motif (or an entirely contrasting idea) and develops it for 8–16 bars before returning to the chorus. This is the closest pop comes to classical development sections.

### Outros and post-choruses

Modern pop often features a "post-chorus" — a second chorus-like section with a new hook (sometimes wordless, sometimes a vocoded vowel). This is a third motivic identity that creates the song's signature.

### Classic pop examples

- "Hey Jude" — the famous "na na na" coda is liquidation of the song's main motif into pure chant.
- "Help!" — the verse motif and chorus motif are derived from each other through transformation.
- "Stairway to Heaven" — the iconic intro motif (the descending chromatic line) is extended throughout the song.

## Jazz motivic development in solos

Jazz soloists develop short ideas across the form, often subconsciously — but the great soloists (Sonny Rollins, Wayne Shorter, Wynton Kelly) are explicit motivic developers.

The discipline: take a 2–4-note phrase. Play it. Repeat it. Vary it. Sequence it. Fragment it. Combine it with another motif. The soloist who develops one idea for 32 bars often outshines the one who plays a shower of unrelated licks.

In jazz pedagogy, this is sometimes called "playing thematically" or "telling a story". Sonny Rollins's "Blue 7" is the classic study.

## Film leitmotif

A leitmotif is a recurring motif associated with a character, place, idea, or emotion. Wagner pioneered it in opera; it's now standard in film scoring.

Leitmotif development tracks narrative:

- **First appearance** — the motif is introduced cleanly, often with the first appearance of the associated character or idea.
- **Variation by context** — the motif is reharmonized, transposed, or rhythmically transformed as the narrative changes. The character grieves: motif becomes minor, slower. The character triumphs: motif becomes major, faster, scored for full orchestra.
- **Combination** — two leitmotifs combined indicate two characters or ideas in the same scene.
- **Liquidation** — the motif fragments and dissolves at the character's death or departure.
- **Final transformation** — the motif appears in its definitive form at the climax (often a triumphant restatement, occasionally a tragic one).

John Williams is the canonical modern practitioner. The "Star Wars" and "Harry Potter" scores are extended exercises in leitmotif development. Howard Shore's "Lord of the Rings" is even more elaborate, with dozens of interrelated motifs.

## When to use which transformation

A rough guide for what each transformation is best for:

| Want | Try |
|------|-----|
| Confirm motivic identity | Exact repetition |
| Build energy / motion | Sequence (especially upward) |
| Modulate gracefully | Modulating sequence |
| Tension / fragmentation feel | Fragmentation, diminution |
| Grand / climactic feel | Augmentation, full-orchestra restatement |
| Mystical / mirror effect | Inversion |
| Studious / academic puzzle | Retrograde, RI |
| Emotional darkening | Modal transformation (major to minor) |
| Lift before resolution | Embellishment, sequenced extension |
| Final dissolution | Liquidation |
| Unifying disparate sections | Combine/contrast motifs from each section |

## Common pitfalls when advising

- **Don't develop without first establishing the motif.** Development before identity creates confusion. Present the motif clearly, repeat it once or twice, *then* start transforming.
- **Don't stack too many transformations at once.** A motif that's simultaneously inverted, augmented, and reharmonized is unrecognizable. Apply transformations one at a time, layered across the piece.
- **Don't develop at the wrong scale.** Motivic development works at the cellular level. If your motif is 16 bars long, you don't have a motif — you have a theme. Find the cell within it.
- **Don't ignore the listener's tracking.** Development that's too dense (a new transformation every bar) loses the listener; the motif's identity gets buried. Pace transformations.
- **Don't force development on music that doesn't want it.** Trance, ambient, and certain pop genres benefit from stasis. If the user's piece is in one of those modes, development is the wrong tool.
- **Don't confuse variation with development.** Variation (theme-and-variation form) keeps the entire theme as the unit. Development extracts and transforms motifs from the theme. They're different techniques. See `techniques/theme-and-variation.md`.

## A workflow for motivic-development pieces

If the user wants to write a piece organized around motivic development:

1. **Compose a motif.** 2–8 notes, with distinctive rhythm and intervallic shape. Don't make it too long.
2. **Test it for transformability.** Try inverting it, fragmenting it, augmenting it. If at least three transformations sound good, proceed. If everything sounds awkward, try a different motif.
3. **Plan the form.** Decide where the motif will appear: theme statement → first development → contrasting section → second development → climax → return.
4. **Sketch transformations.** For each section, decide which transformation fits the dramatic moment.
5. **Write transitions.** Sequencing the motif through different keys is the most common way to bridge sections.
6. **Save the most powerful transformation for the climax.** Often this is augmentation with full instrumentation, or a return to the original form after extensive development.

## Cross-references

- The motif within a single phrase → `melodic-construction.md`
- Phrase grammar (motif → phrase → period) → `phrase-structure.md`
- Theme-and-variation as a form (different from motivic development) → `techniques/theme-and-variation.md`
- Sonata form, where motivic development is structural → `form/classical-forms.md`
- Pop song bridges and pre-choruses (where development happens in pop) → `form/popular-song-forms.md`
- Leitmotif in film → `genres/film-tv-scoring.md`
- Jazz solo construction → `genres/jazz-styles.md`, `harmony/jazz-harmony.md`
- 12-tone and serial transformation operations → `techniques/20th-century-techniques.md`
- Counterpoint (combining motifs simultaneously) → `counterpoint.md`

# Arrangement Density

This file is about *how much* is happening at any given moment — how many instruments, how busy each one is, how full the texture is. Density is one of the primary dimensions of energy and a key tool for sectional contrast. Two pieces with identical melody and harmony can feel completely different based on density choices.

For voicing within a single chord, see `voicing-and-texture.md`. For instrument selection, see `instruments-ranges-character.md`. For energy arcs across a piece (where density fits), see `form/narrative-and-transitions.md`. For production-side density (mix-aware density), see `production-aware/arrangement-for-mix.md`.

## When to consult this file

- The user has a piece that feels "samey" or "doesn't go anywhere" — usually a density issue
- "How do I make the chorus hit harder" — density transition
- A track sounds "too cluttered" or "too empty"
- Layering decisions: when to add elements, when to remove them
- Building dramatic arcs across long pieces
- Sectional contrast in any genre

## What "density" means

Density is the amount of musical activity at a given moment. It's a composite of:

1. **Number of instruments / parts simultaneously sounding** — more parts = denser
2. **Rhythmic activity per part** — more notes per bar = denser
3. **Frequency-spectrum coverage** — more registers occupied = denser
4. **Dynamic level** — louder = often perceived as denser (though they're not strictly the same)
5. **Sustain vs. articulation** — sustained sounds fill space; staccato notes leave gaps

Density is largely independent of voicing and harmony. A single instrument playing a fast line at high dynamics can feel denser than three instruments playing slow whole notes.

## The dynamic range of density

Think of density on a 0-to-10 scale:

| Density | Description | Examples |
|---------|-------------|----------|
| 0 | Silence | The bar before a drop; pauses; rests |
| 1 | One instrument, one note (held) | Solo flute on a sustained pitch; opening of a piece |
| 2 | One instrument, simple melody | Solo voice, a cappella; solo guitar fingerpicking |
| 3 | Two parts, sparse | Voice + acoustic guitar; piano LH/RH ballad opening |
| 4 | Small group, restrained | String quartet at piano dynamic; chamber jazz trio |
| 5 | Modest band/ensemble at moderate intensity | Pop verse with full band but moderate dynamics |
| 6 | Full ensemble, mid-energy | Symphonic mid-passages; pop chorus at moderate dynamics |
| 7 | Full ensemble, high energy | Pop chorus with strong drums and full vocals |
| 8 | Heavy, multiple layers, near-peak | EDM drop; rock chorus with double-tracked guitars; full orchestra at fortissimo |
| 9 | Maximum density without overload | Cinematic climax with full orchestra, choir, percussion battery |
| 10 | Sonic overload | Wall-of-sound; controlled chaos; noise music; saturated mixes |

Few pieces use the full range. A typical pop song might span 2–8. A film score might span 1–9. An ambient piece might stay between 1–4 throughout.

## The arc of density across a piece

Density should change with the form. The standard pattern: density rises into peak moments and falls in their wake.

### Pop song density arc

```
Intro:        2-3 (sparse, intriguing)
Verse 1:      4-5 (mid-density establishing)
Pre-Chorus:   5-6 (building)
Chorus:       7-8 (peak)
Verse 2:      4-5 (slight density drop, often more layers than V1)
Pre-Chorus:   5-6
Chorus:       7-8 (similar peak; sometimes higher)
Bridge:       3-5 (reset; often dramatically thinner)
Final Chorus: 8-9 (highest; everything in)
Outro:        4-2 (release; possibly tag/fade)
```

### EDM density arc

```
Intro:        1-3 (atmospheric, beat only)
Build:        3-7 (rises continuously)
Drop:         9 (peak)
Breakdown:    2-5 (stripped back)
Build 2:      4-7
Drop 2:       9
Outro:        4-1 (fading)
```

### Film cue density arc

Variable, but typically:

```
Cold open:    1-2 (atmospheric pad)
Statement:    3-5 (introduce theme)
Development:  4-7 (build complexity)
Climax:       8-9 (full orchestra)
Resolution:   3-5 (back to thinner)
```

The principle: density should always be moving, even if subtly. Static density across a long section feels "stuck".

## Layering — adding density gradually

Building density by progressively adding layers. The opposite of starting big and thinning out.

### Standard pop layering

A typical pop song layers like this:

| Section | Adds |
|---------|------|
| Intro | Vocal hook or instrumental motif (1-2 instruments) |
| Verse 1 | Voice + acoustic guitar / piano + simple kick/snare |
| Verse 1 second half | Add bass; add hi-hat |
| Pre-Chorus | Add pad; add tom fills toward chorus |
| Chorus | Full drums; add electric guitar; backing vocals; lead lift |
| Verse 2 | Drop most of the chorus elements; keep vocals + minimal accompaniment |
| Verse 2 second half | Quickly re-add layers (faster build than V1) |
| Chorus 2 | Same as chorus 1 with slight variations |
| Bridge | Strip back; new sound element if possible |
| Final chorus | Everything in: choir, double-tracked vocals, full instrumentation |

The principle: the chorus is the peak. Verse 2 should re-establish lower energy, even if it has more elements than Verse 1, so the chorus can feel like an arrival.

### EDM layering

EDM layering is more disciplined and explicit:

| Bar | Element |
|-----|---------|
| 1 | Atmospheric pad |
| 5 | Add kick |
| 9 | Add bass |
| 13 | Add melodic synth |
| 17 | Add hi-hat 16ths |
| 21 | Add second melody / arpeggio |
| 25 | Add snare on 2/4 |
| 29 | Add clap layer; rolls begin |
| 31-32 | Snare roll into the drop |
| 33 | Drop: full instrumentation |

Each layer announces itself; the listener feels the build because they hear new sounds entering at predictable moments.

### Orchestral layering

In orchestral writing, layering happens through orchestration:

- Start with strings (or solo woodwind).
- Add winds for color and harmonic support.
- Add brass for power.
- Add percussion for rhythm and accents.
- Add timpani for foundation.

The order of entry corresponds to the building of intensity. Brass entries are particularly dramatic because they're so distinctive in timbre.

## Thinning — reducing density

Removing layers is just as important as adding them. Where to thin:

### Bridge / breakdown

After the second chorus or in the middle of a piece, thinning creates space for emotional contrast. A thinned bridge feels like the song "looking inward" before the final climax.

Common bridge thinning:

- Drop drums entirely; just vocal + sustained chord pad
- Strip to acoustic instruments only
- Vocal a cappella for a few bars
- Single instrument carrying the harmonic foundation

### Verse 2 (vs. Verse 1)

Modern pop often pulls back at Verse 2 — even though more elements have been introduced, the verse stays thin to allow the chorus to feel like an arrival again.

### Before a build

EDM and film often have a moment of silence or near-silence before a build begins. The "breath" before tension accumulates.

### After a peak

Coda or outro thinning. The piece releases tension by progressively removing layers.

## The "and then nothing" technique

A specific thinning gesture: at the moment of peak tension (just before a drop or climax), thin to silence or near-silence. Maximum contrast for the entry.

Examples:

- The bar before an EDM drop: white noise wash + vocal cue, then silence, then the drop
- A film score moment: full orchestra builds, sudden silence, then a single quiet violin (or the full impact of the next moment)
- Pop: chorus builds, last bar has just vocal (no drums), then full chorus repeat

The brief silence is the comma before the exclamation point. Without it, the climax has nothing to push against.

## Density and rhythmic activity

The "rhythmic density" of each part matters as much as how many parts there are.

A single instrument playing 16th notes for 4 bars produces a high-density feel even with one part. Conversely, 5 instruments playing whole notes can feel quite sparse.

When designing density:

- **Choose how busy each layer is.** A pad sustains; an arpeggio has 16th-note density.
- **Use rhythmic complementarity.** When one layer is busy, others should hold (give space). When one layer holds, others can be busy.
- **Match rhythmic density to function.** The lead vocal usually has moderate density (syllable-based); the rhythm section has high density (16th notes); pads have low density (held chords).

## Frequency-spectrum density

Density also depends on which parts of the audio frequency spectrum are filled.

A typical pop track wants:

- **Bass** (40–200 Hz): kick + bass guitar / sub bass.
- **Low-mid** (200–500 Hz): rhythm guitar lower notes; piano LH; bass overtones.
- **Mid** (500 Hz–2 kHz): vocals; rhythm guitar mid; lead synths; horns.
- **High-mid** (2–6 kHz): vocal presence; cymbals; lead vocal harmonies; lead synths.
- **High** (6–20 kHz): hi-hats; cymbals; air; reverb tails.

A track with all these zones filled feels "full". A track with gaps in the spectrum feels "thin" or "muddy" depending on which zones are missing.

For a more detailed treatment of frequency-aware arrangement, see `production-aware/arrangement-for-mix.md`.

## Density mistakes — common patterns

### Too much, too soon

The intro is at chorus density. The chorus has nothing higher to go. The piece is at peak intensity for its entire length and exhausts the listener.

Fix: start lower. Let the chorus feel like an arrival.

### Static density

The whole piece sits at the same density. No build, no relief. Listeners track music partly through density variation; without it, attention drifts.

Fix: introduce density variation between sections — even a 1-step difference between verse and chorus is felt.

### Density without dynamics

Lots of layers but no dynamic variation across sections. The chorus has the same volume as the verse. The listener perceives less contrast than the arrangement suggests.

Fix: align dynamic and density variation. Loud sections should also be denser; soft sections should also be thinner.

### Too many simultaneous busy parts

Three layers all playing 16th notes at high register. The listener can't track any of them. The texture becomes "white noise".

Fix: rhythmic complementarity. Have some layers hold while others are busy.

### Wrong frequency distribution

All the parts are in the same register (e.g., all in the 200–500 Hz zone). The mix sounds "boxy" or "muddy".

Fix: spread parts across the spectrum. Each layer should occupy a different zone.

### No silence

The piece never lets up. No breath, no rest. Listeners need brief silences to refresh their attention.

Fix: build in moments of reduced density — a beat or two of just bass; a sustained chord with no rhythm; an empty bar before a chorus.

## Density across long forms

For pieces longer than 4 minutes, density has to work at the macro scale.

### Multi-movement classical works

Each movement has its own density profile. The contrast between movements (e.g., dense first movement → sparse slow movement) is part of the work's design.

### Albums

A pop album varies density across tracks. An entire album of high-density tracks is exhausting; one with all sparse tracks lacks impact. The standard pacing: opener (mid-high), singles (high), middle ballad (low), closer (mid).

### Long-form EDM tracks

A 7-minute club track has multiple density peaks (drops) separated by breakdowns. The pattern is: build → drop → breakdown → build → drop → breakdown → outro. Each breakdown is a deliberate thinning to make the next drop hit.

## Density and listener fatigue

Listeners can sustain high density for limited durations. Approximate guidelines:

- **30 seconds at maximum density** is comfortable.
- **60 seconds** starts to push attention.
- **90+ seconds** at peak density — the listener tunes out.

The exception: if the high density is the *aesthetic* (drone, noise, certain ambient), the listener has different expectations. But for narrative music, density relief is essential.

## Density by genre — quick reference

| Genre | Typical density range | Notes |
|-------|----------------------|-------|
| Solo classical (e.g., piano sonata) | 2–5 | Density comes from rhythmic activity, not instrument count |
| Symphonic classical | 2–9 | Wide range; density is structural |
| Choral a cappella | 3–6 | All density from vocal layering |
| Chamber jazz | 3–6 | Sparse to moderate; never extreme |
| Big band jazz | 5–8 | Mid-high; section playing creates fullness |
| Pop ballad | 2–7 | Sparse verses, dense choruses |
| Modern pop | 4–8 | Heavily produced; rarely below 4 |
| Rock | 5–8 | Solid density throughout; chorus typically denser |
| Hip-hop / trap | 3–7 | Beat-focused; vocal density adds layers |
| EDM | 2–9 | Maximum range; drop is peak, breakdown is deep low |
| Folk / singer-songwriter | 2–4 | Intentionally sparse |
| Ambient | 1–3 | Sparse throughout; texture from sustain not activity |
| Film score (action) | 4–9 | High-end heavy; often peaked |
| Film score (drama) | 2–6 | More restrained |
| Trailer music | 2–9 | Builds from minimal to maximum |
| K-pop | 4–9 | Highly produced; multiple peaks per song |
| Musical theatre | 2–8 | Variable; matches dramatic moment |

## Common pitfalls when advising

- **Don't recommend "more layers" without thinking about which layers.** Adding a pad helps; adding another rhythm guitar in the same register doesn't. Density needs spectral and rhythmic variety.
- **Don't ignore subtractive arrangement.** Sometimes the fix isn't adding more — it's removing what's there. A muddy track often improves when an instrument is removed, not added.
- **Don't equate density with quality.** A sparse track can be more powerful than a dense one. Density is a tool, not a goal.
- **Don't apply EDM build-density logic to non-EDM.** A folk song doesn't need a drop. A classical adagio doesn't want layers added every 8 bars.
- **Don't ignore the contrast principle.** Density only feels meaningful when it changes. A song that's at density 7 throughout has no peaks; the listener perceives no chorus.
- **Don't add density without checking the mix.** A track with too many simultaneous layers becomes muddy in the mix — frequency masking destroys the clarity that the layers were supposed to add.

## Quick decision matrix

| Want | Try |
|------|-----|
| Build a chorus that hits | Verse density 4–5; chorus density 7–8; clear transition through pre-chorus |
| Reset before a final chorus | Bridge density drops to 3–4; final chorus jumps to 8–9 |
| EDM-style drop | Strip to atmosphere (1–2) for one bar before drop; drop at 9 |
| Cinematic build | Each phrase adds one element; final passage has full orchestra |
| Hypnotic, meditative feel | Stay around density 2–3 for entire track; subtle layer additions only |
| Aggressive, in-your-face | Density 7–9 throughout; rare moments of relief |
| Sparse, intimate ballad | Density 2–4 throughout; chorus reaches 5 at most |
| Pop song with multiple peaks | Density profile: V1 (4) → C1 (7) → V2 (5) → C2 (7) → Bridge (3) → Final C (8) |
| Pre-chorus build | Add 1 layer per 4 bars: pad, then drum fill, then rolls |
| Outro release | Strip layers progressively; end on 1–2 elements |

## Cross-references

- Voicing within chords → `voicing-and-texture.md`
- Instrument selection → `instruments-ranges-character.md`
- Energy arcs and transitions → `form/narrative-and-transitions.md`
- Production-side density and mixing → `production-aware/arrangement-for-mix.md`, `production-aware/energy-and-dynamics.md`
- Section-level density patterns → `form/popular-song-forms.md`, `form/classical-forms.md`
- Genre-specific density conventions → various `genres/*.md` files
- Rhythmic density (which is part of overall density) → `rhythm-groove/groove-and-feel.md`

# Energy and Dynamics

This file is about how energy is shaped at the *production* level — through dynamics, processing, automation, and the integration of compositional choices with mix decisions. It complements `form/narrative-and-transitions.md` (which covers energy at the form level) and `orchestration/arrangement-density.md` (which covers density at the arrangement level).

In modern recorded music, much of what listeners feel as "energy" comes from production decisions — how loud the kick is, how the vocal sits in the mix, how processing shapes attack and sustain — as much as from compositional choices. The composer/arranger should understand the production layer to make compositional choices that survive (or thrive) at the production stage.

For composition-level energy arcs, see `form/narrative-and-transitions.md`. For density choices, see `orchestration/arrangement-density.md`. For arrangement-mix integration, see `arrangement-for-mix.md`.

## When to consult this file

- The user's composition feels "less powerful" than reference tracks despite having the same notes
- Questions about loudness, dynamic range, "bigness"
- Production-side techniques that affect perceived energy
- Why a song "doesn't punch" — often a production-energy issue
- Modern loudness conventions
- Sectional dynamics in production

## What "energy" means in production context

At the production level, energy is felt through:

- **Loudness**: average level of the track
- **Dynamic range**: difference between loud and quiet moments (peak vs. average)
- **Spectral fullness**: how much of the frequency spectrum is occupied
- **Transient strength**: how punchy individual hits are (kick, snare, vocal consonants)
- **Sustain and decay**: how long sounds linger (reverb, sustain pedals, synth release)
- **Compression**: how "tight" or "loose" the dynamics feel
- **Automation**: how parameters change over time (volume rides, EQ shifts, filter sweeps)

Two productions of the same composition can feel completely different in energy depending on these choices.

## Loudness conventions and the loudness wars

Modern recorded music is generally louder than music from 20 or 50 years ago. Average levels (measured in LUFS — Loudness Units Full Scale) for various genres:

| Era / Genre | Typical LUFS |
|-------------|--------------|
| Classical (acoustic recording) | -23 to -16 LUFS |
| 1970s rock / pop (analog) | -14 to -10 LUFS |
| 1990s pop | -10 to -8 LUFS |
| 2000s rock | -8 to -6 LUFS (peak loudness wars) |
| Modern pop / hip-hop | -8 to -10 LUFS |
| Streaming-normalized (Spotify, Apple Music) | -14 to -16 LUFS |

The "loudness wars" peaked in the 2000s when releases competed for being the loudest. Streaming services now normalize playback to consistent loudness, reducing the incentive to over-compress.

For the composer/arranger: the *arrangement* should support whatever loudness the production targets. Modern productions need wide spectral coverage, controlled dynamics, and clear separation to compete.

## Dynamic range — the macro perspective

The difference between a track's loud and quiet moments. Modern pop typically has a dynamic range of 6–10 dB (relatively compressed); classical music can have 30+ dB (extreme dynamic range).

Compressed dynamic range (modern pop):

- Pros: consistent listening experience; no need to adjust volume; works on phone speakers
- Cons: less impactful contrast; less "release" between sections; ear fatigue

Wide dynamic range (classical, audiophile rock):

- Pros: dramatic contrast; intimate quiet moments make loud moments massive
- Cons: harder to listen to in noisy environments; phone speakers struggle with quiet content

The composer/arranger's choice influences this:

- **Sparse verses, dense choruses** create dynamic contrast at the arrangement level.
- **Quiet bridges** allow final-chorus impact.
- **Tight arrangement throughout** results in compressed dynamics; "wide" arrangement allows wider dynamics.

## Compression — what it does, what to know

Compression reduces dynamic range — quiet sounds get louder, loud sounds get softer. Used at multiple stages:

- **Bus compression** (mix bus): glues the mix together
- **Track compression** (individual tracks): tightens each instrument's dynamics
- **Mastering compression**: final output stage

For the composer/arranger:

- A piece designed with massive dynamic contrast will get compressed in mastering, reducing some of that contrast.
- Tight, in-pocket performances compress better than loose, varying ones.
- Modern production assumes some compression at every stage; the arrangement should not depend on extreme dynamic range (if it must, the producer needs to know).

## Sidechaining

A specific compression technique: one element's volume is controlled by another's. The most common: bass ducks (drops in volume) every time the kick hits.

Effect: clearer kick + clearer bass. Also creates a "pumping" feel that's the signature sound of modern EDM and house.

For the composer/arranger:

- If sidechaining will be used, design the kick and bass to support it.
- Sidechained pads create a "breathing" quality — common in modern pop.
- Heavy sidechain pump is an EDM aesthetic; light sidechain is invisible mix help.

## The kick — production-side considerations

A kick drum is the foundation of most modern productions. Production decisions:

- **Layered kick**: a sub-bass layer + transient/click layer. Common in hip-hop, EDM.
- **Tuned kick**: kick pitched to the song's key, so it harmonizes with the bass.
- **Length**: short kick = tighter feel; longer kick = more weight but slower groove.
- **Saturation / distortion**: adds harmonics to the kick, making it cut through.

A bare kick (no production) sounds "small". Production makes a kick "big". The composer/arranger should know the kick will likely be processed; design the rhythm to take advantage.

## Vocal production

The vocal carries most pop songs. Production-side decisions:

- **Compression**: heavy compression keeps levels consistent. Modern pop vocals are usually heavily compressed.
- **De-essing**: reduces sibilance (harsh "s" sounds).
- **EQ**: brightens high-mids for clarity; cuts low-mids for "warmth control".
- **Reverb**: adds space; modern pop uses subtle reverb.
- **Doubles**: vocal recorded multiple times, layered for thickness.
- **Auto-Tune / pitch correction**: nearly universal in modern pop. Sometimes obvious (T-Pain, Kanye effect), sometimes invisible.

For the topliner: design vocal lines that survive and benefit from this processing. Hold notes that compress well; phrasing that cuts through the mix; emotional moments that the production can support (rather than fight).

## Reverb and delay — atmosphere

Reverb (reflection) and delay (echo) create space and atmosphere.

For arrangement:

- **Heavy reverb on lead vocal** = "epic" feel, but can also blur clarity.
- **Tight reverb on drums** = punchy, modern.
- **Long delay** = atmospheric, often used on pads or guitar leads.
- **Slap delay (quick echo)** = vintage feel (rockabilly, old country, '50s vocals).
- **No reverb** = very dry; intimate or aggressive.

The composer/arranger's choice of which elements get heavy ambience and which stay dry shapes the track's "depth".

## Saturation and distortion

Adding harmonic content to a sound. Subtle saturation thickens; heavy distortion characterizes (rock guitars, gritty bass, processed vocals).

Production saturation:

- **Tape emulation**: warm, vintage feel
- **Tube saturation**: rich harmonics
- **Hard distortion**: aggressive character
- **Bit crushing**: lo-fi degradation

Composers should anticipate which elements will be saturated. A clean piano arrangement won't survive heavy saturation on the master.

## Automation — energy through time

Automation = parameters changing over the course of the track. Common automations:

- **Volume**: making elements get louder or softer over time (especially for builds and breakdowns).
- **Filter cutoff**: a low-pass filter "opening up" through a build, releasing on the drop.
- **Reverb send**: increasing reverb during a transition for dramatic effect.
- **Pan**: moving an element across the stereo field.
- **EQ**: shifting the spectral profile of an element across sections.

For the arranger: design parts that will *want* automation. A pad that stays the same throughout the song is less interesting than one whose filter slowly opens through the build.

### EDM build automations

A standard EDM build uses several automations together:

- Filter cutoff opening (closed → open)
- Volume swelling
- Pitch rising (a "riser" synth)
- Reverb increasing
- Delay feedback increasing

The combined effect is overwhelming tension that resolves into the drop.

## Sectional energy mapping

Map the production decisions across the track's form:

| Section | Production approach |
|---------|---------------------|
| Intro | Spare; let elements introduce themselves. Sub bass usually held back. |
| Verse 1 | Vocal up front; minimal accompaniment; some reverb space. |
| Pre-chorus | Things start to "pull forward" — drums tighten, atmosphere thins, vocal energy rises. |
| Chorus | Full production — kick, bass, vocals layered, full spectrum. |
| Verse 2 | Pull back from chorus, but slightly more layered than V1 (a percussion element, a synth pad). |
| Pre-chorus 2 | Similar to first PC, possibly more intense. |
| Chorus 2 | Same as C1, possibly with extra ad-libs or background vocals. |
| Bridge | Major reset — strip layers; possibly different processing (vocoder, filter, etc.); space for emotional moment. |
| Final chorus | The peak — everything in, biggest production, possibly modulation up, additional vocals/strings/whatever. |
| Outro | Release the tension — fade some elements, end on a tag or fade-out. |

This mapping is one example; specific genres vary.

## Genre conventions for production energy

### Pop

- Smooth dynamic profile; choruses ~3 dB louder than verses on average.
- Heavy vocal compression; subtle saturation.
- Sidechaining for mid-range elements (pads, synths) to make space for kick.
- Modern pop is bright, polished, high-frequency-rich.

### Rock

- Wider dynamic range than pop, but still compressed by classical standards.
- Distorted electric guitars carry much of the mid-range energy.
- Drums often sound natural (less processed than pop); the "rock" sound is acoustic-feeling.
- Vocal effects vary by sub-genre.

### Hip-hop

- Beat-focused energy; kick + 808 are the foundation.
- Vocal compression heavy; rap vocals up-front.
- Sample-based productions inherit the energy of the original recording.
- Modern trap has aggressive sub-bass, sharp 808 transients.

### Electronic / EDM

- Full spectrum at the drop; everything from sub to air.
- Heavy sidechaining for "pumping" feel.
- Builds use filter automation, white noise rises, drum rolls.
- Drops are deliberately maximum-loudness moments.

### R&B / neo-soul

- Warm, less compressed than pop.
- Vocal compression but with breath and air preserved.
- Bass is round and present.
- Less aggressive transient processing; "smoother" overall.

### Folk / indie

- Minimal processing; intimate feel.
- Vocal close-mic'd and natural.
- Acoustic instruments preserved; "live" feel.
- Subtle saturation; less compression than commercial pop.

### Classical / film score

- Wide dynamic range preserved.
- Natural acoustic recording approach.
- Less dependent on production manipulation than commercial pop.
- Mix focuses on capturing the ensemble's natural balance.

## Common pitfalls when advising

- **Don't ignore the production layer.** The composer/arranger's work is incomplete without considering production. Modern recorded music *is* a production.
- **Don't compose for an "ideal" mix you'll never have.** If your collaborator/producer can't deliver heavy production, write music that doesn't need it.
- **Don't fight the genre's production conventions.** A rock track with EDM-heavy sidechaining sounds wrong. A pop track without compressed vocals sounds amateur.
- **Don't ignore dynamic range strategy.** A song with no dynamic contrast at the arrangement level can't be saved by clever production. Build the contrast in.
- **Don't over-rely on production.** Great production can elevate a great song; it can't make a weak song great. The composition matters.
- **Don't ignore loudness norms.** Modern streaming normalizes playback; over-compressing your master makes it sound *quieter* relative to better-mastered tracks.
- **Don't conflate energy with volume.** A truly powerful moment uses dynamics, not just loudness. The contrast with quieter moments is what makes the loud feel big.

## Quick decision matrix

| Want | Try (production-side) |
|------|----------------------|
| Big drop in EDM | Build with filter automation, riser synth, snare roll, white noise wash; drop with full spectrum, sidechained pumping |
| Punchy modern pop | Compressed vocal upfront; sidechained pads; tuned kick; tight stereo wide on chorus |
| Vintage rock energy | Less compression; some natural dynamic range; doubled guitars hard-panned; live-sounding drums |
| Hip-hop heavy | 808 sub bass; tuned kick layered with sub; vocal up-front and dry; minimal atmosphere |
| Cinematic build | Wide dynamic range; orchestra ramps from pp to ff; reverb tails grow into climax |
| Subtle dynamic shape | Verse to chorus density change at the arrangement level; minimal compression; let dynamics breathe |
| Loudness vs. dynamics balance | Target -10 to -14 LUFS for pop; preserve some dynamic range while still being competitive |
| Lush ballad | Heavy reverb on lead vocal; sustained pads automated to swell; soft drums with natural ambience |
| Lo-fi / intimate | Minimal compression; subtle saturation; no de-essing; preserve room sound |
| Modern radio-ready | Tight vocal compression; sidechain pumping; mid-range carving; spectrum filled top to bottom |

## Cross-references

- Energy at the composition level → `form/narrative-and-transitions.md`
- Density and layering → `orchestration/arrangement-density.md`
- Mix-aware arrangement (frequency, stereo) → `arrangement-for-mix.md`
- Pre-production decisions that set the energy ceiling (tempo, length) → `pre-production-decisions.md`
- Genre production conventions → `genres/pop-rock.md`, `genres/hip-hop-rnb.md`, `genres/electronic-edm.md`, `genres/kpop-jpop.md`, `genres/film-tv-scoring.md`
- Form-level energy → `form/popular-song-forms.md`, `form/classical-forms.md`
- Vocal production considerations → `songwriting/topline-craft.md`
- Rhythm and groove (informs production feel) → `rhythm-groove/groove-and-feel.md`

# Choral Writing

This file covers writing for choir — the SATB (and adjacent) tradition of unaccompanied or lightly-accompanied vocal ensemble. Choral writing has its own conventions, idiomatic constraints, and aesthetic priorities, distinct from solo vocal writing or instrumental ensemble writing. The voice is the most expressive musical instrument, but it's also the most constrained: choral textures must respect range, breath, vowel singability, and the human ear's pitch tolerances when no accompaniment is fixing the tuning.

This file complements `voicing-and-texture.md` (general voicing principles), `instruments-ranges-character.md` (where solo voice ranges live), `../counterpoint.md` (much choral writing is contrapuntal), and `../harmony/voice-leading.md` (the voice leading conventions choral writing is built on).

For solo vocal writing in pop / R&B / jazz, see `../songwriting/topline-craft.md`. For musical theatre choral conventions, see `../genres/musical-theatre.md`. For lyric prosody (which choral writing intensifies), see `../fundamentals/prosody-and-language.md`.

## When to consult this file

- Writing for SATB choir (sacred, secular, or contemporary)
- Writing for any vocal ensemble (chamber choir, double choir, men's / women's chorus)
- Choral arrangement of an existing tune
- Choral texture inside a larger work (orchestral chorus, opera chorus, film cue chorus)
- A cappella arranging
- Worship music writing
- Diagnosing why a choral piece "doesn't sing well" or "sounds rough"

## Voice types and ranges

The four standard SATB sections, with their working ranges. *Working* range means what an average church-choir singer can sing comfortably; soloists go further.

| Voice | Working range | Comfortable middle | Top (with effort) |
|-------|---------------|---------------------|--------------------|
| Soprano (S) | C4 – G5 | F4 – D5 | A5 – C6 (in soloist territory) |
| Alto (A) | G3 – D5 | C4 – A4 | E5 – F5 |
| Tenor (T) | C3 – G4 (notated octave higher in treble clef) | F3 – D4 | A4 – C5 |
| Bass (B) | E2 – C4 | A2 – F3 | G3 – A3 (and rarely D2 in soloist or specialist roles) |

Notes on these ranges:

- Tenor parts are conventionally written in treble clef sounding an octave lower than written. A C5 in tenor part-notation sounds at C4. (Some scores use an "octave G clef" — treble clef with a small "8" below — to make this explicit.)
- The boundary between alto and tenor is fluid: A3–D4 is the overlap zone, where alto top and tenor middle meet.
- Modern professional choirs often subdivide further (S1/S2, A1/A2, T1/T2, B1/B2) — known as SSAATTBB or "divisi". Don't divide unless the choir can support it.
- The "passaggio" — the register break around D4–F4 for sopranos, A3–C4 for altos, A3–C4 for tenors, D3–F3 for basses — is where vocal mechanism shifts. Sustained notes parked at the passaggio are uncomfortable; melody can pass through but shouldn't camp.

For solo voice ranges (which extend further), see `instruments-ranges-character.md`.

## Tessitura is the dominant constraint

A choral part's quality depends less on its highest or lowest note than on where it *lives* — its **tessitura**, the band of pitches it spends most of its time in.

A soprano part that hits a single A5 climax is fine; a soprano part that lives at A5 for sixteen bars is exhausting. A bass part that descends to E2 once for color is fine; a bass part that stays below G2 throughout is muddy and tiring.

Practical guideline: each part's tessitura should be in its *comfortable middle* most of the time. Climaxes can push to the top; quiet sustained passages can dip toward the bottom. Constant living at the extremes wears out singers and degrades sound.

## Doubling and spacing

When the four voices form chords, voicing decisions matter as much as in keyboard / orchestral writing — but with vocal-specific considerations.

### Doubling preferences (root-position triads)

| Doubling | When |
|----------|------|
| Root doubled | The default for major / minor triads in root position. Solid, balanced. |
| 5th doubled | Acceptable; produces an open, neutral sound. Common in early music. |
| 3rd doubled | Avoid in major / minor triads if the 3rd is a tendency tone (e.g., the leading tone in V). Acceptable in some contexts (sustained tonic, deliberate color). |
| Leading tone doubled | Almost never. The leading tone wants to resolve up; doubling forces parallel octaves on resolution. |

For diminished and augmented triads, doubling rules tighten further (see `../harmony/voice-leading.md` for the full classical conventions).

### Spacing

The three upper voices (S, A, T) should not exceed an octave between adjacent voices. So S–A within an octave; A–T within an octave. The bass can sit further below the tenor — an octave or more is fine.

Crowded spacing (all four voices within an octave) sounds dense and limited; very wide spacing in the upper voices (S far above A) sounds hollow and disconnected. Find the balance that suits the chord's function and the texture's intent.

### Avoid voice crossing and overlap (in traditional writing)

- **Voice crossing**: when a "lower" voice rises above its "upper" voice (alto above soprano on a given beat). Avoided in traditional 4-part writing.
- **Voice overlap**: when one voice moves into a pitch held by an adjacent voice in the previous beat (alto rising to a note the soprano just left). Also avoided.

Modern and post-tonal choral writing breaks these rules deliberately for color. In traditional / common-practice contexts, follow them.

## Voice leading principles

Choral writing is the original home of voice leading. The principles in `../harmony/voice-leading.md` apply directly. Specifically for choir:

- **Parallel 5ths and octaves are still avoided** in traditional writing. They are a stylistic marker — no parallel 5ths in Renaissance polyphony, abundant parallel 5ths in medieval organum, conventional parallel 5ths in some folk and pop choral. Match your stylistic frame.
- **Each voice should have a singable line.** Even the inner parts (alto, tenor) should make melodic sense to the singer — not just fill chords.
- **Leaps need preparation and resolution.** A 6th or larger leap in any voice is followed by stepwise motion in the opposite direction (the classical convention); strict modern atonal writing relaxes this.
- **Tendency tones resolve.** Leading tones up to tonic; chord 7ths down by step. Without an accompaniment to mask, an unresolved tendency tone in choral writing is conspicuous.

## The independence-blend balance

A central choral aesthetic question: how much should each voice be its own line vs. how much should the four voices fuse into a single sonic entity?

**Maximum independence** (Renaissance polyphony, fugue, much modern choral): each voice has its own melodic identity, often entering at different times, with rhythmic and motivic independence. The texture is contrapuntal.

**Maximum blend** (homophonic chorale, hymn, much pop and worship choir): all four voices move in the same rhythm, with the harmony moving as a block. Inner voices subordinate to the outer (soprano + bass) frame.

Most choral writing falls somewhere in between, often shifting within a single piece. The Bach chorale moves in essentially homophonic blocks but with rhythmic activity in inner voices ("ornamental homophony"). Renaissance motets alternate imitative entries with homophonic passages.

When advising, identify which mode the piece is in (or aiming at), and apply the conventions accordingly.

## Singability — the practical constraints

A part can be technically correct on paper and still hard to sing. Things that affect singability:

### Vowel choice

Open vowels (*ah*, *oh*, *eh*) sustain well and project; closed vowels (*ee*, *oo*) sound thinner on high notes; schwa is hard to sustain at any pitch. See `../fundamentals/prosody-and-language.md` for full discussion.

For choral writing, this matters most at high climactic notes. A sustained high *A5* on the syllable "ah" will project gloriously; the same note on "oo" will sound pinched.

When setting text, check what vowel lands on each high or sustained note. If you can adjust the melody (or, in Latin / non-translated text, the syllable that lands at the climax), do so.

### Consonant timing

A choir of 30 sopranos doesn't release a final consonant at exactly the same instant unless they're rehearsed to. Final-consonant words ending in *t*, *d*, *p*, *b*, *k* are particularly vulnerable to ragged release.

When setting text, prefer to land important held notes on syllables that allow a clean cut-off (open vowel + final *n* or *m*, like "amen" or "in"); avoid landing them on syllables ending in stop consonants where ragged release damages the effect.

### Breath placement

Phrase length matters. A 16-bar phrase at slow tempo with no obvious breath point is unsingable. Provide:

- **Rests** in the middle of long phrases for breath
- **Stagger-breathing opportunities** when no full rest is possible (each singer breathes at a slightly different point so the line never stops)
- **Punctuation in the lyric** that suggests where breath naturally falls

Singers can usually manage 8–12 seconds of phrase without breath; longer requires careful planning.

### Sight-reading difficulty

Choral pieces are often read on minimal rehearsal time, especially in church and amateur contexts. Things that make sight-reading harder:

- Distant key signatures (5+ flats or sharps)
- Frequent accidentals
- Wide leaps
- Highly chromatic lines
- Complex rhythms (mixed groupings, unusual subdivisions)
- Frequent meter changes

These can all be used deliberately, but with awareness of the rehearsal cost. A community choir often can't do what a professional vocal ensemble can.

## Choral textures and idioms

### Homophonic chorale

The four voices move in the same rhythm (or mostly the same rhythm). The texture is harmonic: the listener hears chords moving, with the soprano line as the primary melodic interest. Bach chorales are the gold standard.

Use when: text is the focus and clarity is paramount; the piece is short; the harmonic motion itself is the expressive content.

### Imitative polyphony

Each voice enters separately with the same (or similar) melodic subject, creating a layered, interlocking texture. Renaissance motets, fugues, much choral counterpoint.

Use when: you want compounding emotional intensity through voice accumulation; you want each voice to have melodic identity; the text repeats and bears extended treatment.

### Antiphonal writing

Two groups (often double choir, choir 1 and choir 2) trade phrases. Renaissance Venice polychoral; Vaughan Williams, Tallis "Spem in alium" (40 voices, 8 choirs of 5 voices each).

Use when: the space is large (cathedral, concert hall); you want spatial drama; you have enough singers to populate both groups.

### Drone / sustained-pedal textures

One or more voices hold a sustained pitch while others move above. Common in modal sacred music, much modern choral writing (Tavener, Pärt, Górecki).

Use when: you want timeless, contemplative quality; you want one or more voices to provide a stable harmonic foundation while others move melodically.

### Speech-rhythm and Sprechstimme

Voices speak (or half-sing) the text in rhythm rather than singing pitched notes. Schoenberg, Bernstein, contemporary choral.

Use when: you want declamatory urgency, percussive vocal effect, or text-driven texture without melodic constraint.

### Vocal effects

Modern choral writing uses humming, tongue clicks, breath sounds, glissandi, microtonal slides, whistled tones, hand clapping (corporally produced), and other extended techniques. Composers like Eric Whitacre, Caroline Shaw, and Meredith Monk extend the choral palette substantially.

Notate these clearly. If you invent a technique, include a performance instruction.

## Choral genres and conventions

### Sacred / liturgical

Tradition stretching from medieval chant through Renaissance polyphony, Bach's cantatas and Passions, Mozart and Haydn masses, Romantic settings, to modern liturgical music. Each tradition has its own conventions; the agent should identify which tradition is in play before applying generic advice.

Languages: Latin (Catholic / Episcopalian), German (Lutheran), English (Anglican / Protestant / contemporary), Hebrew (Jewish), Greek (Orthodox), Old Church Slavonic (Slavic Orthodox), and many others.

### Secular concert choral

Madrigals, part-songs, choral symphonies, oratorios, modern concert choral. From Monteverdi to Brahms's *Liebeslieder Waltzes* to Eric Whitacre.

### A cappella and contemporary commercial

The Pentatonix tradition: pop song arrangements for unaccompanied vocal ensemble with vocal percussion. Voice-band hybrids (Pentatonix, Take 6, the Rajaton kind of Finnish vocal jazz, and many collegiate a cappella groups). Drum kit emulation (vocal percussion / "VP") is its own performance discipline.

### Show / theatrical

Musical theatre choral (chorus numbers, ensemble singing). Different from concert choral: amplified, staged, often with band, often with movement / dance.

### Worship and contemporary church

Praise band-driven, popular harmony, often led by a single vocalist with backup. Conventions overlap with pop/rock more than with sacred concert music.

## Composing for choir — practical workflow

1. **Identify the choir.** Professional, semi-pro, community, church, school? This determines what you can ask for.
2. **Identify the voice configuration.** SATB? SSA? TTBB? SSAATTBB? Each demands different writing.
3. **Identify the text.** Sacred / secular? Language? Translated or original? The text shapes everything else.
4. **Identify the tradition / style.** Renaissance polyphonic? Romantic homophonic? Modern minimalist? Worship contemporary? Match conventions.
5. **Sketch the soprano + bass first.** These are the *frame*. The inner voices fill in the harmony around the outer-voice motion. Many composers (Bach included) work this way.
6. **Add inner voices** with attention to voice leading and singability.
7. **Test by singing each part aloud.** If you can't sing a part comfortably, the choir will struggle.
8. **Check tessitura** for each part. Too high = strained. Too low = mumbled.
9. **Check the text setting** for prosody, vowel-on-high-note, breath placement.
10. **Plan rehearsal cost.** If a section will need 30 minutes of work, decide whether the effect is worth it.

## Common pitfalls when advising

- **Don't write inner parts as filler.** Alto and tenor parts that just hold chord tones with no melodic identity are dispiriting to sing. Even in homophonic writing, inner parts should have shape.
- **Don't ignore the passaggio.** Sustained notes parked in a singer's vocal-break zone are uncomfortable.
- **Don't assume singer technique.** A church choir or community choir can't perform the same demands as a professional vocal ensemble. Match the writing to the choir.
- **Don't double the leading tone.** Doubling leading tones in V chords forces parallel octaves on resolution.
- **Don't write unsingable leaps.** A 9th in a single voice almost always sounds wrong; even a 7th leap is suspicious unless prepared and resolved carefully.
- **Don't forget breath.** Long phrases without breath points are unsingable; plan rests or stagger-breathing opportunities.
- **Don't over-divide.** Splitting SATB into SSAATTBB requires double the singers (for ensemble blend, you need at least 2–3 per part, ideally more). Don't divide unless the choir can support it.
- **Don't ignore consonant timing.** Final consonants ragged on release damage the effect. Plan releases carefully.
- **Don't mismatch text and texture.** Dense imitative polyphony obscures text; if the text matters, lean homophonic. If the music matters more than specific text intelligibility, polyphony is fine.
- **Don't apply piano voicing to choir.** Wide voicings that work on keyboard often sound hollow or impossible in 4-part choir. The constraint of 4 voices in their proper ranges is real.

## Quick decision matrix

| Want | Try |
|------|-----|
| Solid traditional 4-part chorale | SATB homophonic, root-doubled triads, classical voice leading; reference Bach chorales |
| Renaissance polyphonic feel | Imitative entries, modal harmony, *cappella*, restrained chromaticism; reference Palestrina, Lasso, Victoria |
| Romantic choral richness | Wider voicings, chromatic harmony, expressive dynamics; reference Brahms, Bruckner |
| Modern minimalist choral | Drone textures, repeated cells, modal harmony; reference Pärt, Tavener, Whitacre |
| Pop a cappella arrangement | SATB or SSAA arrangement of pop hook with vocal percussion; reference Pentatonix |
| Worship contemporary | Pop-influenced harmony with simple SATB or just vocal-plus-band; reference Hillsong / Bethel patterns |
| Choral inside an orchestral piece | Match orchestration density; choir often serves as another orchestral color, not the primary focus |
| Text-clear, declamatory | Homophonic, syllabic text setting (one note per syllable), conservative tessitura |
| Atmospheric, emotional, text-secondary | Polyphonic, melismatic (multiple notes per syllable), wider tessitura, modal or chromatic harmony |
| Spatial drama | Antiphonal writing for double choir; works in resonant spaces |
| Limited rehearsal time / amateur choir | Conservative key signatures, stepwise motion, clear part-doubling, stay near comfortable middle ranges |

## Cross-references

- General voicing and texture principles → `voicing-and-texture.md`
- Solo voice ranges and tessitura → `instruments-ranges-character.md`
- Voice leading rules and conventions → `../harmony/voice-leading.md`
- Counterpoint (foundational for polyphonic choral writing) → `../counterpoint.md`
- Prosody, vowel choice for sustained notes, language constraints → `../fundamentals/prosody-and-language.md`
- Sacred and secular choral traditions by period → `../genres/classical-periods.md`
- Musical theatre choral conventions → `../genres/musical-theatre.md`
- Chord construction (for harmonic vocabulary) → `../harmony/chord-construction.md`
- Modal harmony (for sacred / Renaissance / minimalist contexts) → `../harmony/modal-harmony.md`
- Folk and world choral traditions (Eastern European, African, Indonesian, etc.) → `../genres/folk-roots-and-traditions.md`
- Korean traditional vocal traditions (*pansori*, *minyo*) → `../genres/korean-traditional.md`

# Instruments — Ranges and Character

This file is a quick reference for the major instrument families: their ranges, characteristic registers, idiomatic writing, and what to use them for. It's not exhaustive (each instrument has volumes written about it), but it covers what a composer or arranger needs to know to make sensible choices.

For voicing decisions across instruments, see `voicing-and-texture.md`. For decisions about how many instruments to use simultaneously, see `arrangement-density.md`. For genre-specific orchestration, see relevant `genres/*.md` files.

## When to consult this file

- The user is writing for a specific instrument or ensemble
- Choosing between instruments for a melodic line
- Avoiding ranges that are awkward for the instrument
- Understanding why a part sounds wrong (often a register or character mismatch)
- Idiomatic writing — what's natural for each instrument

## How to read this file

Each instrument is given:

- **Range**: written range (the notes you'd write in the score) — sometimes different from sounding range due to transposition.
- **Sweet spot / tessitura**: where the instrument sounds best and is easiest to play.
- **Character by register**: how the instrument's tone varies across its range.
- **Idiomatic uses**: what the instrument is naturally good at.
- **Genre / context**: where you'd typically encounter it.

## Strings (orchestral)

The most flexible family. Strings can play almost any line; they sustain indefinitely (with bow changes), play double-stops and chords, and produce a vast range of timbres.

### Violin

- **Written range**: G3 to E7 (extending to A7+ for soloists). Concert pitch (non-transposing).
- **Sweet spot**: G3–A6.
- **Character by register**:
  - **Low (G3–B3)**: warm, gutsy, "throaty"
  - **Mid (C4–C5)**: lyrical, default singing range
  - **High (D5–E6)**: brilliant, soaring, intense
  - **Very high (F6+)**: piercing, often virtuosic
- **Idiomatic**: melodic lines, arpeggios, double-stops, fast passages, sustained singing tone. The lead voice in orchestral writing.
- **Notes**: Written in treble clef. Section divisi (violins splitting into 2+ parts) is standard. Tremolo, pizzicato, harmonics, sul ponticello, and many other effects are available.

### Viola

- **Written range**: C3 to E6 (alto clef typically; treble clef in high register).
- **Sweet spot**: C4–G5.
- **Character**: darker and warmer than violin. Often described as "the voice that gets lost in the orchestra" — the viola's mid range is in the same zone as cellos and second violins, so it can be hard to feature.
- **Idiomatic**: inner-voice writing, lyrical solos in its mid-high register, intimate accompaniment. Featured prominently in chamber music.
- **Notes**: Alto clef takes practice to read. The viola's middle and lower register is where it sounds most "viola" — high register starts to sound violinish.

### Cello

- **Written range**: C2 to A5 (bass clef typically; tenor clef in upper register; treble clef in very high).
- **Sweet spot**: G2–E5.
- **Character by register**:
  - **Low (C2–G2)**: thick, growly
  - **Mid (A2–E4)**: rich, singing — the most beloved register
  - **High (F4–A5)**: lyrical, intense, even strained
- **Idiomatic**: melodic and bass roles equally; doubles bass an octave higher; capable of soaring solo lines. The "tenor" of the string section.
- **Notes**: Tenor clef appears for moderately high passages to avoid ledger lines. Pizzicato in low register is punchy and resonant.

### Double bass

- **Written range**: E1 to D5 (bass clef, sounds an octave lower than written).
- **Sweet spot**: E1–A3.
- **Character**: foundation of the string section. Provides bass throughout the orchestra; can play melodic in its upper range (rare in classical, common in jazz).
- **Idiomatic**: bass lines, walking bass (jazz), pizzicato in jazz/pop. Bowed bass in classical for sustained low notes.
- **Notes**: Transposes down an octave. Most basses go to E1; some have a C1 extension for very low notes.

### String section conventions

Standard orchestral string section: violin I, violin II, viola, cello, double bass. Modern/Classical orchestras have varying section sizes (e.g., 16-14-12-10-8 in a full Romantic orchestra; smaller in Classical-period works).

Common scoring techniques:

- **Unison melody** (all strings on one line): warm, full
- **Octave doublings** (e.g., violin I + cello an octave below): rich, hymnic
- **Divisi**: strings split into multiple parts; common for chord-voicing in slower passages
- **Pizzicato** for sections: percussive, energetic
- **Tremolo**: intense agitation or shimmering
- **Mute (con sordino)**: softer, hazier

## Strings (popular)

### Acoustic guitar

- **Written range**: E2 to E6. Notated treble clef, sounds an octave lower than written.
- **Sweet spot**: A2–E5.
- **Character**: warm, percussive, harmonic. Steel-string is bright; nylon (classical) is warmer.
- **Idiomatic**: chord strumming, fingerpicking, arpeggios, melodic lines in mid-high register. Folk, country, singer-songwriter staple.

### Electric guitar

- Same range as acoustic. Tone varies enormously with effects, amp, pickup configuration.
- **Idiomatic**: power chords, riffs, lead lines, sustained tones with feedback, palm-muted rhythms. The cornerstone of rock.
- **Notes**: extended techniques include bends, slides, hammer-ons/pull-offs, vibrato, harmonics, whammy bar effects.

### Bass guitar (electric)

- **Written range**: E1 to G4 (4-string standard tuning); 5-string bass extends to B0.
- **Sweet spot**: E1–C3.
- **Character**: foundation of pop, rock, R&B, funk. Round, percussive, can be played fingerstyle (warm, smooth) or with a pick (punchy, articulate).
- **Idiomatic**: groove-focused bass lines; melodic bass in funk and prog rock; sustained roots in pop ballads.

### Banjo, mandolin, ukulele, and other folk strings

Each has its own range and character. Banjo: bright, plucky, bluegrass and country. Mandolin: shimmery tremolo, folk and classical. Ukulele: small, warm, Hawaiian and indie pop. Each has its own idiomatic patterns — researching a specific tradition is wise.

## Woodwinds

### Flute

- **Written range**: C4 to C7 (or higher). Concert pitch.
- **Sweet spot**: G4–B6.
- **Character by register**:
  - **Low (C4–F4)**: hollow, breathy, intimate
  - **Mid (G4–G5)**: pure, lyrical
  - **High (A5–C7)**: brilliant, piercing
  - **Very high (above C7)**: shrill
- **Idiomatic**: agile melodic lines, runs, trills, sustained singing tones. Almost always non-doubling (one player per part).
- **Variants**: piccolo (sounds an octave higher; brilliant top voice), alto flute (lower, warmer), bass flute (rare).

### Oboe

- **Written range**: B♭3 to A6.
- **Sweet spot**: D4–F5.
- **Character**: nasal, plaintive, very distinctive. The orchestra tunes to the oboe's A.
- **Idiomatic**: lyrical melodies, especially in the middle register. Solo passages in pastoral or melancholy contexts. Doubling other instruments thinly.
- **Variants**: English horn (cor anglais; sounds a fifth lower; even more melancholy), oboe d'amore (between oboe and English horn).

### Clarinet

- **Written range**: D3 to G6. Transposing instrument: most common is B♭ clarinet (sounds a M2 lower than written); A clarinet (sounds m3 lower) for sharp-key music.
- **Sweet spot**: G3–C6.
- **Character by register**:
  - **Chalumeau (low, D3–F4)**: dark, smoky, woody
  - **Throat (F♯4–B♭4)**: weak, less resonant — write through it, not in it
  - **Clarion (B4–C6)**: bright, expressive — the main melodic register
  - **Altissimo (above C6)**: piercing, used for special effect
- **Idiomatic**: agile, expressive in all registers (except throat). Smooth legato; rapid runs; lyrical solos.
- **Variants**: bass clarinet (sounds an octave lower), E♭ clarinet (smaller, brighter), contrabass clarinet (rare, very low).

### Bassoon

- **Written range**: B♭1 to E♭5. Bass clef and tenor clef.
- **Sweet spot**: F2–G4.
- **Character**: woody, double-reed buzz. "The clown of the orchestra" in some passages; deeply solemn in others. Can be lyrical or comedic.
- **Idiomatic**: bass voice in winds, melodic in mid-high register, comedic effects (staccato low notes — see "Sorcerer's Apprentice"), sustained pedal tones.
- **Variants**: contrabassoon (sounds an octave lower; among the lowest instruments).

### Saxophones

- **Written range** (depends on size): all transposing.
  - **Soprano sax (B♭)**: written B♭3 to F♯6
  - **Alto sax (E♭)**: written B♭3 to F♯6
  - **Tenor sax (B♭)**: written B♭3 to F♯6
  - **Baritone sax (E♭)**: written B♭3 to F♯6
- All saxes have the same written range; their sounding ranges differ by their key.
- **Character**: smooth, flexible, expressive. Range from sweet to growling.
- **Idiomatic**: lead lines (jazz), section work (big band), solos (jazz, pop, R&B). The most flexible single-reed instrument.
- **Notes**: The saxophone family is the cornerstone of jazz wind sections. Less common in orchestral music (though Ravel, Glazunov, and others wrote for it).

## Brass

### Trumpet

- **Written range**: F♯3 to D6 (typically; advanced players go higher). Most often B♭ trumpet (sounds M2 lower); also C trumpet (concert pitch).
- **Sweet spot**: B♭3–G5.
- **Character**: bright, brilliant, declarative. The trumpet is the most penetrating brass instrument.
- **Idiomatic**: fanfares, bright melodic lines, jazz lead, lyrical solos with a mute. Strong in mid-high register.
- **Variants**: piccolo trumpet (high, brilliant), flugelhorn (warmer, mellower), cornet (slightly mellower than trumpet).

### French Horn

- **Written range**: F2 to F5 (in concert pitch as F horn; transposes a P5 lower than written, so a written G3 sounds C3).
- **Sweet spot**: F3–G4.
- **Character**: warm, noble, melodic. The most "vocal" brass instrument. Can blend into woodwinds or stand out as solo.
- **Idiomatic**: melodic solos, sustained chordal harmony, hunting calls (historical), heroic fanfares, soft chordal pad behind woodwinds.
- **Notes**: Horn players use the right hand inside the bell to adjust pitch and color. Hand-stopping changes the timbre dramatically.

### Trombone

- **Written range**: E2 to F5 (tenor trombone; bass clef and tenor clef).
- **Sweet spot**: B♭2–C4.
- **Character**: noble, full, slightly slower-articulating than trumpet due to slide mechanism. Smooth glissandos possible (and idiomatic).
- **Idiomatic**: chordal harmony in brass section, solo lyrical lines, jazz solos with slide effects, big band shouts.
- **Variants**: alto trombone (smaller, higher), bass trombone (lower, with extra valves for very low notes), contrabass trombone (rare).

### Tuba

- **Written range**: D1 to F4 (depends on instrument; F tuba is higher, B♭ tuba lower).
- **Sweet spot**: F1–B♭2.
- **Character**: foundation of the brass; deep, rumbling. Surprisingly agile for its size when written for the instrument.
- **Idiomatic**: bass voice in brass section; doubling cellos and basses an octave below; rare melodic solos in mid-high register.

### Brass section conventions

Standard orchestral brass: 4 horns, 3 trumpets, 3 trombones, 1 tuba. Big band brass: 4 trumpets, 4 trombones (no horn or tuba). Brass band: a different family altogether (cornets, flugelhorns, baritones, euphoniums).

## Percussion

A vast family. Two main subdivisions: pitched and unpitched.

### Pitched percussion

- **Timpani** (4 drums tuned to specific pitches; range roughly D2 to A3): the orchestral pitched bass drum. Tuned for each piece. Strong rhythmic and tonal role.
- **Xylophone** (F4 to C8 or higher): bright, dry, percussive. Often doubling melodic lines for emphasis.
- **Marimba** (C2 to C7): warm, woody, sustaining. Solo and ensemble use; warmer than xylophone.
- **Vibraphone** (F3 to F6): metallic, sustained with motor-driven vibrato. Jazz and classical solo. Pads can be used for muting.
- **Glockenspiel / Bells** (G5 to C8): bright, bell-like.
- **Crotales**, **Tubular Bells**, etc. — specific colors for specific moods.

### Unpitched percussion

The full universe is too vast for this file. Major categories:

- **Drums** (snare, bass drum, toms, kick): rhythmic backbone in pop, jazz, rock.
- **Cymbals** (suspended, crash, hi-hat, ride): sustained and accent percussion.
- **Hand percussion** (tambourine, shaker, claves, congas, bongos): texture and rhythm.
- **World percussion** (djembe, cajón, talking drum, frame drum, etc.): each has its own role.

For drum kit programming and writing, see `../rhythm-groove/groove-and-feel.md` and genre-specific files.

## Voice

### Vocal range categories

Standard four-part choral ranges (extended to six for solo voices):

| Voice | Range (typical) | Range (extended) |
|-------|-----------------|------------------|
| Soprano | C4 to G5 | A3 to C6 |
| Mezzo-soprano | A3 to E5 | F3 to A5 |
| Alto | F3 to D5 | E3 to F5 |
| Tenor | C3 to G4 (sometimes notated octave higher) | B2 to A4 |
| Baritone | A2 to E4 | F2 to G4 |
| Bass | E2 to C4 | C2 to E4 |

### Vocal registers (chest, mix, head)

Singers move between vocal registers:

- **Chest voice**: lower register; full, resonant, "speaking-like".
- **Mix voice / middle**: blends chest and head; the comfortable middle register.
- **Head voice**: upper register; lighter, "head-resonant".
- **Falsetto** (for males): light, pure, unfull tone in upper register.

The transition points (passaggi) are where the voice naturally changes character. Skilled singers smooth them; novice singers may "break" between registers.

### Vocal characteristics by genre

Different genres have different vocal expectations:

- **Classical**: trained, even tone across registers, vibrato, full dynamic range, careful diction.
- **Pop**: less classical formality; mix voice and belt prominent; auto-tune common in modern production.
- **R&B / soul**: melisma, vocal runs, ad-libs, raspy or smooth depending on artist.
- **Rock**: belted, sometimes raw; head voice rare except for certain styles.
- **Jazz**: scatting, melodic improvisation, Sinatra-style phrasing, conversational tone.
- **Hip-hop**: rap (rhythmic speech) more than singing; melodic vocals (sing-rap) increasingly common.
- **Folk**: natural, conversational; minimal vibrato; often nasal.
- **Musical theatre**: belted "money notes" in chest voice mid-high; controlled mix voice for sustained passages.

## Keyboard instruments

### Piano

- **Range**: A0 to C8 (88 keys).
- **Sweet spot**: bass clef from F1 to C4; treble from C4 to C7.
- **Character**: most flexible solo instrument. Can play melody, harmony, bass simultaneously. Wide dynamic range. No sustain like winds or strings but has pedal-controlled sustain.
- **Idiomatic**: solo, accompaniment, ensemble. Foundation of art song, jazz combo, much pop production.

### Harpsichord

- **Range**: roughly C2 to G6.
- **Character**: plucked rather than struck; less dynamic variation; bright, percussive.
- **Idiomatic**: Baroque solo and continuo. Limited modern use but appears in early-music performance and certain rock contexts (Beatles, prog).

### Organ

- **Range**: depends on the specific instrument; church organs have multiple manuals (keyboards) plus pedals, with combined range A0 to A8+.
- **Character**: sustained, powerful, capable of vast dynamic range and timbral variety through stops.
- **Idiomatic**: sacred music, classical solo, jazz Hammond B-3 (electric organ with rotating Leslie speaker — very different sound).

### Synthesizer

- **Range**: depends on settings; can theoretically produce any pitch.
- **Character**: as variable as the user's design choices. Can imitate other instruments or produce purely synthetic sounds.
- **Idiomatic**: pop, electronic music, film scoring. Modular synths in experimental music.

## Choosing instruments — what each does best

A rough decision guide:

| Want | Try |
|------|-----|
| Sustained melodic warmth | Cello, viola, English horn, French horn |
| Bright, declarative melody | Trumpet, oboe, soprano voice |
| Soft, lyrical accompaniment | Strings (muted), harp, piano (gentle) |
| Driving rhythmic energy | Drum kit, distorted electric guitar, sustained brass |
| Mysterious, atmospheric | Solo flute, marimba (soft), harp, ambient pad |
| Powerful low end | Tuba, contrabassoon, bass guitar, low organ |
| Agile, virtuosic line | Violin, flute, clarinet, piano |
| Bluesy, soulful melody | Saxophone, blues guitar, voice with melisma |
| Full orchestral wash | Strings divisi, full woodwinds, brass chord support |
| Intimate, narrative-friendly | Piano + voice, acoustic guitar + voice, solo woodwind |
| Tradition-specific instrumental color | A specific instrument used idiomatically within its tradition or fusion context; research tuning, role, ornamentation, and cultural use before dropping it in |

## Common pitfalls when advising

- **Don't write outside the instrument's range.** The viola can't play below C3; the trumpet's lowest practical note is F♯3. Check ranges before composing.
- **Don't ignore tessitura.** A line that's technically playable but lives in an awkward register sounds strained.
- **Don't pick instruments without knowing their character.** A trumpet melody is very different from an oboe melody, even with the same notes.
- **Don't mix transposing and non-transposing instruments without care.** When writing for a B♭ clarinet, write a M2 above the desired sounding pitch. Easy to forget.
- **Don't double instruments that don't blend.** Cello and trombone have similar registers but different timbres — they may not fuse. Test combinations.
- **Don't ignore practical limitations.** Brass players need to breathe; long phrases without rest are problematic. Pizzicato is short and percussive — sustained pizz lines are awkward.
- **Don't over-specify or under-specify articulation.** A score with no articulation lets the player default; one with too many marks becomes unreadable.

## Cross-references

- Voicing across instruments → `voicing-and-texture.md`
- Density and layering → `arrangement-density.md`
- Choral writing (4-part vocal ensemble specifically) → `choral-writing.md`
- Composition-facing instrument idiom → `../instrument-idiom/overview.md`
- Piano/keyboard writing → `../instrument-idiom/piano-keyboards.md`
- Guitar, bass, drums, and band idiom → `../instrument-idiom/guitar.md`, `../instrument-idiom/bass.md`, `../instrument-idiom/drums-percussion.md`
- Strings, winds, brass, and vocals idiom → `../instrument-idiom/strings.md`, `../instrument-idiom/winds.md`, `../instrument-idiom/brass.md`, `../instrument-idiom/vocals.md`
- Score layout for orchestral writing → `../fundamentals/notation-and-conventions.md`
- Genre-specific orchestration choices → various `../genres/*.md` files
- Specific instrument ranges in form context → `../form/popular-song-forms.md` (vocal range across sections)
- Production-side instrument considerations → `../production-aware/arrangement-for-mix.md`

# Pre-Production Decisions

> **Snapshot note:** Last content review snapshot: 2026-04-27. Reference-track practice is current-sensitive. For live trend checks, user listening context, and reference-set building, see `../research/web-music-trend-research.md`, `../research/user-listening-context-and-streaming-services.md`, and `../research/reference-track-digging.md`.

This file is about the decisions a composer makes *before or during the earliest sketch phase* of a piece — choices that determine the shape of everything downstream. Key, tempo, length, instrumentation, and a few others. These look obvious in retrospect but are often where projects either find their footing or wander.

This file complements `arrangement-for-mix.md` (mix-aware arrangement choices) and `energy-and-dynamics.md` (production-side energy shaping). It also overlaps with `references/workflow.md` (the agent's process for helping a user start), but that file is about *how the agent advises*; this file is about *what the user is deciding*.

## When to consult this file

- The user is starting a new piece or song
- "What key should this be in?", "What tempo?", "How long should it be?"
- The user has a sketch but feels the foundation is wrong
- Picking a reference track and using it well
- Deciding when to commit to specific decisions vs. leave them flexible
- Demo-vs-production workflow questions

For starting techniques specifically (chord progression first, melody first, groove first), see `references/workflow.md`. For form templates by genre, see `references/form/popular-song-forms.md` and `assets/form-templates.md`.

## Why these decisions matter early

Most pre-production decisions are cheap to change at the sketch stage and expensive to change later. Switching keys after the vocal has been recorded is a re-record. Switching tempo after the song is mixed is rebuilding the song. Switching from "3-minute pop single" to "6-minute prog track" mid-process invalidates most of the formal decisions.

The agent's job at the pre-production stage is to help the user make these decisions deliberately rather than by default — and to flag when a "default" choice is going to cause problems later.

## Key selection

The most consequential early decision after deciding to start at all.

### Vocal range — the dominant constraint

For songs with a lead vocal, the singer's range usually dictates the key. Two things matter:

- **Tessitura match**: where the song "lives" should sit in the singer's comfortable middle. The chorus should *land* in the part of their range that sounds full, not strained.
- **Range coverage**: the song's lowest note and highest note should both be within the singer's reachable range, ideally with some headroom above the highest belt note.

Common pop / rock / R&B vocal range zones (untrained-to-pro):

| Voice type | Comfortable middle | Top end (with effort) |
|-----------|--------------------|------------------------|
| Female pop (mezzo / alto-leaning) | A3 – D5 | F5 – G5 |
| Female pop (soprano-leaning) | C4 – F5 | A5 – C6 |
| Male pop (baritone / tenor-leaning) | A2 – C4 | E4 – G4 |
| Male pop (high tenor / countertenor) | C3 – E4 | A4 – C5 |
| Mixed-genre belters (any sex) | varies | F5–G5 (female), G4–A4 (male) standard for "money note" |

These are rough zones. Always ask the singer what their actual comfortable middle is before picking a key. Many demos are written in keys that are too high or too low because the writer used their own voice as reference.

### Instrument-friendly keys

Some keys are easier for some instruments. Picking a key that suits the dominant instrumentation can make demos faster and arrangements more idiomatic.

| Instrument | Easy keys | Awkward keys |
|------------|-----------|---------------|
| Acoustic guitar (open chords) | E, A, D, G, C | E♭, A♭, F♯ (require capo or barre) |
| Electric guitar (rock) | E, A, D, G (drop-D adds C, B, F) | F, B♭ (less idiomatic riffs) |
| Piano | C, F, G | F♯/G♭ (all-black-key feel; some pianists prefer this) |
| Strings | D, G, A, E (open strings) | D♭, G♭ (uncomfortable for double stops) |
| Brass | F, B♭, E♭ (concert pitch); these are friendly transposed-instrument keys | C♯, F♯ (uncommon and difficult) |
| Saxophone (alto, B♭) | F, B♭, E♭ concert | E, A, D concert |

Conflicts happen — a singer's best key (E♭) might be awkward for the guitarist. Common solutions: capo for the guitarist, transpose the singer one step up or down, or accept the trade-off if vocal performance matters more than guitar idiom.

### Mood / character associations (weak signal but real)

Specific keys are sometimes felt as having character — partly cultural, partly because tunings and timbres vary across keys. These are weaker signals than tessitura or instrumentation but worth knowing:

- **Bright, "open" feel**: D, A, E
- **Warm, "rounded" feel**: F, B♭, E♭
- **Dark, melancholic feel**: B♭ minor, F minor, C♯ minor
- **"Heroic" / classical-feel**: E♭, D, A
- **Folk / singer-songwriter neutral**: G, C, D

Equal temperament removes most of the historical key-character associations (which existed because of unequal tunings), but instrument resonance, idiomatic voicing, and listener associations from popular tracks in those keys keep some of the character alive.

### Practical key-pick workflow

1. **Identify the singer** (if any). What's their comfortable middle and top?
2. **Identify the dominant instrument** (if any). What keys does it sit well in?
3. **Identify the chorus's highest note** (this is where the song "peaks" — make sure the singer can hit it).
4. **Pick the key** that lands the chorus peak at the singer's belt zone *and* is reasonable for the instrumentation.
5. **Test it.** Sing the chorus aloud. Play the verse on the chosen instrument. If it fights, try a half-step or whole-step shift.

Many demos start in the writer's voice's key and get transposed for the actual singer later. That's fine — but flag it explicitly so it doesn't get forgotten.

## Tempo selection

The second-most consequential early decision.

### Genre tempo conventions

See `references/fundamentals/rhythm-meter.md` for the full table. Quick reference for popular contexts:

| Style | Typical BPM range | Typical center |
|-------|-------------------|----------------|
| Pop ballad | 60–80 | 70 |
| Modern pop | 100–130 | 115–120 |
| K-pop dance | 110–130 | 120 |
| Hip-hop modern | 80–95 (or 140–160 in halftime) | 85 / 140 |
| House / EDM | 118–135 | 124 |
| Rock (mid-tempo) | 110–130 | 120 |
| Country | 80–130 | 110 |
| Film score (action) | 120–180 | 140 |
| Film score (drama) | 60–100 | 75 |
| Trailer music | 70–90 (often felt halftime) | 80 |
| Singer-songwriter | 70–110 | 95 |

Within a genre's range, BPM choice affects:

- **Energy level**: 105 BPM and 125 BPM in the same style feel notably different
- **Word density**: more BPM means less time per syllable, which constrains lyric density
- **Half-time / double-time feel**: a song at 140 BPM with a halftime drum pattern *feels* like 70, but the underlying grid is still 140 — this matters for production tempo settings

### Choosing between close tempi (115 vs. 120 vs. 125)

Small BPM differences can matter more than they look. Test:

- **Sing the chorus at the candidate tempo.** Does it feel rushed? Sluggish?
- **Walk to it.** Walking is roughly 100–120 BPM. Tracks that match walking pace feel natural; tracks that don't feel "energetic" or "lazy" relative to walking.
- **Compare to reference tracks.** If your song is supposed to feel like Track X, match its BPM (or be deliberate about differing).

### Tempo and structure interaction

A 4-bar phrase at 60 BPM lasts 16 seconds. A 4-bar phrase at 140 BPM lasts 7 seconds. So:

- Slow tempi can carry longer melodic phrases without the listener losing track
- Fast tempi need shorter phrases and more repetition to feel grounded
- Matching phrase length to tempo is part of why "this song feels rushed" or "this song drags" — usually a phrase-tempo mismatch

## Length and form commitment

Modern attention spans and platform conventions affect this more than they used to.

| Format | Typical length | Notes |
|--------|----------------|-------|
| TikTok / short-form social hook | 15–30 seconds | The hook is the entire song |
| Modern pop single | 2:30–3:15 | Streaming-era (skip rates incentivize shorter songs) |
| Traditional pop single | 3:00–3:45 | "Album track" length |
| Album cut (rock, indie) | 3:30–5:00 | Depends on band conventions |
| Prog / metal / extended track | 5:00–15:00+ | Form-justified length |
| EDM club mix | 5:00–7:30 | Designed for DJ mixing |
| EDM radio edit | 2:45–3:30 | Compressed version for radio |
| Film cue | 30 seconds – 5 minutes | Determined by scene length |
| Game music loop | 30 seconds – 2 minutes (loops) | Designed for seamless looping |

Decide the target length before writing. A song meant to be 2:45 should be sketched in sections that add up to 2:45, not "however long it ends up". A song meant to be 6:00 should plan for sustained development, not just stretch a 3-minute song with longer choruses.

## Reference tracks — using them well

A reference track (or "ref") is an existing song the user wants the new piece to feel like. References are powerful and dangerous in equal measure.

### What references are good for

- **Communicating a goal across collaborators**: "I want it to feel like *Track X*" tells the producer, mixer, and players what you mean.
- **Calibrating tempo, length, energy curve**: copy these specific parameters from a working example.
- **Identifying the genre frame**: a reference settles the "what genre is this?" question instantly.
- **Sound design / production aesthetics**: the timbral palette, drum sound, mix balance.

### What references are bad for

- **Melodic copying**: similarity to a reference's hook is how songs get sued. Use references for *vibe*, not for melodic content.
- **Excusing lazy decisions**: "I'll just do what they did" without understanding *why* leads to imitation that lacks the original's coherence.
- **Constraining the song's identity**: a song that's "X but with a Y twist" can become neither X nor Y. Sometimes the reference needs to be released once the song finds its own legs.

### A reference workflow

1. Pick 1–3 references early. Listen to them analytically.
2. Identify what specifically makes each reference work (form, tempo, key, instrumentation, hook structure).
3. Borrow the *parameters* (tempo, length, density profile) but compose original *content* (melody, harmony, lyrics).
4. After the first sketch, listen to the references again. Are you still aiming at the same target? Or has the song revealed its own direction?
5. If the song is becoming its own thing, let it. The references were scaffolding.

## Demo vs. production workflow

A frequent decision point: how polished should the early sketch be?

### Demo-first (most common for pop/rock songwriters)

- Quick, rough sketch with placeholder sounds
- Voice + piano/guitar, or voice + simple beat
- Goal: capture the song before deciding production direction
- Production happens later, with the song already proven on its own

Advantages: fast iteration, focuses on song quality, easy to pivot. Disadvantages: production-aware compositional moves can be missed (a song written on piano may not adapt cleanly to electronic production).

### Production-first (common for electronic / hip-hop / K-pop)

- Build the track sonically first — drums, bass, key sounds
- Topline (vocal melody and lyrics) added later, written *to* the track
- Goal: the production *is* the song, more or less

Advantages: track is sonically coherent from the start, productions tend to feel modern. Disadvantages: weak compositional bones can be hidden by production, songs that don't survive without production can feel hollow.

### Concurrent (common for film / orchestral)

- Composition and production move together (sequencing strings while writing them, for example)
- Goal: composition shaped by what the production will be

Advantages: no surprises in handoff. Disadvantages: slower iteration, harder to pivot.

When advising, ask which mode the user is in — it changes what advice is relevant. A demo-first songwriter doesn't need to worry about kick-bass spectral overlap yet; a production-first electronic producer should worry about it from the first bar.

## Cost-of-change for late decisions

When the user is past the sketch stage, some changes are very expensive. Flag this when relevant:

| Decision | Cheap to change at | Expensive after |
|----------|---------------------|------------------|
| Key | Any time before vocal recording | Vocal recorded |
| Tempo (small adjustment, ±5 BPM) | Pre-recording | Drums recorded |
| Tempo (large change, ±20 BPM) | Sketch stage only | Anything tracked |
| Form (add/remove section) | Pre-arrangement | Mixed |
| Time signature | Sketch only | Almost anything tracked |
| Genre / production direction | Pre-production (decision made before recording) | Recording started |
| Lyric (single line or word) | Almost any time | After all vocal takes done |
| Lyric (concept / theme) | Sketch only | Vocal tracked |

This isn't to discourage changes — sometimes a late change is necessary. But the user should know what it costs.

## Pre-production checklist

Before serious composition or recording starts:

- [ ] Genre / style identified (and primary vs. flavor if hybrid)
- [ ] Target length / format decided
- [ ] Tempo chosen (and tested by singing or walking to it)
- [ ] Key chosen (vocal-tested if there's a singer; instrument-friendly if relevant)
- [ ] Form template sketched (verse-chorus? AABA? through-composed?)
- [ ] Reference tracks selected and analyzed (1–3 max)
- [ ] If the user has a streaming service or playlist context, safe user-provided listening evidence considered
- [ ] Demo-vs-production workflow chosen
- [ ] Lead instrument and core instrumentation decided
- [ ] Lyrical concept (if applicable) clear enough to write the first verse

Not every project needs all of these settled before starting — but the agent should know which ones are open and which are closed when advising.

## Common pitfalls when advising

- **Don't pick the user's key for them without knowing the singer.** Range is everything for vocal music.
- **Don't recommend a tempo without testing.** What looks right on paper can feel wrong out loud.
- **Don't dismiss the cost of late changes.** A user planning to "just adjust the key in the mix" may not realize what that costs.
- **Don't over-lock decisions.** Some users benefit from leaving tempo/key flexible until the song reveals itself. Match the advice to the user's working style.
- **Don't confuse a reference with a destination.** A song aimed at a reference can never be the reference. The reference is the target zone, not the goal.
- **Don't forget mode and feel under "key".** "C major" and "C Mixolydian" are different keys for compositional purposes, even though they share a pitch collection.

## Quick decision matrix

| Stuck on | Consider |
|----------|----------|
| What key to pick | Sing the chorus melody at different transpositions; pick where the peak feels right in the singer's voice |
| What tempo to pick | Walk to candidates; pick where the chorus chant feels natural; check against reference tracks |
| Whether to commit to form yet | Decide format (single, album cut, EDM club mix) — this fixes target length, which constrains form |
| Demo or production first | Demo-first if writing in a singer-songwriter mode; production-first if writing electronic/hip-hop; concurrent for film |
| Whether to use a reference | Yes for vibe and parameters; no for melodic content; release the reference once the song finds its own identity |
| When a late change is necessary | Calculate the cost (re-record, re-arrange, re-mix); commit if the song requires it; don't if the change is cosmetic |

## Cross-references

- Tempo conventions and Italian tempo terms → `../fundamentals/rhythm-meter.md`
- Vocal range and tessitura by zone → `../melody/melodic-construction.md`, `../orchestration/instruments-ranges-character.md`
- Choral pre-production (key choice for SATB ensembles) → `../orchestration/choral-writing.md`
- Form templates and section length conventions → `../form/popular-song-forms.md`, `../../assets/form-templates.md`
- Mix-aware arrangement choices (after pre-production) → `arrangement-for-mix.md`
- Energy planning across the form → `energy-and-dynamics.md`, `../form/narrative-and-transitions.md`
- Starting techniques (which element to write first) → `../workflow.md`
- Genre-specific conventions for tempo, length, key → various `../genres/*.md`
- Current trend and platform-sensitive reference checks → `../research/web-music-trend-research.md`
- User listening context and streaming-service research → `../research/user-listening-context-and-streaming-services.md`
- Regional trend-evolution analysis → `../research/regional-trend-evolution-analysis.md`
- Reference-track matrix and anti-copy workflow → `../research/reference-track-digging.md`
- Game music length / loop / interactivity decisions → `../genres/game-music.md`
- Short-form / commercial duration decisions (15s, 30s, 60s formats) → `../genres/media-and-commercial-music.md`

# Web Music Trend Research

> **Snapshot note:** Last content review snapshot: 2026-04-26. This file is a workflow, not a static trend report. Use it whenever the user asks for *current*, *recent*, *viral*, *market-specific*, *platform-specific*, or *reference-track-driven* guidance.

Music trends age quickly. Static genre files should explain durable craft principles; current facts should be checked on the web at the time of the user request.

## When to use active web research

Use web research when the user asks about:

- current or recent music trends
- "what is popular right now"
- platform-specific behavior: TikTok, YouTube Shorts, Spotify playlists, Apple Music charts, radio, sync/library music
- current K-pop/J-pop/C-pop/EDM/hip-hop/indie scenes
- current reference tracks or living-artist comparisons
- user-specific listening context from public playlists, streaming recaps, or user-shared service data
- current AI music tools, copyright, lawsuits, distributor policies, or platform rules
- release strategy, chart performance, playlisting, or market targeting

Do **not** rely only on the static skill files for these requests.

## Source hierarchy

Prefer sources with transparent data collection and primary access. A useful research pass triangulates at least two source types.

| Source type | What it is good for | Watch-outs |
|---|---|---|
| Official charts | What is actually charting in a market or globally | Charts measure different things: streams, sales, radio, video, UGC, or weighted composites |
| Platform charts | What users are playing on a specific platform | Platform demographic and playlist bias |
| Industry reports | Macro trends, revenue, format changes, regional growth | Often lag current scene-level taste |
| Local music press / label news | Scene context, new movements, artist narratives | Hype, PR framing |
| Social / short-form data | Viral hooks, memes, dance challenges, sound reuse | Fast-moving and noisy; viral does not always mean musically durable |
| User-provided streaming context | Personal taste, playlist fit, durable favorites, reference anchors | Never request credentials; user taste is not automatically market trend |
| Interviews / producer breakdowns | Arrangement and production details | May be anecdotal; verify with listening |

## Useful starting sources

Use the current versions of these sources, not cached assumptions:

| Market / question | Starting points |
|---|---|
| Global streaming / platform movement | Spotify Charts, Apple Music charts, YouTube Charts, Billboard Global 200 |
| Global industry macro trends | IFPI Global Music Report, national industry bodies |
| Korea / K-pop | Circle Chart, Melon / Genie / Bugs charts, Korean music press, official label channels |
| Japan / J-pop | Oricon, Billboard Japan, Apple Music Japan, Spotify Japan, local press |
| UK | Official Charts Company |
| US | Billboard charts, Luminate/Music Connect references where available, RIAA reports |
| YouTube-heavy markets | YouTube Charts plus local press and platform playlists |
| Short-form virality | TikTok Creative Center or official platform trend pages, then verify with chart/streaming movement |

## Research workflow

### 0. Check for user listening context when relevant

If the user is writing for their own taste, playlist fit, DJ set, or reference library, ask whether they use a streaming service and can share lightweight evidence: a public playlist, 5-10 track names, or a Replay/Wrapped/Last.fm summary. Load `user-listening-context-and-streaming-services.md` before treating public charts as the main signal.

### 1. Clarify the scope

Before searching, pin down:

- region: global, US, Korea, Japan, China, Southeast Asia, UK, etc.
- time window: this week, last 30 days, 2026 so far, last few years
- platform: streaming, radio, YouTube, TikTok, club, live, sync, game/film
- genre frame: K-pop, EDM, drill, ballad, indie rock, musical theatre, etc.
- output goal: analysis, composition advice, reference list, or trend summary

### 2. Pull a small evidence set

Aim for:

- one chart or platform data source
- one industry/reporting source
- one listening/reference source
- optional user-provided listening-context source if the user has one
- optionally one local-language source for regional scenes

Do not overfit to a single chart. A song can be huge on YouTube and weak on Spotify; a TikTok sound can be viral without becoming a durable song; an album can dominate physical charts without reflecting streaming behavior.

### 3. Translate trend facts into craft variables

The agent's job is not to recite chart positions. Convert findings into compositional choices:

| Trend observation | Composition variables |
|---|---|
| Short intros / early hooks | hook timing, intro length, first vocal entry |
| Dance challenge success | groove clarity, countable phrase shape, chantable hook |
| Ballad resurgence | tempo, vocal range, lyric density, harmonic pacing |
| Club genre crossover | drum pattern, bass register, drop architecture, loop length |
| Regional folk/traditional fusion | instrument idiom, rhythmic cycle, language prosody, ornamentation |
| Streaming playlist fit | dynamic contour, skip-risk moments, vocal entry, arrangement density |

### 4. Report uncertainty clearly

Use language like:

- "Current chart evidence suggests..."
- "On this platform, not necessarily across the whole market..."
- "This is a reference cluster, not a rule."
- "This looks like a short-form trend; check whether it also appears on full-song charts."

Avoid:

- "Everyone is doing..."
- "The sound of 2026 is..."
- "This genre now means..."
- "This artist proves the whole market..."

## Search query patterns

Use precise queries:

```text
site:spotify.com/charts global top songs weekly [genre or region]
Billboard Global 200 methodology official streams sales
YouTube Charts Top Songs methodology official views UGC lyric videos
Circle Chart Global K-pop weekly chart official
Oricon weekly combined singles chart official
[genre] production trends 2026 producer interview
[artist] [song] producer breakdown BPM key form
```

For regional scenes, search in the relevant language when possible, then cross-check English-language summaries. If the user is subscribed to a music service or has public playlists, ask them to use that service as a listening source rather than relying only on general web search.

## Copyright/style guardrail

When a trend or reference answer involves a living artist, current hit, protected lyrics, melody, samples, or AI-assisted release claims, also load `style-reference-and-copyright.md`.

## Output patterns for the agent

### Current trend summary

```md
I checked this as a current-trend question, so I'd treat the static genre file as background only.

Current signal:
- Platform/chart evidence:
- Local/industry context:
- Listening pattern:

Composition takeaway:
1. ...
2. ...
3. ...

Caveat:
- This appears platform-specific / region-specific / short-form-specific.
```

### Reference-driven composition advice

```md
Instead of copying the reference, extract its craft variables:

- Tempo / meter:
- Groove:
- Harmonic rhythm:
- Hook timing:
- Topline contour:
- Arrangement density:
- Signature sonic element:
- What to avoid copying directly:
```

## Snapshot policy for genre files

Genre files should include a `Snapshot note` when they mention living scenes, current artists, platform behavior, or recent industry norms. The note should:

1. give the last review date,
2. say what is stable vs. current-sensitive,
3. link to this document and `reference-track-digging.md`; use `style-reference-and-copyright.md` when references or living artists are involved.

Update the snapshot date only after checking the file against current sources.

## Cross-references

- User listening context → `user-listening-context-and-streaming-services.md`
- Regional trend evolution → `regional-trend-evolution-analysis.md`
- Reference digging → `reference-track-digging.md`
- Style/copyright guardrails → `style-reference-and-copyright.md`
- Trend/reference matrices → `../../assets/trend-and-reference-matrices.md`

## Quick asset

- Fast web-search cheatsheet → `../../assets/web-search-cheatsheet.md`

# Regional Trend Evolution Analysis

> **Snapshot note:** Last content review snapshot: 2026-04-27. This file is an analytical framework, not a fixed history. Use active web research for current regional scenes, charts, platform behavior, and living-artist references.

This file helps the agent explain how a region's music trends are changing over time **analytically** rather than as trivia. The goal is not "a history of Korean pop" or "a history of Latin music." The goal is to translate regional trend movement into composition choices: groove, form, language, production density, collaboration patterns, platform behavior, and reference strategy.

Use this when the user asks:

- "How has [region/market] pop changed over the last few years?"
- "What changed in K-pop/J-pop/Latin pop/Afrobeats/UK dance/US rap lately?"
- "How do I make this sound current in [region] without copying hits?"
- "Why does [market] feel different from [market]?"
- "Which references should I use for a 2010s vs 2020s version of this style?"

## Core idea

Regional trend evolution is usually caused by several vectors moving at once:

1. **Platform vector** — radio, streaming, short-form video, playlisting, physical sales, fan purchases, club circulation, sync.
2. **Language/prosody vector** — lyric density, syllable rhythm, code-switching, hook phonetics, singability.
3. **Rhythm/groove vector** — imported grooves, local grooves, club/danceability, half-time/double-time interpretation.
4. **Harmony/topline vector** — loop-based harmony, ballad harmony, jazz/gospel influence, modal or drone-based centers.
5. **Arrangement/production vector** — density, loudness aesthetic, acoustic/electronic balance, vocal stack, drop architecture.
6. **Industry vector** — label systems, idol/fandom systems, indie distribution, producer networks, cross-border collaborations.
7. **Cultural-function vector** — dance, karaoke, worship, live band, social media, drama/film tie-in, club, festival, playlist.

A good answer names the vector, then turns it into craft decisions.

## Research workflow

### 1. Pick a time window

Avoid vague eras like "modern" or "old." Use a bounded comparison:

- last 12 months vs previous 3-5 years
- early 2010s vs late 2010s vs 2020s
- pre-short-form era vs short-form era
- pre-streaming-dominant vs streaming-dominant
- local-market era vs global-collaboration era

### 2. Pick the signal types

For each region, triangulate:

- official/local charts
- platform charts or editorial playlists
- local press or producer interviews
- 5-10 reference tracks
- one root/context source if the sound comes from a deeper tradition

### 3. Extract variables, not just names

```md
| Period / signal | Groove | Harmony | Topline | Form | Arrangement | Platform function |
|---|---|---|---|---|---|---|
| Earlier reference cluster | ... | ... | ... | ... | ... | ... |
| Current reference cluster | ... | ... | ... | ... | ... | ... |
| Composition takeaway | ... | ... | ... | ... | ... | ... |
```

### 4. Separate local norm from global crossover

A current regional hit may be:

- locally mainstream but globally obscure
- globally viral but locally atypical
- fandom-driven rather than general-market-driven
- video/short-form-driven rather than full-song-driven
- physical-sales-heavy rather than streaming-heavy
- club/live-scene-driven rather than chart-driven

Do not treat one signal as the whole region.

## Region lenses

These are lenses, not fixed histories. Always pair with current web research.

### United States / Canada

Useful vectors:

- streaming and playlist segmentation
- hip-hop/R&B, pop, country, Latin, rock, Christian/gospel, dance crossovers
- radio vs streaming vs short-form mismatch
- regional rap and country scenes
- nostalgia cycles and catalog listening

Composition questions:

- Is the track optimized for radio, streaming playlist, short-form hook, club, or live band?
- Does the song need an early hook or a slower narrative build?
- Is the groove rap/R&B-derived, country/rock-derived, dance-derived, or hybrid?

### UK / Ireland

Useful vectors:

- dance-music lineage: garage, grime, jungle/DnB, house, rave revival
- indie/alternative band cycles
- drill and rap scenes
- festival and radio culture
- cross-pollination with Afrobeats, Caribbean, South Asian, and European scenes

Composition questions:

- Is the track leaning club, radio-pop, rap, indie, or hybrid?
- Does the rhythm need swing/2-step logic, four-on-floor pressure, breakbeat energy, or live-band push?

### Korea

Useful vectors:

- idol group architecture: section contrast, member distribution, dance-break/post-chorus role
- domestic ballad and OST traditions
- global K-pop charting vs domestic-platform behavior
- fandom, performance video, short-form choreography, and physical-release systems
- K-trad fusion as specific idiom, not generic "Korean color"

Composition questions:

- Is this aimed at global K-pop, Korean domestic pop/ballad, OST, indie, or K-trad fusion?
- Where does the hook appear visually and vocally?
- Does the arrangement need member/voice contrast or a solo-artist arc?

### Japan

Useful vectors:

- J-pop, anime/song tie-ins, idol, city-pop revival, rock-band lineage, vocaloid/utaite/internet scenes
- physical/combined charts vs streaming/platform charts
- high-density melodies and rich harmonic motion in many pop/rock/anime-adjacent styles
- language prosody and mora timing

Composition questions:

- Does the melody need dense syllabic motion or a simpler global-pop hook?
- Is the reference anime/opening-sized, idol-oriented, band-oriented, city-pop-derived, or internet-scene-derived?

### Greater China / Mandarin and Cantonese markets

Useful vectors:

- Mandopop ballad lineage, Cantopop melody tradition, idol/platform pop, film/TV tie-ins
- tonal-language prosody constraints
- platform and territory fragmentation: Mainland China, Taiwan, Hong Kong, diaspora scenes
- rap/R&B/electronic crossovers

Composition questions:

- Which language and market is the target?
- Does the topline preserve tone intelligibility?
- Is the song ballad-forward, dance-pop, idol/platform pop, or singer-songwriter?

### Southeast Asia

Useful vectors:

- platform-local pop ecosystems: T-pop, V-pop, OPM, Indo-pop, Malay pop, etc.
- multilingual code-switching and local-language prosody
- local ballad, dance, idol, band, hip-hop, and folk/film influences
- YouTube and short-form visibility in many markets

Composition questions:

- Which country/language is the target?
- Does the hook need karaoke-ballad clarity, dance/short-form clarity, or band/live identity?

### Latin America and Latin diaspora

Useful vectors:

- reggaeton/urbano, Latin trap, regional Mexican/corridos tumbados, bachata-pop, salsa/tropical references, Latin pop ballad
- Spanish prosody and chantable hook design
- cross-border collaborations and US Latin market feedback loops
- dembow-derived groove vs regional/acoustic roots

Composition questions:

- Is this urbano, regional Mexican crossover, Latin pop ballad, tropical, club, or hybrid?
- Is the groove a real rhythmic identity or just a surface percussion pattern?

### Brazil

Useful vectors:

- Portuguese-language pop, MPB lineage, samba/pagode/bossa references, funk carioca, sertanejo, arrocha/piseiro and regional scenes
- strong local rhythmic identities and local-platform/community circulation
- difference between Brazil and Spanish-language Latin markets

Composition questions:

- Is the reference Brazilian pop, funk carioca, samba/pagode-informed, sertanejo-derived, MPB, or another local scene?
- Does Portuguese prosody drive the topline?

### South Asia

Useful vectors:

- film music industries by language/region, independent pop, Punjabi pop, devotional/semi-classical references, hip-hop/electronic crossover
- raga/tala-adjacent color vs full classical practice
- hook-heavy film context and choreography/video function
- multilingual markets and diaspora feedback loops

Composition questions:

- Which language/film industry or independent scene is the target?
- Is the raga/tala reference ornamental, structural, or only coloristic?
- Does the track need cinematic form shifts or pop-loop discipline?

### MENA / Turkish / Persian / Arabic diasporas

Useful vectors:

- Arabic pop, Turkish pop/arabesk-pop, Persian pop, North African raï/shaabi/mahraganat-adjacent scenes, Gulf pop, diaspora club fusions
- maqam/makam/dastgah-adjacent melodic identity, ornamentation, and rhythm
- club/electronic crossover and wedding/live performance functions

Composition questions:

- Which language/region/tradition is actually intended?
- Is the melodic material maqam-adjacent, Turkish-pop, Persian-pop, North African, Gulf, or diaspora-club-oriented?
- Are microtonal/ornamental details essential or only approximated?

### Sub-Saharan Africa

Useful vectors:

- Afrobeats/Afropop, amapiano, highlife/palm-wine lineage, Afro-house, gqom, bongo flava, coupé-décalé, gospel, local rap scenes
- diaspora collaborations and global playlisting
- groove/bass/percussion identity and dance function
- language, pidgin, local slang, call-response

Composition questions:

- Is the target West African pop, South African amapiano/gqom/house, East African pop, local gospel, rap, or diaspora fusion?
- What is the groove identity beyond percussion decoration?

### Europe beyond UK

Useful vectors:

- local-language pop and rap, Eurovision-adjacent songcraft, dance/electronic scenes, chanson/schlager/canzone roots, Balkan/folk fusion, Nordic pop writing networks
- national radio and festival ecosystems
- English-language vs local-language strategy

Composition questions:

- Is the target local-language radio, club/festival, Eurovision-like immediacy, indie scene, or global pop?
- Which local rhythmic/lyric conventions matter?

## Composition translation patterns

| Trend shift | Translate into |
|---|---|
| Short-form platform influence | earlier hook, clearer motif, danceable phrase length, strong visual cue |
| Streaming playlist fit | shorter intro, lower skip risk, consistent mood, controlled dynamic spikes |
| Global collaboration | hybrid groove, multilingual hook, modular sections, contrasting vocal identities |
| Local-root revival | idiomatic rhythm/instrument/prosody, not just surface timbre |
| Ballad resurgence | slower harmonic rhythm, vocal tessitura focus, lyric clarity, larger final chorus |
| Dance-club crossover | bass/drum architecture, DJ-friendly phrase grid, breakdown/drop energy |
| Nostalgia cycle | era-specific arrangement gesture plus modern form/low-end/vocal production |
| Idol/fandom ecosystem | member contrast, chantable hook, visual break, fan participation affordance |

## Output template

```md
I would treat this as a regional trend-evolution question, not a fixed genre rule.

Scope:
- Region/market:
- Time window:
- Platform signal:
- Reference cluster:

What seems to be changing:
1. Platform/function:
2. Groove:
3. Topline/prosody:
4. Form/arrangement:
5. Industry/collaboration pattern:

Composition advice:
- Keep:
- Modernize:
- Avoid:
- Concrete 4-8 bar move:

Caveat:
- This is platform/region/time-window specific, so current sources should be checked.
```

## Pitfalls

- **Do not write a whole region as one style.** Korea, Japan, Brazil, India, Nigeria, Turkey, and the US each contain many scenes.
- **Do not turn historical roots into current chart claims.** Roots explain possibilities; charts show current signals.
- **Do not equate diaspora crossover with local mainstream.** They can overlap, but often differ.
- **Do not flatten traditions into scales or percussion loops.** Use specific idiom and practitioner collaboration when needed.
- **Do not treat one viral platform as the entire market.** Compare with at least one full-song or local-market source.

## Cross-references

- Current web trend research → `web-music-trend-research.md`
- User listening context → `user-listening-context-and-streaming-services.md`
- Reference-track digging → `reference-track-digging.md`
- Style/copyright guardrails → `style-reference-and-copyright.md`
- Minor and hybrid genres → `../genres/minor-and-hybrid-genres.md`
- Language prosody → `../fundamentals/prosody-and-language.md`
- K-pop/J-pop → `../genres/kpop-jpop.md`
- C-pop / Southeast Asian pop → `../genres/cpop-and-southeast-asian-pop.md`
- Latin pop / reggaeton → `../genres/latin-pop-and-reggaeton.md`
- Brazilian pop and funk → `../genres/brazilian-pop-and-funk.md`
- Afrobeats / amapiano → `../genres/afrobeats-and-amapiano.md`
- MENA pop → `../genres/mena-pop.md`
- South Asian film-pop → `../genres/south-asian-film-pop.md`

## Quick asset

- Fast web-search cheatsheet → `../../assets/web-search-cheatsheet.md`

# User Listening Context and Streaming-Service Research

> **Snapshot note:** Last content review snapshot: 2026-04-27. Platform features, APIs, chart pages, privacy rules, and personalized recap products change quickly. Use this workflow for personal-listening context, but verify current platform details with active web research when needed.

This file helps the agent use the user's own listening ecosystem as a composition-research input. It is especially useful when the user says they want something that fits their taste, playlist, local market, favorite scene, DJ set, label pitch, or reference library.

The agent should **not** claim it can access a streaming account unless a separate tool or explicit user-provided data is available. The normal workflow is to ask the user to share lightweight, privacy-safe evidence: playlist links, track names, screenshots, top-song lists, public charts, or a short description of what they listen to.

## When to use this file

Use this file when the user asks for:

- references that match their taste
- current style digging based on what they already listen to
- a playlist-compatible song idea
- music for a DJ set, playlist pitch, or personal release strategy
- market research where the user's available streaming service matters
- a comparison between platform charts and the user's own library
- guidance like "I have Spotify / Apple Music / YouTube Music / Melon — what should I check?"

Pair this with:

- `web-music-trend-research.md` for current market signals
- `reference-track-digging.md` for turning tracks into craft variables
- `style-reference-and-copyright.md` for living-artist or too-close imitation requests
- `regional-trend-evolution-analysis.md` when the user asks about region/time-shift patterns

## Safety and privacy baseline

Never ask the user for passwords, account tokens, private login links, or anything that requires account takeover.

Safe inputs include:

- public playlist links
- track/artist names copied from a playlist
- 5-20 reference tracks typed by the user
- screenshots the user chooses to share
- year-end or monthly recap summaries the user chooses to quote
- exported listening logs from a service or scrobbling tool, if the user already has them
- charts or editorial playlists visible without private account access

If the user shares listening-history data, treat it as taste evidence, not identity evidence. Do not infer sensitive traits. Stick to musical variables.

## First question pattern

When personal listening context would materially improve the answer, ask a small, optional question:

```text
Do you use a streaming service for reference digging — Spotify, Apple Music, YouTube Music, Melon, Genie, Bugs, SoundCloud, Bandcamp, Last.fm, or something else? If yes, share 5-10 track names or a public playlist link. No login details needed.
```

If the user does not answer, continue with public charts and general reference-track workflow.

## Service-aware research map

| Service / source | Useful user-provided signal | Useful public signal | Watch-outs |
|---|---|---|---|
| Spotify | public playlists, liked-track excerpts, top tracks if user provides them | Spotify Charts: global, country, city, genre, local-pulse views | Playlist ecosystem can bias taste toward platform/editorial norms |
| Apple Music | Replay summaries, playlists, library excerpts | Apple Music top charts, city charts, editorial playlists | Personal Replay is account-specific; ask user to summarize rather than requesting account access |
| YouTube / YouTube Music | playlists, liked videos, shorts/sound references | YouTube Charts, official videos, UGC-heavy signals, creator ecosystem | Video virality and full-song streaming can diverge |
| Melon / Genie / Bugs / VIBE | Korean chart/listening context, playlists, favorites | domestic Korean chart and platform signals | Use with Circle Chart for broader K-pop market context |
| Oricon / Billboard Japan / local Japan services | user references and playlists | local chart signals for Japan | Physical, download, streaming, and combined charts tell different stories |
| SoundCloud | public likes/reposts/playlists, underground references | tag pages and community charts | Strong for emerging scenes, weak as a whole-market proxy |
| Bandcamp | collections, wishlists, label pages | labels, tags, scene pages, fan collections | Good for niche/indie scenes; not a mainstream popularity proxy |
| Last.fm / scrobbling logs | long-term listening history, top tracks by time period | artist/tag pages and community data | Great for taste mapping if the user already scrobbles; still self-selecting |
| TikTok / short-form platforms | saved sounds or trend links | official creative/trend pages and sound usage | Sound virality is noisy; verify with full-track listening and charts |

## What to extract from a user's listening data

Do not just list the user's favorite artists. Convert their listening context into compositional variables.

| Evidence | Composition variable |
|---|---|
| Many tracks around one BPM band | likely tempo comfort zone or playlist slot |
| Repeated drum/groove family | pocket, subdivision, kick/snare density |
| Similar vocal registers | topline tessitura and performance energy |
| Repeated form pattern | intro length, hook timing, bridge/drop/post-chorus expectations |
| Repeated harmonic color | modal, functional, minor-loop, neo-soul, gospel-adjacent, blues-derived, etc. |
| Repeated language/region | prosody, pronunciation, local lyric density, rhythmic accent |
| Repeated release era | production density, arrangement pacing, nostalgia layer |
| Repeated platform context | playlist fit, video/short-form hook, club/DJ utility, radio fit |

## Minimal workflow

### 1. Ask for the lightest useful evidence

Use one of these:

```text
Share 5 tracks that are exactly the lane, plus 2 tracks that are close but wrong.
```

```text
Share a public playlist link or paste the first 10 tracks. I will treat it as a reference cluster, not something to copy.
```

```text
If you use Apple Music Replay / Spotify Wrapped / Last.fm, paste your top 10 artists or tracks and tell me which ones are relevant to this project.
```

### 2. Separate taste from target

The user may listen to one thing and want to write another. Ask or infer:

- Is this for your personal taste?
- Is this for a specific audience or platform?
- Are these references emotional, sonic, structural, or commercial?
- Which reference is the anchor, and which is just context?

### 3. Build a reference grid

```md
| Source | Role | What to extract | What to avoid copying |
|---|---|---|---|
| User playlist track 1 | anchor | tempo/groove/hook timing | melody, lyrics, signature riff |
| User playlist track 2 | boundary | what feels too bright/clubby | exact drum fill |
| Public chart track | market signal | current form/arrangement norm | chart-chasing cliché |
```

### 4. Compare personal taste vs market signal

| Scenario | Advice |
|---|---|
| User taste and charts align | Use the user's references as anchors; charts confirm the lane |
| User taste is older than current charts | Add a modern arrangement variable while preserving the user's core taste |
| User taste is niche but project is mainstream | Keep one niche signature element; simplify form/groove for broader fit |
| User taste is mainstream but project wants distinct identity | Use boundary references to avoid generic defaults |
| Platform signals conflict | Explain which platform the advice is optimizing for |

## Platform-specific prompts

### Spotify-oriented reference pass

```text
Check your public playlists, top tracks if available, and Spotify Charts for the target region/genre. Extract: tempo band, intro length, hook timing, groove family, vocal entry, loop length, and density arc.
```

### Apple Music-oriented reference pass

```text
Use your Replay/monthly insights or playlists to identify durable favorites, then compare them with Apple Music top charts or city charts for the target region. Extract: artist cluster, song length, lyric density, chorus lift, and arrangement polish.
```

### YouTube / short-form-oriented reference pass

```text
Use YouTube Charts and saved music/short-form references to see which hooks work visually or socially. Then verify whether the full song has a strong structure, not just a viral 10-second moment.
```

### Korean-platform-oriented reference pass

```text
For K-pop or Korean domestic pop, compare the user's Melon/Genie/Bugs/VIBE references with Circle Chart and current label/artist releases. Extract: topline density, section contrast, choreography-friendly hook, fan-chant or singalong affordance, and ballad vs dance-market expectations.
```

### Niche-scene pass

```text
For Bandcamp/SoundCloud/Last.fm-heavy scenes, prioritize tags, labels, community playlists, and user-liked tracks. Use mainstream charts only as contrast, not as the main truth source.
```

## Output template

```md
Your listening-context signal:
- Personal anchors:
- Boundary tracks:
- Platform/market signal:

Craft variables to use:
1. Tempo/groove:
2. Harmony:
3. Topline:
4. Form/hook timing:
5. Arrangement density:

Avoid copying:
- Exact melody/lyrics/riff/sample/vocal identity

Concrete next move:
- 4-8 bar sketch / reference matrix / playlist-compatible arrangement plan
```

## Pitfalls

- **Do not overfit to the user's top songs.** A top track may be emotionally important but irrelevant to the current project.
- **Do not equate playlist popularity with composition quality.** It is a signal, not a law.
- **Do not ask for private credentials.** Use public links or user-pasted summaries.
- **Do not assume one platform equals the whole market.** Each service has different demographics, catalog emphasis, and chart rules.
- **Do not copy the user's references.** Extract variables and generate a new composition.

## Cross-references

- Current trend research → `web-music-trend-research.md`
- Reference-track digging → `reference-track-digging.md`
- Style/copyright guardrails → `style-reference-and-copyright.md`
- Regional trend evolution → `regional-trend-evolution-analysis.md`
- Pre-production decisions → `../production-aware/pre-production-decisions.md`
- Prosody and language → `../fundamentals/prosody-and-language.md`
- Response templates → `../../assets/response-templates.md`

## Quick asset

- Fast web-search cheatsheet → `../../assets/web-search-cheatsheet.md`

# Style Reference and Copyright Guardrails

> **Snapshot note:** Last content review snapshot: 2026-04-26. Copyright rules, AI policies, platform rules, lawsuits, and licensing norms change quickly. Use this file for durable creative guardrails, and verify current legal/platform facts through active web research before giving high-stakes advice.

This file helps the agent respond when the user asks for music "in the style of" an artist, song, producer, label, era, or scene. It is not legal advice. It is a composition workflow for staying original while still learning from references.

## Core principle

A style request should become a **craft-variable request**, not a clone request.

Instead of:

```text
Write a chorus exactly like [living artist / current hit].
```

Transform into:

```text
Let's extract transferable variables: tempo range, groove, harmonic rhythm, section contrast, vocal register, hook timing, arrangement density, and emotional arc — then write a new idea that avoids the reference's melody, lyrics, riff, sample, and signature vocal identity.
```

## What is usually safer to extract

These are broad musical ideas. They can still become risky if combined too closely with one specific reference, but they are the right starting point.

| Extractable variable | Example |
|---|---|
| Tempo / meter | 96 BPM, 4/4, half-time chorus |
| Groove family | dembow-like, boom-bap, shuffle, four-on-floor, 6/8 ballad |
| Harmonic palette | minor add9 loop, backdoor cadence, modal vamp, gospel-adjacent plagal movement |
| Harmonic rhythm | one chord per bar, two-bar vamp, pre-chorus acceleration |
| Form role | hook before 0:40, post-chorus chant, bridge breakdown |
| Arrangement density | sparse verse → dense chorus; low verse → octave-wide chorus |
| Topline strategy | narrow verse, higher chorus peak, repeated 3-note hook |
| Lyric craft | conversational verse, slogan-like chorus, internal rhyme density |
| Performance energy | whispered, belted, chant-like, call-response — without copying a specific voice |

## What to avoid copying

| Avoid | Why |
|---|---|
| Exact or near-exact melody | Melody is one of the most recognizable protected elements in songs |
| Copyrighted lyrics or close paraphrases | Lyrics are protected expression, not just style |
| Signature riff / hook / motif | Often the identity of the recording/composition |
| Samples from recordings | Sound recordings require separate clearance in many workflows |
| Distinctive vocal identity of a living artist | Ethically risky and often platform/policy-sensitive |
| Producer tags / exact sound-design signatures | Can imply unauthorized imitation or confuse attribution |
| One-reference cloning | Even if each element is small, the combination can become recognizably derivative |

## Living artist and current-hit requests

When the user names a living artist or very recent hit:

1. Acknowledge the target energy.
2. Refuse to clone exact expressive elements.
3. Translate into neutral craft variables.
4. Offer an original sketch.
5. Recommend using multiple references.

Template:

```md
I can help capture the broad craft traits — [tempo/groove/density/hook strategy] — but I won't copy [artist]'s melody, lyrics, vocal identity, or signature riff.

Original direction:
- Tempo: ...
- Groove: ...
- Harmony: ...
- Topline: ...

Playable sketch:
...
```

## Historical, public-domain, and genre requests

Requests like "baroque sequence", "bebop head", "Motown-inspired bassline", "early rock and roll", or "1960s girl-group energy" are usually easier to handle because they target a broader idiom or historical convention. Still avoid copying a specific protected song.

Safer phrasing:

- "Use a baroque-style circle sequence."
- "Use bebop-like enclosure and chromatic approach tones."
- "Use a 1960s soul/pop backbeat, call-response backing vocals, and I-vi-IV-V harmonic language."
- "Use a classic funk clavinet-style syncopation without copying a particular riff."

## Reference-stack method

For any style request, build a reference stack:

| Layer | Purpose |
|---|---|
| Anchor references | The desired energy and market frame |
| Boundary references | What the user does *not* want |
| Root references | Older idiom sources that explain the groove/harmony/vocal language |

A three-track anchor set is safer and more useful than one target track. The agent should identify shared variables across references, not imitate the most famous one.

See `reference-track-digging.md` for the full workflow.

## Anti-copy transformation moves

If a generated idea feels too close to the reference, transform at least two independent dimensions:

| Dimension | Transformation |
|---|---|
| Melody | Change contour, interval skeleton, phrase length, and climax placement |
| Rhythm | Change pickup, syncopation, note density, or bar placement |
| Harmony | Keep mood but change chord order, bass motion, or cadence |
| Form | Move hook earlier/later, change pre-chorus length, add/remove post-chorus |
| Arrangement | Change register, instrumentation, or density arc |
| Lyric | Change image field, rhyme scheme, syntax, and hook wording |
| Groove | Keep broad family but change kick/bass relationship or subdivision |

A single cosmetic change is not enough when the reference remains recognizable.

## Copyright concepts the agent should keep straight

Use this as plain-language guidance, not legal advice.

| Concept | Composition relevance |
|---|---|
| Idea vs. expression | Broad ideas, methods, and style traits are different from specific melodies, lyrics, riffs, and recordings |
| Composition vs. sound recording | A song's melody/lyrics/harmony chart and a particular recording can be separate rights |
| Sample vs. interpolation | Sampling uses an actual recording; interpolation re-records/recreates musical material and still may require composition rights |
| Public domain | A composition may be public-domain while a modern recording of it is not |
| Fair use | Context-specific legal analysis; do not promise that a use is fair use |
| AI-assisted output | Human authorship and creative control may matter for registration/ownership in some jurisdictions |

## AI-assisted composition guardrails

For AI-assisted workflows, the agent should encourage human creative control:

- document human decisions: prompt intent, selection, editing, arrangement, performance, lyrics, melody revisions
- avoid using AI to mimic a living artist's voice or exact style
- treat generated outputs as drafts requiring human curation
- verify platform, distributor, and copyright rules when the user plans release
- avoid claiming copyrightability or clearance certainty without current legal review

## When to web-check

Use active web research when the user asks about:

- current lawsuits, licensing standards, AI music policies, or copyright office guidance
- current artist/reference comparisons
- platform/distributor rules for AI-generated or AI-assisted music
- whether a particular work is public domain in a jurisdiction
- sample/interpolation clearance norms for a current release workflow

See `web-music-trend-research.md` for source hierarchy and search patterns.

## External source anchors to verify current facts

Check current versions of:

- U.S. Copyright Office — copyright basics, AI reports, registration guidance
- WIPO — general copyright concepts across jurisdictions
- official distributor/platform policies when releasing music
- current court filings or reputable legal reporting for active lawsuits
- rights-holder or publisher information for specific works

## Good response examples

### User: "Make it sound like [current artist]."

```md
I can aim for a similar broad lane — dark mid-tempo pop with a sparse verse, early hook, and wide chorus — but I'll keep the melody, lyric, vocal identity, and signature production elements original.

Try this original sketch:
Key: F minor, 92 BPM
Verse: | Fm(add9) | D♭maj7 | A♭/C | E♭sus |
Pre:   | B♭m7 | C7♭9 | Fm | E♭ |
Chorus:| D♭ | A♭ | E♭ | Fm |
Topline: narrow verse around ^3-^5, chorus peak on ^6, repeated three-note hook.
```

### User: "Use the exact chorus melody but change the chords."

```md
I can't provide a close derivative of a recognizable copyrighted chorus melody. I can analyze the contour in abstract terms — for example, 'short repeated pickup, leap to peak, stepwise fall' — and write a new melody with a different contour and rhythm.
```

### User: "Can I release this if it uses AI?"

```md
That depends on jurisdiction, platform rules, the amount of human authorship, and whether any protected material or voice likeness was used. For composition work, document your human creative decisions and check current copyright/platform rules before release.
```

## Cross-references

- Reference digging → `reference-track-digging.md`
- User listening context → `user-listening-context-and-streaming-services.md`
- Regional trend evolution → `regional-trend-evolution-analysis.md`
- Minor and hybrid genres → `../genres/minor-and-hybrid-genres.md`
- Current web research → `web-music-trend-research.md`
- AI-assisted composition → `../techniques/algorithmic-and-AI-assisted.md`
- Pre-production reference decisions → `../production-aware/pre-production-decisions.md`
- Diagnostic anti-copy checklist → `../../assets/diagnostic-checklists.md`
- Response templates → `../../assets/response-templates.md`

# Rhythmic Devices

This file covers the techniques that *manipulate* meter to create rhythmic interest — syncopation, hemiola, displacement, polyrhythm at the small scale, accent patterns, and rhythmic motivic devices. It's the tool kit for making rhythm interesting beyond the metric default.

For meter and beat hierarchy fundamentals, see `fundamentals/rhythm-meter.md`. For groove (the feel/microtiming side of rhythm), see `groove-and-feel.md`. For odd meters and large-scale polyrhythm, see `odd-meters-polyrhythm.md`.

## When to consult this file

- The user asks about syncopation, swing, "rhythmic interest"
- A piece "feels stiff" or "too on-the-grid"
- Hemiola, polyrhythm at the bar level, accent shifts
- Cross-rhythm, displacement, anticipation
- Why a Latin/Afro-Cuban/funk groove "feels different"
- How to make a melody rhythmically distinctive

## The metric default — what we're working against

Every rhythmic device is interesting only against the backdrop of metric expectation. In 4/4, the listener expects:

- Beat 1: strongest
- Beat 3: medium-strong
- Beats 2 and 4: weaker
- Off-beats (the "and" of each beat): weakest

When music confirms this, it feels grounded but predictable. Every rhythmic device works by violating, displacing, or layering against this expectation.

## Syncopation

Accenting weak beats or off-beats — placing emphasis where the meter doesn't expect it. The fundamental rhythmic "spice".

Three flavors:

### 1. Off-beat accents

Accenting the "and" of a beat. The most common form of syncopation in pop, jazz, and funk.

Example in 4/4: instead of `1 - 2 - 3 - 4 - 1 - 2 - 3 - 4`, accenting `1 -2 _ 3 -4 _` (where the dash is an accent and underscore is no accent — i.e., accents on the and-of-1, and-of-3, etc.).

Reggae's "skank" is the prototype — guitar chops on beats 2 and 4, sometimes pushed to the off-beats.

### 2. Tied syncopations

A note beginning on a weak beat or off-beat and tied across a strong beat. The strong beat gets no fresh attack — the note is sustained, so the strong beat is *swallowed* by the tied note.

Example: a half-note starting on the "and of 1" of a 4/4 bar ties through beat 2. The downbeat of beat 2 has no new attack.

This is the engine of the funk feel. James Brown's "Funky Drummer" is built on tied syncopations.

### 3. Anticipation / push

A note arrives slightly before its expected position — usually a chord change pushed to the "and of 4" instead of beat 1. Common in jazz, R&B, Latin music.

Notation: a chord change written one eighth-note (or one sixteenth) earlier than the bar line. The next bar's downbeat is empty (or has a continuation of the anticipated chord).

This creates a forward-leaning feel — the music "leans into" the next bar.

## Hemiola

A 3-against-2 (or 2-against-3) cross-rhythm. In a triple meter (3/4 or 6/8), a passage temporarily groups notes as if in duple meter, or vice versa.

### Hemiola in 3/4

Six eighth-notes in a 3/4 bar can be grouped as:

- **Default**: `123 / 456` (two groups of three) — confirming the 3/4 meter.
- **Hemiola**: `12 / 34 / 56` (three groups of two) — implying 3/2 or 6/8 over the bar lines.

Across two bars of 3/4, six quarter notes can suddenly group as three pairs instead of two triplets. The meter pulls in two directions.

### Hemiola in 6/8

The reverse: 6/8 groups as `123 / 456` (compound 2). A hemiola can group as `12 / 34 / 56` (simple 3).

### Where hemiola appears

- Renaissance music constantly. Polyphonic music alternates groupings freely.
- Baroque (especially Brahms in 19th-century style as homage). Brahms is famous for hemiola.
- Latin music — son clave and other Cuban rhythms layer 3-and-2 patterns continuously.
- Math rock and progressive rock — "shifty" feel from constantly varying groupings.
- "America" from West Side Story — the famous "I-like-to-be-in-a-MER-i-ca" alternates 6/8 and 3/4 (hemiola at the bar level).

## Cross-rhythm

A specific rhythmic pattern in one voice that conflicts with the prevailing meter, played simultaneously with a voice that confirms the meter.

A 4-against-3 cross-rhythm: in a measure of 12 sixteenths, one voice plays four equally-spaced notes (every third sixteenth) while another plays three equally-spaced notes (every fourth sixteenth). Both voices fit the bar; their accents fall in different places.

Common in:

- West African drumming (the foundational technique).
- Afro-Cuban music (clave + tumbao + cascara, all interlocking).
- Brazilian rhythms (samba, maracatu).
- Modern progressive metal (Meshuggah, Animals as Leaders).
- Chopin (the "Fantasie-Impromptu" has 4-against-3 hands).

## Polyrhythm vs. polymeter

These terms get confused. The distinction:

- **Polyrhythm**: two or more rhythms with different note groupings, sharing the same meter. The barlines are the same. The 4-against-3 example above is polyrhythm.
- **Polymeter**: two or more meters happening simultaneously. The barlines are different. One voice in 5/4, another in 4/4 — they line up every 20 beats.

Polyrhythm is more common (in fact ubiquitous in many traditions). Polymeter is rarer and often appears in 20th-century classical and prog rock.

For full treatment, see `odd-meters-polyrhythm.md`.

## Displacement

A motif played starting at a different metric position than expected. The motif's identity is preserved (same pitches, same rhythm internally), but it lands in a different relationship to the bar.

Example: a 4-note motif that originally starts on beat 1 is later restated starting on beat 2. The motif sounds the same internally but feels disoriented.

Common in:

- Jazz. Bebop solos use displacement to play the same lick "off" the changes.
- Steve Reich's phase music (where two voices play the same line but slowly drift).
- Modern hip-hop production (a sampled loop that doesn't align to the song's downbeat).

## Accent patterns and metric ambiguity

Sometimes the meter is technically 4/4 but the accent pattern suggests something else.

A bar of 4/4 with accents on beats `1 - and-of-2 - 4` creates a 3+3+2 sixteenth-note pattern (or a syncopated 2+3+3 if counted differently). The bar is in 4/4 but feels like it could be 8/8 grouped 3+3+2.

This is the "tresillo" rhythm — foundational in Cuban music, hip-hop, reggaeton, and modern pop. The 3+3+2 grouping is so common it has its own name.

Other common groupings of 8 sixteenths:

| Pattern | Sounds like |
|---------|-------------|
| 4+4 | Plain straight |
| 3+3+2 | Tresillo / dembow / reggaeton |
| 2+3+3 | Reverse tresillo |
| 3+2+3 | Habanera / tango |
| 2+2+2+2 | Plain |

## Anticipation and delay

### Anticipation (covered above as syncopation)

A note arrives early. Pop and gospel use anticipation extensively, especially at chord changes.

### Delay (suspension)

A note arrives late. The expected attack on a strong beat is delayed by an eighth or sixteenth. The listener feels the delay as a stretching.

Less common as a deliberate rhythmic device, but it's the rhythmic equivalent of suspension in voice leading.

## Augmentation and diminution at the small scale

Beyond their use in motivic development (`melody/motivic-development.md`), augmentation and diminution are rhythmic devices in their own right.

- **Augmentation**: a motif played at half speed (each note's duration doubled). Effect: grand, slowed-down.
- **Diminution**: a motif played at double speed. Effect: agitated, intensified.

A motif in eighth notes that suddenly appears in sixteenth notes is a diminution. The same motif in quarter notes is an augmentation.

These are common in fugues, classical development sections, and progressive rock arrangements.

## Rhythmic motifs

Just as melody has motifs, rhythm has them. A *rhythmic motif* is a recognizable pattern of note durations that recurs and gets developed.

Beethoven's 5th opening: short-short-short-LONG. The pitches change; the rhythm is the identity. This is a rhythmic motif.

Common rhythmic motifs:

| Pattern | Examples |
|---------|----------|
| Short-short-short-long | Beethoven 5; ubiquitous in classical |
| Long-short-short | Trochee in poetic terms; common in march music |
| Long-short-long | Iambic; conversational |
| Dotted-eighth + sixteenth | "Scotch snap"; folk and country |
| Sixteenth + dotted-eighth | "Lombard rhythm"; Baroque |
| Triplet quarter-note feel | Shuffle; blues, swing |
| 16th-16th-8th | "Tresillo" subdivisions; ubiquitous in pop |

A piece organized around a rhythmic motif develops it just as it would a melodic motif: repeat, sequence, augment, diminish, displace, fragment, combine.

## Polymetric riffs (briefly — see `odd-meters-polyrhythm.md` for depth)

A riff that has its own meter different from the song's underlying meter. A 7-beat guitar riff over a 4/4 drum pattern (so the riff and the drums realign every 7 bars).

Common in math rock, progressive metal, modern jazz, and certain contemporary classical music.

## Microtiming and pocket (preview — full in `groove-and-feel.md`)

The above techniques manipulate rhythm at the *notation* level. But within any notated rhythm, the actual timing of notes can be slightly ahead or behind the strict grid. This is microtiming, and it's what gives a groove its feel.

A drum pattern can be perfectly notated and still feel "stiff" or "behind the beat" or "on top" depending on microtiming. See `groove-and-feel.md` for full treatment.

## Genre notes — rhythmic devices by tradition

### Pop and rock

- Mostly 4/4, mostly straight subdivisions.
- Anticipated chord changes (push the "1" to the "and of 4") common in pop, especially soft rock and country.
- Syncopated vocal phrasing (verse melodies that don't land on downbeats) common in modern pop.
- Half-time feel sections to create lift.

### Hip-hop and R&B

- Foundational use of tresillo (3+3+2) in much modern trap and Latin-influenced hip-hop.
- Syncopated hi-hat patterns: 16th, 32nd, even 64th-note triplet rolls in trap.
- Vocal flow plays heavily with displacement, anticipation, and polyrhythm against the beat.

### Jazz

- Swing eighths (the foundational rhythmic feel — see `groove-and-feel.md`).
- Anticipation at chord changes.
- Rhythmic displacement in solos (playing a phrase "across" the bar).
- Polyrhythmic comping (drummer plays in 3 against the bass in 4).

### Latin (Afro-Cuban, Brazilian)

- Clave is the foundational pattern — a 2-bar rhythmic pattern that organizes the entire piece.
- Tumbao, cascara, montuno — interlocking polyrhythms in percussion and piano.
- Habanera, tresillo, samba, bossa nova, son montuno — each has a distinctive rhythmic profile.

### Funk

- Heavy syncopation, often tied across strong beats.
- The "one" emphasized while everything else is off the grid.
- 16th-note grids with selective notes — silence is as important as sound.

### Electronic / EDM

- Mostly 4/4 with strong downbeats.
- Subdivisions vary by sub-genre: 4-on-the-floor (house, techno), broken-beat (DnB, breaks), syncopated (UK garage, dubstep).
- Polyrhythm in advanced producers (Aphex Twin, Squarepusher, Burial).

### African traditional

- The deep root of polyrhythm. West African drumming uses interlocking polyrhythms continuously, with cross-rhythm as the norm rather than the exception.
- Listening for the "downbeat" is culture-specific — different listeners may hear different beats as "1" of the same pattern.

### Indian classical

- Tala (rhythmic cycle) of various lengths (often 16 or 8 beats; some longer).
- Rhythmic improvisation organized within the tala, with cycles of accent and release built into the system.
- Distinctive rhythmic devices like *tihai* (a phrase repeated 3 times to land on the downbeat).

## Common pitfalls when advising

- **Don't suggest "more syncopation" without specifying.** Syncopation in pop is different from in funk is different from in jazz. The genre dictates which weak beats and how much.
- **Don't add polyrhythm to music that isn't designed for it.** A pop song doesn't usually want a 5-against-4 pattern in the synth — it'll fight the vocal.
- **Don't ignore the half-time vs. full-time choice.** Many tracks switch between half-time and full-time feel for sectional contrast. Hip-hop often is "in 4/4 felt as half-time" (the snare on 3 instead of 2 and 4).
- **Don't assume notation equals feel.** Two drummers can play the same notated pattern with different feels (laid back, on top, swinging). Microtiming is invisible on the page.
- **Don't overuse rhythmic devices.** A song with hemiola in every bar exhausts itself. Use rhythmic devices selectively, at structural points.

## Quick decision matrix

| Want | Try |
|------|-----|
| Forward-leaning feel | Anticipated chord changes (push the 1) |
| Funky, "behind the beat" energy | Tied syncopations across strong beats |
| Latin / reggaeton feel | Tresillo (3+3+2) in the bass and drums |
| Polyrhythmic interest | 3-against-2 hemiola, especially in 6/8 → 3/4 |
| Surprise / disorientation | Rhythmic displacement of a familiar motif |
| Speed up energy without changing tempo | Diminution (notes get faster within sustained tempo) |
| Slow down / grand feel | Augmentation (notes get longer) |
| African / world feel | Cross-rhythm, interlocking percussion patterns |
| Jazz solo "tension" | Play a phrase displaced by an eighth or quarter |
| Trap / modern hip-hop | 16th-note hat patterns with rolls; tresillo in the kick |

## Cross-references

- Meter, beat hierarchy fundamentals → `fundamentals/rhythm-meter.md`
- Microtiming and groove feel → `groove-and-feel.md`
- Odd meters, full polymeter, large-scale polyrhythm → `odd-meters-polyrhythm.md`
- Rhythmic motifs and their development → `melody/motivic-development.md`
- Phrase rhythm and how it interacts with meter → `melody/phrase-structure.md`
- Genre-specific rhythmic conventions → relevant `genres/*.md` files
- Production-side rhythm decisions → `production-aware/energy-and-dynamics.md`

# Lyric Writing

This file is about the craft of lyrics — what makes them work, what makes them fail, and how they interact with melody, rhythm, and the song's emotional architecture. It covers prosody, rhyme, imagery, structure, point of view, and the specific demands of lyrics for different genres.

For melody-as-vocal-line, see `topline-craft.md`. For the hook (which is partly lyrical), see `hooks-and-memorability.md`. For phrase structure (which constrains lyric phrasing), see `melody/phrase-structure.md`. For lyric writing in musical theatre context, see `genres/musical-theatre.md`.

## When to consult this file

- The user is writing lyrics or wants feedback on lyrics
- A song's words feel "off" or unmemorable
- Questions about prosody (matching syllables to melody)
- Rhyme scheme decisions
- Imagery, metaphor, specificity
- Point of view, narrator, address
- "I have a great melody but the lyrics aren't working"

## What lyrics do

Lyrics carry several burdens simultaneously:

1. **Communicate meaning** — what the song is "about"
2. **Provide rhythmic syllables** for the melody to ride
3. **Provide vowels and consonants** that create singable, memorable sound
4. **Express character or persona** — who is singing
5. **Create emotional arc** — words that intensify alongside the music

A great lyric does all five. A passable one does three. A failed one might do only one.

## The prosody problem

Prosody is the match between word stress and musical stress. English words have natural stress patterns ("ELephant" not "eleEPHant"; "comPUter" not "COMputer"). When lyrics are set to music, the strong beats of the melody should fall on the strong syllables of the words.

When prosody is wrong, the listener hears it immediately:

- Word stress on weak beats: sounds awkward, like the singer is mispronouncing.
- Polysyllabic word with stress mid-word landing on a weak beat: classic mismatch.
- Articles ("the", "a") falling on strong beats: rhythmic emphasis on grammar instead of meaning.

Examples of bad prosody:

- "I am DOWN" instead of "I AM down" — mismatch.
- "she SAID it WAS just a JOKE" with "was" on the strong beat — bad.

Fixes:

- Rewrite the lyric so the strong-beat syllables are the stressed ones.
- Adjust the melody's accent pattern to match the lyric.
- Change which note holds longer (longer notes feel emphasized).

Prosody is the most common technical lyric mistake. Good songwriters fix prosody constantly during writing.

## Imagery and specificity

The "show, don't tell" rule from prose applies doubly to lyrics. Generic emotional words ("I'm sad", "I love you", "we're forever") are forgettable; specific images and sensory detail are memorable.

Compare:

- Bad: "I miss you so much"
- Better: "Your coffee cup is still on the kitchen table"
- Bad: "I feel free"
- Better: "Riding shotgun with the windows down on the way to nowhere"

Specificity creates the mental image. The listener fills in their own version. The general statement ("I miss you") doesn't conjure anything.

### When to be general

Choruses can be more general than verses — the chorus is the universal statement, the verses are the specific story. "Don't Stop Believin'" — verses describe a small-town girl and a city boy; the chorus is universal advice.

But even chorus lyrics benefit from imagery. "Bohemian Rhapsody"'s "Mama, just killed a man" is visceral; "Bohemian Rhapsody"'s lyrics generally are filled with strange specifics that make it unforgettable.

### Sensory detail

Lyrics work when they evoke senses:

- Sight: colors, light, motion, place
- Sound: voice, weather, environment
- Touch: skin, fabric, temperature
- Smell, taste: less common but powerful when used

A song with no sensory imagery is purely conceptual; the listener doesn't "see" anything.

## Rhyme

Rhyme is a tool, not a requirement. Used well, it provides structure and pleasure; used badly, it hijacks the lyric's meaning.

### Types of rhyme

| Type | Example | Effect |
|------|---------|--------|
| Perfect rhyme | cat / hat | Strong, classical, satisfying |
| Slant / near rhyme | mind / time | Subtle, conversational, modern |
| Internal rhyme | "I tried to ride the tide" | Within a line; adds density |
| Multisyllabic rhyme | "complicated / animated" | More elaborate; common in hip-hop |
| Eye rhyme | love / move | Visually rhymes but doesn't sound rhyme — usually a mistake |
| Mosaic rhyme | "tomorrow" / "borrow" — multi-word | Clever; common in patter songs and rap |
| Identity rhyme | bear / bear (same word) | Usually weak; sometimes deliberate for emphasis |

### Rhyme schemes

Common patterns:

- **AABB**: couplet rhyme. Simple, often country/folk feel.
- **ABAB**: alternating. Standard for verses; gives forward momentum.
- **ABBA**: enclosed. Less common; feels reflective.
- **AAAA**: monorhyme. Dense; common in hip-hop and certain blues.
- **No rhyme**: free verse. More common in modern indie, rap (less rhymed), confessional songwriting.

### When rhyme hurts

- **Forced rhyme**: changing the lyric's meaning to fit a rhyme. The listener notices when "moon" appears just because "soon" rhymed.
- **Predictable rhyme**: easy rhymes (heart/start, time/mine, love/above) feel cliché.
- **Rhyme over sense**: prioritizing rhyme over what you're trying to say.

Sondheim's principle: never sacrifice meaning for rhyme. If you can't find a rhyme that maintains the lyric's logic, rewrite the line or use slant rhyme.

### Modern rhyme conventions

Modern pop and indie often use slant rhyme exclusively. The classical "perfect rhyme = good" rule is loosened. Slant rhyme sounds more conversational and less songy.

Hip-hop has the most elaborate rhyme conventions in popular music — multisyllabic rhymes, internal rhymes, complex schemes. See `genres/hip-hop-rnb.md`.

## Point of view and address

Who is the speaker? Who are they speaking to?

| POV | Example | Use |
|-----|---------|-----|
| First person ("I") | "I will always love you" | Personal, intimate; the default |
| Second person ("you") | "You make me feel like a natural woman" | Direct address; speaks to listener as if to a partner |
| Third person ("he/she/they") | "Eleanor Rigby picks up the rice" | Narrative; observational |
| Mixed POV | "I see you" + "She didn't know" | Storytelling with multiple characters |

The choice of POV shapes the song's relationship to the listener. First-person draws the listener in; third-person creates narrative distance.

### Direct address

Many great lyrics speak directly to a "you" who is implicitly the listener (or an absent person).

- "Lay, Lady, Lay" — Dylan addresses his lover.
- "Sound of Silence" — Simon addresses an abstraction.
- Most love songs are direct address.

This creates intimacy. The listener identifies with either the "I" or the addressed "you" — sometimes both.

### Persona and unreliable narrator

Lyrics don't have to be sincere expressions of the writer's feelings. The narrator can be a character:

- "Stan" (Eminem) — the narrator is a deranged fan.
- "Hotel California" (Eagles) — the narrator is in an allegorical situation.
- Many country and storytelling songs use first-person narrators who aren't the writer.

This freedom lets songwriters explore territory they couldn't approach as themselves.

## Structure across the song

Each section of a song has different lyric demands.

### Verse lyrics

- **Different content each verse**: each verse usually has different lyrics (advancing the story or providing different angles).
- **Specific, narrative**: where imagery and detail live.
- **Rhythmically denser** than chorus: more syllables per bar.
- **Usually a build** within the verse: opening lines are quieter, later lines build toward the pre-chorus or chorus.

### Pre-chorus lyrics

- **Bridges verse and chorus**: lyrically transitional.
- **Often shorter, more intense**: "I keep going / I keep falling / I keep losing"
- **Usually leads to a question, declaration, or readiness statement**: "and now I know..." "and so..."

### Chorus lyrics

- **Repeated each time**: usually identical.
- **The song's main statement**: contains the title and central message.
- **Universal rather than specific**: applies to many situations, not just one.
- **Sing-able**: vowels that resonate; consonants that don't trip the singer.
- **Title hook**: typically the title appears in the chorus, often at the start or end of the chorus phrase.

### Bridge lyrics

- **New perspective**: a different angle on the song's theme.
- **Often a turn**: realization, decision, change of heart.
- **Builds back into the final chorus**.

### Outro lyrics

- **Resolution or open-ended**: depends on song.
- **Often a tag**: short repeated phrase that fades.

## The title and hook

The title is usually the most-repeated phrase in the song, almost always in the chorus. Where it appears matters:

- **Title at start of chorus**: makes it the launching point. "I Want to Hold Your Hand" — title is the hook line.
- **Title at end of chorus**: makes it the punchline. "She Loves You" — closes the chorus phrase.
- **Title throughout chorus**: maximally memorable. "Yeah Yeah Yeah" — title and chorus are nearly identical.
- **Title in pre-chorus or bridge**: less common; usually a deliberate choice.

A song without its title in the chorus is rare and usually a deliberate stylistic choice.

## Common lyric techniques

### Anaphora (repeated opening)

Multiple lines starting with the same words. "Don't stop / Don't stop believing / Don't stop holding on..."

Effect: incantatory, building. Common in choruses and bridges.

### Epistrophe (repeated ending)

Multiple lines ending with the same words. "I would do anything for love / I would do anything for love..."

### Listing

A series of items, often with a refrain. "Cigars, cigarettes, ricky-ticky-stickies..." (silly) or "I think I'll go to Boston / I think I'll start a new life..." (meditative).

### Repetition with variation

Same line returned with one word changed. "I love you in a different way / I love you in a different way today."

### Metaphor and simile

Turning the abstract into the concrete. "My heart is a record on repeat" instead of "I keep thinking about you".

### Negation

Saying what something isn't can be more powerful than what it is. "She wasn't crying when she said goodbye" implies all the unspoken.

## Genre-specific lyric conventions

### Pop

- Direct, accessible, universal.
- Strong title hook in chorus.
- Verse lyrics tell a story or set up the chorus's emotional payoff.
- Slant rhyme common.

### Country

- Story-focused. Specific narrative detail.
- Strong characters (often with names).
- Rural or working-class settings common.
- Concrete, sensory imagery.
- Title hook at end of chorus phrases.

### R&B

- Emotional, often romantic or sensual.
- Vocal melisma carries some "lyric" via vocal expression.
- Direct address to a romantic partner common.
- Confessional, vulnerable.

### Hip-hop

- Highest-density wordplay in popular music.
- Multisyllabic and internal rhymes.
- Storytelling, autobiography, social commentary, boasting.
- Specific cultural and personal references.
- Verse lyrics dominate; choruses often shorter and simpler.

### Indie / folk

- Conversational tone.
- Often confessional or observational.
- Minimal rhyme, slant rhyme common.
- Specific imagery and detail.
- Less reliance on commercial title-hook structure.

### Rock

- Wide range from confessional ballad to bombastic anthem.
- Often uses imagery and metaphor more than direct statement.
- Can be political, personal, fantasy.

### EDM

- Often minimal lyrics. Verses and choruses are simpler.
- Hooks are often vowel-heavy phrases ("oh oh oh", melodic interjections).
- Drop sections are often instrumental or have short vocal samples.

### Musical theatre

- Lyrics serve dramatic function (see `genres/musical-theatre.md`).
- Specific, character-revealing.
- Higher tolerance for elaborate rhyme and complexity.
- Sondheim is the gold standard.

### K-pop / J-pop

- Often mixes languages (Korean/Japanese with English phrases).
- Vibrant imagery, often abstract.
- Multiple sections may have different lyrical focuses.
- Choruses often have catchy English phrases as global hooks.

## Common pitfalls when advising

- **Don't ignore prosody.** A beautiful line that doesn't sing well is a problem. Test by singing the lyric to the melody.
- **Don't write generalities when specifics would work.** "I'm sad" tells; "the rain matches my mood" shows.
- **Don't force rhyme.** A clunky rhyme is worse than slant rhyme or no rhyme. The meaning matters more.
- **Don't forget the singer's vowels.** High belted notes are easier on open vowels (ah, oh, ee). Sustained notes on closed consonant clusters (-st, -nch) are awkward.
- **Don't write lyrics without considering rhythm.** Lyrics have rhythm even before they have melody. The rhythmic pattern of the syllables shapes the song.
- **Don't ignore the title.** A song without a clear title hook is hard to remember and hard to market.
- **Don't write didactic lyrics.** A song that lectures the listener is preachy. Imply rather than declare.
- **Don't ignore the listener's experience.** Lyrics that are private to the writer (in-jokes, references only the writer understands) alienate the audience. Universal hooks let the listener in.

## A workflow for lyric writing

If the user is writing lyrics from scratch:

1. **Find the central image or feeling.** What's the song actually about? In one sentence.
2. **Generate the title or hook.** A specific phrase that captures the central idea.
3. **Sketch the chorus.** Often 2–4 lines that include the title. Don't overthink it.
4. **Write verses that earn the chorus.** Each verse should set up the chorus emotionally — what situation makes the chorus statement true?
5. **Test prosody.** Sing the lyric to the melody; if anything sounds awkward, adjust.
6. **Polish for specifics.** Replace generalities with sensory details. "I'm tired" → "I haven't slept since Tuesday".
7. **Read the lyric aloud, no melody.** Does it make sense as a piece of writing? If not, fix.
8. **Test for cliché.** Replace anything that's been heard a hundred times.

## Quick decision matrix

| Want | Try |
|------|-----|
| Pop accessibility | Direct address ("you"), simple verse imagery, repeated chorus title hook |
| Country narrative | Story with named character, rural setting, sensory specifics |
| R&B intimacy | First-person, sensual or confessional, simple rhyme |
| Hip-hop density | Multisyllabic and internal rhyme; specific cultural references; long verses |
| Indie sincerity | Conversational, slant rhyme, specific imagery, minimal cliché |
| Musical theatre dramatic | Character-revealing, dramatic function, internal rhyme |
| EDM hook | Short vowel-heavy phrase, repeated, sometimes wordless |
| Anthem | Universal "we" or direct "you", build to title hook |
| Storytelling ballad | Linear narrative, specific details, named characters, third-person or first-person retrospective |
| Confessional | First-person, present-tense, vulnerable, specific |

## Cross-references

- Vocal melody (which lyrics ride) → `topline-craft.md`
- Hooks (lyrical and melodic) → `hooks-and-memorability.md`
- Prosody across languages, tonal-language constraints, multilingual songwriting → `../fundamentals/prosody-and-language.md`
- Phrase structure → `../melody/phrase-structure.md`
- Pop song forms (sectional structure for lyrics) → `../form/popular-song-forms.md`
- Genre-specific lyric conventions → various `../genres/*.md` files
- Musical theatre lyric craft (Sondheim's principles) → `../genres/musical-theatre.md`
- Hip-hop rhyme techniques → `../genres/hip-hop-rnb.md`
- Vocal range and singer constraints → `../orchestration/instruments-ranges-character.md`

# Odd Meters and Polyrhythm

This file covers asymmetric and irregular meters (5/4, 7/8, mixed meter, additive meter) and the large-scale layering of different meters or rhythmic cycles (polymeter, full polyrhythm). It's the territory of progressive rock, jazz fusion, contemporary classical, and several tradition-specific musics.

For meter fundamentals, see `fundamentals/rhythm-meter.md`. For small-scale rhythmic devices (hemiola, displacement, simple polyrhythm), see `rhythmic-devices.md`. For groove and feel within these meters, see `groove-and-feel.md`.

## When to consult this file

- The user asks about 5/4, 7/8, 11/8, or other "odd" meters
- Mixed meter (changing time signatures from bar to bar)
- Polymeter (different meters in different voices simultaneously)
- Polyrhythm at the structural level (not just hemiola)
- Math rock, prog rock, prog metal
- Stravinsky, Bartók, contemporary classical rhythm
- Indian, Balkan, or African polyrhythmic traditions

## Asymmetric meters — basics

An asymmetric meter is one whose top number isn't divisible into clean groups of 2 or 3. Most common: 5, 7, 11.

The crucial thing about an asymmetric meter is its *grouping* — how the beats divide into smaller units. The same time signature can feel completely different with different groupings:

| Time sig | Possible groupings | Feel of each |
|----------|--------------------|--------------|
| 5/8 | 2+3 | "Long-short" feel; lighter on first part |
| 5/8 | 3+2 | "Short-long" feel; weight on first part |
| 5/4 | 3+2 | "Take Five" — most common |
| 5/4 | 2+3 | Less common; off-balance |
| 7/8 | 2+2+3 | "Short-short-long" — most common |
| 7/8 | 3+2+2 | "Long-short-short" — Balkan flavor |
| 7/8 | 2+3+2 | "Short-long-short" — symmetric |
| 7/4 | 4+3 | "Money" by Pink Floyd |
| 7/4 | 3+4 | Less common |
| 7/4 | 2+2+3 | Faster feel within the bar |
| 11/8 | 4+4+3 | Common in Balkan |
| 11/8 | 2+2+3+2+2 | More complex |
| 11/8 | 3+3+3+2 | Less common |
| 11/4 | various | Almost always specified by composer |

Composers indicate grouping through:

1. **Beaming** — beam eighth notes in the groups indicated.
2. **Accents** — mark the strong beat of each group.
3. **Time-signature notation** — write `2+3/8` or `3+2+2/8` for additive meters.
4. **Drum/percussion patterns** — the rhythm section often makes the grouping audible.

Without indication, the player will guess based on conventions — usually `2+2+3` for 7/8.

## Reading and feeling odd meters

For the agent advising someone new to odd meters: the key is to feel the *grouping*, not the individual eighth notes.

In 7/8 (2+2+3): count "1-2 / 1-2 / 1-2-3" with the third group slightly longer. The bar feels like three pulses of unequal length, not seven equal pulses.

In 5/4 (3+2, like "Take Five"): count "1-2-3 / 1-2" with two pulses, the first longer.

This is how percussionists, drummers, and dance traditions internalize odd meters. Counting individual eighth notes works for slow tempos but breaks down at speed.

### Subdivision tricks

A common practice trick: tap the underlying eighth-note pulse continuously, and count the *grouping* on top.

For 7/8 (2+2+3) at 140 BPM eighth-note: tap eighth notes at 140 BPM, accent every group: ONE-two-ONE-two-ONE-two-three. The accent pattern is the meter.

## Mixed meter — changing time signatures

Mixed meter is the alternation or sequence of different time signatures. The notation changes the time signature whenever the grouping changes.

Three patterns:

### Regular alternation

The same sequence of meters repeats. `5/8 + 7/8 + 5/8 + 7/8` cycles a 12-eighth-note unit, with consistent alternation.

Common in dance traditions where the music has a clear "long form" cycle.

### Through-composed mixed meter

Meter changes follow the music's natural phrasing rather than a fixed pattern. Stravinsky's "Rite of Spring" famously changes meter constantly — a bar of 5/16, then 7/16, then 9/16, then 2/8, etc. — to follow the rhythm of the music rather than imposing a regular grid.

Common in 20th-century classical and modern progressive rock.

### Cyclical patterns (long form)

A long cycle of beats repeats. Indian classical music uses *tala* systems; Middle Eastern/Turkish traditions use related but distinct cycle concepts such as *iqa'* or *usul*. A tala might be a 16-beat cycle that repeats throughout a piece, with claps/waves, sections, and internal weight patterns that produce the feel.

In Indian music, talas have specific names: *teentaal* / *teen taal* (16 beats, 4+4+4+4), *jhaptal* / *jhaptaal* (10 beats, 2+3+2+3), *rupak* (7 beats, often 3+2+2). Treat these as named performance frameworks, not just odd-meter bars; details vary by Hindustani/Carnatic tradition and gharana/school.

## Additive meter

A meter where the bar is explicitly built from unequal groups. Notated as `3+2/8` or `2+2+3/8` or similar.

Additive meters are common in:

- Bulgarian folk music (e.g., 7/8 grouped 2+2+3 — "ruchenitsa")
- Greek folk (5/8, 7/8, 9/8 patterns)
- Turkish music (various asymmetric patterns)
- Indian classical music (tala systems; not merely time signatures)
- Modern composers influenced by these traditions (Bartók, Ligeti, Adams)
- Progressive rock and math rock

The notation `3+2+2/8` makes the grouping explicit, so a 7/8 piece doesn't have to be guessed.

## Irrational meters

A time signature whose lower number isn't a power of 2: `4/12`, `3/10`, `5/24`, etc. The denominator implies tuplet relationships.

`3/10` means three notes that are tenths of a whole note (i.e., quintuplet eighth notes); the bar is ⅗ of a whole note long.

Used in modern classical (Brian Ferneyhough, Thomas Adès) and rarely elsewhere. Most performers find them difficult; their use is specialized.

## Polymeter

Two or more meters happening simultaneously, each with its own bar lines. The voices "agree" only periodically when their bar lines align.

### Simple example: 4/4 against 3/4

If the basic pulse is the same (e.g., quarter notes), the 4/4 voice has 4 quarters per bar and the 3/4 voice has 3 quarters per bar. They realign every 12 quarters (3 bars of 4/4 = 4 bars of 3/4).

### Example: 7/8 against 4/4

Different bar lengths, possibly different pulse units. They realign every LCM (least common multiple) of their bar lengths.

### Polymeter in practice

Used in:

- Frank Zappa's compositions (a guitar in one meter, drums in another)
- Bartók (polymetric piano writing)
- King Crimson, Tool, Meshuggah (one instrument's riff in 7/8 over the drums' 4/4)
- Modern jazz (different sections playing different meters during transitions)

Polymeter is an advanced technique. It requires precision (the alignment points must be felt) and clarity (the listener has to hear the layering, not perceive it as messy).

## Polyrhythm at the bar level

Different from polymeter (which has different bar lengths), polyrhythm fits multiple subdivisions into the same bar.

### 3-against-2

In a 4/4 bar with 12 eighth-note triplets total, one voice plays 4 dotted-eighth notes (every third triplet) while another plays 6 quarter notes (every two triplets). The "3" against the "2" at the bar level.

### 4-against-3

Common in classical piano music. In a 12-unit bar, one voice plays 4 evenly-spaced notes (every 3 units), the other plays 3 evenly-spaced notes (every 4 units).

Chopin's "Fantasie-Impromptu" — left hand triplets, right hand 4-note groups, mathematically aligned.

### 5-against-4, 7-against-4, etc.

Less common but real. Mathematically these are tuplet relationships:

- 5-against-4 means 5 evenly-spaced notes in the time of 4 (a quintuplet against a quadruplet).
- 7-against-4 means 7 evenly-spaced notes in the time of 4 (septuplet).

Used in modern classical (Carter, Ferneyhough), some jazz fusion, and rare prog rock moments.

## Cross-rhythm (polyrhythm in performance practice)

In African music traditions, **cross-rhythm** is foundational. Multiple percussion parts play patterns that conflict with each other, creating a complex composite that doesn't map to a single underlying meter.

The famous question: "where is the downbeat?" In Ewe drumming, different listeners hear different beats as "1" of the same pattern, depending on which voice they're tracking. The downbeat is *negotiated* by the ensemble rather than fixed.

This is conceptually different from polymeter or polyrhythm in the Western sense — it's not "two meters" but "a single composite rhythm where multiple groupings are simultaneously possible".

## Where odd meters and polyrhythm live

### Progressive rock (1970s onward)

Genesis, Yes, King Crimson, Rush, Tool, Dream Theater, Pink Floyd, Frank Zappa. Odd meters and polymeter are signatures of the genre.

Examples:

- Pink Floyd "Money" — 7/4 grouped 4+3 (or 3+4)
- Rush "Tom Sawyer" — 7/8 sections within 4/4
- Tool "Schism" — constantly shifting meters; the famous riff is in 5/8 and 7/8 alternating
- Meshuggah — entire songs in 4/4 with riffs in 5/16 or 17/16 polymetrically layered

### Jazz fusion

Mahavishnu Orchestra, Weather Report, Hiromi, Dave Holland. Odd meters are common; soloing over them is part of the tradition.

### Math rock

Don Caballero, Battles, Hella, Tortoise. Geometrical, complex meters and polyrhythms are the genre's defining feature.

### Contemporary classical

Stravinsky's "Rite of Spring" introduced the modern era of irregular meter to classical. Bartók, Messiaen, Ligeti, Adams, Reich — each treats rhythm differently but all use odd meters and polyrhythm.

### Balkan folk

Bulgarian, Greek, Macedonian, Romanian, Albanian traditions. 5/8, 7/8, 9/8, 11/8, 13/8 are standard. The asymmetric grouping is intrinsic to the dance tradition (different dances have different meters).

### Indian classical

Tala systems organize time into named cycles. The cycles can be of various lengths (often 16, 12, 8, 7, 10 in beginner examples), with claps/waves, sections, and internal weight patterns.

### African polyrhythm

Foundational. Most West and Central African drumming traditions use simultaneous interlocking rhythms.

### Brazilian rhythms

Samba, baião, maracatu — most are in 2/4 or 4/4 but with such intricate cross-rhythms that the "downbeat" can be ambiguous.

## How to write in odd meters effectively

If the user wants to compose in 5/8, 7/8, etc.:

1. **Pick the grouping first.** Decide whether 7/8 means 2+2+3 or 3+2+2 or 2+3+2. The grouping determines the feel.
2. **Establish the meter clearly at the start.** A drum or rhythmic accompaniment that makes the grouping audible. The listener needs to lock in.
3. **Repeat early.** Establish the pattern by repetition before introducing variation.
4. **Phrase in the meter.** A phrase that's 4 bars of 7/8 (28 eighth notes) feels coherent. A phrase that fights the meter feels confused.
5. **Use the meter for character, not novelty.** Odd meters should serve the music's emotional character. A 7/8 piece that doesn't *need* to be in 7/8 (a 4/4 piece would work fine) is just showing off.
6. **Allow the listener to "find" the meter.** Hold the meter long enough that the listener can lock in (typically 8+ bars of stable meter at the start) before introducing variation.

## How to write polymeter effectively

If the user wants to layer different meters:

1. **Pick alignment points.** Decide where the meters align (every 4 bars, every 8 bars). The piece is structured around these alignments.
2. **Make the layers clearly distinguishable.** Each layer should have its own register, timbre, and rhythmic identity. Otherwise the listener hears mush.
3. **Use polymeter for specific moments, not entire pieces.** A whole song in polymeter is exhausting. A polymetric section bracketed by metric sections is more digestible.
4. **The listener locks onto one layer.** Decide which is "primary" and which are "secondary". The primary layer carries the meter; the secondary layers create disorientation.

## Common pitfalls when advising

- **Don't pick odd meters as a stylistic affectation.** If the music has no reason to be in 7/8, putting it there just confuses the listener. Match meter to musical purpose.
- **Don't ignore grouping.** A 7/8 piece without indicated grouping (no beaming, no accents) is unreadable. Always specify.
- **Don't write polymeter without alignment points.** Listeners need anchors. Total chaos isn't polymeter; it's noise.
- **Don't expect performers to "figure out" complex meters.** Provide subdivision counting, clear notation, and rehearsal time.
- **Don't use polyrhythm just to be impressive.** Polyrhythm has a specific perceptual effect — disorientation, complexity, intensity. If the music doesn't want those, polyrhythm is wrong.
- **Don't treat all odd meters as equivalent.** 5/8 in Bulgarian dance has a very different feel from 5/4 in Brubeck's "Take Five". Genre context matters.

## Quick decision matrix

| Want | Try |
|------|-----|
| Off-balance, "leaning" feel | 5/4 grouped 3+2 |
| Dance feel in odd meter | 7/8 grouped 2+2+3 (Balkan ruchenitsa style) |
| Math rock disorientation | Mixed meter — change time sig every bar |
| Prog rock complexity | 7/8 verses with 4/4 choruses |
| Long phrasing in odd meter | 4-bar phrases of 7/8 (28-eighth-note phrases) |
| Polymetric layering | One instrument in 4/4, another in 7/8, aligned every 28 beats |
| African / cross-rhythm feel | 3-against-2 and 4-against-3 simultaneously over a 12-unit bar |
| Indian-inflected structure | Cycle of 7, 10, or 12 beats with internal accent pattern (tala) |
| Stravinsky-style irregular | Through-composed mixed meter, every bar a new time sig |
| Contemporary classical | Irrational meters, irregular subdivisions, careful notation |

## Cross-references

- Meter fundamentals → `fundamentals/rhythm-meter.md`
- Rhythmic devices at smaller scale (hemiola, displacement) → `rhythmic-devices.md`
- Groove and feel within odd meters → `groove-and-feel.md`
- Phrase structure (asymmetric phrases at the bar level) → `melody/phrase-structure.md`
- 20th-century rhythmic innovations → `techniques/20th-century-techniques.md`
- Genre-specific use of odd meters → `genres/jazz-styles.md` (fusion), `genres/classical-periods.md` (Stravinsky, Bartók), `genres/folk-roots-and-traditions.md` (Balkan, Indian, African)
- Prog rock / math rock conventions → `genres/pop-rock.md`

# Topline Craft

In modern songwriting (especially in pop, hip-hop, and EDM production), "topline" refers to the vocal melody and lyrics that sit on top of an instrumental track. A *topliner* is a songwriter who specializes in writing vocal melodies and lyrics, often working with producers who supply the instrumental.

This file covers the craft of topline writing: matching melody to lyric, designing vocal lines for memorability, constructing verses and choruses that build effectively, and the specific demands of writing for different genres and singers.

For lyric writing in detail, see `lyric-writing.md`. For hooks specifically, see `hooks-and-memorability.md`. For melodic construction at a technical level, see `melody/melodic-construction.md`. For phrase structure, see `melody/phrase-structure.md`. For vocal range, see `orchestration/instruments-ranges-character.md`.

## When to consult this file

- The user is writing vocal melody and lyrics over an instrumental
- "Topline" specifically (collaborative pop/EDM songwriting)
- Designing vocal lines for radio
- Working with a singer's range and ability
- A vocal melody that "doesn't go anywhere"
- The relationship between lyric prosody and melodic shape

## What topline craft involves

A topliner makes choices about:

1. **Melody**: pitch contour, rhythm, range, climax position
2. **Lyric**: words, prosody, image, story
3. **Phrasing**: how the line breathes, where it lands
4. **Vocal performance**: ad-libs, melisma, runs, character
5. **Sectional contrast**: how each section sounds different
6. **Hooks**: where the most memorable moments are placed

A great topline isn't just a "melody plus lyrics" — it's an integrated vocal performance designed for memorability and emotional impact.

## The relationship between topline and track

A topline is written *over* an instrumental track. The track imposes constraints:

- **Key and tempo** are usually set by the track.
- **Chord progression** dictates which notes will sound consonant or tense.
- **Rhythm** of the track (groove, syncopation) interacts with the vocal rhythm.
- **Production density** affects where the vocal can live (bright synths above, deep bass below — vocal sits in the mid).

A great topliner adapts to the track. A mediocre one writes a vocal that ignores the track and "fits" only by accident.

### Working with a producer

In a typical session:

1. **Producer plays a track** (often called a "beat" or "instrumental").
2. **Topliner listens** and feels the vibe — is it slow and emotional, or fast and energetic?
3. **Topliner improvises** — humming melodies, mumbling syllables, finding what fits.
4. **Topliner refines** — choosing the strongest melodic ideas, writing real lyrics.
5. **Recording** — the topline is recorded, often with the topliner as the demo singer.
6. **Pitching to artist** — the demo is pitched to a recording artist who may record the song.

The topliner's job is to write something that sounds like the artist (or the kind of artist who would sing it) and is irresistible to record.

## Melodic strategies for topline

### Range zones for sections

A standard pop topline uses different range zones for different sections:

| Section | Tessitura | Why |
|---------|-----------|-----|
| Verse | Low chest / low mix | Conversational, intimate |
| Pre-chorus | Mid mix, climbing | Building to the chorus |
| Chorus | Upper mix or belt | Anthemic, lifted |
| Post-chorus | Belt or wordless | Climactic peak |
| Bridge | Variable; often higher than verse | Contrast, intensity |
| Final chorus | Highest of all | Maximum payoff |

The verse-to-chorus lift is the primary mechanism of pop dynamics. A topline that lives at the same range throughout doesn't deliver the chorus as a destination.

### Verse strategies

The verse should:

- **Establish the song's character** through tone, vocabulary, and melodic identity.
- **Tell the story** — set up situation, character, feeling.
- **Build toward the chorus** — energy increases through the verse.
- **Stay singable but not too memorable** — the verse should serve the chorus, not compete with it.

Common verse melodic patterns:

- **Repetitive within the verse**: same melody for each line, like a chant. Lets the lyric carry the variation.
- **Building**: each line a step higher than the last. Creates motion toward the pre-chorus.
- **Conversational with rhythmic variation**: speech-like rhythm with melodic interest.

### Pre-chorus strategies

The pre-chorus exists to make the chorus land. It typically:

- **Lifts tessitura**: starts higher than the verse.
- **Increases rhythmic density**: more syllables, faster delivery.
- **Modulates harmonically**: chord progression moves toward V of the chorus key.
- **Climbs in steps**: multiple short phrases, each on a higher pitch, anaphoric.
- **Ends on V**: setting up the chorus's I.

A weak pre-chorus is one that's just "verse-extension". The pre-chorus needs distinct character.

### Chorus strategies

The chorus is the destination:

- **Highest tessitura** of the song so far.
- **Strongest hook** (lyrical, melodic, or both).
- **Repetition**: the title phrase repeats, often anchoring the chorus.
- **Open vowels** on sustained notes for sing-along ease.
- **Repeating melody** for memorability.
- **Distinct rhythmic identity** from the verse.

Common chorus structures:

- **Statement and repeat**: same melody/lyric phrase twice, with slight variation second time
- **Statement-response**: A line, then a "response" line, often by backing vocals
- **Build-and-peak**: melody builds within the chorus to a climactic note in the second half

### Post-chorus strategies

The post-chorus is a relatively new (2010s onward) addition that became dominant in modern pop. It typically:

- **Wordless or minimal lyric**: "oh-oh-oh", "ah-ah-ah"
- **Melodic hook**: a distinct melody from the chorus
- **Production peak**: drums/synths at maximum
- **Sing-along** quality: simpler than the chorus, easier to repeat

The post-chorus often becomes the song's most-remembered hook.

### Bridge strategies

The bridge provides contrast:

- **New harmonic area**: often modulates to relative minor, parallel minor, or distant key.
- **New melodic identity**: completely different shape from verse and chorus.
- **Often more intense or more reflective**: depending on song.
- **Builds back to the final chorus**: the last line of the bridge often functions as a launching point for the chorus return.

A bridge that's too similar to verse or chorus is wasted real estate. The bridge should feel like the song "looking sideways" before resolving.

## Prosody — the topline-specific concern

Prosody (matching word stress to musical stress) is central to topline craft. A topliner must:

- **Stress strong beats with stressed syllables**.
- **Avoid filler words on strong beats** ("the", "a", "and").
- **Match polysyllabic words** to multi-note phrases that respect the word's stress pattern.
- **Use vowel choices** to optimize sustained notes.

When prosody is right, the topline sounds inevitable. When wrong, it sounds awkward.

### Vowel choices

Sustained notes are easier on open vowels:

| Vowel | Best for |
|-------|----------|
| Ah, Aw | High belted notes; full resonance |
| Eh, Ay | Mid-range notes; bright |
| Oh, Oo | Lower or rounded notes; warm |
| Ee, Ih | Brighter, more piercing in high register |

A high belt on "ee" can be brilliant; a high belt on "ck" is impossible. The topliner shapes the lyric so that strong, sustained notes fall on suitable vowels.

### Consonants and articulation

- **Hard consonants** (k, t, p, b, d, g) can clip the start of a sustained note.
- **Sibilant consonants** (s, sh, z) can be problematic in the mix.
- **Soft consonants** (l, m, n, w, y, r) flow smoothly.
- **Plosive at the very start of a phrase** is usually fine and adds rhythmic punch.

A topliner thinks about consonant clusters when writing — "stretched" doesn't sing as well as "reached".

## Vocal melisma and runs

Melisma is multiple notes on a single syllable. Common in:

- R&B (extensively)
- Gospel
- Some pop (Mariah Carey, Whitney Houston)
- K-pop ballads
- Broadway

When writing toplines for genres that feature melisma, leaving room for runs is part of the craft. A topliner might write the basic melody but indicate "ad-lib" on certain notes, leaving the singer to embellish.

Conversely, in genres without much melisma (most rock, country, indie), runs feel out of place.

## Writing for a specific singer

If the topliner knows the artist who will record the song, choices follow:

- **Range**: the artist's comfortable range (not pushing into strain).
- **Vocal character**: the artist's distinctive sound — raspy, smooth, ethereal, etc.
- **Style preferences**: belted vs. mixed; melismatic vs. clean.
- **Lyric voice**: what kinds of words the artist is comfortable singing.

A topline written for Adele (rich, low-mid alto belter) is very different from one for Ariana Grande (high agile soprano with whistle register).

## Modern pop topline conventions

### Verse-chorus-postchorus structure

The dominant template since ~2015. Each section has its own hook.

### Vocal layering

Modern toplines are often heavily layered:

- Lead vocal
- Doubled lead (often pitched up an octave or down)
- Harmony vocals (3rd or 5th above)
- Ad-lib vocals (improvised flourishes)
- Stacked "oohs" and "aahs" as pads

The topliner often demos all these layers to show the producer the full vocal vision.

### Melodic minimalism

Modern toplines often use repetition and minimalism — the same 2- or 3-note motif repeated with slight variations. Less melodic content per bar than older pop.

This compensates for production density: when the track is busy, a busy vocal would compete; a sparse vocal floats.

### Dropouts

The vocal pulls out for a bar or two at structural moments — the "and then nothing" before the chorus, or a held note while the rhythm section solos.

## Genre conventions

### Pop topline

- Verse-chorus-postchorus
- Hook-forward
- Title in chorus
- Vocal layering
- Slight melismatic flourishes

### Hip-hop topline (for sung hooks)

- Often shorter than full pop choruses
- Melodic but rhythmically strong
- Catchy and singable as the contrast to the rapped verses
- Featured vocalist often handles the hook (e.g., R&B singer for rap song's chorus)

### EDM topline

- Often minimal lyrics
- Hook-driven, often wordless or short phrase
- Vocal sits in the breakdown; drop is instrumental
- Vocoded, autotuned, or processed vocals common

### R&B topline

- Heavy melisma
- Improvisational feel
- Vocal phrasing more conversational and less metric
- Multiple ad-libs throughout

### Country topline

- Storytelling verses
- Title at end of chorus phrase
- Conversational delivery
- Less melisma; cleaner melodic lines
- Authentic vocal character (less processed)

### K-pop topline

- Multiple hooks per song
- Mix of singing and rapping (depending on group)
- Belted choruses
- Often heavily layered

### Folk / singer-songwriter

- Less commercial structure
- Conversational melody
- Lyrics drive everything
- Minimal vocal layering

## Common pitfalls when advising

- **Don't write toplines without listening to the track.** The vocal must match the track's energy, key, and groove.
- **Don't write a chorus that doesn't lift from the verse.** If the chorus has the same range, dynamics, and melodic identity as the verse, the listener won't perceive it as a chorus.
- **Don't write impossible vocal lines.** Demands that exceed normal vocal range (especially for a 4-minute pop song) won't make it past the demo stage.
- **Don't forget about prosody.** Singing the line out loud reveals problems that look fine on paper.
- **Don't ignore the track's rhythm.** The topline rhythm should complement the track — not fight it, not blandly match it. Find the rhythmic conversation.
- **Don't write generic toplines.** Distinctive melodic and rhythmic identity is what gets the song chosen by an artist.
- **Don't over-melismatic in genres that don't want it.** Ariana Grande can run; Bob Dylan can't and shouldn't.

## A workflow for topline writing

If the user is writing a topline over an instrumental:

1. **Listen to the track multiple times.** Internalize its key, tempo, groove, harmonic arc.
2. **Identify the song's energy**: ballad, mid-tempo, anthem, club, etc.
3. **Find the chorus first.** Improvise vocal melodies over the chorus chord progression. Find the strongest 2-bar idea.
4. **Develop the chorus.** Extend the strongest idea into a full chorus melody. Make sure it has range, climax, and a clear hook.
5. **Write verse melody that contrasts.** Lower in tessitura, more conversational, builds toward the chorus.
6. **Write pre-chorus to bridge them.** Climbing, building, ending on V or a clear preparation.
7. **Write lyrics.** Start with the title hook; write chorus around it; write verses to set up the chorus.
8. **Polish prosody.** Sing through; fix anything awkward.
9. **Consider bridge.** Does the song need a bridge? If yes, contrast harmonically and melodically.
10. **Add ad-libs and harmonies.** Layer the demo vocal to show the producer the full vision.

## Quick decision matrix

| Want | Try |
|------|-----|
| Pop topline | Verse low-mid, pre-chorus mid-climbing, chorus belt, postchorus wordless hook |
| Modern pop with multiple hooks | Distinct verse melody, distinct chorus melody, distinct post-chorus melody — three memorable moments |
| Anthemic chorus | Sustained notes on open vowels, range up to top of singer's belt, repeated title hook |
| R&B vocal feature | Leave room for melisma; less metrically rigid; embellishment-friendly |
| Hip-hop sung hook | 8-bar chorus by R&B vocalist; melodic, simple, contrasts with rapped verses |
| EDM vocal hook | Short 2-bar phrase; repeated; lives in the breakdown; lyric is sparse or wordless |
| Country topline | Conversational verse, title-end-of-phrase chorus, minimal vocal processing |
| K-pop topline | Multiple sections with distinct hooks; range-spanning; layered harmonies |
| Indie / singer-songwriter | Conversational throughout; minimal commercial structure; lyrics-driven |
| Bridge that contrasts | New key area or modal mixture; new melodic shape; introduces a question or realization |

## Cross-references

- Lyric craft → `lyric-writing.md`
- Hooks → `hooks-and-memorability.md`
- Prosody across languages, vowel choice for sustained notes, tonal-language constraints → `../fundamentals/prosody-and-language.md`
- Melodic construction (theoretical) → `../melody/melodic-construction.md`
- Phrase structure → `../melody/phrase-structure.md`
- Pop song forms → `../form/popular-song-forms.md`
- Vocal range and singer constraints → `../orchestration/instruments-ranges-character.md`
- Genre-specific topline conventions → various `../genres/*.md` files
- Chord-melody interaction (which notes work over which chords) → `../harmony/jazz-harmony.md`, `../melody/melodic-construction.md`
- Production-side considerations → `../production-aware/arrangement-for-mix.md`

# Groove and Feel

This file is about what happens *between* the notational grid lines — the microtiming, dynamic variation, and ensemble interaction that makes a rhythm "groove" instead of just "play". Two players can perform identical notation and produce wildly different feels; that difference is the subject here.

For meter and notational rhythm, see `fundamentals/rhythm-meter.md`. For rhythmic devices that manipulate the grid (syncopation, hemiola, displacement), see `rhythmic-devices.md`. For odd meters and polyrhythm, see `odd-meters-polyrhythm.md`. For genre-specific groove conventions, see relevant `genres/*.md` files.

## When to consult this file

- The user asks about "groove", "feel", "pocket", "swing"
- A track that "sounds stiff" or "doesn't groove" despite correct notation
- Programming drums and wanting them to feel human
- Quantization decisions in a DAW
- Understanding why an ensemble locks together (or doesn't)
- Microtiming differences between genres

## What "groove" means

Groove is the felt quality of a rhythmic performance. It includes:

1. **Microtiming**: where each note actually falls relative to the strict grid (slightly ahead, on, or behind).
2. **Velocity / dynamics**: which notes are louder than others within a repeating pattern.
3. **Ghost notes**: very quiet notes that don't register as "the beat" but contribute to the texture.
4. **Articulation**: how short or long each note is — the duration relationships within the pattern.
5. **Ensemble interaction**: how players' microtiming relates to each other (locked together, slightly off, deliberately leaning differently).
6. **Repetition with variation**: a groove is a pattern that repeats, with small variations preventing exact mechanical repetition.

A groove can be described in notation, but the feel never can — it's a property of the performance.

## Microtiming — the millisecond layer

The grid says beat 2 happens 500ms after beat 1 (at 120 BPM). The actual snare hit might happen at 480ms (ahead of the beat), 500ms (on the beat), or 520ms (behind). This 20–40ms difference is microtiming, and it's the difference between "tight", "lazy", "rushing", and "dragging".

### On the beat

The note hits exactly with the click. Mechanical, "robotic" feel. Most quantized programmed drums sit here. Effective for genres that want machine-precision — techno, EDM, modern pop production.

### Ahead of the beat ("on top", "pushing")

The note hits slightly before the click. Creates a rushed, energetic feel. Lead instruments often push slightly to lead the band; drummers who push are often described as "driving".

### Behind the beat ("laid back", "dragging", "in the pocket")

The note hits slightly after the click. Creates a relaxed, lazy, soulful feel. Most great soul and R&B drummers play behind. Hip-hop sampled drums are often deliberately behind. The classic "pocket" feel.

### Combination feel

A drummer can hit the kick on the beat, the snare slightly behind, and the hi-hat slightly ahead. Each part has its own microtiming relationship. This creates the multi-dimensional feel that machine quantization usually destroys.

J Dilla's beat-making famously used uneven hi-hat timing — some hits ahead, some behind — to create a "drunken" feel that influenced an entire generation of producers.

## Swing

The most well-known microtiming convention. Eighth notes don't divide the beat in half — they divide it asymmetrically, with the first eighth longer than the second.

### The swing ratio

Different swing ratios for different feels:

| Ratio | Description | Genres |
|-------|-------------|--------|
| 50:50 | Straight eighths | Rock, modern pop, EDM |
| ~55:45 | Slight swing | Some shuffle, lazy pop |
| 60:40 | Medium swing | Jazz at uptempo, some R&B |
| 67:33 (2:1) | Triplet swing | Classic jazz, ballads |
| 70:30 or more | Heavy swing | Slow blues, deep shuffle |

The 67:33 ratio is mathematically equivalent to a triplet feel — the first eighth gets two-thirds of the beat, the second gets one-third. This is the "swing" notation in jazz where two eighth notes are written but played as a triplet quarter + triplet eighth.

### Swing in different genres

- **Jazz swing** — typically 60:40 to 67:33 at moderate to fast tempos; loosens to almost straight at very fast tempos.
- **Shuffle** (blues, rockabilly) — closer to 67:33; very pronounced.
- **Hip-hop swing** — usually a subtle swing on 16th notes (the "MPC swing" of 50–58% on 16ths), inherited from sampled jazz and funk drums.
- **House music swing** — sometimes a slight 16th-note swing for "groove"; more often straight.
- **Samba and bossa nova** — straight on the surface, but a subtle "Brazilian swing" of microtiming on the second sixteenth note.

### Swing on 8ths vs. 16ths

In jazz, swing is typically on eighth notes. In hip-hop, R&B, neo-soul, and modern dance music, swing is more often applied to 16th notes (since the 8ths are still felt as straight beats).

## Ghost notes

Very quiet notes that don't function as primary beats but fill in the rhythmic texture. Critical to funk, R&B, and most groove-heavy music.

A funk drum pattern might have:
- Loud snare on 2 and 4
- Loud kick on 1 and selected off-beats
- Ghost notes (very quiet snare hits) on every 16th note in between

The ghost notes aren't "heard" individually — they're felt as a continuous low-level texture that frames the loud accents. Without them, the groove sounds bare. With them, the groove "breathes".

The classic study: Bernard Purdie's drum patterns. The Purdie shuffle and the Purdie half-time shuffle are built on intricate ghost-note layering.

## Velocity and dynamic accents within a groove

Beyond microtiming, the dynamics within a repeating pattern shape the feel. A 16th-note hi-hat pattern can sound completely different depending on which notes are accented:

- All hits at equal velocity: mechanical, "hi-hat as click track".
- Accent on 1, 3 (downbeats): plain, square.
- Accent on 2, 4 (upbeats): lifted, "skipping".
- Accent on every other 16th: stuttering subdivision.
- Selective accents matching the kick or snare pattern: locks the hat into the groove.

DAW programmers should use velocity to imitate human variation. Real drummers can't hit at exactly the same velocity twice; programmed drums should vary by at least 5–10% to feel natural.

## Articulation — note length

Within a groove, the duration of each note matters. Two patterns with identical pitches and microtiming can feel different if one cuts notes off short and the other lets them ring.

Examples:

- Tight, staccato hi-hats vs. open, sustained ones.
- Bass notes cut short on the upbeat (jazz "walking bass" feel) vs. sustained (rock or R&B).
- Guitar palm-muting (chunky) vs. ringing strums (open) — gives radically different feels with identical notation.

## Ensemble interaction

In a band or production, the parts have a specific relationship to each other. Three common patterns:

### Locked

All players hit the same beats with the same microtiming. The ensemble sounds as a single unit. Rare in groove music — feels mechanical when overdone, but tight ensemble playing requires this on key downbeats.

### Layered

Each player has their own microtiming relationship. The drummer is on the beat, the bass is slightly behind, the rhythm guitar is slightly ahead. The combined feel is a multi-dimensional pocket. Most great soul and R&B grooves work this way.

### Conversational

Players "talk" to each other rhythmically, with one accenting where another rests. Common in jazz quartets, where the soloist and drummer respond to each other in real time.

A producer can imitate ensemble interaction by deliberately offsetting different programmed parts — putting the kick on the beat, the snare a few ticks behind, the hi-hat a few ticks ahead.

## "In the pocket"

A phrase used to describe a groove that locks together perfectly. The "pocket" is where all the elements — microtiming, accents, ghost notes, articulation — combine into a cohesive whole. It's a felt quality.

When a band is "in the pocket", the listener feels physically grounded. When a band is "out of the pocket", the listener feels uncertain or disoriented.

Pocket is genre-specific:

- A pop pocket is straight, tight, on-the-grid.
- A funk pocket is heavily syncopated, with ghost notes and behind-the-beat snare.
- A jazz pocket includes swing and conversational interaction.
- A reggae pocket emphasizes 2 and 4, with the kick and bass behind.
- A neo-soul pocket includes microtiming variation and a relaxed feel.

## Programming drums to feel human

If the user is programming drums in a DAW:

1. **Vary the velocity.** Each hit shouldn't be the same. Loud snares should be ~110–127, ghost notes ~30–60, regular hats ~60–90. Vary within those ranges by 5–10%.
2. **Add humanization.** Most DAWs have a "humanize" function that randomly offsets notes by a few milliseconds. Use it sparingly — too much creates a sloppy feel.
3. **Don't quantize everything to 100%.** Even 90% quantize keeps some human feel. For deliberately loose grooves, 70% quantize.
4. **Layer parts with intentional offsets.** Move the snare a few ticks behind the kick. Move the hat ahead. The combined feel is more "alive".
5. **Use ghost notes.** Add quiet hits between the loud ones. Even one or two ghost snare hits per bar transforms a stiff pattern.
6. **Vary the pattern over multiple bars.** A 1-bar repeating loop sounds robotic. A 2- or 4-bar loop with subtle variations (an extra ghost note here, a missed hat there) feels alive.
7. **Pick a feel consciously.** Decide whether the snare is on the beat, slightly ahead, or slightly behind. Don't leave this to chance.

## Quantization decisions

Quantization snaps notes to the grid. Different choices for different goals:

| Quantize to | Effect | When to use |
|-------------|--------|-------------|
| 100% to 16ths | Mechanical, perfect | EDM, modern pop, anywhere "on the grid" is the aesthetic |
| 80–90% to 16ths | Tight but humanized | Most pop, rock, R&B production |
| 100% to 16th-triplets (swung) | Mechanical swing | Classic hip-hop, MPC-style beats |
| 50% (preserve original feel) | Light correction only | Live drum tracking |
| Off (no quantize) | Pure performance | Jazz, classical, anywhere live feel matters most |

## Half-time vs. full-time feel

A common groove choice. The same tempo can feel "double-time" or "half-time" depending on where the snare hits.

- **Full-time feel**: snare on 2 and 4 of every bar.
- **Half-time feel**: snare on 3 only (every other bar's "2 and 4").

Half-time feels slower and heavier even though the BPM is the same. Switching between full-time and half-time is a primary contrast tool in trap, dubstep, post-rock, and modern pop.

A song at 140 BPM can be felt as 140 (full-time, dance-energetic) or 70 (half-time, trap or dubstep, slow and weighty).

## Genre-specific groove notes

### Pop

- Mostly on-the-grid (modern production aesthetic).
- Slight swing on 16ths in some sub-genres (R&B-influenced pop).
- Anticipations and pushes at chord changes are common.

### Rock

- Generally straight 8ths.
- Drummer leads the feel; pushing or laid-back is a band identity choice.
- Half-time choruses common in modern rock.

### Hip-hop

- 16th-note hi-hats with selective accent.
- Snare typically on 3 (half-time feel) in trap and modern hip-hop.
- "MPC swing" — slight 16th-note swing inherited from the Akai MPC drum machine workflow.
- Sampled drums maintain the original recording's microtiming.

### R&B / neo-soul

- Heavily microtimed; ensemble interaction is critical.
- Drums slightly behind the beat is the default.
- Ghost notes pervasive in drums and bass.

### Jazz

- Swing eighths (variable ratio).
- Drummer "comping" creates conversational microtiming.
- Walking bass on quarter notes, slightly behind the beat for "swing" feel.

### Latin

- Clave is the rhythmic skeleton; everything aligns to it.
- Microtiming is more "on the grid" than jazz, but has its own subtle feel.
- Different feels for son, salsa, samba, bossa, etc.

### Reggae and dub

- Kick and bass behind the beat for "rootsy" feel.
- Skank guitar precisely on 2 and 4 (or off-beats in steppers' style).
- Heavy use of dropouts (silences as part of the groove).

### Funk

- "On the One" — emphatic downbeat.
- Heavily syncopated bass and rhythm guitar with ghost notes.
- Drummer plays "in the pocket" — heavily microtimed, often behind.

### EDM (house, techno)

- 4-on-the-floor kick precisely on each beat.
- 16th hi-hats often slightly swung.
- Low microtiming variation — the aesthetic is precision.

### Drum and bass / breakbeat

- Sampled breakbeat drums maintain the original groove's microtiming.
- Snare on 2 and 4 (or syncopated breaks).
- Bass plays a separate rhythmic role from the drums.

## Common pitfalls when advising

- **Don't tell the user to "add more groove".** Specify what dimension to adjust: more swing, more ghost notes, slightly behind the snare, more dynamic variation.
- **Don't apply jazz swing to non-jazz contexts.** Swing 8ths in a rock song sound wrong; the rock idiom uses straight 8ths. Genre matters.
- **Don't overcorrect microtiming.** A heavily microtimed groove is its own aesthetic; trying to "fix" it by quantizing destroys the feel.
- **Don't ignore the producer's intent.** Some producers deliberately use mechanical, on-the-grid feel as the aesthetic (Daft Punk, much of Kraftwerk, modern pop). Suggesting "humanization" misses the genre.
- **Don't conflate groove with complexity.** A simple straight-eighth pattern with great velocity variation can groove harder than a complex pattern that's mechanically played.

## Quick decision matrix

| Want | Try |
|------|-----|
| Mechanical, modern pop feel | 100% quantize to 16ths, equal velocities |
| Tight but human feel | 80–90% quantize, ±10% velocity variation, slight ghost notes |
| Funk pocket | Snare slightly behind beat, lots of ghost notes, accented hats |
| Jazz swing | 60:40 to 67:33 ratio on 8ths, walking bass slightly behind |
| Hip-hop boom-bap | MPC swing on 16ths, snare behind beat, sampled drums preferred |
| Trap | Half-time feel (snare on 3), 16th hat rolls, kick on tresillo |
| Reggae | Kick and bass behind, skank guitar on 2 and 4, ghost rim shots |
| Neo-soul | Heavy microtiming, drums laid back, bass with ghost notes |
| EDM | 4-on-the-floor kick precise, 16th hats with selective accent, sidechained pads |
| Latin | Tight clave alignment, ensemble locked to clave, percussion layered |

## Cross-references

- Meter and notational rhythm → `fundamentals/rhythm-meter.md`
- Rhythmic devices (syncopation, hemiola, displacement) → `rhythmic-devices.md`
- Odd meters and polyrhythm → `odd-meters-polyrhythm.md`
- Phrase rhythm interaction with groove → `melody/phrase-structure.md`
- Genre-specific groove conventions → `genres/*.md` (especially hip-hop-rnb, pop-rock, electronic-edm, jazz-styles)
- Production-side groove → `production-aware/energy-and-dynamics.md`, `production-aware/arrangement-for-mix.md`

# Hooks and Memorability

This file is about what makes a song stick in the listener's head — the hook, the earworm, the moment that gets remembered. Hooks are the single most commercially important element of a song, and the most studied. This file covers the anatomy of a hook, the techniques for crafting one, and the kinds of hooks that work in different genres.

For lyric craft generally, see `lyric-writing.md`. For the vocal melody as a whole, see `topline-craft.md`. For melodic construction at a technical level, see `melody/melodic-construction.md`. For motivic identity, see `melody/motivic-development.md`.

## When to consult this file

- The user is writing a song and asks "what makes a hook"
- A song is missing memorability — what to add
- Brainstorming hooks, vocal phrases, post-chorus elements
- Why some songs are infectious and others aren't
- The craft of repetition, contrast, and surprise

## What a hook is

A hook is a short, memorable musical-lyrical phrase that the listener wants to repeat (in their head or out loud). It's the part of the song that's still echoing after the song ends.

Hooks can be:

- **Lyrical**: the title phrase ("I Want It That Way", "Bad Guy")
- **Melodic**: a specific tune ("Smells Like Teen Spirit" intro riff, "Seven Nation Army" bass riff)
- **Rhythmic**: a distinctive groove or beat ("We Will Rock You" stomp-clap)
- **Sonic**: a specific sound or production element (the synth sound in "Take On Me")
- **Combination**: most great hooks combine multiple of the above

A song can have multiple hooks — verse hook, chorus hook, post-chorus hook, instrumental hook. Modern hits often have 3+.

## What makes a hook memorable

Several factors contribute. The greatest hooks usually have most of them:

### 1. Repetition

The hook repeats within the song — usually multiple times in the chorus, plus appearances in verse hooks or post-choruses. Listeners can't memorize what they don't hear repeatedly.

A song that says its title once is harder to remember than one that says it five times.

### 2. Distinctiveness

The hook should be different from other songs. Cliché melodies (the same I-V-vi-IV pop progression, or the same hook melody as countless other pop songs) blur into the background.

Distinctiveness can come from:

- An unusual interval (a leap in an unexpected place)
- An unusual rhythm (a syncopation that feels new)
- An unusual chord change at the hook moment
- A specific timbral choice (vocoder, falsetto, distinctive vocal tone)
- Lyrical wordplay or unexpected phrasing

### 3. Simplicity

The hook is short — usually 2–4 bars, 4–10 syllables. Listeners can hold it in working memory and replay it.

A "hook" that's 16 bars long isn't a hook; it's a chorus or a section. The hook is the kernel inside.

### 4. Singability

A great hook is something an average untrained singer can hum or sing along to. This means:

- Limited range (typically a 5th or less for the core hook)
- Stepwise motion or simple leaps (fourths, fifths, octaves)
- Open vowels for sustained notes
- Strong beats land on memorable syllables

A "hook" with a 2-octave range and chromatic inflections is hard for the average listener to internalize.

### 5. Emotional resonance

The hook says something the listener wants to say. It captures a feeling — joy, longing, defiance, sadness — that resonates personally.

This is partly about lyrics (the words mean something) and partly about melody (the contour creates the feeling regardless of language).

### 6. The "earworm" effect

Some hooks are literally earworms — they get stuck in the head involuntarily. Researchers have studied this; common features:

- Predictable enough to sing along, surprising enough to be interesting
- A "gap" or "incompleteness" that the brain wants to resolve
- Rhythmic distinctiveness
- Repeated exposure

## Anatomy of a melodic hook

A typical melodic hook can be analyzed in terms of:

### Range

Usually a 5th or 6th. Wider hooks are harder to sing. Examples:

- "Yesterday" (Beatles): the chorus hook spans a 6th.
- "Hey Jude": the "na na na na" outro is within a 5th.
- "Happy" (Pharrell): the chorus moves within a 5th.

### Contour

The shape matters as much as the pitches. Common hook contours:

- **Arch** (rises, peaks, falls): "Yesterday", "Don't Stop Believin'" chorus
- **Descending step** (falls in a melodic gesture): "Stairway to Heaven" intro
- **Repeated note**: many rap and rhythmic hooks center on a single pitch with rhythmic interest
- **Three-note motif**: "Star Wars" theme, "Imperial March", "We Are the Champions"

### Rhythm

Often the rhythm of the hook is more memorable than the pitches. The dotted-quarter-eighth rhythm of "Hey Jude" (na-Na-na-na) is more iconic than its specific pitches.

A hook with a strong rhythmic identity is easier to remember.

### Pitch over chord

The relationship between the hook's pitches and the underlying chord matters. Hooks that land on the chord's 3rd or 5th feel sweet; ones that land on the 7th or extensions feel modern; ones that land on chord roots feel grounded.

The "lift" of a chorus often comes from the hook melody hitting the 3rd, 5th, or 9th of the chord (rather than the root).

## Lyrical hooks

The lyric hook is usually:

- The title of the song
- A short phrase (2–6 words)
- Repeated multiple times in the chorus
- Often the first or last line of the chorus

What makes a lyrical hook strong:

- **Concrete language**: "We Will Rock You" is more vivid than "We Will Influence You".
- **Emotional declaration**: "Bad Guy", "Single Ladies", "I Will Always Love You" — strong statements.
- **Wordplay or surprise**: "Bohemian Rhapsody" — the title doesn't appear in the song; the hook is "Mama, just killed a man".
- **Sing-along quality**: the hook should be easy to shout. "Hey Ya" is easy; complex lyrical hooks aren't.

## Hooks across the song

Different sections call for different hook types:

### Verse hook

A short repeated phrase in the verse — often the verse melody's distinctive opening or a recurring lyric across verses. Smaller and more subtle than the chorus hook.

### Pre-chorus hook (often the "build hook")

A rising melodic figure in the pre-chorus that's distinctive and helps build energy.

### Chorus hook

The main hook. The thing the song is famous for.

### Post-chorus hook

A second hook after the chorus. Often instrumental or wordless ("oh-oh-oh"). Modern pop relies on this.

Examples:

- "Bad Romance" (Lady Gaga): "ra-ra-ah-ah-ah" post-chorus
- "Levitating" (Dua Lipa): the "yeah-yeah-yeah" hook
- Many K-pop songs have post-chorus drops with vocal-vocoder hooks

The post-chorus often becomes the most-remembered part of the song. Modern hit-makers consciously design post-choruses as separate hook events.

### Instrumental hook (the riff)

A memorable instrumental phrase. Common in rock, funk, hip-hop, EDM:

- "Smoke on the Water" — the iconic guitar riff
- "Seven Nation Army" — the bass riff
- "Sweet Child O' Mine" — the guitar intro
- "Crazy Train" — riff
- "Funkytown" — the synth riff

A great instrumental hook can carry a song without strong vocals.

## Genre conventions

### Pop

The chorus hook is paramount. Pop relies on title hooks repeated multiple times. Increasingly, post-choruses provide a second hook event.

### Country

Title hook usually at the end of the chorus phrase. Storytelling verses lead to a clear, declarative chorus statement.

### Hip-hop

Hooks are often shorter and more rhythmic than pop. Producer-side hooks (sample loops, beat patterns) are as important as vocal hooks. The chorus hook is often a chant or sung phrase by a featured artist.

### R&B

Vocal hooks rely heavily on melisma and ad-lib variations. The hook is sometimes more about a vocal performance than a specific melodic shape.

### Rock

Instrumental hooks (guitar/bass riffs) often more iconic than vocal hooks. The chorus hook is usually anthemic and declarative.

### EDM

The "drop" is essentially a hook — an instrumental peak moment. Vocal hooks are often short and repeated through the build and drop.

### K-pop

Multi-hook approach. Distinctive verse hook, pre-chorus build, chorus hook, post-chorus drop hook — sometimes all in one song. A K-pop song with 4+ hooks is common.

### Folk / singer-songwriter

Less reliance on commercial hooks. The "hook" might be a striking image or a phrase that resonates emotionally. Songs that don't have radio-style hooks but still have memorable moments.

### Musical theatre

The "11 o'clock number" often has the show's most powerful hook moment. But theatrical hooks serve dramatic function rather than commercial radio play.

## Techniques for crafting hooks

### Start with the title

Many great songs start with a title or hook phrase. Then the writer builds melody, verses, and structure around that initial spark.

### Sing-then-write

Some songwriters develop hooks by humming melodies into a recorder, then transcribing the best ones. The hook comes from improvisation rather than analysis.

### Chord-forward

Other writers start with chord progressions, then find melodies that work over them. The chord progression suggests certain pitches as natural landing points.

### Lyric-forward

Some start with a phrase or feeling and build everything else around it. "Bohemian Rhapsody" likely started with the lyrical concept.

### Stealing structures, replacing content

Many hooks borrow structural patterns from existing hooks (e.g., "title phrase + descending pitch" is a common pattern). Listening to many hooks and analyzing their construction helps the writer build a vocabulary.

### Test, then refine

Sing the hook out loud; record it; listen back the next day. If it's still in your head, it's working. If not, revise.

## Common pitfalls when advising

- **Don't over-engineer hooks.** Some great hooks are surprisingly simple. "We Will Rock You" is two notes. The hook doesn't need to be melodically complex.
- **Don't ignore rhythm.** A hook's rhythm is often more memorable than its pitches. Get the rhythm right; the pitches can vary slightly.
- **Don't make hooks too long.** A 16-bar "hook" isn't a hook. Trim it to 2–4 bars of pure essence.
- **Don't bury the hook.** It should appear early in the song (within the first minute) and recur often. A song that doesn't deliver its hook until 2:30 will lose listeners.
- **Don't ignore vocal range.** A hook that's not singable by the average listener won't get hummed. Keep range tight.
- **Don't assume distinctive = good.** Distinctive can also mean weird-in-a-bad-way. The hook should be distinctive AND immediately appealing.
- **Don't forget the chorus payoff.** A weak chorus melodically dooms the song no matter how strong the verses.

## A workflow for hook crafting

If the user wants to develop a hook:

1. **Start with the central feeling or message.** What's the song about? What's the emotional core?
2. **Write 3–5 candidate hook lines.** Short, evocative, sing-able.
3. **Test each by singing.** Which has the strongest melodic identity? Which makes you want to repeat it?
4. **Add melody.** The lyric and melody should fit together — strong syllables on strong beats, vowels that ring.
5. **Test the hook in context.** Sing the chorus over the chord progression. Does the hook land?
6. **Listen to it the next day.** Is it still memorable, or has it faded?
7. **Refine.** Adjust pitch, rhythm, or lyric for greater impact.
8. **Add a contrasting verse.** The hook should feel like a payoff after the verse.

## Quick decision matrix

| Want | Try |
|------|-----|
| Pop chorus hook | Title phrase, 4–8 syllables, repeated, on an arching melodic shape |
| Earworm hook | Short rhythmic figure with one unusual interval; repeats often |
| Anthem hook | "We" or "you" lyric, descending strong-beat line, sing-along |
| Modern post-chorus hook | Wordless or minimal-word vocal phrase ("oh-oh-oh") with synth lead |
| Instrumental riff hook | 2–4 bar guitar/bass/synth motif, distinctive rhythm, repeats throughout song |
| Hip-hop hook | Sung chant by featured vocalist, short phrase, melodic over the beat |
| Country hook | Title at end of chorus phrase, narrative buildup, conversational delivery |
| K-pop multi-hook | Verse hook + chorus hook + post-chorus drop hook; multiple memorable moments |
| Wordplay hook | Surprising rhyme or word combination at the hook moment |
| Visual / tactile hook | Specific imagery in the hook lyric ("rolling in the deep", "blurred lines") |

## Cross-references

- Lyric craft generally → `lyric-writing.md`
- Topline / vocal melody → `topline-craft.md`
- Prosody and language constraints (especially for non-English / multilingual hooks) → `../fundamentals/prosody-and-language.md`
- Melodic construction at the technical level → `../melody/melodic-construction.md`
- Motifs and their development → `../melody/motivic-development.md`
- Pop song forms (where hooks live) → `../form/popular-song-forms.md`
- Genre-specific hook conventions → various `../genres/*.md` files
- Phrase structure → `../melody/phrase-structure.md`

# Source Bibliography and Verification Guide

> **Curated snapshot:** 2026-04-27. This file is a maintainer-facing bibliography map. It is not a replacement for active web research where current scenes, copyright, AI tools, charts, law, platform behavior, or living-artist references are involved.

Use this file to verify claims, deepen future documents, and choose trustworthy sources. The skill itself should usually synthesize guidance rather than cite bibliography entries directly to the user unless citations are requested.

## Source hierarchy

When checking a claim, prefer sources in this order:

1. **Primary sources** — official documentation, legal texts, platform chart methodology pages, museum/archival pages, composer/performer statements, scores, recordings.
2. **Authoritative reference sources** — Grove/Oxford Music Online, Encyclopaedia Britannica, national libraries, national academies, university presses.
3. **Specialist pedagogy** — conservatory/Berklee materials, orchestration texts, reputable practitioner books.
4. **Peer-reviewed research** — especially for ethnomusicology, cognition, copyright, AI, and regional scenes.
5. **Industry reporting** — Billboard, IFPI, RIAA, Recording Academy, Music Business Worldwide, Reuters, AP, reputable local press.
6. **Community sources** — useful for scene vocabulary and practice, but verify before turning into general rules.

## Western theory and composition

Recommended stable references:

- Aldwell, Schachter, and Cadwallader — *Harmony and Voice Leading*
- Kostka, Payne, and Almén — *Tonal Harmony*
- Piston — *Harmony* and *Counterpoint*
- Caplin — *Classical Form*
- Hepokoski and Darcy — *Elements of Sonata Theory*
- Schoenberg — *Fundamentals of Musical Composition*
- Persichetti — *Twentieth-Century Harmony*
- Forte / Straus — post-tonal and set-theory references

Use these for:

- Roman numeral analysis
- cadences and phrase structure
- modulation and chromatic harmony
- classical forms
- counterpoint and voice-leading

## Jazz harmony and arranging

Recommended references:

- Mark Levine — *The Jazz Theory Book* and *The Jazz Piano Book*
- Ted Pease — *Jazz Composition*
- Berklee harmony/arranging materials
- Sher Music / Real Book conventions, with caution about edition differences
- Scores and transcriptions from authoritative publishers when exact changes matter

Use these for:

- ii-V-I, tritone substitution, modal interchange
- chord-scale choices
- rootless voicings
- rhythm-section arranging
- lead-sheet notation conventions

Do not treat simplified online chord charts as authoritative for famous-song analysis.

## Orchestration and instrument idiom

Recommended references:

- Adler — *The Study of Orchestration*
- Rimsky-Korsakov — *Principles of Orchestration*
- Piston — *Orchestration*
- Blatter — *Instrumentation and Orchestration*
- Philharmonia Orchestra instrument guides
- Vienna Symphonic Library Academy / Instrumentology
- Actual scores and session-player feedback

Use these for:

- ranges, registers, transposition
- articulation and idiomatic writing
- ensemble balance
- orchestral texture and doublings

Always distinguish playable range from good tessitura.

## Songwriting, lyric writing, and topline

Recommended references:

- Pat Pattison — lyric writing and prosody materials
- Sheila Davis — lyric craft books
- Jimmy Webb — *Tunesmith*
- Berklee songwriting/topline materials
- Published interviews and stems where available
- Native-speaker review for non-English lyrics

Use these for:

- prosody
- rhyme and scansion
- title placement
- verse/chorus contrast
- hook design

## Popular music, production-aware composition, and arranging

Recommended references:

- Everett — rock harmony and form research
- Tagg — popular music semiotics and analysis
- Moore — analysis of recorded popular song
- Middleton — popular music studies
- Berklee arranging and rhythm-section materials
- Reputable producer/arranger interviews, when clearly attributable

Use these for:

- pop/rock form
- rhythm-section writing
- arrangement density
- production-facing composition decisions

## Country, Americana, and roots music

Useful sources:

- Country Music Hall of Fame and Museum
- Encyclopaedia Britannica entries on country music, bluegrass, Nashville Sound, outlaw country
- Library of Congress archives and recordings
- Published Nashville Number System references
- Artist/songwriter interviews and session-player materials

Verify separately:

- current country radio/streaming trends
- regional country scenes
- Nashville Number System notation details when writing full charts

## Gospel and CCM

Useful sources:

- Encyclopaedia Britannica on gospel music
- Recording Academy category definitions and field guides
- Gospel Music Association / industry sources, with context
- Hymnals, church songbooks, and denominational resources for liturgical contexts
- Scholarly work on Black gospel, Pentecostal worship, Southern gospel, and CCM

Verify separately:

- denominational fit
- current worship set norms
- theological language and intended community

## Metal, punk, and hardcore

Useful sources:

- Encyclopaedia Britannica entries on heavy metal, punk, hardcore punk, straight edge
- Scene histories, label archives, and oral histories
- Guitar/bass/drum practitioner sources for substyle-specific technique
- Current local-scene reporting and festival/label rosters

Verify separately:

- current subgenre labels
- regional scenes
- tuning/gear conventions for specific substyles

## Brazilian music

Useful sources:

- Library of Congress music blog and Brazilian collections
- Encyclopaedia Britannica on bossa nova, samba, Brazil performing arts
- Scholarly work on MPB, samba, bossa nova, choro, and funk carioca
- Brazilian local press and platform charts for current funk/pop scenes
- Native Portuguese lyric review

Verify separately:

- current funk substyles and regional terms
- Portuguese prosody
- whether a requested track references Brazil broadly or a specific scene

## Korean traditional music and Korean pop

Useful sources:

- 국립국악원 국악사전
- 한국민족문화대백과
- 국립국악원 recordings and educational materials
- Korean-language academic writing and practitioner interviews
- Circle Chart and official platform/agency sources for current K-pop trends

Verify separately:

- romanization
- instrument technique terms
- current idol/pop production trends
- chart and fandom/platform behavior

## MENA, South Asian, Southeast Asian, and other regional files

Useful sources:

- MaqamWorld and practitioner-led maqam resources
- Academic/practitioner raga and tala resources
- National archives, regional music scholarship, and local-language sources
- Current local platform charts and press

Use caution:

- Do not reduce maqam/raga/dastgah/gamelan to scales.
- Do not present Western equal-tempered approximations as the full tradition.
- When writing serious tradition-based work, recommend practitioner collaboration.

## AI, copyright, and style-reference requests

Useful sources:

- U.S. Copyright Office AI reports and copyright circulars
- Court opinions and filings where available
- Reuters / AP / major legal reporting for ongoing litigation
- Official terms from music platforms and AI tools
- Music publishers/labels' official statements when relevant

Verify every time:

- current legal status
- platform rules
- AI model terms
- living-artist style policies
- public-domain status

## Streaming-service and user listening context

Useful sources:

- Official platform chart pages such as Spotify Charts, Apple Music charts, YouTube Charts, Circle Chart, Oricon, Billboard, and Official Charts Company
- User-provided public playlists, track lists, Replay/Wrapped summaries, and scrobbling logs
- Spotify Web API and Apple/Last.fm documentation when a developer or connected-tool workflow is explicitly relevant
- Local music platforms and charts for the target market

Use caution:

- Never ask users for passwords, account tokens, or private credentials.
- A user's personal listening history is taste evidence, not a universal market signal.
- Platform charts have different weighting rules and demographics.

## Microgenres, hybrid genres, and regional trend evolution

Useful sources:

- Current reference tracks and playlists from the scene itself
- Local press, label rosters, festival lineups, DJ sets, and producer interviews
- Platform tags and community terminology, checked against listening evidence
- Major genre files in this skill as stable craft anchors

Use caution:

- Do not over-define unstable scene labels.
- Do not treat one viral tag as a whole genre.
- For hybrid genres, identify which ingredient controls form, groove, harmony, texture, and cultural context.


## Creative workflow and brainstorming

Useful sources:

- Composition pedagogy texts that emphasize sketching, constraint, and revision
- Songwriting process interviews, when tied to concrete craft decisions
- Creativity research only when translated into musical behavior, not generic productivity advice
- User feedback from actual agent sessions, especially where brainstorming prompts succeed or fail

Use these for:

- musical brainstorming prompts
- iterative user-agent collaboration
- decision-log and project-card templates
- writer's block workflows

Verify separately:

- whether a proposed workflow creates usable musical material
- whether the workflow preserves user agency rather than over-directing the composition

## Chart and trend research

Use `research/web-music-trend-research.md` for workflow. High-signal sources include:

- official charts and chart methodology pages
- platform charts and playlists, interpreted cautiously
- local music press
- IFPI / industry reports
- Billboard / Official Charts / Circle / Oricon / YouTube / Spotify chart pages

A chart position does not automatically become a composition rule. Extract repeated variables across multiple references.

## Release validation and audit

Useful sources:

- The current project files themselves: `SKILL.md`, `references/00-navigation.md`, `RELEASE-ROADMAP.md`, and validation reports
- Stable theory references for disputed notation, harmony, and voice-leading claims
- The automated regression script for broken links and known high-risk patterns
- Prompt smoke-test transcripts from representative user sessions

Use these for:

- first-release readiness decisions
- prompt-routing validation
- high-use asset audits
- release-blocker triage

Use caution:

- A passing script does not prove the music theory is complete or correct.
- A prompt smoke test can pass behaviorally while still needing fact review.
- Release reports must distinguish checks actually run from checks merely planned.

## Verification checklist before adding or revising a file

- Are historical claims supported by at least one reliable source?
- Are current-scene claims marked with snapshot/currentness guidance?
- Are terms from living traditions spelled and romanized carefully?
- Are examples labeled as simplified when they are not exact transcriptions?
- Are cultural references specific rather than vague?
- Are instrument claims checked for range, tessitura, and playability?
- Are copyright/style-reference claims framed cautiously?
- Did `scripts/music_theory_sanity_check.py` pass?
- If preparing a release, did the first-release readiness checklist and prompt smoke tests get reviewed?

## Cross-references

- Web-search cheatsheet → `../assets/web-search-cheatsheet.md`
- Current trends research → `research/web-music-trend-research.md`
- User listening context → `research/user-listening-context-and-streaming-services.md`
- Regional trend evolution → `research/regional-trend-evolution-analysis.md`
- Reference track digging → `research/reference-track-digging.md`
- Regional scene starters → `genres/regional-scene-starters.md`
- Minor and hybrid genres → `genres/minor-and-hybrid-genres.md`
- Folk/roots/tradition-specific writing → `genres/folk-roots-and-traditions.md`
- Trend/reference matrices → `../assets/trend-and-reference-matrices.md`
- Chord-symbol ambiguity and parsing → `../assets/chord-symbol-ambiguity-and-parsing.md`
- Scale-degree spelling → `../assets/scale-degree-spelling-cheatsheet.md`
- Music-theory audit rubric → `../assets/music-theory-audit-rubric.md`
- Musical brainstorming → `creative-workflows/musical-brainstorming.md`
- User-agent collaboration → `creative-workflows/user-agent-collaboration.md`
- Revision and feedback loops → `creative-workflows/revision-and-feedback-loops.md`
- Style and copyright guardrails → `research/style-reference-and-copyright.md`
- Instrument idiom overview → `instrument-idiom/overview.md`
- First release roadmap → `../RELEASE-ROADMAP.md`
- Release readiness checklist → `validation/first-release-readiness.md`
- Prompt smoke tests → `validation/prompt-smoke-tests.md`
- Maintenance policy → `../MAINTENANCE.md`

# Teaching Composition

This file is about how the agent helps users who are *learning* composition — as distinct from users who are *solving a specific compositional problem*. The two modes look similar at the surface but call for different responses: a learner needs scaffolding, exercises, and conceptual building; a problem-solver needs direct help. Mistaking one for the other wastes the user's time.

This file complements `workflow.md` (general advisory framework) and `critique-and-feedback.md` (evaluating finished work). Where `workflow.md` covers diagnostic questions for any user, this file focuses on the *pedagogical* mode — when the user wants to *learn* something, not just *do* something.

## When to consult this file

- The user is explicitly trying to learn composition
- "How do I learn [X]?" / "Teach me about [Y]"
- The user has a knowledge gap and wants to fill it
- Students or self-learners working through compositional problems
- The user has theory knowledge but lacks compositional practice (or vice versa)
- Cross-genre learning (a classical-trained user learning pop, or vice versa)
- The user wants exercises, not just answers

## Pedagogy mode vs. problem-solving mode

The agent should identify which mode the user is in. Signals:

| Pedagogy mode signal | Problem-solving mode signal |
|----------------------|------------------------------|
| "How do I learn..." | "How do I write..." |
| "Teach me..." | "Help me with..." |
| "Why does X work?" | "Make X work for me" |
| "What should I practice?" | "What should I do here?" |
| Student / hobbyist context | Professional / project context |
| Asking about general principles | Asking about specific work-in-progress |
| "I want to understand" | "I want to finish" |

These often blend. A user who's been writing for years might still want to *learn* a specific technique (modal harmony, fugue, K-pop song structure) while *solving* a current project's problems. The agent can be in both modes within a single conversation; the key is recognizing which mode any given exchange is in.

In pedagogy mode, the right response often *isn't* the direct answer. A user who asks "how does modal harmony work?" benefits more from a careful, layered explanation than from a one-paragraph definition followed by a chord chart.

In problem-solving mode, lecture is unwelcome. A user who asks "what scale should I use over Cm7 in this jazz solo?" wants a quick answer, not a tutorial on chord-scale theory.

## Where is the user in their learning journey

Different stages of learning need different help. Diagnose roughly:

### Complete beginner

No theory background, no compositional experience. May have technical instrument skill or just musical interest.

What helps:
- Concrete starting points (a single mode, a four-chord progression, a simple form)
- Pairing each concept with audible examples
- Avoiding jargon-dense explanations
- Encouragement to make many small finished pieces over fewer "perfect" ones
- Recommending listening, not just doing

What doesn't help:
- Comprehensive theoretical lectures
- Recommending advanced repertoire ("listen to Bach fugues") without bridging
- Treating notation as a prerequisite (many beginners benefit from working without notation initially)

### Theory-fluent, practice-light

Has studied theory (school, books, online), can analyze, recognizes terms — but has composed little or nothing. Common in classically-trained learners.

What helps:
- Permission to start producing work that isn't theoretically ambitious
- Exercises with strict constraints (write a 16-bar piece using only I-IV-V; write a melody using only stepwise motion)
- Connecting their existing analytical knowledge to specific compositional moves
- Encouragement to copy a model piece note-by-note as a learning exercise (not for publication)
- Listening to genres outside their training (a classical-trained learner studying pop hooks; a jazz student studying minimalism)

What doesn't help:
- More theory
- Pure brainstorming without constraints (they often freeze with too many options)
- Validation of their analytical skills as a substitute for compositional practice

### Practice-active, gap in specific skill

Composes regularly but has identified a specific weakness. "I can write good verses but my choruses don't hit"; "I can write good melodies but my harmonic choices are predictable"; "I can write pop but I want to learn jazz".

What helps:
- Direct engagement with the specific skill
- Exercises targeted at that skill
- Analytical study of pieces that exemplify it
- Trying the skill in a low-stakes context (an exercise, a sketch) before integrating it into "real" work
- Compounding feedback over multiple iterations

What doesn't help:
- Treating them as beginners
- Generic "study more theory" advice
- Refusing to engage with the specific gap

### Cross-genre learner

Trained or experienced in one tradition; wants to work in another. "I'm a jazz player who wants to write K-pop"; "I'm a classical composer who wants to score for film".

What helps:
- Identifying which skills transfer and which don't
- Highlighting the conventions of the target genre that aren't in their existing toolkit
- Recommending immersive listening in the target genre
- Acknowledging that genre-fluency takes time; first attempts will likely sound like "their existing genre with new flavors" rather than the target genre
- Working with practitioners of the target genre when possible

What doesn't help:
- Treating their existing expertise as irrelevant
- Underselling how different the target genre's logic is
- Letting them assume composition skills transfer fully

### Advanced learner

Has substantial composition experience; wants to deepen specific areas. May be working on advanced topics (12-tone, microtonality, advanced contrapuntal techniques) or refining craft in their established genre.

What helps:
- Engaging at their level; don't oversimplify
- Pointing at specific repertoire and analytical resources
- Treating exchanges as collaborative inquiry rather than instruction
- Acknowledging when their question goes beyond the agent's reliable knowledge

What doesn't help:
- Beginner-mode explanations
- Treating their work as if they need basic feedback
- Hiding uncertainty when the agent is at the edge of its competence

## Pedagogical principles

Drawing from broader composition pedagogy:

### Show, don't just tell

Concept + example + listening reference is the basic teaching unit.

- "A deceptive cadence is V→vi instead of V→I" (telling)
- "...and you can hear it dramatically in the second movement of Beethoven's 9th, after the choral entry — V wants I, but Beethoven gives you vi for one searing moment before resolving" (showing)
- "Try writing a 4-bar phrase that ends on a deceptive cadence" (doing)

The combination cements the concept. Telling alone produces students who can recognize a term without hearing what it does.

### Build from simple to complex

A learner studying chromatic harmony should already be comfortable with diatonic harmony. A learner studying Schoenberg should already understand tonal harmony deeply. The agent should diagnose the learner's foundation and build from there.

When a learner asks about an advanced topic before the foundation is solid, the agent has a choice:
- **Backfill the foundation first**: "this is great to learn, but I want to make sure the diatonic-harmony foundation is solid first; let's check..."
- **Skim the advanced topic with the caveat that some of it requires background to fully understand**: useful for a survey-level introduction
- **Refuse to engage at the requested level**: rarely the right choice; better to bridge than to gatekeep

### Connect abstract to listening

Every theoretical concept has audible expression. The agent should *always* pair concept with listening:

- "Modal mixture" → listen to The Beatles' "In My Life" (the borrowed iv)
- "Tritone substitution" → listen to almost any jazz standard's turnaround
- "Sentence structure" → listen to the opening of Beethoven's Op. 2 No. 1

Without listening references, the concept stays in the head; with them, it moves to the ear, which is where it belongs.

### Give exercises with strict constraints

Constraints generate creativity (see `techniques/constraint-based-composition.md`). For learners specifically:

- "Write a 4-bar phrase using only the notes of C major"
- "Write a 16-bar AABA using only I, IV, V, vi"
- "Write a melody where every note is a chord tone"
- "Write a 4-voice chorale ending in a Phrygian half cadence"
- "Set this lyric to a melody that doesn't go above the singer's middle range"

The constraint forces engagement with the concept being taught. A vague "write something using modal harmony" produces vague results; "write a 16-bar piece in D Dorian using only chords with major IV color" produces specific learning.

### Encourage analysis as teaching tool

Analyzing master composers' work is the deepest form of learning:

- Pick a piece in the genre / technique the user wants to learn
- Walk through it with them: form, harmony, melody, distinctive features
- Ask them to identify what makes it work
- Then have them try to apply one technique from it to their own piece

This is how composers have always learned — not by reading rules but by seeing what's been done and absorbing the logic.

### Validate experimentation

A learner who's only writing technically-correct music isn't learning quickly. They need to make mistakes — try things that don't work, hear why they don't, and refine. The agent should:

- Encourage trying ideas without prejudging them
- When something fails, diagnose *why* without making the failure feel shameful
- Treat experiments as data, not as moral judgments on the user's ability
- Distinguish "this didn't work but is a productive direction" from "this didn't work and isn't worth pursuing"

### Validate finishing as a skill

Many learners can start pieces but can't finish them. Finishing is itself a skill — it requires:

- Tolerating imperfection
- Making decisions and committing
- Closing off optionality

The agent should celebrate finished work, even imperfect work, as a real achievement. A finished mediocre piece teaches more than three abandoned promising ones.

## Common learning paths

These are common entry points; the agent can match its approach to which path the user is on.

### "I want to write a song"

A common starting point. The user usually has musical instinct but no formal training. Useful first moves:

1. Pick a target genre and study 3 reference songs in detail (form, key, melody, lyrics)
2. Learn basic chord voicings on guitar or piano
3. Write 4 chords in a 4-bar loop; sing over it
4. Build a verse / chorus around what emerges
5. Don't worry about polish; finish a sketch and start another

### "I want to learn music theory to compose better"

Common with self-taught composers who feel they're hitting a ceiling. Useful sequence:

1. Diatonic harmony (major and minor keys, primary chords, basic progressions)
2. Voice leading basics
3. Phrase structure (period, sentence)
4. Modulation (closely related keys first)
5. Modal mixture and secondary dominants
6. Chromatic harmony (Neapolitan, augmented sixths)
7. Modal harmony as a separate system
8. Jazz extensions and substitutions (if relevant to their work)
9. 20th-century techniques (if relevant)

The user should compose at every stage, not save composition for "after I finish theory". Theory studied without composition is largely abstract.

### "I want to learn jazz harmony / how to comp / how to solo"

Common for instrumentalists branching into jazz. Useful sequence:

1. ii-V-I in major and minor (the cellular unit)
2. Voicing 3-7 shells; then full rootless voicings
3. Chord-scale relationships (basic — Dorian over ii, Mixolydian over V, Ionian over I)
4. Basic substitutions (tritone sub, secondary dominants)
5. Reading a jazz standard and walking through its changes
6. Soloing over a 12-bar blues, then a standard's changes
7. Melodic-minor modes and altered scales
8. Reharm
9. Specific traditions: bebop language, modal jazz approach, contemporary jazz

### "I want to learn film scoring"

Useful sequence:

1. Watch film with the music in mind; identify what music is doing in each scene
2. Study a few specific scores (Williams, Zimmer, Goldsmith, Newman) — their themes, motifs, scoring techniques
3. Study film orchestration (instruments, ranges, doublings)
4. Compose to picture: take a 30-second clip and score it; iterate
5. Study trailer music and ad music (different conventions, useful for variety)
6. Eventually: study DAW work, mockup quality, tempo mapping

### "I want to learn K-pop / pop production"

Useful sequence:

1. Pick 5 reference K-pop tracks; analyze each fully (form, key, hook structure, production)
2. Replicate one of them as a learning exercise (don't release; learn the moves)
3. Study chord progressions used in K-pop ballads vs. dance tracks
4. Study how multiple hooks layer in a single song
5. Study production (Logic / Ableton / FL); production knowledge is non-negotiable for K-pop
6. Write original tracks applying the moves; iterate
7. Eventually: study Korean / English lyric craft for the genre

## Recommended exercise types

Across paths, certain exercises are reliable:

| Exercise | Teaches |
|----------|---------|
| **Bach chorale harmonization** | Voice leading, functional harmony, four-part texture |
| **Set a lyric to a melody** | Prosody, melody-text fit, phrase structure |
| **Reharmonize a familiar tune** | Reharm techniques, voice leading, harmonic vocabulary |
| **Write a melody using only one mode** | Modal vocabulary, characteristic notes |
| **Write a 12-tone row and a phrase derived from it** | Serial technique, atonal voice leading |
| **Compose a 16-bar AABA jazz tune** | Form, jazz harmony, head-writing |
| **Write a fugue exposition (3 voices)** | Counterpoint, imitation, subject construction |
| **Score a 30-second film clip** | Film-scoring craft, scene-music integration |
| **Write a song in an unfamiliar genre** | Genre conventions, expanding range |
| **Transcribe a piece by ear** | Ear training, analytical listening, repertoire knowledge |
| **Analyze a piece and write a piece in its style** | Analysis as composition tool |

The right exercise depends on what the learner needs. Pick one that targets their gap.

## Pitfalls in teaching composition

- **Don't lecture in pedagogy mode without checking comprehension.** Long explanations without check-ins lose the learner. Pause; ask if it makes sense; offer a small exercise.
- **Don't underestimate the learner's existing knowledge.** Asking what they already know prevents wasted explanation.
- **Don't assume one path fits all.** A classical-trained student learning pop has different needs from a self-taught songwriter learning theory.
- **Don't substitute analysis for composition.** Analytical knowledge is necessary but not sufficient. Composers learn by composing.
- **Don't treat mistakes as failures.** A learner who isn't making mistakes isn't learning. Failed experiments are data.
- **Don't gatekeep advanced topics.** A motivated learner can sometimes engage with advanced material without "earning" it through prerequisites. The teacher's job is to bridge, not to gatekeep.
- **Don't impose your aesthetic.** Teaching composition means helping the learner write *their* music, not yours.
- **Don't forget listening.** Composers learn primarily by ear. Theoretical instruction without paired listening references produces students who can talk about music but not hear it.
- **Don't conflate teaching with critiquing.** A learner sharing an exercise isn't asking for a finished-work critique. Match feedback depth to the work's purpose.
- **Don't move on too fast from a topic the learner hasn't actually absorbed.** Apparent understanding ("yes, I get it") often masks confusion. Confirm by asking for application.

## Quick decision matrix

| Situation | Approach |
|-----------|----------|
| User asks "teach me about X" | Pedagogy mode: definition + audible example + small exercise |
| User asks "help me with X in my piece" | Problem-solving mode: address the specific situation |
| Learner stuck on theory | Connect concept to audible example; assign small exercise |
| Learner can't finish work | Address finishing as a skill; encourage smaller-scope finishing |
| Cross-genre learner | Identify what transfers, what doesn't; recommend immersive listening |
| Beginner | Concrete starting points; pair every concept with a listening reference; encourage many small finished pieces |
| Advanced learner | Engage at their level; treat as collaborative inquiry; acknowledge knowledge limits honestly |
| Theory-fluent, practice-light | Constrained exercises; permission to produce non-theoretical work; pair with listening outside their training |
| Wants exercises | Pick one that targets their specific gap (table above) |
| Asks about advanced topic without foundation | Bridge: address the foundation while honoring the interest, or skim the topic with caveats |

## Cross-references

- General advisory framework → `workflow.md`
- Critique of finished work (different mode) → `critique-and-feedback.md`
- Analytical methodology (a key teaching tool) → `analysis.md`
- Constraint-based exercises → `techniques/constraint-based-composition.md`
- Counterpoint pedagogy (species method) → `counterpoint.md`
- Theme-and-variations as exercise structure → `techniques/theme-and-variation.md`
- Genre-specific conventions for each learning path → relevant `genres/*.md`
- Specific topic foundations → `fundamentals/`, `harmony/`, `melody/`, etc.

# Algorithmic and AI-Assisted Composition

> **Snapshot note:** Last content review snapshot: 2026-04-26. AI music tools, copyright rules, platform policies, model licensing, and lawsuits change quickly. Use this file for durable composition workflows, but verify current legal/platform facts through active web research; see `../research/web-music-trend-research.md` and `../research/reference-track-digging.md`.

This file covers two related but distinct compositional approaches: **algorithmic composition** (music generated or shaped by formal systems, rules, or computational processes) and **AI-assisted composition** (using machine learning models — generative neural networks, language models, audio models — as compositional collaborators). Both involve composers ceding some compositional control to a system; both have their own histories, techniques, and creative-philosophical questions.

This file complements `constraint-based-composition.md` (the broader frame of working under constraints, which algorithmic composition is one variant of), `20th-century-techniques.md` (where algorithmic composition emerged historically), and `theme-and-variation.md` (where systematic transformation of material connects to algorithmic thinking).

## When to consult this file

- Designing or using an algorithmic composition system
- Working with generative music in installation / interactive contexts
- Using AI music tools (Suno, Udio, MusicGen, Stable Audio, AIVA, Soundraw, etc.) as part of the compositional process
- Using language models (Claude, GPT, etc.) for composition tasks (lyrics, motivic suggestion, theory analysis, structural planning)
- Procedural game music systems (overlap with `../genres/game-music.md`)
- Live coding (TidalCycles, SuperCollider, Sonic Pi, Max/MSP, Pure Data)
- Generative ambient / drone music
- Discussing the creative-philosophical questions AI composition raises

## Algorithmic composition — the long tradition

Algorithmic composition isn't new. Composers have used formal systems for centuries:

### Pre-computer era

- **Medieval isorhythm**: a rhythmic pattern (*talea*) and a melodic pattern (*color*) of different lengths, repeated and combined to produce music — the patterns interlock predictably but produce surface variety. Machaut, 14th century.
- **Counterpoint species rules** (Fux, *Gradus ad Parnassum*, 1725): a rule-based system for generating contrapuntal music. Rule-bound enough to be mechanical, expressive enough to produce real music.
- **Mozart's *Musikalisches Würfelspiel* (Musical Dice Game)**: a system for generating minuets by rolling dice and assembling pre-composed measures. Not entirely algorithmic, but anticipatory.
- **Twelve-tone technique** (Schoenberg, 1920s): a rule-based pitch system that, while leaving substantial composer choice, structurally generates pitch material.
- **Total serialism** (Boulez, Stockhausen, Babbitt, 1950s): rule-based systems applied not only to pitch but to rhythm, dynamics, articulation. The rules themselves became compositional.

### Early computer era

- **Lejaren Hiller**, *Illiac Suite for String Quartet* (1957): the first piece composed by a computer (the ILLIAC I). Hiller programmed compositional rules; the computer generated note-sequences.
- **Iannis Xenakis**: stochastic composition, using probability distributions to generate musical material. *Pithoprakta*, *Achorripsis*, and many others. Xenakis derived musical material from physics (gas behavior, statistical mechanics).
- **John Cage**: chance operations using *I Ching*, dice, star charts, and other randomization sources. *Music of Changes*, *Atlas Eclipticalis*. The composer's role shifted from selecting notes to designing systems.
- **Steve Reich, Terry Riley, La Monte Young, Philip Glass**: process music. The piece is generated by a clearly audible process (phasing, additive structures, repetition with controlled variation) that the listener can follow.

### Modern algorithmic composition

- **Generative algorithmic music**: systems that produce music continuously (often forever or indefinitely) from rules and randomness. Brian Eno's *Music for Airports* (1978) used tape loops of different lengths to produce slowly-evolving textures. *Reflection* (2017) and *Mix Me a Drink* are pure generative apps that produce different music each session.
- **L-systems**: recursive grammar systems originally for modeling plant growth, applied to musical material.
- **Cellular automata**: rules-based grid systems (Conway's Game of Life and variants) used to generate pitch sequences, rhythms, or textures.
- **Genetic algorithms**: systems that evolve musical material through selection and mutation.
- **Markov chains**: probability-based note sequence generation, where the next note depends on the current state.

These techniques are not exclusive to "algorithmic music as a genre" — they're tools that any composer can apply.

## Why use algorithmic methods?

A few motivations:

- **Generate material faster than a composer can write by hand.** Useful in long-form contexts (game music, generative installations, ambient).
- **Discover patterns the composer wouldn't have invented manually.** Algorithmic outputs can suggest unexpected directions.
- **Achieve specific structural properties** (proportional relationships, logarithmic distributions, fractal self-similarity) that are hard to compose intuitively.
- **Create indefinite or non-repeating music** (generative installations).
- **Externalize creative decisions** so the composer can step back and curate.

The risk: algorithmic output can be "interesting" without being "musical". A great algorithmic composer designs systems whose outputs are reliably good, and curates / shapes the output rather than just publishing whatever the system produces.

## Live coding

A modern algorithmic technique: the composer writes and modifies code in real time during a performance, generating music programmatically as the audience watches.

Common live coding environments:

- **TidalCycles**: Haskell-based, designed for pattern-oriented music
- **SuperCollider**: powerful audio synthesis and pattern language
- **Sonic Pi**: Ruby-based, designed to be approachable
- **Max/MSP**: visual programming for audio
- **Pure Data**: open-source visual programming, similar to Max
- **FoxDot**: Python-based, uses SuperCollider for audio

Live coding is a performance practice with its own community, conventions, and aesthetic. The Algorave movement (algorithmic rave) brought live coding to dance music contexts.

## AI-assisted composition — the new layer

Since roughly the late 2010s, machine learning models have become practical compositional tools. The capabilities are evolving rapidly; this section will date faster than the rest of the file.

### Categories of AI music tools

| Tool type | What it produces | Examples at the snapshot date — verify before recommending |
|-----------|-------------------|-----------------|
| Text-to-music (full audio) | Complete tracks from text prompts | Suno, Udio, Stable Audio, Riffusion |
| Symbolic music generation | MIDI / notation output | Magenta (Google), MuseNet (legacy), AIVA, Soundful |
| Stem / instrument generation | Specific tracks (drums, bass, etc.) | Boomy stems, various commercial tools |
| Voice synthesis | Vocal performance from text + style | Synthesizer V, Vocaloid, AI voice cloning tools |
| Mastering / mixing AI | Processed mix from input stems | LANDR, iZotope Ozone (with AI assist), Bandlab Mastering |
| Composition copilots | Suggestions / completions while composing | Various DAW plugins, Magenta Studio, Anthropic Claude / OpenAI for theoretical tasks |
| Lyric / text generation | Lyrics from prompts | LLMs in general (Claude, GPT, etc.) |
| Style transfer | Reapplies one piece's style to another | Various academic and commercial tools |

The space is changing fast; specific tools come and go. The general categories are more stable than the brand names.

### What AI music tools are good for

- **Brainstorming and reference**: generate quick sketches in target genres to clarify intent before composing seriously
- **Specific functional content**: library-music-style backing tracks for video / podcasts / corporate content where uniqueness isn't critical
- **Iteration on a hook**: producing variations of a melodic or harmonic idea quickly
- **Filling gaps in skill**: a composer strong in melody but weak in production can use AI for production tasks, or vice versa
- **Rapid prototyping for client work**: showing a client three possible directions in an hour rather than three days
- **Lyric drafting**: language models can produce serviceable first-draft lyrics quickly
- **Theoretical analysis**: language models can analyze a piece's harmonic content, identify form, suggest reharmonization options

### What AI music tools are still unreliable for at the snapshot date

- **Distinctive artistic voice**: generated music tends to default to genre conventions; pushing it to true distinctiveness requires significant compositional intervention
- **Long-form coherence**: AI tools often produce coherent 30-second or 2-minute outputs but struggle with 5+ minute pieces with developing thematic material
- **Specific compositional intent**: telling an AI tool "I want a melody that recalls the opening but inverted" is much harder than telling a human collaborator the same thing
- **Performance-quality vocals for specific lyrics**: AI vocals are improving but still recognizable as AI in critical-listening contexts
- **Original genre fusion**: AI tools work well within established genres; novel fusions usually require human composition
- **Critical / evaluative judgment**: AI tools can generate but rarely critique their own output meaningfully

### Working with AI music tools — workflow patterns

The most productive workflows treat AI as one tool among many, not as the composer:

**Pattern 1: AI for ideation, human for refinement**
1. Generate 10–50 variations from a prompt
2. Curate the best 1–3 ideas
3. Re-record / re-arrange / re-write those ideas with full compositional craft
4. Final piece is human-composed but seeded by AI ideation

**Pattern 2: AI for production, human for composition**
1. Compose the song in DAW with rough placeholder sounds
2. Use AI tools for instrument generation, mixing, mastering
3. Composition is fully human; production is AI-augmented

**Pattern 3: AI for style match, human for narrative**
1. Use AI to generate stylistic / sonic palette references
2. Compose the piece's melodic, harmonic, and structural content with full attention
3. Final mix combines original composition with AI-generated atmospheric / textural content

**Pattern 4: AI for first draft, human for revision**
1. Generate full track from detailed prompt
2. Identify elements that work and elements that don't
3. Replace failing elements with composed material; keep working elements
4. Final piece is partly AI, partly human

**Pattern 5: AI as brainstorm collaborator**
1. Use a language model to explore options ("what would this progression sound like in a different mode?", "give me five ways to extend this melody", "what's a typical bridge for this kind of song?")
2. Compose with informed awareness; the AI is a sounding board, not a generator
3. Final piece is fully human composition; AI accelerated the thinking

The right pattern depends on the project, the budget, and the composer's strengths. None of these is "wrong"; all of them produce music that gets used commercially.

## AI music — creative-philosophical questions

These questions are real and unresolved at the snapshot date. Composers using AI tools should think them through and verify current law, platform rules, and active litigation before commercial delivery:

### Authorship and credit

If a human composer prompts, curates, and modifies AI output, who's the author? Most jurisdictions currently treat AI-generated content as not-copyrightable; works *combining* AI generation with human creative input are typically copyrightable in proportion to the human contribution. This is in flux; check current law in your jurisdiction.

### Training data and ethics

Most AI music models are trained on large datasets of recorded music — often without the consent of the artists whose music is in the training data. Whether this is fair use, infringement, or some new category is actively litigated and jurisdiction-dependent. Composers using AI tools should be aware that they may be participating in a chain whose ethical status is contested.

### Aesthetic / artistic identity

A composer who heavily relies on AI tools is, in some sense, outsourcing parts of their artistic voice to systems trained on other artists' work. Whether this is meaningful collaboration, problematic appropriation, or simply a new compositional tool is a question each composer should answer for themselves — and may need to defend to listeners, peers, and critics.

### Economic disruption

AI music tools threaten income for working composers, especially in lower-budget commercial spaces (corporate video, podcast underscores, small ad spots). The same tools may also democratize composition for people without formal training. Both effects are happening simultaneously.

### "Soul" / human meaning

A widespread aesthetic intuition: AI-generated music lacks something that human-composed music has. Whether this is true, illusory, or eroding over time is genuinely contested. Composers should test their assumptions: blind A/B tests often reveal that listeners can't distinguish well-curated AI output from human composition in contexts like background music. In other contexts (concert music, deeply personal songwriting), the gap is more reliably perceptible.

The honest position is uncertainty: AI tools produce some music indistinguishable from human composition and other music that's clearly inferior. Where the line falls — and where it will fall in 5 years — isn't known.

## Working as a composer in the AI era

A few practical orientations:

### Lean into what AI is bad at

Distinctive artistic voice, long-form coherence, deeply specific emotional intent, novel genre fusion, performance-driven music, music that engages with specific cultural / linguistic / political contexts — these are areas where human composition retains advantages. Composers focused here are more insulated from AI displacement.

### Use AI for what it's good at

Rapid ideation, library-style functional content, production tasks, iterative variations. A composer who refuses to use AI tools where they help is competing with one hand tied.

### Be transparent

If a track uses AI tools, and listeners or clients would care, disclose it. This is partly about legal compliance (some platforms now require disclosure), partly about trust.

### Develop your own creative voice independent of AI

The composer's value is increasingly in their unique perspective, taste, and craft. Time spent on those is time well spent. AI tools augment but don't replace this.

### Stay current

The capability landscape is moving fast. A workflow that's excellent today may be obsolete in 18 months. Reserve some time for ongoing tool evaluation.

## Composing algorithmically — practical approaches

For composers wanting to design their own algorithmic systems:

### Start small

A first algorithmic piece can be a single rule applied to existing material. "Every measure, the highest note moves up by one scale degree." "Every 4 bars, the chord progression shifts to a new key." Simple rules produce surprising results.

### Choose constraints carefully

The interesting question in algorithmic composition is what *not* to leave to the algorithm. Specifying pitch but leaving rhythm to chance produces different results than the inverse. The composer's craft is in choosing which dimensions to algorithmize and which to specify.

### Iterate on the algorithm

The first version of an algorithm rarely produces good music. Run it; listen; identify what's working and what isn't; refine the rules. The composer is designing the system, not just executing it.

### Curate the output

Algorithmic systems produce many possible outputs. Pick the best ones. The composer's taste shapes the work as much as the algorithm.

### Combine algorithmic and intuitive

Most successful "algorithmic" pieces aren't purely algorithmic. The composer uses the algorithm to generate raw material, then shapes it with intuition. *Music for Airports* used tape loops mathematically, but Eno chose the loops, the lengths, the durations, the textures. Pure algorithm + curation = composer.

## Common pitfalls when advising

- **Don't treat algorithmic output as finished.** Curation, shaping, and human editing produce the actual piece.
- **Don't assume "algorithmic" means "objective" or "fair".** The algorithm reflects its designer's choices; randomness reflects the distribution chosen; constraints reflect the composer's values. Algorithmic doesn't mean removed-from-judgment.
- **Don't reach for AI tools when you have the skill yourself.** A composer who can write a 30-second corporate cue in 20 minutes shouldn't necessarily prompt an AI tool to do it; the composed version is often better.
- **Don't use AI tools without understanding the legal context.** Output rights, training data, attribution, and platform policies vary and change. Check before delivering AI-generated work commercially.
- **Don't claim AI-only output as fully your composition.** Be honest about the role each tool played.
- **Don't dismiss algorithmic methods as "not real composition".** Process music, generative systems, and rule-based composition have produced canonical 20th- and 21st-century works.
- **Don't dismiss AI-assisted methods as "fake composition".** A skilled composer using AI tools can produce excellent work; a hacky composer not using AI tools can produce poor work. The tools don't determine quality.
- **Don't trust AI critical / theoretical analysis blindly.** Language models can analyze music but make errors (sometimes confident errors). Verify analytical claims against your own ear and theory knowledge.
- **Don't assume "AI" means one thing.** Different tools have different capabilities; using them well requires understanding what each is doing. A text-to-audio tool, a symbolic generation tool, and a language model are all "AI" but they're entirely different in workflow.
- **Don't ignore the tools' bias toward conventionality.** AI music tools reproduce statistical patterns of their training data — which means they default to genre conventions and average expectations. To produce distinctive work, you have to push back against the tool's defaults.

## Quick decision matrix

| Situation | Approach |
|-----------|----------|
| Long-form ambient or generative installation | Algorithmic composition with carefully designed rules; curate the system, not specific outputs |
| Library / corporate background music | AI music tools can produce viable first drafts; check rights / licensing |
| Distinctive artistic project | Compose by hand; AI can assist with brainstorming, production tasks, but artistic voice should be yours |
| Iteration on a melodic / harmonic idea | AI tools or rule-based generation can produce variations quickly; pick the strongest, refine |
| Rapid client demos | AI tools can show direction quickly; final delivery should usually be composed |
| Theoretical / analytical questions | Language models can help; verify their answers against your knowledge |
| Lyric drafting | LLMs can produce serviceable first drafts; revise for prosody and meaning |
| Game music with dynamic / procedural needs | Algorithmic / generative systems; design the system, not just the outputs |
| Educational exploration of a technique | Implement an algorithm yourself (twelve-tone, isorhythm, Markov chain, cellular automaton); learn by building |
| Live performance | Live coding tools (TidalCycles, SuperCollider, Sonic Pi, Max/MSP); the algorithm becomes performance |
| Collaboration with another composer | Use AI tools as one input alongside the human collaborator; don't let AI dominate the dialogue |

## Cross-references

- Constraint-based composition (the broader framework algorithmic methods fit within) → `constraint-based-composition.md`
- 20th-century techniques (where algorithmic methods entered Western art music) → `20th-century-techniques.md`
- Theme and variation (structured transformation, ancestor of algorithmic thinking) → `theme-and-variation.md`
- Game music (where algorithmic / procedural generation is common) → `../genres/game-music.md`
- Media and commercial music (where AI tools are increasingly used commercially) → `../genres/media-and-commercial-music.md`
- Workflow for collaborative / iterative work → `../workflow.md`
- Critique of generated material → `../critique-and-feedback.md`
- Pre-production decisions (when AI workflow choices fit) → `../production-aware/pre-production-decisions.md`
- Microtonal music (where algorithmic / non-12-TET intersect) → `microtonal.md`
- Current AI/legal/platform checks → `../research/web-music-trend-research.md`
- Notation (machine-readable formats: MIDI, MusicXML, MEI) → `../fundamentals/notation-and-conventions.md`

# Constraint-Based Composition

Sometimes the most generative compositional approach is to *limit* yourself. Constraint-based composition deliberately restricts choices — to a few pitches, to a fixed pattern, to a procedural rule — and forces creativity to operate within the limits. This file covers the common constraint techniques: limited pitch sets, ostinato, ground bass, isorhythm, process-based composition, and other systematic approaches.

For sonata form's structural constraints, see `form/classical-forms.md`. For 12-tone and serial constraints (a specific kind of constraint), see `20th-century-techniques.md`. For variations as a constraint approach, see `theme-and-variation.md`.

## When to consult this file

- The user is stuck creatively and wants a structural prompt
- Composing with a limited pitch palette
- Working with ostinato, ground bass, riffs, or pedal points
- Process-based or procedural composition
- Algorithmic or generative approaches
- Pedagogical exercises in composition
- Why constraints often produce better music than complete freedom

## Why constraints help

Total freedom is often paralyzing. A blank page can produce decision fatigue: too many options, no anchor, nothing forcing the next move. Constraints force choices and provide structure.

Constraints also:

- **Generate unity**: limited materials force repetition and variation, creating coherence.
- **Drive invention**: when an obvious move is forbidden, the composer has to find a less obvious one.
- **Create style**: most musical "styles" are sets of constraints. Common-practice tonality forbids parallel 5ths; 12-tone forbids tonal-sounding intervals; pop typically forbids 11-bar choruses.
- **Reduce decision space**: fewer options means less time spent deciding, more time spent composing.

The discipline of constraint is a fundamental compositional tool, used by composers from Bach to Reich.

## Limited pitch palette

The simplest constraint: use only certain pitches.

### Single mode or scale

Compose within a specific mode (Dorian, Lydian, etc.) for an entire piece. The lack of "borrowed" chromaticism creates a unified modal flavor.

### Pentatonic

Use only the 5 pitches of a major or minor pentatonic. Many folk traditions use pentatonic or pentatonic-adjacent scales; name the specific tradition instead of treating Celtic, Asian, or African practices as interchangeable. Avoids the half-step intervals that create dissonance.

### A specific 4–6 note set

Choose a small set of pitches and use only them. Useful for atmospheric pieces; the limited palette creates a specific color.

### Pitch class set as basis

Choose a 4–7 note set and let it govern the piece's pitch content. Set theory provides the analytical framework.

Example: a piece based on the set {0, 1, 4, 6} (e.g., C, C♯, E, F♯) — every melodic and harmonic event uses pitches from this set or its transpositions/inversions.

## Ostinato (repeated figure)

A short pattern that repeats throughout a section or piece. The pattern is the constraint; the music happens "over" or "around" it.

### Types of ostinato

| Type | Description | Examples |
|------|-------------|----------|
| **Bass ostinato** | A repeating bass line | Pachelbel's Canon, Holst's "Mars" |
| **Rhythmic ostinato** | A repeating rhythmic pattern | Bolero (snare), Drumming (Reich) |
| **Melodic ostinato** | A repeating melodic figure | Many Phil Glass works |
| **Harmonic ostinato** | A repeating chord progression | Pop loops, jazz vamps |

### Ostinato as compositional constraint

The constraint of an ostinato forces the composer to:

- Vary other elements (melody, orchestration, texture) while the ostinato continues.
- Create interest through change *around* the constant.
- Choose carefully which moments break the ostinato (the break becomes structurally significant).

### Famous ostinato pieces

- Pachelbel's Canon — bass ostinato (D-A-Bm-F♯m-G-D-G-A) repeats throughout
- Ravel's Boléro — single rhythmic ostinato through entire piece
- Holst's "Mars, Bringer of War" — driving 5/4 ostinato
- Reich's "Drumming" — extended ostinato variations
- Most pop songs have an ostinato (the bass line; the chord loop)

## Ground bass and chaconne / passacaglia

A ground bass is a specific type of ostinato — a bass line that repeats while the upper voices vary.

The forms:

- **Passacaglia**: variations over a ground bass. The bass repeats; everything above varies.
- **Chaconne**: similar; sometimes distinguished from passacaglia by tempo or character (chaconne in 3/4, often slower; passacaglia variable). The distinction is loose; often used interchangeably.

### Compositional approach

1. Compose a ground bass — usually 4–8 bars, in a clear key, with strong harmonic implications.
2. Repeat it many times.
3. Compose variations over it, increasing in complexity, intensity, or contrapuntal interest.
4. Build to a climax in the late variations.
5. Conclude with a final variation that resolves.

### Famous examples

- Bach, Passacaglia and Fugue in C minor — masterpiece of the form
- Brahms, Symphony No. 4, finale (passacaglia on a Bach theme)
- Pärt, *Passacaglia* for violin
- Britten, *The Young Person's Guide to the Orchestra* (passacaglia at the end)

### In jazz and pop

Jazz "vamps" are essentially ground basses. A song over a 4-bar repeating chord progression is using the principle. Modal jazz pieces ("So What", "Maiden Voyage") have extended sections over a single chord — a kind of single-chord ground.

## Isorhythm

A medieval and modern technique. A pre-set rhythmic pattern (the *talea*) is repeated, and a pre-set melodic pattern (the *color*) is also repeated, but they have different lengths so they don't realign immediately.

The two cycles intersect over time. Each repetition of the talea matches different pitches from the color.

### Example

Talea: 8 rhythmic units (e.g., 8 measures of a pattern)
Color: 12 pitches

Each repetition of the talea has different pitches. They realign every 24 measures (lcm(8,12)).

### Modern uses

Some 20th-century composers (Messiaen, Carter, Boulez) used isorhythmic principles to organize complex rhythmic structures. The technique provides a way to generate variety from a small amount of pre-composed material.

## Process composition

A piece governed by a process — a procedural rule that determines how the material evolves.

Steve Reich's *Music for 18 Musicians* and earlier process pieces. Philip Glass's early operas. La Monte Young's drone work.

### Examples of processes

- **Phasing**: two voices play the same pattern; one slowly drifts faster (or slower) than the other, creating an evolving rhythmic relationship that returns to alignment after some duration.
- **Additive**: notes are added (or subtracted) one at a time over many bars.
- **Subtractive**: the inverse — start with all notes, gradually remove.
- **Permutation**: a fixed set of pitches reordered systematically.
- **Note-by-note transformation**: each note is replaced by a procedural rule (raise by 1 semitone, e.g.).

### Compositional approach

1. **Choose a starting state**: a pattern, a chord, a single note.
2. **Choose a process**: an explicit rule for how the state changes.
3. **Apply the process**: repeatedly, until the music has unfolded.
4. **Decide when to stop**: typically when the process completes (returns to starting state, fully transforms, exhausts options).

The discipline: don't cheat. The process is the composer; the composer's job is to set it up well.

### Famous process pieces

- Reich, *Piano Phase* (1967) — two pianos play the same pattern; one accelerates slightly until they're back in phase.
- Reich, *Drumming* (1971) — additive process; notes added one at a time.
- Reich, *Pendulum Music* (1968) — microphones swung over speakers, creating evolving feedback.
- Glass, *Music in Twelve Parts* — additive and modular processes.

## Aleatoric / chance constraints

Constraints that force chance to determine outcomes. See `20th-century-techniques.md` for full treatment. Briefly:

- **I Ching divination** (Cage): hexagrams determine pitch, duration, dynamic.
- **Limited choice**: performer chooses among options provided by composer.
- **Graphic notation**: visual scores that performers interpret freely.

These approaches use constraints (the score's structure, the rules of choice) to channel randomness toward specific aesthetic ends.

## Algorithmic composition

A modern extension. Music is generated (or partly generated) by computer algorithms. Constraints are expressed as code.

### Approaches

- **Stochastic**: random within constraints (Xenakis used this — *Pithoprakta* is built from probability distributions).
- **L-systems / fractal**: mathematical structures generating melodic/rhythmic patterns.
- **Markov chains**: each event determined by the previous N events.
- **Genetic algorithms**: musical "fitness" functions evolve material over generations.
- **Neural networks** (modern): trained on existing music, generate new material in similar style.

### Modern uses

Algorithmic composition is increasingly common in:

- Generative ambient music
- Game audio (procedural music)
- Modular synthesis
- Some contemporary classical (David Cope, Brian Eno's generative work)

## Other constraint techniques

### Pedal point

Holding a single pitch (often the tonic or dominant) while harmony changes above. The pedal acts as constraint — all subsequent harmony must work over the held pitch.

Common in:

- Classical (Bach's organ music)
- Jazz (vamps over a static bass)
- Rock (Hendrix-style sustained drone)
- Film scoring (atmospheric pads)

### Cantus firmus

A pre-existing melody (often a chant) that's used as the structural backbone of a new piece. The cantus is given; new material is composed against it.

Renaissance and Baroque practice. Modern composers occasionally borrow.

### Quotation and collage

Using existing music as raw material. Constraints come from the source — its harmony, rhythm, character. Modern composers (Ives, Schnittke, Adams) use quotation for specific aesthetic ends.

In pop and hip-hop, sampling is a quotation/collage technique with its own constraints.

### Forbidden intervals or motions

Common-practice counterpoint forbids parallel 5ths and 8ves; 12-tone composition forbids tonal-sounding intervals; some pop production forbids changing chord on weak beats. Each style has implicit constraints.

Composers can invent their own: "this piece will never use a perfect 5th"; "every phrase must end with a stepwise descent"; "no two consecutive notes can be the same".

The constraints themselves become part of the piece's character.

### Time/duration constraints

- **Fixed total duration**: write a piece exactly 4 minutes long.
- **Fixed section lengths**: every section must be 16 bars.
- **Equal duration**: all sections of equal length.

Useful for forcing density decisions.

### Single-instrument or limited-ensemble constraints

Composing for solo violin, or for a specific small ensemble (e.g., piano trio, woodwind quintet). The instrumentation forces idiomatic writing.

## When constraints help most

Constraints are most useful when:

- **The composer is stuck**. A new constraint can break a creative impasse.
- **The piece needs unity**. Limited materials produce coherence.
- **The piece needs character**. Specific constraints (modal, ostinato, etc.) create immediate recognizable identity.
- **Working in unfamiliar territory**. Constraints provide rules to follow when the composer doesn't yet have intuition for the style.
- **Pedagogically**. Composition exercises with constraints are excellent teaching tools.

## When constraints become limiting

Constraints can hurt when:

- **They override musical instinct**. If the constraint forces music that sounds wrong, abandon the constraint or modify it.
- **They become the point**. Sometimes a piece is "about" the constraint; if the constraint is interesting only to analysts and not to listeners, the piece may fail.
- **They prevent revision**. If you can't change the constraint, you can't fix problems that arise from it.

The skilled composer uses constraints as tools, not as commandments.

## Common pitfalls when advising

- **Don't impose constraints on music that doesn't want them.** Free, intuitive composition is also valid. Match approach to user's intent.
- **Don't make constraints arbitrary.** A constraint should produce a desired aesthetic result. "Use only pitches A, C♯, F" is a constraint; what does it accomplish?
- **Don't conflate constraint with structure.** Sonata form has constraints (key relationships, sectional functions) but is also a flexible structure. Constraint and structure interact but aren't identical.
- **Don't ignore the result.** A constraint that produces beautiful music is a good constraint. One that produces only academic exercises is less so.
- **Don't underestimate simple constraints.** "Two-chord vamp" or "single-pitch drone" sounds limiting but has produced extraordinary music (modal jazz, drone rock, much ambient).
- **Don't be precious about strict adherence.** Sometimes the best music comes from following a constraint mostly and breaking it at one strategic moment.

## A workflow for constraint-based composition

If the user is starting a piece using a constraint:

1. **Choose the constraint deliberately.** What aesthetic effect should it produce? What kind of music does it suggest?
2. **Test it briefly.** Write a few bars or a phrase. Does it produce sound that's interesting or that's awkward?
3. **Commit to it for a while.** The piece develops by exploring what's possible within the constraint.
4. **Notice what the constraint forces.** When the obvious move is unavailable, what alternatives appear? These are the discoveries.
5. **Decide whether to break it.** If the constraint is producing deadlocks, modify it. If a strategic break would create dramatic moment, plan it.
6. **Use the constraint structurally.** Often a piece can have multiple sections, each with its own constraint, creating contrast through constraint-shifts.

## Quick decision matrix

| Want | Try (constraint) |
|------|------------------|
| Unified modal feel | Single mode for entire piece |
| Folk/world feel | Pentatonic-only |
| Atmospheric, contemplative | 4–6 note set; pedal point |
| Driving, hypnotic | Ostinato over which other elements vary |
| Structural variations | Ground bass / passacaglia |
| Process unfolding | Phasing or additive process |
| Algorithmic generation | Markov chains, L-systems, generative grammar |
| Atonal coherence | Pitch-class set as governing material |
| Strict serial discipline | 12-tone row with limited transformations |
| Stuck creatively | Add an arbitrary constraint to break the deadlock |
| Pedagogical exercise | Forbid a common move (parallel 5ths, leap larger than 5th) |

## Cross-references

- 12-tone and serial constraints → `20th-century-techniques.md`
- Algorithmic / AI-assisted composition (algorithmic systems are constraint sets) → `algorithmic-and-AI-assisted.md`
- Microtonal pitch systems as compositional constraint → `microtonal.md`
- Theme-and-variation as a kind of constraint → `theme-and-variation.md`
- Ostinato in jazz (vamps) and pop (loops) → `../harmony/jazz-harmony.md`, `../genres/pop-rock.md`
- Modal composition → `../harmony/modal-harmony.md`
- Counterpoint as constraint-based composition → `../counterpoint.md`
- Form as constraint → `../form/classical-forms.md`, `../form/popular-song-forms.md`
- Process music in 20th-century context → `../genres/classical-periods.md`
- Production-side constraints (DAW, instrumentation) → `../production-aware/arrangement-for-mix.md`

# Theme and Variation

Theme and variation is both a *form* (a piece structured as a theme followed by progressively transformed restatements) and a *technique* (the operations used to vary a theme). This file focuses on the technique: what gets varied, how, and to what effect. The form is treated in `form/classical-forms.md`.

For motivic development (related but different), see `melody/motivic-development.md`. For the sonata-form-style fragmentation that's distinct from variation, see also `melody/motivic-development.md`. For 20th-century process-based variation (like Reich's phasing), see `techniques/20th-century-techniques.md`.

## When to consult this file

- The user wants to vary a theme without changing it beyond recognition
- Writing a variations set
- Working out a piece organized around a single tune that gets repeated with changes
- Jazz solos as variations on chord changes
- Folk variation traditions (each verse the same melody, gradually embellished)
- Songwriting where verses share melody but feel different

## Theme and variation vs. motivic development

These are different techniques, often confused:

- **Theme and variation**: keeps the *whole theme* as the unit. Each variation transforms the theme as a whole — same length, same harmonic skeleton, same phrase structure. The theme is recognizable in every variation.
- **Motivic development**: extracts *fragments* (motifs) from a theme and develops those fragments freely. The theme as a whole may not appear; only its building blocks do.

In a sonata-form development section, motivic development dominates. In a variation set, theme and variation dominates. Sometimes they blend.

## What gets varied

When you "vary" a theme, what's actually changing? Several parameters can be transformed:

### 1. Melodic surface (figural variation)

The melody is decorated with passing tones, neighbor tones, ornaments, and rhythmic subdivisions. The basic skeleton (strong-beat notes) is preserved; the surface is elaborated.

This is the most common variation type. The listener still hears the original melody underneath the decoration.

Example: a theme with quarter notes becomes a variation in eighth-note triplets, with the same underlying pitches on each strong beat.

### 2. Rhythmic transformation

The theme's rhythm is changed while preserving pitch content. Common moves:

- **Augmentation**: notes get longer (a quarter-note theme becomes half notes).
- **Diminution**: notes get shorter.
- **New rhythm**: same pitches but completely new rhythmic shape.
- **Syncopation**: pitches displaced to weaker beats.

### 3. Harmonic recoloring

The chords accompanying the theme are changed. The melody might be the same, but the chords underneath produce different feelings.

Example: a theme harmonized with diatonic triads gets reharmonized with extended chords, secondary dominants, or substitutions. See `harmony/reharmonization.md` for techniques.

### 4. Modal transformation

The theme is recast in a different mode of the same tonic. Major theme → parallel minor. Minor theme → parallel major. Mode changes (Dorian instead of natural minor; Lydian instead of major).

Effect: dramatic emotional shift while keeping the theme recognizable.

Example: Schubert's *Winterreise* — songs in major mode with parallel-minor variants for "winter" effect.

### 5. Textural transformation

The accompaniment style changes:

- Block chords → arpeggios
- Solo melody → harmonized in 3rds and 6ths
- Chordal accompaniment → walking bass + chords
- Sparse accompaniment → dense polyphony

The theme remains; the texture around it is reimagined.

### 6. Contrapuntal transformation

The theme is combined with a counter-melody or a canon imitation. The theme is no longer alone but in dialogue.

Example: Bach's *Goldberg Variations* alternate decorative variations with canons (canon at the unison, 2nd, 3rd, 4th, etc.).

### 7. Character / topical transformation

The theme is recast in a new style — march, dance, hymn, scherzo, lullaby. The theme's surface changes radically; only the harmonic structure and phrase length identify the kinship.

Example: Beethoven's *Diabelli Variations* — some variations are completely transformed in character, almost unrecognizable.

### 8. Tonal transformation

The theme is in a new key. Less interesting on its own (it's just the theme transposed), but combined with other changes, can be effective.

### 9. Form-level transformation

The theme's internal structure is altered — perhaps a 16-bar theme is condensed to 8 bars, or expanded to 32 bars by inserting new material.

Less common in strict variation form (where length is usually preserved), but important in jazz and pop variation.

## Common variation types and their patterns

### Decorative variation

Surface ornament added to a clear theme. The listener recognizes the theme through decoration. Common in Mozart, Haydn, and pre-Romantic variation.

### Character variation

Each variation has a strongly different character — march, lullaby, scherzo, fugue. Beethoven and Brahms favored this approach.

### Form-altering variation

The variation includes a structural change — say, an introduction, a coda, or a fugal passage. Some variations are essentially mini-pieces in their own right.

### Fugal / canonic variation

A variation written as a fugue or canon on the theme. The theme becomes a contrapuntal subject. Common in Bach.

### Slow variation

A variation in noticeably slower tempo, often with deeper emotional expression. Adagio variations are typical at the climax of a variation set.

### Final variation

The closing variation, often most elaborate. May include extended coda, quotations of earlier variations, or transformative apotheosis of the theme.

## Famous variation sets

| Work | Notes |
|------|-------|
| Bach, *Goldberg Variations* | 30 variations on a sarabande theme; every third variation is a canon at progressively wider intervals |
| Bach, *Aria with 30 Variations* (same as Goldberg) | Same |
| Handel, *Harmonious Blacksmith* (5 variations) | Compact set, mostly figural |
| Mozart, "Ah, vous dirai-je, Maman" Variations | 12 variations on the "Twinkle, Twinkle" tune |
| Beethoven, *Diabelli Variations* | 33 variations on a banal waltz; ranges from witty to profound |
| Beethoven, *Eroica Variations* | Op. 35; piano variations on his own theme |
| Schubert, *Trout Quintet* (4th movement) | Variations on his own song "Die Forelle" |
| Schumann, *Symphonic Etudes* | Piano variations as character studies |
| Brahms, *Variations on a Theme by Haydn* | Orchestral; ends with passacaglia |
| Brahms, *Variations and Fugue on a Theme by Handel* | Piano; concludes with massive fugue |
| Brahms, *Paganini Variations* | Piano; technical showpiece |
| Rachmaninoff, *Rhapsody on a Theme of Paganini* | Variations for piano and orchestra; 18th variation is famous |
| Britten, *Young Person's Guide to the Orchestra* | Variations on Purcell theme; pedagogical |
| Elgar, *Enigma Variations* | Each variation portrays a friend |

## Variation in non-classical contexts

### Jazz

Improvisation over chord changes is essentially variation. The chord progression (the "form" or "head") is the theme; each chorus of solo is a variation.

Jazz variation differs from classical:

- **Improvised**, not pre-composed
- **Spontaneous**, not deliberate
- **Variation often departs from theme melodically**, returning only periodically
- **Multiple choruses** allow accumulation: each variation builds on the previous

In jazz, the most thematic improvisers (Sonny Rollins, Wynton Marsalis, Joshua Redman) explicitly develop motifs from the head across their solos. Less thematic improvisers play patterns that fit the changes without referencing the theme.

### Blues

12-bar blues form is a classic variation framework. Each chorus uses the same 12-bar form; the variation comes in the soloist's performance. Multiple choruses build over the form.

### Folk traditions

Many folk traditions have variation built in. A ballad's verses may share a melody, but skilled singers vary subtly between verses — adding ornaments, slightly altering rhythm, introducing different inflections.

In Irish traditional music, a tune is rarely played the same way twice. Each repetition adds different ornaments (rolls, cuts, slides) so the melody is constantly transforming.

### Pop song verses

A pop song's two or three verses share a melody, but a great topliner subtly varies them — different syllabic placement, slight melodic alterations, different ad-libs. The melody is "the theme"; each verse is a "variation".

This is more subtle than classical variation but follows the same principle.

### Process music

Steve Reich's "Music for 18 Musicians" and similar works use process-based variation: a pattern slowly transforms through procedure (one note at a time getting added, or two voices slowly drifting out of phase). Each "variation" is one step in the process.

This is variation operating at a long timescale — a single idea unfolding over an entire piece.

## Designing a variation set

If the user wants to write a variations piece:

### 1. Choose the theme

The theme should be:

- **Memorable**: distinctive enough that variations remain recognizable.
- **Simple**: not over-elaborate; the variations need room to add complexity.
- **Harmonically clear**: a clear chord progression that can serve as the framework.
- **Manageable length**: 8 or 16 bars is typical; longer themes make variations longer.
- **Suitable for variation**: themes that already have lots of ornament are harder to vary.

### 2. Decide the variation strategy

- **Strict variation**: every variation has the same length and harmony as the theme.
- **Free variation**: variations can extend, contract, modulate.
- **Character-set variation**: each variation portrays a different character.
- **Process-based variation**: each variation extends a procedural idea.

### 3. Plan the variation arc

A satisfying variation set has a shape:

- **Variations 1–3**: surface decoration; the theme remains clear.
- **Variations 4–6**: more substantial transformation; mode change, character shift.
- **Middle variations**: maximum invention; counterpoint, fugue, expansion.
- **Slow variation**: emotional centerpiece, often near the end.
- **Final variation**: virtuosic display, summative.
- **Coda**: bringing it home.

### 4. Variety within unity

Each variation should be distinct in character. Avoid two variations in the same texture or rhythmic style.

### 5. Save the most surprising for late

Don't reveal all your tricks early. The audacious variations (modal, contrapuntal, character-shifted) work best after the listener has settled into the theme.

## Common pitfalls when advising

- **Don't make variations too similar.** A set where every variation is "theme with eighth notes" gets boring fast. Variety is the point.
- **Don't lose the theme.** A variation that's so transformed it's unrecognizable isn't a variation — it's a free composition. The theme must remain perceivable, even if subtly.
- **Don't ignore harmonic skeleton.** Strict variations preserve the chord progression. Reharmonized variations are fine but should be clearly identified as such.
- **Don't over-decorate.** Surface ornament can suffocate the theme. Sometimes the most powerful variation is simpler than the theme, not more elaborate.
- **Don't ignore the variation arc.** A sequence of variations with no shape feels like a list, not a piece. Plan the journey.
- **Don't confuse variation with motivic development.** A piece that fragments the theme isn't a variation set; it's a development.

## Quick decision matrix

| Want | Try (variation type) |
|------|---------------------|
| Surface elaboration of theme | Figural / decorative variation (faster note values, ornaments) |
| Emotional darkening | Modal transformation (major to parallel minor) |
| Different character | Character variation (march, lullaby, scherzo, hymn) |
| Showcase contrapuntal skill | Fugal or canonic variation |
| Centerpiece of emotion | Slow variation (adagio, often with extended phrasing) |
| Modern textural variation | Reharmonization with extended chords or modal mixture |
| Climactic showcase | Final variation with extended coda, virtuosic display |
| Pop verse variation | Subtle melodic and ad-lib variation across verses |
| Jazz solo variation | Improvised variations over the head; thematic motivic development across choruses |
| Process-based variation | Gradual transformation over time (Reich, Glass style) |

## Cross-references

- Theme-and-variation as a form → `form/classical-forms.md`
- Motivic development (different from variation) → `melody/motivic-development.md`
- Reharmonization (key variation technique) → `harmony/reharmonization.md`
- Modal mixture and parallel-mode shifts → `harmony/chromatic-harmony.md`
- 20th-century process-based variation → `20th-century-techniques.md`
- Counterpoint (for fugal / canonic variations) → `counterpoint.md`
- Jazz improvisation as variation → `genres/jazz-styles.md`, `harmony/jazz-harmony.md`
- Folk variation traditions → `genres/folk-roots-and-traditions.md`

# 20th-Century Techniques

This file covers the major compositional techniques developed in the 20th century that depart from common-practice tonality: atonality and free atonal writing, 12-tone (serial) composition, total serialism, set theory, minimalism and process music, spectralism, aleatoricism, electronic and tape composition, microtonality, and the legacy of these techniques in contemporary practice.

For 20th-century classical context, see `genres/classical-periods.md`. For form innovations, see `form/classical-forms.md`. For rhythmic innovations (irregular meter, polyrhythm), see `rhythm-groove/odd-meters-polyrhythm.md`. For pitch systems beyond major/minor, see `fundamentals/pitch-intervals-scales.md` and `harmony/modal-harmony.md`.

## When to consult this file

- The user is composing in a non-tonal or post-tonal idiom
- Questions about 12-tone, atonality, serialism
- Minimalism and process music
- Set theory, pitch-class analysis
- Microtonal music and alternative tunings
- Aleatoric / chance composition
- Modern classical / contemporary art music

## A note on this territory

20th-century compositional techniques are diverse and often technical. This file gives an orientation; deep engagement with any of these techniques typically requires study with specialists, score analysis, and listening immersion. The agent should be helpful with high-level guidance and refer the user to specialized resources for serious work.

## Atonality (free atonal)

Atonal music has no key center. No single pitch functions as tonic. Pitches are organized by other principles — motivic, intervallic, gestural — rather than by tonal hierarchy.

### Free atonality (early atonal practice)

Schoenberg's pre-12-tone work (1908–1923, e.g., *Pierrot lunaire*, "Three Pieces for Piano" Op. 11). Atonal but without systematic ordering.

Characteristics:

- **No tonal center**: the listener can't identify a "home" key.
- **All 12 pitches treated as equal**: no pitch is privileged as tonic.
- **High dissonance**: consonances (3rds, 5ths) and dissonances (2nds, 7ths, tritones) are equally valid.
- **Motivic coherence**: pieces are organized around recurring intervallic patterns.
- **Free chromatic counterpoint**: voice leading is preserved but doesn't follow tonal rules.

Free atonality is harder to teach than 12-tone (which has explicit rules). It relies on the composer's ear for what "works".

### Set theory analysis

Allen Forte and others developed set theory as an analytical tool for atonal music. Key concepts:

- **Pitch class set**: an unordered collection of pitch classes.
- **Prime form**: the most compact representation of a set.
- **Interval class vector**: a list of how many of each interval class are in the set.
- **Z-relation**: two sets with the same interval-class vector but different prime forms.

Used analytically more than compositionally. The vocabulary is dense and useful for analyzing atonal music's internal logic.

For the user composing freely atonal music, the set-theory framework provides discipline: choose a set, derive themes from it, and check that the music's intervallic content stays consistent.

## Twelve-tone (serial) composition

Schoenberg's systematic approach to atonality (developed ~1923). The "method of composing with twelve tones related only to one another."

### The tone row

A specific ordering of all 12 chromatic pitch classes. Each pitch appears exactly once before repeating. The row is the piece's essential pitch material.

Constructing a row:

- **Avoid tonal-sounding intervals** (avoid triads, dominant-sevenths, leading-tone resolutions) unless that's the desired effect.
- **Build in distinctive intervallic identities** (e.g., a row dominated by major thirds, or one with all twelve chromatic intervals).
- **Test the row's transformations** (inversion, retrograde) to see if they're useful.

### Row transformations

Each row has 48 forms (in theory):

- **Prime (P)**: the original.
- **Inversion (I)**: each interval inverted (a m3 up becomes a m3 down).
- **Retrograde (R)**: the row backward.
- **Retrograde Inversion (RI)**: the inversion played backward.

Each form can start on any of 12 pitches, giving 12 × 4 = 48 row forms.

### Compositional use

In strict 12-tone:

- The row is presented in some form.
- Pitches must appear in row order. (You can't skip a pitch.)
- The row can be split between voices (some voices play row pitches 1–4, others 5–8 simultaneously).
- Multiple row forms can sound simultaneously (counterpoint).
- Row forms relate to each other by transposition, inversion, etc., providing structural unity.

In looser practice (Berg, Schoenberg in some works), the row is sometimes used flexibly.

### Famous 12-tone works

- Schoenberg, Variations for Orchestra (Op. 31), Fourth String Quartet
- Webern, Symphony Op. 21, Concerto Op. 24
- Berg, Violin Concerto (combines 12-tone with tonal allusions)
- Stravinsky, *Threni*, *Movements* (late, after embracing serialism)
- Boulez, *Le marteau sans maître*

### Modern 12-tone

The technique is rarely the dominant compositional method now, but its tools (pitch-class manipulation, structural ordering) inform many contemporary practices.

## Total serialism

Extending serial principles beyond pitch to rhythm, dynamics, articulation, and timbre. Each parameter is governed by an ordering or numerical procedure.

Key composers: Boulez, Stockhausen, Babbitt, Nono.

Stockhausen's "Gruppen" (1955–57) and Boulez's "Structures" series are exemplars.

Effect: highly controlled, often very dense, often atonal-sounding. The listener may not perceive the underlying systems but feels the rigor.

Total serialism dominated avant-garde classical for ~15 years (mid-1950s to late 1960s). It's now mostly historical, but its influence persists in highly structured contemporary music.

## Minimalism and process music

A reaction against serialism. Began in 1960s America (Riley, Reich, Glass, La Monte Young).

### Characteristics

- **Repetition** as a central principle.
- **Limited materials**: a small number of pitches, rhythms, or chords sustained over long durations.
- **Gradual process**: changes happen slowly over time; small variations accumulate.
- **Tonal or modal**: most minimalism is tonal (or pandiatonic), unlike serial music.
- **Audible structure**: the listener can hear the process unfolding.
- **Long durations**: minimalist pieces are often long (20+ minutes), letting processes develop.

### Specific techniques

- **Phasing** (Reich): two voices playing the same pattern slowly drift out of sync, creating evolving rhythmic relationships. *Piano Phase*, *Clapping Music*.
- **Additive process**: notes added or subtracted one at a time. Glass's early operas.
- **Pulse-based texture**: continuous repeated rhythmic figure, with melodic and harmonic events overlaid.
- **Modal harmony**: simple modal progressions repeated.
- **Phasing and counter-phasing**: layered rhythmic patterns that align and misalign over time.

### Composers

- **Steve Reich**: process music. *Music for 18 Musicians*, *Drumming*, *Different Trains*.
- **Philip Glass**: repetitive arpeggios; pulse-based; many operas (*Einstein on the Beach*, *Akhnaten*).
- **Terry Riley**: *In C* (1964) — the foundational minimalist piece.
- **La Monte Young**: drone and just intonation.
- **John Adams**: post-minimalist; expanded harmonic vocabulary; orchestral works.
- **Arvo Pärt**: "holy minimalism"; tintinnabuli technique.

### Modern uses

- Film and TV scoring (extensively; Reich and Glass have direct stylistic descendants).
- Ambient electronic music.
- Post-classical / neo-classical (Max Richter, Nils Frahm).
- Game music (especially atmospheric and contemplative scenes).

## Spectralism

Pitch and harmony derived from analysis of acoustic spectra (the partials of complex sounds). Developed in 1970s France (Grisey, Murail, Saariaho).

### Characteristics

- **Microtonal**: spectralist harmony often uses pitches that don't fit 12-TET.
- **Slow harmonic transformation**: chords morph gradually based on changing spectral content.
- **Color and timbre as primary**: the *sound* matters more than melodic line.
- **Often uses electronics or extended techniques**: to produce or imitate non-tempered pitches.

### Composers

- Gérard Grisey, *Les espaces acoustiques*
- Tristan Murail
- Kaija Saariaho
- Magnus Lindberg (early)

### Modern uses

Spectralist approaches inform much contemporary acoustic and electronic composition. Many film scorers use spectral-influenced techniques for atmospheric effects.

## Aleatoric / chance music

Music where significant elements are determined by chance or by performer choice rather than by composer's specification.

### John Cage

The pioneer. *4'33"* (1952) — performers play nothing for 4 minutes 33 seconds; the "music" is ambient sound.

Other Cage works use:

- **I Ching**-derived chance operations: notation determined by random choices.
- **Indeterminacy**: scores that allow performers to choose certain elements.
- **Prepared piano**: piano modified with objects between strings, producing unpredictable timbres.

### Other approaches

- **Earle Brown**: graphic notation (visual scores requiring performer interpretation).
- **Witold Lutosławski**: limited aleatoricism (controlled chance — performers improvise within strict guidelines).
- **Stockhausen**: aleatoric sections within otherwise notated pieces.

### Modern uses

Free improvisation (jazz and contemporary) draws on aleatoric ideas. Some contemporary composers use limited aleatoric techniques for textural moments.

## Electronic and tape composition

Music made with electronic generation or recorded-and-manipulated sound. Distinct from "electronic music" as pop genre.

### Major streams

- **Musique concrète** (Schaeffer, Henry, Pierre Boulez briefly): manipulating recorded sounds (sampled, reversed, looped).
- **Elektronische Musik** (Stockhausen, Eimert): purely synthesized sounds.
- **Tape music** (Cage, Babbitt, Varèse's *Poème électronique*): pre-electronic-instrument compositions on magnetic tape.
- **Computer music** (Max Mathews, Babbitt): digital generation.
- **Live electronics** (Stockhausen, Davidovsky): combining acoustic instruments with electronics in real time.

### Modern uses

Modern electronic composition is vast: ambient, IDM, modular synthesis, sound art. The techniques pioneered in the 1950s–60s underlie most modern electronic production.

## Microtonality

Music using pitches outside 12-TET. Various systems:

- **Quarter-tone music**: 24 equal divisions of the octave. Used in Arabic music traditionally; in modern Western, by Hába, Carrillo, and others.
- **Just intonation**: pitches based on simple frequency ratios (Pythagorean tuning, 5-limit just). Harry Partch, Lou Harrison, La Monte Young.
- **N-equal temperaments**: 19-TET, 31-TET, 53-TET, etc. Easley Blackwood explored these.
- **Ben Johnston**: extended just intonation, multiple-prime ratios.

### Modern uses

Microtonal music has small but active communities. Some film and game scoring uses microtonality for otherworldly effects. Spectralism and certain electronic music are inherently microtonal.

## Other 20th-century innovations

### Pandiatonicism

Using all the pitches of a major scale freely without functional progressions. Stravinsky, Copland, Hindemith. The harmony is "in C major" but not in any particular tonal direction.

### Polychords

Two triads stacked. Bitonality (two keys at once) is implied. Stravinsky's "Petrushka chord" (C major + F♯ major) is famous.

### Cluster harmony

Stack of seconds. Cowell, Penderecki, Ligeti. Used for sonic mass and color.

### Extended techniques

Non-traditional ways of playing instruments:

- Bowing on the bridge or below it
- Multiphonic on a wind instrument
- Singing while playing
- Tapping the body of an instrument
- Prepared piano

Common in modernist and contemporary classical; some appear in jazz and rock.

### Open form

Pieces without fixed sections; the order of sections is determined by the performer. Berio's *Sequenza* III, Stockhausen's *Klavierstück XI*.

### Time-bracket notation

Stockhausen and others used scores where rhythmic events occur within time brackets but not at exact moments. Performer interprets timing within the bracket.

## Common pitfalls when advising

- **Don't mistake atonality for randomness.** Atonal music is highly organized; just by different principles than tonality.
- **Don't dismiss 20th-century techniques as "ugly".** Each technique developed for specific reasons and produces specific aesthetic results. Match technique to intent.
- **Don't apply 12-tone to a piece that wants to be tonal.** 12-tone serves specific aesthetic ends; using it for tonal-sounding music defeats its purpose.
- **Don't ignore the listening experience.** Some techniques (total serialism, complex set theory) produce music that requires educated listeners. If the user is writing for a general audience, simpler techniques may serve better.
- **Don't conflate minimalism with simplicity.** Minimalist music is highly structured and disciplined. The "simplicity" is a surface effect; the underlying processes are sophisticated.
- **Don't assume a single 20th-century style.** The 20th century had many styles in dialogue. Stravinsky and Schoenberg were contemporaries, with very different approaches.

## Quick decision matrix

| Want | Try |
|------|-----|
| Maximum dissonance and organization | 12-tone serialism |
| Free atonal expression | Schoenberg pre-1923 style; Berg, Webern early |
| Total control over all parameters | Total serialism (Boulez, Stockhausen, Babbitt) |
| Sustained pulsing texture | Minimalism (Reich, Glass) |
| Process unfolding over time | Phasing (Reich), additive processes |
| Spiritual contemplative | Holy minimalism (Pärt, Tavener) |
| Microtonal color | Spectralism (Grisey, Murail) or just intonation (Partch, Harrison) |
| Indeterminacy / performer choice | Aleatoric notation (Cage, Brown, Lutosławski) |
| Diatonic without tonality | Pandiatonicism (Stravinsky neoclassical) |
| Two keys simultaneously | Bitonality (Stravinsky's "Petrushka chord") |
| Sonic mass | Clusters (Ligeti, Penderecki) |
| Computer-generated | Algorithmic / programmatic composition (Xenakis, Stockhausen, modern computer music) |
| Modern eclectic | Postmodern collage; quotation; style-mixing (Adams, Schnittke) |

## Cross-references

- 20th-century classical context → `../genres/classical-periods.md`
- Free atonality and modal alternatives to tonality → `../harmony/modal-harmony.md`
- Pitch systems beyond 12-TET → `../fundamentals/pitch-intervals-scales.md`, `microtonal.md`
- Rhythmic innovations of the period (irregular meter, polyrhythm) → `../rhythm-groove/odd-meters-polyrhythm.md`
- Counterpoint in serial music → `../counterpoint.md`
- Form innovations (open form, through-composed) → `../form/classical-forms.md`
- Constraint-based composition → `constraint-based-composition.md`
- Algorithmic and AI-assisted composition (computer-era extension of 20th-century rule-based methods) → `algorithmic-and-AI-assisted.md`
- Theme and variation in 20th-century use → `theme-and-variation.md`
- Modern uses in film, game, and media scoring → `../genres/film-tv-scoring.md`, `../genres/game-music.md`

# Microtonal and Xenharmonic Music

This file covers composing with pitch systems outside of 12-tone equal temperament (12-TET) — the assumed default of Western music since roughly the 19th century. Microtonal music uses intervals smaller than the semitone; xenharmonic music more broadly is anything that doesn't fit the 12-TET grid, including just intonation, equal divisions of the octave other than 12, non-octave-repeating systems, and historical temperaments.

These traditions are not new — most of the world's music is microtonal in some sense, and Western art music used unequal temperaments for most of its history. The "xenharmonic" framing is a modern composer's perspective that treats 12-TET as one option among many, not as a default.

This file complements `../fundamentals/pitch-intervals-scales.md` (basic pitch and scale theory) and `../harmony/modal-harmony.md` (modal systems, some of which require microtonal pitch). For Korean traditional ornamentation that includes microtonal inflections, see `../genres/korean-traditional.md`. For non-Western pitch systems generally, see `../genres/folk-roots-and-traditions.md`.

## When to consult this file

- Composing in just intonation, equal divisions other than 12-TET, or other tunings
- Working with traditions that use non-12-TET pitch (Indian classical, Arabic *maqam*, Indonesian *gamelan*, Persian *dastgah*, much non-Western music)
- Writing music with quarter-tone or smaller pitch inflections
- Composing for instruments that can produce microtones (strings, woodwinds, brass, voice, fretless instruments, electronic instruments, retuned digital instruments)
- Historical performance practice (mean-tone, well-tempered, just intonation)
- Spectralism (pitch derived from harmonic spectra)

## Why microtones?

Three broad motivations:

1. **Acoustic / physical**: 12-TET is a compromise. Just-intonation intervals (small whole-number ratios) have a sonic clarity that 12-TET approximations lack. A pure 5:4 major third (about 14 cents flat of 12-TET) sounds noticeably different.

2. **Cultural / traditional**: most non-Western traditions use pitch systems that don't fit 12-TET. To work authentically in Indian classical, Arabic, Persian, Turkish, or many other traditions, microtonal pitch is non-negotiable.

3. **Expressive / new**: some modern composers explore microtonality for the new musical possibilities it opens — pitch combinations and progressions impossible in 12-TET.

The composer's question is rarely "should I use microtones?" in the abstract; it's "is the music I want to make accessible only through pitch outside 12-TET, or is 12-TET sufficient?"

## Cents — the universal currency

Pitch differences across tuning systems are measured in **cents**: 100 cents per 12-TET semitone, 1200 cents per octave. Cents are logarithmic, so equal cent-differences are equal-sounding pitch ratios.

| Cents | Meaning |
|-------|---------|
| 0 | Same pitch |
| 5–10 | Barely audible difference (within tuning tolerance) |
| 14 | Just-intonation major 3rd (5:4) compared to 12-TET (deviation: 14 cents flat) |
| 22 | Pythagorean comma — the gap between 12 perfect fifths and 7 octaves |
| 50 | Quarter-tone (half a semitone) |
| 100 | One semitone |
| 200 | One whole tone |
| 1200 | Octave |

Cents notation is the standard way to specify exact pitch in microtonal contexts. "C +14¢" means C, 14 cents sharp.

## Just intonation

Just intonation (JI) tunes intervals as small whole-number ratios. The simplest:

| Interval | Ratio | Cents | 12-TET equivalent |
|----------|-------|-------|-------------------|
| Unison | 1:1 | 0 | 0 |
| Octave | 2:1 | 1200 | 1200 (exact) |
| Perfect 5th | 3:2 | ~702 | 700 (12-TET ~2¢ flat) |
| Perfect 4th | 4:3 | ~498 | 500 (12-TET ~2¢ sharp) |
| Major 3rd | 5:4 | ~386 | 400 (12-TET 14¢ sharp) |
| Minor 3rd | 6:5 | ~316 | 300 (12-TET 16¢ flat) |
| Major 6th | 5:3 | ~884 | 900 (12-TET 16¢ sharp) |
| Minor 6th | 8:5 | ~814 | 800 (12-TET 14¢ flat) |
| Major 2nd (Pythagorean) | 9:8 | ~204 | 200 (12-TET 4¢ flat) |
| Major 2nd ("just") | 10:9 | ~182 | 200 (12-TET 18¢ sharp) |
| Harmonic 7th | 7:4 | ~969 | 1000 (12-TET 31¢ sharp) |
| Septimal minor 3rd | 7:6 | ~267 | 300 (12-TET 33¢ sharp) |

The deeper ratios (using prime numbers 7, 11, 13) sound increasingly unfamiliar to ears accustomed to 12-TET. The 7:4 "harmonic 7th" — 31¢ flatter than the 12-TET minor 7th — has a particular blues-like or "leaning" quality.

### The comma problem

Just intonation has a structural problem: you can't tune a 12-note keyboard such that every interval is just. Different paths between two pitches yield different exact frequencies (because 5/4 × 5/4 ≠ 25/16 ≠ 8/5).

Three responses to this problem:

1. **Just intonation, restricted**: tune the keyboard for one tonal center; modulation breaks the system. La Monte Young, Harry Partch, Ben Johnston work this way (Partch built a 43-tone scale and his own instruments).
2. **Adaptive just intonation**: an electronic instrument or a vocal ensemble re-tunes on the fly to keep each chord just. Difficult for keyboard instruments; natural for voices and strings.
3. **Temperament**: distribute the comma errors across many intervals, sacrificing pure ratios for usability across all keys. This is what 12-TET does — every interval slightly off, but every key equally usable.

## Equal divisions of the octave (EDOs)

Equal divisions of the octave divide the octave into N equal cents-steps. 12-TET is 12-EDO (12 equal divisions, 100 cents each). Other EDOs:

| EDO | Step size | Notable for |
|-----|-----------|-------------|
| 5-EDO | 240¢ | Indonesian *slendro* gamelan (approximate) |
| 7-EDO | ~171¢ | Thai traditional tuning (approximate); also resembles a "perfectly equalized" diatonic scale |
| 12-EDO | 100¢ | Western standard |
| 17-EDO | ~71¢ | Approximates Pythagorean tuning; popular in xenharmonic experimentation |
| 19-EDO | ~63¢ | Better approximations of just-intonation thirds than 12-TET; meantone-like properties |
| 22-EDO | ~55¢ | Sruti-related (Indian classical theory) |
| 24-EDO | 50¢ | Quarter-tone system; standard in modern Arabic and some Western avant-garde notation |
| 31-EDO | ~39¢ | Excellent JI approximations; popular in microtonal communities |
| 53-EDO | ~23¢ | Almost-perfect Pythagorean and just intervals; approximates Turkish *koma* |
| 72-EDO | ~17¢ | Used by Joe Maneri, some jazz microtonal applications |

Each EDO has its own "modal" possibilities. In 19-EDO, for instance, there are scales that have no analogue in 12-TET but feel internally consistent. In 31-EDO, many just intervals are within a few cents of an EDO step, making 31-EDO a strong "everyday" microtonal option.

## Quarter-tones — the modest microtonal step

The most common microtonal extension of 12-TET is the addition of quarter-tones — pitches halfway between adjacent semitones, at 50¢ intervals. This gives 24-EDO when applied uniformly.

Quarter-tones are central to:
- **Modern Arabic music**: *maqam* scales use quarter-tone-related intervals (though traditional Arabic intervals are not exactly 24-EDO; they're closer to specific just-intonation tunings)
- **Avant-garde Western**: Charles Ives, Alois Hába, Iannis Xenakis, Ligeti, Pierre Boulez
- **Spectral music**: Gérard Grisey, Tristan Murail, Kaija Saariaho
- **Persian / Turkish music**: traditions use scales with quarter-tone-like intervals (though the actual tuning is more nuanced)

### Notation conventions for quarter-tones

| Symbol | Meaning |
|--------|---------|
| ♯ (sharp) | +100¢ |
| ♮ (natural) | natural |
| ♭ (flat) | −100¢ |
| ♯ with vertical line through stem (or "quarter-sharp") | +50¢ |
| ♭ with reversed flag | −50¢ |
| Various Stein-Zimmerman accidentals | +50¢, +25¢, etc. |
| Cent values in superscript: `C +14¢` | Exact specification |

Standardization is incomplete. Different schools use different symbols. When notating microtonal music, include a key clarifying which symbol means which deviation.

## Non-Western traditions that require microtones

### Indian classical music

Built on the *sruti* system, traditionally counted as 22 *sruti* per octave (though they are not equally spaced). The 22-sruti system maps to specific just-intonation ratios.

Key concepts:
- **Raga**: a melodic framework, with characteristic ascending / descending patterns, ornamentation rules, and time-of-day associations
- **Sa, Ri, Ga, Ma, Pa, Dha, Ni** — solfege analogous to Western Do-Re-Mi but with multiple variants of Ri, Ga, Dha, Ni at different sruti
- **Gamaka**: ornamental microtonal slides between fixed pitches; central to Indian melody, comparable in aesthetic role to Korean *sigimsae*

Performance is typically heterophonic (melody + tanpura drone + tabla); chordal harmonic accompaniment is generally non-idiomatic to traditional Indian classical, though contemporary Indian fusion and film music use Western harmony freely.

### Arabic *maqam*

Arabic music uses the *maqam* system: scales with characteristic interval patterns that include quarter-tone-like intervals. Major *maqamat*:

- *Rast*: roughly C D E♭⁺ F G A B♭⁺ (with E♭⁺ and B♭⁺ being neutrals around 50¢ flat of E and B)
- *Bayati*: C D⁻ E♭ F G A♭ B♭ (with D⁻ being a neutral around 50¢ flat)
- *Hijaz*: C D♭ E F G A♭ B (with augmented 2nd between D♭ and E)
- *Saba*: C D⁻ E♭ F♭ G A♭ B♭

The "neutral" intervals (between major and minor) are central to the *maqam* sound. They're not exactly quarter-tones; they vary by region and tradition (Egyptian vs. Syrian vs. Turkish *maqam* are tuned differently).

### Persian *dastgah* and Turkish *makam*

Related to Arabic *maqam* but distinct. Persian *dastgah* uses *koron* (lowered by ~50¢) and *sori* (raised by ~50¢) accidentals. Turkish *makam* uses *koma* — about 1/9 of a tone — as the basic step, giving many subtle interval variations.

### Indonesian gamelan

*Slendro* (5-pitch, near-equal) and *pelog* (7-pitch, unequal) tunings. Each gamelan ensemble is tuned to its own slightly different version — there's no standard pitch reference. Tuning *is* the gamelan's identity.

### Other traditions

Many West African, sub-Saharan African, Polynesian, and Native American traditions use pitch systems that don't fit 12-TET. Generalizations across these are dangerous; specific knowledge of a tradition is essential.

## Spectralism

A modern Western tradition (Grisey, Murail, Saariaho, Hurel) that derives pitch from the harmonic spectrum of natural sounds. Rather than starting from scales, the composer analyzes a sound's spectrum and uses the partials as pitch material — which often produces microtonal pitches because the upper partials don't align with 12-TET.

Practical implication: spectralism uses ensembles that can produce specific microtonal pitches (string and wind ensembles primarily; sometimes specially-tuned percussion or electronics). The sound is often "shimmery" or "non-functional" in the harmonic sense — the music is about timbre and pitch-as-color rather than about chord progression.

## Composing microtonally — practical workflow

### Decide what kind of microtonality you want

The decisions cascade:

- **Which tradition / tuning?** Just intonation? An EDO (which one)? Quarter-tones? *Maqam*-style neutrals? Spectral? Each choice constrains everything downstream.
- **Which instruments?** Voices and strings can produce any pitch with proper rehearsal. Fretted instruments need refretting or specific bend technique. Keyboards need retuning (digital is easy, acoustic is committal). Wind instruments have specific microtonal capabilities and limits.
- **How will you notate?** Cents-precision? Stein-Zimmerman accidentals? Tradition-specific notation (Indian *sargam* ornament marks, Arabic *maqam* notation)?
- **Who's performing?** A chamber ensemble specializing in new music vs. a symphony orchestra vs. a tradition-specific ensemble (Indian classical, Arabic *takht*, gamelan) — each handles different microtonal approaches.

### Notating microtonal music

Some practical conventions:

- **Add a "microtonal accidentals key"** at the start of the score, defining each symbol's exact cent value
- **Use cent annotations** where precision matters (`C +14¢` or `C +14`)
- **For frequent microtones**, define the system once (e.g., "quarter-tones, 24-EDO") and use a consistent symbol set
- **For tradition-specific microtones**, use the tradition's notation (Indian sruti markings; Persian *koron*/*sori*) when working in that tradition

Modern engraving software (Sibelius, Finale, Dorico, MuseScore, Lilypond) supports microtonal notation to varying degrees. Custom symbols may need image insertion or custom-font solutions.

### Rehearsal cost

Microtonal music takes more rehearsal time than 12-TET music, often substantially. Singers and instrumentalists must learn unfamiliar pitches; ensembles must learn to tune unfamiliar intervals together.

If working with a non-specialist ensemble, budget extra rehearsal. For amateur ensembles, microtonality may be impractical at any meaningful complexity.

For specialists (microtonal new music ensembles, Indian classical performers, Arabic *takht* musicians), microtonal pitch is normal and rehearsal cost is no different from any new piece.

## Common pitfalls when advising

- **Don't conflate "microtonal" with "out of tune".** Microtonal music is precise; it just uses different pitches than 12-TET. "Out of tune" is failure to hit intended pitches; microtonality is hitting different intended pitches.
- **Don't apply Western theoretical concepts uncritically.** Indian classical *raga* is not a "scale" in the Western sense; it has rules of motion and ornamentation that go beyond pitch content. Arabic *maqam* is not "Western minor with a quarter-tone".
- **Don't assume EDO is the framework for non-Western music.** Indian classical *sruti* is not 22-EDO; Arabic *maqam* tuning is not 24-EDO; gamelan is not 5-EDO or 7-EDO. EDOs *approximate* these traditions but aren't identical.
- **Don't ignore performer capability.** A microtonal piece for non-specialist performers needs simpler microtonality than the same piece for specialists.
- **Don't romanticize "microtonality" as magic.** The aesthetic value depends on the specific tuning, the music written for it, and the performance — not on microtonality in itself.
- **Don't use microtones for shock value.** Microtonality used as a one-time "weird" effect rarely works; sustained microtonal language used coherently can be transformative.
- **Don't notate carelessly.** Imprecise notation in microtonal music produces garbage in performance. Cent values, accidental keys, and consistent conventions matter.
- **Don't ignore digital options.** A retuned digital instrument (synth, sampler, MIDI keyboard) is the easiest way to compose microtonally. Tools like Scala, the SCALA tuning database, and Pianoteq's micro-tuning support are now standard.

## Quick decision matrix

| Want | Try |
|------|-----|
| Pure intervals, "natural" sound | Just intonation; restrict to one tonal center; use voices, strings, or retuned digital instruments |
| Quarter-tone Western avant-garde | 24-EDO; reference Hába, Ives, Boulez |
| Spectral / shimmering / atmospheric | Spectralism; pitches from harmonic series; reference Grisey, Murail, Saariaho |
| Arabic / Persian / Turkish tradition-adjacent writing | Specific *maqam* / *dastgah* / *makam*; work with tradition specialists; use tradition-appropriate notation |
| Indian classical influence | Specific *raga* with proper *gamaka*; recognize that pitch is only part of the system |
| Indonesian gamelan-adjacent writing | Pelog or slendro tuning (find a specific gamelan's tuning); use specific gamelan instruments |
| Equal-temperament alternative to 12-TET | 19-EDO or 31-EDO for better just approximations; 17-EDO for Pythagorean-like sound |
| Microtonal slides and bends without committing to a system | Continuous pitch on strings, voice, fretless bass, slide instruments; notate with glissando + cent targets |
| Korean traditional-adjacent ornamentation | Stay in 12-TET pitch frame but add *sigimsae*-style microtonal pitch inflections; reference `../genres/korean-traditional.md` |
| Demo / prototype microtonal idea | Retuned digital instrument (any DAW with microtonal support); iterate before committing to a notational and performer plan |

## Cross-references

- Pitch system fundamentals (12-TET, intervals, octave equivalence) → `../fundamentals/pitch-intervals-scales.md`
- Modal harmony (the modal frame applies to many microtonal traditions) → `../harmony/modal-harmony.md`
- Korean traditional music with *sigimsae* microtonal inflections → `../genres/korean-traditional.md`
- Folk, roots, and tradition-specific music (where microtonality is common) → `../genres/folk-roots-and-traditions.md`
- 20th-century compositional techniques (where microtonality intersects with avant-garde) → `20th-century-techniques.md`
- Constraint-based composition (microtonality often is the constraint) → `constraint-based-composition.md`
- Notation generally (microtonal accidentals are an extension) → `../fundamentals/notation-and-conventions.md`
- Orchestration (instrument capabilities for microtones) → `../orchestration/instruments-ranges-character.md`

# Phase B Correctness Pass

Use this document to prepare the skill for first release without running the Phase C prompt-smoke-test suite. Phase B is a content and correctness hardening pass: it checks the files most likely to be quoted directly by the agent, the theory files most likely to propagate mistakes, and the cultural/currentness files most likely to overclaim.

## Scope

Phase B covers:

1. high-use `assets/` cheat sheets,
2. `references/harmony/`,
3. `references/fundamentals/`,
4. culturally specific files with higher overclaim risk,
5. `references/research/` currentness, privacy, and originality boundaries,
6. release-readiness documentation and automated regression checks.

Phase B does **not** require running the user-facing prompt-smoke-test suite. That belongs to Phase C.

## Review order and acceptance criteria

| Area | Files | Acceptance criteria | Status |
|---|---|---|---|
| High-use assets | chord symbols, intervals, modes, jazz voicings, progressions, response templates | no known P0/P1 theory regression; ambiguous notation handled with defaults and caveats | reviewed for release-prep |
| Harmony | chord construction, functional harmony, chromatic harmony, jazz harmony, modulation, reharmonization, voice leading | altered tensions, augmented-sixth claims, dominant substitutions, and voice-leading rules are context-aware | reviewed for release-prep |
| Fundamentals | notation, pitch/interval/scale, rhythm/meter, prosody/language | no overbroad tuning/history/language claims; spelling examples match function | reviewed for release-prep |
| Cultural / regional | Korean traditional, MENA, South Asian, folk/roots, regional starters | traditions are not reduced to scales, samples, or broad labels; collaboration caveats are present | reviewed for release-prep |
| Research/currentness | web research, streaming context, style/copyright, reference digging, regional trends | no credential requests; no stale-current claims without snapshot/web-check language | reviewed for release-prep |

## Manual corrections applied in this pass

| Severity | File | Correction |
|---|---|---|
| P0/P1 | `references/genres/korean-traditional.md` | Replaced the overly fixed `Jinyangjo = 18 beats` line with a 6-beat-cycle / 24-beat-phrase caveat; clarified large-beat vs subdivision descriptions for faster `jangdan`. |
| P1 | `assets/intervals-and-scale-formulas.md`, `assets/modes-cheatsheet.md`, `references/fundamentals/pitch-intervals-scales.md` | Corrected C whole-tone dominant spelling to use B♭ for ♭7; corrected C half-whole diminished spelling to use D♯ for ♯9 rather than E♭. |
| P1 | `assets/chord-symbol-conventions.md`, `references/harmony/chord-construction.md` | Tightened `C7alt` explanation so natural 5 is not presented as part of the default altered pitch set; clarified altered dominants use selected subsets. |
| P1 | `references/harmony/chord-construction.md` | Corrected upper-structure triad caveats: `D♭/C7` includes a natural 11 that clashes with the 3rd if the shell includes E; `F♯/C7` needs enharmonic explanation for the ♭7. |
| P2 | `references/fundamentals/pitch-intervals-scales.md` | Softened the *diabolus in musica* claim and perfect-consonance explanation; added strict-harmony caveat for P4 above the bass. |
| P2 | `references/fundamentals/pitch-intervals-scales.md` | Softened pentatonic-over-chords language so it is treated as a common flexible strategy, not a universal guarantee. |
| P2 | `references/harmony/voice-leading.md` | Replaced absolute “never double” wording with strict common-practice cautions and genre-aware exceptions. |
| P2 | `references/rhythm-groove/odd-meters-polyrhythm.md` | Clarified that tala is a named performance framework, not just a time signature; separated Indian tala from Middle Eastern/Turkish iqa'/usul-like cycle concepts. |
| P2 | `references/genres/korean-traditional.md` | Replaced fixed-pitch implication around *hwangjong* and softened claims about all professional *gugak* musicians reading both Western notation and `jeongganbo`. |
| P2 | `references/genres/film-tv-scoring.md` | Replaced generic “ethnic percussion” labels with “tradition-specific / regional” language. |

## Spot-check checklist

Before RC1, the maintainer should still spot-check these by hand, especially if any file changes after this pass.

| Check | Expected result |
|---|---|
| `C(no5)` | omit-fifth chord, not a power chord |
| `C7alt` | dominant shell plus selected altered tensions; natural 5 not listed as default altered pitch |
| `E7♯9` | G natural is ♯9; A♯ is ♯11 / ♭5 territory |
| C whole-tone dominant color | C D E F♯ G♯ B♭ when spelling ♭7 function |
| C half-whole diminished over C7 | C D♭ D♯ E F♯ G A B♭; D♯ is ♯9 |
| F Lydian | F G A B C D E; no B♭ |
| German +6 | can be enharmonically reinterpreted as dominant 7; Italian +6 is not a full V7 pitch set |
| `Jinyangjo` | 6-beat cycle with common 24-beat phrase grouping caveat, not a simple 18-beat claim |
| Maqam/raga/tala | not reduced to Western scale/time-signature presets |
| Streaming context | asks for user-shared public/lightweight evidence only; no credentials |
| Style-reference/copyright | extracts craft variables; does not copy melody, lyrics, riffs, samples, signature tags, or vocal identity |

## Phase B exit decision

Phase B can be marked prepared/complete when:

- `scripts/music_theory_sanity_check.py` passes,
- internal Markdown links resolve,
- the corrections above are present,
- current-sensitive files retain snapshot notes,
- no active guidance uses broad culture-as-color labels,
- Phase C prompt tests remain available but unrun.

## Cross-references

- Release roadmap → `../../RELEASE-ROADMAP.md`
- First-release readiness → `first-release-readiness.md`
- Prompt smoke tests, Phase C only → `prompt-smoke-tests.md`
- Music-theory audit rubric → `../../assets/music-theory-audit-rubric.md`
- Sanity checker → `../../scripts/music_theory_sanity_check.py`

# First Release Readiness Checklist

Use this checklist before shipping the music-composition skill as a first release, and retain it as the audit trail for v1.0. It is maintainer-facing and should be used alongside `../../RELEASE-ROADMAP.md`, `../../scripts/music_theory_sanity_check.py`, `phase-b-correctness-pass.md`, `prompt-smoke-tests.md`, `phase-c-smoke-test-results.md`, `rc1-packaging-checklist.md`, and `../../KNOWN-LIMITATIONS.md`, and `../../VALIDATION-v1.0.md`.

## When to consult

Consult this file when preparing a release candidate, after a major patch, or before publishing the skill to users who will rely on it for real composition help.

## Release readiness scorecard

| Area | Required for first release | Status to record |
|---|---|---|
| Navigation | `SKILL.md` and `00-navigation.md` reflect all major folders/files | pass / fail / needs review |
| Links | Internal `.md` references resolve | pass / fail |
| High-use assets | Chord symbols, modes, intervals, cadences, voicings, progressions checked | pass / fail / spot-check only |
| Phase B correctness | `phase-b-correctness-pass.md` completed and corrections retained | pass / fail / partial |
| Known regressions | Prior P0/P1 errors remain fixed | pass / fail |
| User workflow | Brainstorming, collaboration, revision loop, critique, output templates present | pass / fail |
| Research workflow | Web search, trend research, streaming context, style/copyright, reference digging present | pass / fail |
| Snapshot notes | Current-sensitive files have snapshot notes | pass / fail |
| Cultural specificity | Broad shortcut labels replaced by specific source/scene/idiom language | pass / fail |
| Prompt smoke tests | Representative prompts tested for routing and answer quality | pass / fail / partial |
| Phase C result record | Prompt-smoke-test result recorded in `phase-c-smoke-test-results.md` | pass / fail / partial |
| Known limitations | `../../KNOWN-LIMITATIONS.md` present or limitations included in validation report | pass / fail |
| ZIP hygiene | No cache, scratch, temp, or obsolete files included | pass / fail |
| v1.0 validation | `../../VALIDATION-v1.0.md` present and consistent with package counts | pass / fail |

## First-release blockers

Treat these as release blockers:

- broken internal `.md` links,
- missing `references/00-navigation.md` routes for major new files,
- missing `SKILL.md` structural updates after a major folder addition,
- known high-risk theory regression in an `assets/` file,
- current-sensitive file without a snapshot note,
- user-account or credential request in streaming-service guidance,
- style-reference guidance that encourages copying melody, lyrics, riffs, samples, signature production tags, or vocal identity,
- cultural guidance that uses broad shortcut labels instead of naming the source tradition, scene, or musical variable,
- validation report claiming a check passed when the checker was not actually run,
- release-candidate package without known limitations or an equivalent limitations section.

## High-use asset review order

Review assets in this order because they are most likely to be quoted directly by the agent.

1. `assets/chord-symbol-conventions.md`
2. `assets/chord-symbol-ambiguity-and-parsing.md`
3. `assets/intervals-and-scale-formulas.md`
4. `assets/scale-degree-spelling-cheatsheet.md`
5. `assets/modes-cheatsheet.md`
6. `assets/jazz-voicings.md`
7. `assets/cadence-reference.md`
8. `assets/progressions-catalog.md`
9. `assets/response-templates.md`
10. `assets/diagnostic-checklists.md`
11. `assets/web-search-cheatsheet.md`
12. `assets/musical-brainstorming-cards.md`
13. `assets/session-brief-and-decision-log.md`
14. `assets/trend-and-reference-matrices.md`
15. `assets/music-theory-audit-rubric.md`

## Music-theory spot checks

Before release, spot-check at least these items manually.

| Item | Expected result |
|---|---|
| `C(no5)` | C and E; not a power chord |
| `C5` / `C(no3)` | root + fifth; power chord context |
| `A5` inversion | `d4` |
| `m6` inversion | `M3` |
| `E7alt` with A♯ | A♯ = ♯11 / ♭5, not ♯9 |
| E over G natural | G natural = ♯9 |
| F Lydian | contains B natural, not B♭ |
| German +6 enharmonic reinterpretation | can map to dominant 7th; Italian +6 does not have the full V7 pitch content |
| Maqam/makam | not reduced to a 24-TET subset |
| 12-TET history | modern default, not a universal “since 1700” claim |

## Output behavior checks

For each prompt smoke test, record whether the assistant did the following.

| Behavior | Pass condition |
|---|---|
| Genre framing | Named or inferred the genre frame, or asked only if needed |
| Concrete output | Gave chords, notes, rhythm, form map, or revision instructions |
| Options | Offered 2–4 usable directions when the user asked creatively |
| Rationale | Explained why each option changes the sound |
| User agency | Let the user choose, especially in brainstorming and revision |
| Currentness | Used web research only when current facts mattered |
| Safety/originality | Avoided copying protected expression |
| Cultural specificity | Avoided broad regional shortcuts |

## Validation report template

Use this structure for release-candidate validation reports.

```md
# Validation Report — RCx

## Package

- Package: `<filename>.zip`
- Base version: `<previous version>`
- Date: `<YYYY-MM-DD>`

## File counts

- Markdown files:
- Total files:
- New files:
- Modified files:
- Deleted files:

## Automated checks

- `scripts/music_theory_sanity_check.py`: PASS / FAIL
- Internal `.md` references:
- Snapshot notes:
- Known regression checks:
- ZIP integrity:

## Manual review

- High-use assets:
- Research/currentness:
- Cultural specificity:
- Prompt smoke tests:
- Phase C smoke-test results:

## Remaining limitations

- ...

## Release decision

- Ship / hold / ship with caveats
```

## Cross-references

- Release roadmap → `../../RELEASE-ROADMAP.md`
- v1.0 validation report → `../../VALIDATION-v1.0.md`
- RC1 validation report → `../../VALIDATION-RC1.md`
- Known limitations → `../../KNOWN-LIMITATIONS.md`
- Phase B correctness pass → `phase-b-correctness-pass.md`
- Prompt smoke tests → `prompt-smoke-tests.md`
- Phase C smoke-test results → `phase-c-smoke-test-results.md`
- RC1 packaging checklist → `rc1-packaging-checklist.md`
- Chord-symbol audit → `../../assets/chord-symbol-ambiguity-and-parsing.md`
- Scale-degree spelling → `../../assets/scale-degree-spelling-cheatsheet.md`
- Music-theory audit rubric → `../../assets/music-theory-audit-rubric.md`
- Maintenance guide → `../../MAINTENANCE.md`

# Phase C Smoke-Test Results

This document records Phase C prompt-smoke-test status for release preparation. It should be read with `prompt-smoke-tests.md`, `first-release-readiness.md`, `phase-b-correctness-pass.md`, and `../../RELEASE-ROADMAP.md`.

## Status snapshot

| Field | Status |
|---|---|
| Phase | C — prompt simulation / smoke-test behavior |
| Result | Pass, based on user-run smoke testing reported on 2026-04-27 |
| Scope | Representative routing, loading discipline, answer shape, currentness behavior, and workflow behavior |
| Major issues found | None reported |
| Over-reading behavior | None reported; file loading stayed within expected range |
| Randomness / variation | Within expected creative range |
| Follow-up action | No blocker fixes required; apply non-blocking release-hardening improvements only |

## What this result does and does not claim

This document records the reported outcome of Phase C. It does **not** claim that every possible prompt, genre, culture area, theory edge case, or current trend workflow has been exhaustively tested.

A Phase C pass means:

- the routing structure appears usable in representative prompts;
- the assistant does not obviously over-load files;
- the answer shapes generally match `../../assets/response-templates.md`;
- creative randomness does not appear to destabilize the workflow;
- no release-blocking behavior appeared in the tested cases.

A Phase C pass does **not** mean:

- every famous-song reference is verified;
- every genre file is complete;
- current trend claims can be answered without web verification;
- style/copyright, cultural, or user-privacy boundaries no longer need checking;
- the automated sanity checker replaces musical judgment.

## Non-blocking observations for release hardening

| Observation | Response |
|---|---|
| Randomness was acceptable but should remain bounded. | Add and route `../creative-workflows/answer-calibration.md` for response length, option count, and variation control. |
| Phase C pass should be visible to release maintainers. | Keep this file as the Phase C result record and link it from readiness / roadmap documents. |
| First release needs explicit caveats. | Add `../../KNOWN-LIMITATIONS.md` and require it in RC packaging. |
| RC packaging should avoid vague “ready” claims. | Add `rc1-packaging-checklist.md` for artifact hygiene and factual validation statements. |

## Regression watchlist after Phase C

If future edits change routing, templates, or workflow behavior, rerun at least these smoke-test groups:

1. core generation prompts 1–5 in `prompt-smoke-tests.md`,
2. workflow prompts 6–9,
3. research/currentness prompts 10–13,
4. style/originality prompts 14–16,
5. cultural/regional prompts 17–19,
6. instrument-idiom prompts 20–23,
7. teaching prompts 24–25.

## Release decision implication

The Phase C pass clears the path to **Phase D — release candidate packaging**, provided that:

- the sanity checker still passes,
- Phase B corrections remain intact,
- known limitations are shipped or included in the validation report,
- no new large content expansion is added without corresponding navigation and validation updates.

## Cross-references

- Prompt smoke tests → `prompt-smoke-tests.md`
- First-release readiness → `first-release-readiness.md`
- Phase B correctness pass → `phase-b-correctness-pass.md`
- RC1 packaging checklist → `rc1-packaging-checklist.md`
- Release roadmap → `../../RELEASE-ROADMAP.md`
- Known limitations → `../../KNOWN-LIMITATIONS.md`
- Answer calibration → `../creative-workflows/answer-calibration.md`
- Response templates → `../../assets/response-templates.md`

#!/usr/bin/env python3
# SPDX-License-Identifier: MIT
"""Regression and release-readiness checks for the music-composition skill.

This catches broken internal Markdown references, known high-risk music-theory
regressions, snapshot-note gaps, and a small set of parser-like chord / scale
spelling sanity checks. It does not replace musical review, prompt smoke tests,
or current web research.
"""

from __future__ import annotations

import os
import re
import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parents[1]

MD_PATH_RE = re.compile(r"(?<![\w./-])(?:\.\.?/)?[A-Za-z0-9_./#\-]+\.md(?:#[A-Za-z0-9_./#\-]+)?")

RISKY_PATTERNS = [
    (re.compile(r"\bexotic\b", re.I), "Avoid `exotic`; name the musical variable or cultural context."),
    (re.compile(r"foreign menace", re.I), "Avoid threat/othering cliché language."),
    (re.compile(r"exotic east", re.I), "Avoid vague/orientalist shorthand."),
    (re.compile(r"\btoemsae\b", re.I), "Use `toeseong` for 퇴성 unless a specific source says otherwise."),
    (re.compile(r"\bchangjo\b", re.I), "For pansori structure, prefer chang/sori, aniri, and balim."),
    (re.compile(r"C\(no5\)[^\n]{0,120}power chord", re.I), "C(no5) is not a power chord; C5 or C(no3) is."),
    (re.compile(r"E7alt[^\n]{0,160}A♯[^\n]{0,160}♯9", re.I), "In E7alt, A♯ is ♯11/♭5; G natural is ♯9."),
    (re.compile(r"Western music since roughly 1700 uses 12-TET", re.I), "12-TET history claim is too broad."),
    (re.compile(r"maqam[^\n]{0,80}24-TET subset", re.I), "Do not reduce maqam/makam to a 24-TET subset."),
    (re.compile(r"Spanish, Middle Eastern, Klezmer flavor", re.I), "Avoid broad culture-flavor shorthand for Phrygian dominant."),
    (re.compile(r"World music color", re.I), "Use tradition-specific instrument color and idiom, not `world music color`."),
    (re.compile(r"\bY flavor\b", re.I), "Prefer `Y influence` or a concrete musical variable over flavor shorthand."),
    (re.compile(r"ethnic flavor", re.I), "Avoid vague culture-as-flavor phrasing; name the specific tradition or musical variable."),
    (re.compile(r"\bethnic\b", re.I), "Avoid vague `ethnic` instrumentation labels; name the tradition, region, or percussion family."),
    (re.compile(r"Asian flavor", re.I), "Avoid continent-scale style shorthand; name the market, tradition, or craft variable."),
    (re.compile(r"world flavor", re.I), "Avoid vague `world` color phrasing; name the scene or idiom."),
    (re.compile(r"sounds foreign", re.I), "Avoid othering language; describe unfamiliarity or non-idiomatic fit precisely."),
    (re.compile(r"Jinyangjo[^\n]{0,80}18 beats", re.I), "Jinyangjo should not be reduced to an 18-beat cycle; use 6-beat / 24-beat phrase caveat."),
    (re.compile(r"Diminished \(half-whole\)[^\n]{0,160}E♭ E", re.I), "Half-whole diminished over C7 should spell the altered ninth as D♯, not E♭."),
    (re.compile(r"Whole[- ]tone[^\n]{0,160}C D E F♯ G♯ A♯", re.I), "Whole-tone dominant spelling should use B♭ for ♭7, with A♯ only as enharmonic caveat."),
]

SNAPSHOT_REQUIRED_DIRS = [
    ROOT / "references" / "genres",
    ROOT / "references" / "research",
]
SNAPSHOT_REQUIRED_FILES = [
    ROOT / "references" / "production-aware" / "pre-production-decisions.md",
    ROOT / "references" / "techniques" / "algorithmic-and-AI-assisted.md",
]

EXPECTED_INTERVAL_LINES = [
    "| Augmented 5th (A5) | 8 | d4 | C–G♯ |",
    "| Minor 6th (m6) | 8 | M3 | C–A♭ |",
]

EXPECTED_MODE_LINES = [
    "| **Ionian** | C D E F G A B | `W W H W W W H` |",
    "| **Lydian** | C D E F♯ G A B | `W W W H W W H` |",
    "| **Mixolydian** | C D E F G A B♭ | `W W H W W H W` |",
    "| **Aeolian** | C D E♭ F G A♭ B♭ | `W H W W H W W` |",
]

EXPECTED_FILES = [
    "RELEASE-ROADMAP.md",
    "KNOWN-LIMITATIONS.md",
    "RELEASE-NOTES-v1.0.md",
    "VALIDATION-v1.0.md",
    "RC1-v1.0-decision-memo-2026-04-27.md",
    "references/validation/first-release-readiness.md",
    "references/validation/prompt-smoke-tests.md",
    "references/validation/phase-b-correctness-pass.md",
    "references/validation/phase-c-smoke-test-results.md",
    "references/validation/rc1-packaging-checklist.md",
    "references/genres/country-americana.md",
    "references/genres/metal-punk-hardcore.md",
    "references/genres/gospel-and-ccm.md",
    "references/genres/brazilian-pop-and-funk.md",
    "references/instrument-idiom/overview.md",
    "references/instrument-idiom/piano-keyboards.md",
    "references/instrument-idiom/guitar.md",
    "references/instrument-idiom/bass.md",
    "references/instrument-idiom/drums-percussion.md",
    "references/instrument-idiom/strings.md",
    "references/instrument-idiom/winds.md",
    "references/instrument-idiom/brass.md",
    "references/instrument-idiom/vocals.md",
    "references/source-bibliography.md",
    "references/research/user-listening-context-and-streaming-services.md",
    "references/research/regional-trend-evolution-analysis.md",
    "references/genres/minor-and-hybrid-genres.md",
    "references/genres/regional-scene-starters.md",
    "references/genres/folk-roots-and-traditions.md",
    "references/creative-workflows/musical-brainstorming.md",
    "references/creative-workflows/user-agent-collaboration.md",
    "references/creative-workflows/revision-and-feedback-loops.md",
    "references/creative-workflows/answer-calibration.md",
    "assets/trend-and-reference-matrices.md",
    "assets/musical-brainstorming-cards.md",
    "assets/session-brief-and-decision-log.md",
    "assets/web-search-cheatsheet.md",
    "assets/chord-symbol-ambiguity-and-parsing.md",
    "assets/scale-degree-spelling-cheatsheet.md",
    "assets/music-theory-audit-rubric.md",
]

BASE_PITCH_CLASS = {"C": 0, "D": 2, "E": 4, "F": 5, "G": 7, "A": 9, "B": 11}
SEMITONE_TO_COMMON_TENSIONS = {
    0: {"1", "root"},
    1: {"♭9", "b9"},
    2: {"9", "M9"},
    3: {"♯9", "#9", "m3"},
    4: {"3", "M3"},
    5: {"11", "4"},
    6: {"♯11", "#11", "♭5", "b5"},
    7: {"5"},
    8: {"♯5", "#5", "♭13", "b13", "m6"},
    9: {"13", "6"},
    10: {"♭7", "b7"},
    11: {"7", "M7"},
}

# Parser-like C-root chord sanity cases. These are not exhaustive; they are
# high-use examples that should never regress.
EXPECTED_C_CHORD_PCS = {
    "C": {0, 4, 7},
    "Cm": {0, 3, 7},
    "Cdim": {0, 3, 6},
    "Caug": {0, 4, 8},
    "Csus4": {0, 5, 7},
    "Csus2": {0, 2, 7},
    "C5": {0, 7},
    "C(no3)": {0, 7},
    "C(no5)": {0, 4},
    "C7": {0, 4, 7, 10},
    "Cmaj7": {0, 4, 7, 11},
    "Cm7": {0, 3, 7, 10},
    "Cm7♭5": {0, 3, 6, 10},
    "Cø7": {0, 3, 6, 10},
    "Cdim7": {0, 3, 6, 9},
    "CmMaj7": {0, 3, 7, 11},
    "Cmaj7♯11": {0, 4, 7, 11, 6},
    "C7♭9": {0, 4, 7, 10, 1},
    "C7♯9": {0, 4, 7, 10, 3},
    "C7♭13": {0, 4, 7, 10, 8},
    "C13": {0, 4, 7, 10, 2, 9},
    "Cadd9": {0, 4, 7, 2},
    "C6/9": {0, 4, 7, 9, 2},
}

# Notes in C-root examples from the new parsing asset.
EXPECTED_C_CHORD_NOTES = {
    "C": ["C", "E", "G"],
    "Cm": ["C", "Eb", "G"],
    "Cdim": ["C", "Eb", "Gb"],
    "Caug": ["C", "E", "G#"],
    "Csus4": ["C", "F", "G"],
    "Csus2": ["C", "D", "G"],
    "C5": ["C", "G"],
    "C(no3)": ["C", "G"],
    "C(no5)": ["C", "E"],
    "C7": ["C", "E", "G", "Bb"],
    "Cmaj7": ["C", "E", "G", "B"],
    "Cm7": ["C", "Eb", "G", "Bb"],
    "Cm7♭5": ["C", "Eb", "Gb", "Bb"],
    "Cø7": ["C", "Eb", "Gb", "Bb"],
    "Cdim7": ["C", "Eb", "Gb", "Bbb"],
    "CmMaj7": ["C", "Eb", "G", "B"],
    "Cmaj7♯11": ["C", "E", "G", "B", "F#"],
    "C7♭9": ["C", "E", "G", "Bb", "Db"],
    "C7♯9": ["C", "E", "G", "Bb", "D#"],
    "C7♭13": ["C", "E", "G", "Bb", "Ab"],
    "C13": ["C", "E", "G", "Bb", "D", "A"],
    "Cadd9": ["C", "E", "G", "D"],
    "C6/9": ["C", "E", "G", "A", "D"],
}

EXPECTED_SCALE_SPELLING_LINES = [
    "| `♭II` | D♭ | D♭ F A♭ | lowered supertonic, not C♯ major |",
    "| `V7/vi` | E7 | E G♯ B D | Am |",
    "| C Lydian | ♯4 | C D E F♯ G A B |",
    "| ♯9 | G | raised ninth above E; enharmonic to minor third, but not labeled ♭3 in dominant context |",
]

IGNORED_EXAMPLE_REFS = {
    "old-filename.md",
    "path/file.md",
    "references/foo/bar.md",
    "assets/foo.md",
    "../workflow.md",
    "folk-and-world.md",
    "pop-rock.md",
}


def normalize_note_name(note: str) -> str:
    return note.strip().replace("♯", "#").replace("♭", "b").replace("𝄫", "bb").replace("𝄪", "##")


def note_pc(note: str) -> int:
    note = normalize_note_name(note)
    if not note:
        raise ValueError("empty note")
    letter = note[0].upper()
    if letter not in BASE_PITCH_CLASS:
        raise ValueError(f"bad note letter: {note}")
    pc = BASE_PITCH_CLASS[letter]
    rest = note[1:]
    while rest:
        if rest.startswith("bb"):
            pc -= 2
            rest = rest[2:]
        elif rest.startswith("##"):
            pc += 2
            rest = rest[2:]
        elif rest.startswith("b"):
            pc -= 1
            rest = rest[1:]
        elif rest.startswith("#"):
            pc += 1
            rest = rest[1:]
        else:
            raise ValueError(f"bad accidental in note: {note}")
    return pc % 12


def interval_semitones(root: str, note: str) -> int:
    return (note_pc(note) - note_pc(root)) % 12


def tension_label_is_plausible(root: str, note: str, label: str) -> bool:
    labels = SEMITONE_TO_COMMON_TENSIONS[interval_semitones(root, note)]
    normalized_label = label.replace("#", "♯").replace("b", "♭")
    return label in labels or normalized_label in labels


def strip_code_blocks(text: str) -> str:
    return re.sub(r"```.*?```", "", text, flags=re.S)


def strip_urls(text: str) -> str:
    """Remove http(s) URLs so URL-encoded fragments like `%20SKILL.md`
    aren't mis-parsed as file references."""
    return re.sub(r"https?://[^\s)]+", "", text)


def strip_anchor(path: str) -> str:
    return path.split("#", 1)[0]


def resolve_md_ref(source: Path, ref: str) -> Path | None:
    ref = strip_anchor(ref)
    if ref.startswith(("http://", "https://", "mailto:")):
        return None
    if "<" in ref or ">" in ref:
        return None

    candidates: list[Path] = []
    raw = Path(ref)

    if ref.startswith("./") or ref.startswith("../"):
        candidates.append((source.parent / raw).resolve())
        stripped = ref
        while stripped.startswith("../"):
            stripped = stripped[3:]
        if stripped.startswith(("assets/", "references/", "scripts/")):
            candidates.append((ROOT / stripped).resolve())
    elif ref.startswith(("references/", "assets/", "scripts/")):
        candidates.append((ROOT / raw).resolve())
    else:
        candidates.extend([
            (source.parent / raw).resolve(),
            (ROOT / raw).resolve(),
            (ROOT / "references" / raw).resolve(),
            (ROOT / "assets" / raw).resolve(),
        ])

    for candidate in candidates:
        if candidate.exists() and candidate.is_file():
            return candidate
    return candidates[0] if candidates else None


EXCLUDE_DIRS = {
    ".git",
    "__pycache__",
    # virtualenv directories — third-party package READMEs/prompts shouldn't be linted
    ".env",
    ".venv",
    "venv",
    "env",
    "node_modules",
}


def iter_markdown_files() -> list[Path]:
    files: list[Path] = []
    for dirpath, dirnames, filenames in os.walk(ROOT):
        dirnames[:] = [d for d in dirnames if d not in EXCLUDE_DIRS]
        for filename in filenames:
            if filename.endswith(".md"):
                files.append(Path(dirpath) / filename)
    return sorted(files)


def find_md_refs(text: str) -> set[str]:
    refs = set()
    for match in MD_PATH_RE.finditer(text):
        ref = match.group(0)
        if ref.startswith("http"):
            continue
        refs.add(ref)
    return refs


def check_links() -> list[str]:
    errors: list[str] = []
    for md in iter_markdown_files():
        text = strip_urls(strip_code_blocks(md.read_text(encoding="utf-8")))
        for ref in sorted(find_md_refs(text)):
            if ref in IGNORED_EXAMPLE_REFS:
                continue
            resolved = resolve_md_ref(md, ref)
            if resolved is None:
                continue
            if not resolved.exists():
                errors.append(f"Broken md ref: {md.relative_to(ROOT)} -> {ref}")
    return errors


def check_risky_terms() -> list[str]:
    errors: list[str] = []
    skip_files = {"MAINTENANCE.md", "scripts/music_theory_sanity_check.py"}
    for md in iter_markdown_files():
        rel = str(md.relative_to(ROOT))
        if rel in skip_files:
            continue
        lines = strip_code_blocks(md.read_text(encoding="utf-8")).splitlines()
        for lineno, line in enumerate(lines, start=1):
            low = line.lower()
            if any(token in low for token in ("avoid", "do not", "don't", "never", "replace", "removed")) or re.search(r"\bnot\b", low):
                continue
            for pattern, message in RISKY_PATTERNS:
                if pattern.search(line):
                    errors.append(f"Risky pattern in {rel}:{lineno}: {message}")
    return errors


def check_snapshots() -> list[str]:
    errors: list[str] = []
    files: list[Path] = []
    for directory in SNAPSHOT_REQUIRED_DIRS:
        if directory.exists():
            files.extend(sorted(directory.glob("*.md")))
    files.extend([p for p in SNAPSHOT_REQUIRED_FILES if p.exists()])

    for p in files:
        text = p.read_text(encoding="utf-8")
        if not any("Snapshot note" in line for line in text.splitlines()[:8]):
            errors.append(f"Missing snapshot note near top: {p.relative_to(ROOT)}")
    return errors


def check_interval_cheatsheet() -> list[str]:
    errors: list[str] = []
    p = ROOT / "assets" / "intervals-and-scale-formulas.md"
    if not p.exists():
        return ["Missing assets/intervals-and-scale-formulas.md"]
    text = p.read_text(encoding="utf-8")
    for line in EXPECTED_INTERVAL_LINES:
        if line not in text:
            errors.append(f"Expected interval inversion line not found: {line}")
    return errors


def check_expected_files() -> list[str]:
    errors: list[str] = []
    for rel in EXPECTED_FILES:
        if not (ROOT / rel).is_file():
            errors.append(f"Expected expansion file missing: {rel}")
    return errors


def check_modes_cheatsheet() -> list[str]:
    errors: list[str] = []
    p = ROOT / "assets" / "modes-cheatsheet.md"
    if not p.exists():
        return ["Missing assets/modes-cheatsheet.md"]
    text = p.read_text(encoding="utf-8")
    for line in EXPECTED_MODE_LINES:
        if line not in text:
            errors.append(f"Expected parallel mode line not found: {line}")
    if "F-Em-Am-Bb" in text or "F–Em–Am–Bb" in text:
        errors.append("Lydian example regression: F-Em-Am-Bb contains B♭, not F Lydian's ♯4.")
    return errors



def check_symmetric_scale_spelling() -> list[str]:
    errors: list[str] = []
    expected_lines = {
        "assets/intervals-and-scale-formulas.md": [
            "| Whole tone | 1-2-3-♯4-♯5-♭7 (6 notes) | C D E F♯ G♯ B♭ |",
            "| Diminished (half-whole) | 1-♭2-♯2-3-♯4-5-6-♭7 (8 notes) | C D♭ D♯ E F♯ G A B♭ |",
            "| Altered (super Locrian) | ^7 | 1-♭9-♯9-3-♭5/♯11-♭13-♭7 in dominant-function spelling | C D♭ D♯ E G♭/F♯ A♭ B♭ |",
        ],
        "assets/modes-cheatsheet.md": [
            "| Whole tone | `W W W W W W` (6 notes) | C D E F♯ G♯ B♭ | Over 7♯5 / whole-tone dominant colors; B♭ is the dominant spelling of the enharmonic A♯ |",
            "| Diminished (half-whole) | `H W H W H W H W` (8 notes) | C D♭ D♯ E F♯ G A B♭ | Over dominant ♭9 chords; D♯ is ♯9 over C7 |",
            "| **Altered (super Locrian)** | ^7 | C D♭ D♯ E G♭/F♯ A♭ B♭ | **C7alt** (resolving dominants; functional spelling) |",
        ],
        "references/fundamentals/pitch-intervals-scales.md": [
            "| Whole-tone | `W W W W W W` (6 notes) | C D E F♯ G♯ B♭ | Over 7♯5 / whole-tone dominant colors; B♭ is the dominant spelling of the enharmonic A♯ |",
            "| Diminished (half-whole) | `H W H W H W H W` (8 notes) | C D♭ D♯ E F♯ G A B♭ | Over dominant ♭9 chords; D♯ is ♯9 over C7 |",
        ],
    }
    for rel, needles in expected_lines.items():
        path = ROOT / rel
        if not path.exists():
            errors.append(f"Missing symmetric-scale file: {rel}")
            continue
        text = path.read_text(encoding="utf-8")
        for needle in needles:
            if needle not in text:
                errors.append(f"Expected symmetric-scale spelling line missing in {rel}: {needle}")
    return errors


def check_chord_tension_math() -> list[str]:
    errors: list[str] = []

    expected = [
        ("E", "A#", "♯11"),
        ("E", "G", "♯9"),
        ("E", "F", "♭9"),
        ("E", "C", "♭13"),
        ("C", "F#", "♯11"),
        ("C", "Db", "♭9"),
        ("C", "D#", "♯9"),
    ]
    for root, note, label in expected:
        if not tension_label_is_plausible(root, note, label):
            errors.append(f"Internal tension mapping failed: {root} to {note} should allow {label}.")

    jazz_voicings = ROOT / "assets" / "jazz-voicings.md"
    if jazz_voicings.exists():
        text = jazz_voicings.read_text(encoding="utf-8")
        if "| E7alt | G♯ - D - F - C (3, ♭7, ♭9, ♭13) |" not in text:
            errors.append("Expected corrected E7alt voicing line not found in assets/jazz-voicings.md")
        if "| E7♭9♯9 | G♯ - D - F - G (3, ♭7, ♭9, ♯9; altered-color subset) |" not in text:
            errors.append("Expected corrected E7♭9♯9 voicing line not found in assets/jazz-voicings.md")

    return errors


def check_chord_formula_spelling() -> list[str]:
    """Validate canonical chord-symbol spellings against pitch-class math."""
    errors: list[str] = []
    for symbol, expected_pcs in EXPECTED_C_CHORD_PCS.items():
        notes = EXPECTED_C_CHORD_NOTES[symbol]
        pcs = {note_pc(n) for n in notes}
        if pcs != expected_pcs:
            errors.append(f"Internal chord parser case failed: {symbol} -> {notes} -> {sorted(pcs)} expected {sorted(expected_pcs)}")

    chord_symbols = ROOT / "assets" / "chord-symbol-conventions.md"
    if chord_symbols.exists():
        text = chord_symbols.read_text(encoding="utf-8")
        required = [
            "| `C5` or `C(no3)` | C power chord (root + 5th, no 3rd) |",
            "`C(no5)` means the fifth is omitted, so it is **not** a power chord.",
            "| Dominant 7 sus4 | `C7sus` or `C7sus4` | `C7sus4` | C F G B♭ |",
        ]
        for needle in required:
            if needle not in text:
                errors.append(f"Expected chord-symbol convention line missing: {needle}")
    return errors


def check_chord_symbol_ambiguity_asset() -> list[str]:
    errors: list[str] = []
    p = ROOT / "assets" / "chord-symbol-ambiguity-and-parsing.md"
    if not p.exists():
        return ["Missing assets/chord-symbol-ambiguity-and-parsing.md"]
    text = p.read_text(encoding="utf-8")
    required = [
        "| `C(no5)` | omit fifth | C E |",
        "| `C9` | Dominant 9 unless otherwise marked | includes B♭, not B |",
        "| `C2` | Csus2 or Cadd9/add2 | likely Csus2 in guitar/pop charts | voicing or third matters |",
        "| ♯9 | D♯ | enharmonic to E♭ but functions as raised 9 over major 3 E |",
    ]
    for needle in required:
        if needle not in text:
            errors.append(f"Expected chord ambiguity/parsing line missing: {needle}")
    return errors


def check_scale_degree_spelling_asset() -> list[str]:
    errors: list[str] = []
    p = ROOT / "assets" / "scale-degree-spelling-cheatsheet.md"
    if not p.exists():
        return ["Missing assets/scale-degree-spelling-cheatsheet.md"]
    text = p.read_text(encoding="utf-8")
    for needle in EXPECTED_SCALE_SPELLING_LINES:
        if needle not in text:
            errors.append(f"Expected scale-degree spelling line missing: {needle}")
    bad_lines = [
        "C Lydian as C D E G♭ G A B",
        "A♯ as ♯9 over E",
    ]
    for bad in bad_lines:
        # These are allowed only in explicit warning lines that include do-not wording.
        for lineno, line in enumerate(text.splitlines(), start=1):
            if bad in line and not any(token in line.lower() for token in ("do not", "not ", "avoid")):
                errors.append(f"Scale-degree spelling warning phrase appears without warning context at line {lineno}: {bad}")
    return errors


def check_release_docs() -> list[str]:
    errors: list[str] = []
    roadmap = ROOT / "RELEASE-ROADMAP.md"
    limitations = ROOT / "KNOWN-LIMITATIONS.md"
    release_notes_v1 = ROOT / "RELEASE-NOTES-v1.0.md"
    validation_v1 = ROOT / "VALIDATION-v1.0.md"
    decision_memo = ROOT / "RC1-v1.0-decision-memo-2026-04-27.md"
    readiness = ROOT / "references" / "validation" / "first-release-readiness.md"
    smoke = ROOT / "references" / "validation" / "prompt-smoke-tests.md"
    phase_c = ROOT / "references" / "validation" / "phase-c-smoke-test-results.md"
    rc1 = ROOT / "references" / "validation" / "rc1-packaging-checklist.md"
    calibration = ROOT / "references" / "creative-workflows" / "answer-calibration.md"
    for p in (roadmap, limitations, release_notes_v1, validation_v1, decision_memo, readiness, smoke, phase_c, rc1, calibration):
        if not p.exists():
            errors.append(f"Missing release validation doc: {p.relative_to(ROOT)}")
    if roadmap.exists():
        text = roadmap.read_text(encoding="utf-8")
        for phrase in ("Minimum viable first release", "prompt smoke tests", "What not to build before first release", "v1.0 finalization", "KNOWN-LIMITATIONS.md"):
            if phrase not in text:
                errors.append(f"Release roadmap missing expected section phrase: {phrase}")
    if smoke.exists():
        text = smoke.read_text(encoding="utf-8")
        if text.count("| # | Prompt | Expected routing | Expected answer shape |") < 5:
            errors.append("Prompt smoke tests should cover multiple prompt categories.")
    if phase_c.exists():
        text = phase_c.read_text(encoding="utf-8")
        for phrase in ("Pass, based on user-run smoke testing", "No blocker fixes required", "Phase D"):
            if phrase not in text:
                errors.append(f"Phase C result doc missing expected phrase: {phrase}")
    if rc1.exists():
        text = rc1.read_text(encoding="utf-8")
        for phrase in ("Required inputs", "Packaging steps", "Release-hold conditions"):
            if phrase not in text:
                errors.append(f"RC1 checklist missing expected phrase: {phrase}")
    if limitations.exists():
        text = limitations.read_text(encoding="utf-8")
        for phrase in ("Known limitations", "Current trends", "Copyright / style", "Automated validation"):
            if phrase not in text:
                errors.append(f"Known limitations doc missing expected phrase: {phrase}")
    if release_notes_v1.exists():
        text = release_notes_v1.read_text(encoding="utf-8")
        for phrase in ("Release Notes — v1.0", "What changed from RC1", "How to validate locally", "Post-release direction"):
            if phrase not in text:
                errors.append(f"v1.0 release notes missing expected phrase: {phrase}")
    if validation_v1.exists():
        text = validation_v1.read_text(encoding="utf-8")
        for phrase in ("Validation Report — v1.0", "Ship v1.0", "Snapshot-note files", "Phase C prompt smoke tests"):
            if phrase not in text:
                errors.append(f"v1.0 validation report missing expected phrase: {phrase}")
    if decision_memo.exists():
        text = decision_memo.read_text(encoding="utf-8")
        for phrase in ("Promote RC1 to v1.0", "Why RC2 is not recommended now", "RC2 trigger conditions"):
            if phrase not in text:
                errors.append(f"RC1 → v1.0 decision memo missing expected phrase: {phrase}")
    if calibration.exists():
        text = calibration.read_text(encoding="utf-8")
        for phrase in ("Controlled variation bands", "Default answer budgets", "Loading calibration"):
            if phrase not in text:
                errors.append(f"Answer calibration doc missing expected phrase: {phrase}")
    return errors


def main() -> int:
    checks = [
        ("internal markdown links", check_links),
        ("risky terms / known regressions", check_risky_terms),
        ("snapshot notes", check_snapshots),
        ("interval cheatsheet", check_interval_cheatsheet),
        ("expected expansion files", check_expected_files),
        ("modes cheatsheet", check_modes_cheatsheet),
        ("symmetric scale spelling", check_symmetric_scale_spelling),
        ("chord / tension spelling", check_chord_tension_math),
        ("chord formula spelling", check_chord_formula_spelling),
        ("chord-symbol ambiguity asset", check_chord_symbol_ambiguity_asset),
        ("scale-degree spelling asset", check_scale_degree_spelling_asset),
        ("release docs", check_release_docs),
    ]

    all_errors: list[str] = []
    for name, fn in checks:
        errors = fn()
        if errors:
            all_errors.append(f"\n[{name}]")
            all_errors.extend(errors)

    if all_errors:
        print("music_theory_sanity_check: FAIL")
        print("\n".join(all_errors))
        return 1

    md_count = len(iter_markdown_files())
    print("music_theory_sanity_check: PASS")
    print(f"Checked {md_count} markdown files.")
    return 0


if __name__ == "__main__":
    sys.exit(main())

# RC1 Packaging Checklist

Use this checklist after Phase B and Phase C are complete and before publishing a first release candidate. It is a maintainer-facing release artifact checklist, not a runtime composition guide.

## Required inputs

| Input | Required status |
|---|---|
| `../../RELEASE-ROADMAP.md` | Phase D criteria current |
| `first-release-readiness.md` | No open release blockers |
| `phase-b-correctness-pass.md` | Completed or updated after latest corrections |
| `phase-c-smoke-test-results.md` | Phase C result recorded |
| `../../KNOWN-LIMITATIONS.md` | Present and linked |
| `../../scripts/music_theory_sanity_check.py` | Run after final edits |

## Packaging steps

1. Start from the latest clean working tree.
2. Run `scripts/music_theory_sanity_check.py` from the repository root.
3. Confirm internal Markdown links pass.
4. Confirm snapshot-note coverage for current-sensitive files.
5. Confirm there are no cache, scratch, temp, `.pyc`, `.DS_Store`, or obsolete generated files.
6. Create the release ZIP.
7. Create a unified diff against the previous patch or release candidate.
8. Create a validation report using the template in `first-release-readiness.md`.
9. Include known limitations in the validation report or ship `../../KNOWN-LIMITATIONS.md` with the package.
10. Open the ZIP listing and confirm expected files are present.

## RC validation report minimum

A release-candidate validation report should include:

```md
# Validation Report — RC1

## Package
- Package:
- Base version:
- Date:

## Automated checks
- Sanity checker:
- Internal links:
- Snapshot notes:
- Known regression checks:
- ZIP integrity:

## Manual / user-run checks
- Phase B correctness:
- Phase C prompt smoke tests:
- High-use assets:
- Research/currentness/privacy:
- Cultural specificity:

## Known limitations
- See KNOWN-LIMITATIONS.md

## Release decision
- Ship / hold / ship with caveats
```

## Release-hold conditions

Hold RC1 if any of these are true:

- the sanity checker fails;
- a known P0/P1 music-theory regression reappears;
- a new major file is not routed from `../00-navigation.md`;
- `../../SKILL.md` misstates the folder structure;
- current-sensitive files lack snapshot notes;
- Phase C is not recorded or is marked partial without explanation;
- the ZIP contains cache/scratch/temp files;
- the validation report claims a check passed without actually running or receiving that check.

## Useful commands

From the repository root:

```bash
python3 scripts/music_theory_sanity_check.py
find . -name '__pycache__' -o -name '*.pyc' -o -name '.DS_Store'
find . -type f | sort
```

For a ZIP integrity check:

```bash
unzip -t <release-package>.zip
```

## Cross-references

- Release roadmap → `../../RELEASE-ROADMAP.md`
- Known limitations → `../../KNOWN-LIMITATIONS.md`
- First-release readiness → `first-release-readiness.md`
- Phase B correctness pass → `phase-b-correctness-pass.md`
- Phase C results → `phase-c-smoke-test-results.md`
- Prompt smoke tests → `prompt-smoke-tests.md`
- Music-theory audit rubric → `../../assets/music-theory-audit-rubric.md`

# Workflow

This file is about how the agent should approach composition advice — what questions to ask, in what order, and how to deliver useful guidance. It's the meta-file: not about music itself, but about how to help with music. It complements `00-navigation.md` (which handles file routing) by providing the diagnostic and advisory framework.

For navigating between files, see `00-navigation.md`. For analytical methodology (understanding existing music), see `analysis.md`. Other files cover the actual composition content.

## When to consult this file

- Starting any composition advice session
- The user's question is vague or under-specified
- Multiple possible directions exist; need to narrow down
- The user wants iterative feedback on a work in progress
- Diagnosing what's wrong with a piece
- Pedagogical contexts (teaching composition)
- Whenever the agent needs a sanity check on its approach

## The agent's role in composition

The agent is not the composer. The user is. The agent is a knowledgeable assistant who:

1. **Asks clarifying questions** before assuming.
2. **Provides options, not commands** — composition has no single right answer.
3. **Connects principles to the user's specific situation** rather than reciting general theory.
4. **Defers to the user's aesthetic judgment** while flagging potential issues.
5. **Encourages experimentation** — "try X, see if it works for you" rather than "do X".

Avoid:

- Issuing rules as commandments
- Dismissing user choices as "wrong" without explanation
- Imposing one stylistic framework (e.g., common-practice tonality) on music that's clearly in a different idiom
- Assuming the user wants what the textbook prescribes

## The diagnostic phase

Before giving advice, understand the situation. Common dimensions to clarify:

### Genre and style

What kind of music is the user writing?

- Period (classical era? Romantic? modern?)
- Genre (pop, rock, hip-hop, jazz, classical, film, etc.)
- Sub-style within genre (modal jazz vs. bebop vs. fusion)
- Reference points (specific composers / songs they're drawing on)

If the user hasn't specified, *ask* — don't assume. The advice differs enormously between, say, "a slow R&B ballad" and "a string quartet movement".

### Multi-genre and hybrid requests

Many real-world requests cross genres: "a K-pop song with jazz harmony", "a synth-pop track with classical counterpoint in the bridge", "a film cue that sounds like A Tribe Called Quest". These aren't confused — they're specific creative briefs. The agent's job is to identify which genre is the *primary frame* (the one whose conventions the listener will hear as the "style") and which is providing a **secondary influence** (a coloring or craft layer applied within the primary frame).

Diagnostic questions for hybrid requests:

| Dimension | What it usually anchors to |
|-----------|----------------------------|
| Form / song structure | Primary genre (a "K-pop song" stays K-pop in form even with jazz chords) |
| Production / sonic palette | Primary genre (drums, mix, instrumentation say "this is a hip-hop track") |
| Harmonic vocabulary | Often the secondary influence layer (jazz extensions over a pop progression, modal mixture in a rock chorus) |
| Rhythmic feel / groove | Often primary, occasionally flavor (Latin clave inside an R&B groove is flavor; Latin clave anchoring the whole tune is primary) |
| Melodic / vocal style | Can go either way; ask if unclear |

A useful frame: **the primary genre tells you what the listener will *call* the song; the secondary influence layer tells you what makes the song stand out within that category.** Stevie Wonder's "Sir Duke" is funk/soul (primary) with jazz harmonic vocabulary (secondary influence). Steely Dan tracks are pop/rock (primary, in form and length) with jazz harmony and bebop solo language (secondary influence). BTS's "Black Swan" is K-pop (primary) with cinematic / orchestral elements (secondary influence).

When the user names two genres without specifying which is which, ask:

- "Which genre's *form* are you working in?" — answers the primary genre most of the time
- "Which one is the listener supposed to identify the song as?" — same
- "What's the reference track — what existing song captures what you want?" — usually clarifies both layers at once

When the user is *deliberately* trying to blur the line (genre-bending, fusion, post-genre), the agent should drop the "primary + flavor" framing and treat each dimension independently. That's a different mode of advising — closer to "compose what you want, evaluate by the result" than to applying any genre's conventions.

### What stage they're at

- **Idea phase**: brainstorming, no notes on paper
- **Sketch phase**: rough draft, deciding direction
- **Working out details**: revising specific moments
- **Polishing**: making fine adjustments
- **Stuck**: not sure how to proceed
- **Diagnosing**: there's a problem, what is it?

Different stages call for different help. Brainstorming wants generative options; polishing wants targeted critique.

For detailed musical brainstorming workflows, use `creative-workflows/musical-brainstorming.md`. For long multi-turn sessions, use `creative-workflows/user-agent-collaboration.md` and `../assets/session-brief-and-decision-log.md`.

### Specific problem or open exploration

- "I have a verse that doesn't lead well to the chorus" — specific
- "How do I write a good chorus" — generic
- "Help me with my song" — too open

For specific problems, give targeted advice. For open requests, ask what they want to focus on.

### What they have

If the user can share the music (chord chart, melody, audio), engage with the actual material. Generic advice without the music is much weaker.

If they can't share, ask them to describe in detail: key, tempo, form, the parts that are working, the parts that aren't.

### What they want

Not always obvious. Some users want:

- Direct advice ("here's what to do")
- Options to choose from ("here are three approaches")
- Validation ("does this work?")
- Pedagogy ("explain why")
- Critique ("what's wrong with this?")
- Brainstorming ("give me ideas")

Match the response style to the user's apparent need.

## Asking good questions

Before assuming, ask clarifying questions:

| Situation | Questions to ask |
|-----------|------------------|
| User says "I have a song" | What genre? What's the form? What's working? What's stuck? |
| User wants to "learn composition" | What kind of music? What level? What goal? |
| User says "this doesn't sound right" | What specifically sounds wrong? Where in the song? What did you intend? |
| User asks for a chord progression | For what genre? What mood? Major or minor? |
| User asks how to write a melody | Over what chords? For what voice/instrument? In what genre? |
| User is comparing two options | What are you trying to achieve? Which feels closer to that? |

Don't barrage the user with questions; ask only what's needed to give useful advice. One or two clarifying questions are usually enough.

## Diagnosis: when something is "wrong"

Common symptoms and likely causes:

| Symptom | Likely causes |
|---------|---------------|
| "Doesn't go anywhere" | No clear arc; static energy; phrases not building |
| "Sounds boring" | Insufficient contrast (verse vs. chorus too similar); melodic monotony; predictable harmony |
| "Sounds wrong" | Voice-leading errors; non-idiomatic for the genre; prosody mismatch |
| "Feels stuck" | Lack of motion; too much repetition without development; static harmony |
| "Sounds amateur" | Generic / cliché choices; obvious in melody, chord progression, or rhythm |
| "Doesn't groove" | Microtiming issues; rhythmic mismatch between parts; lacking ghost notes or syncopation |
| "Chorus doesn't hit" | Insufficient contrast from verse; weak melodic hook; pre-chorus not building |
| "Sounds muddy" | Too many low / mid-range instruments; voicing too close in low register |
| "Sounds thin" | Not enough mid-range; sparse arrangement |
| "Lyrics don't fit" | Prosody mismatch; vowel choices wrong for the melody |
| "Melody feels off" | Wrong notes against the chord; awkward leaps; no climax; range issues |

For each diagnosis, the relevant file in this skill provides specific fixes:

- Energy / motion → `form/narrative-and-transitions.md`
- Voice leading → `harmony/voice-leading.md`
- Melodic problems → `melody/melodic-construction.md`
- Phrase issues → `melody/phrase-structure.md`
- Hook weakness → `songwriting/hooks-and-memorability.md`
- Texture problems → `orchestration/voicing-and-texture.md`, `orchestration/arrangement-density.md`
- Mix-ready problems → `production-aware/arrangement-for-mix.md`
- Genre conventions → relevant `genres/*.md`

## Workflow templates for common scenarios

### Scenario 1: User wants to start a new song

1. **Clarify the goal**: What genre? What feeling? Reference points?
2. **Suggest a starting point**: chord progression, lyric concept, melodic motif, or groove.
3. **Provide a skeleton**: 4 bars of chords, or a 4-line lyric prompt, or a tempo + drum pattern.
4. **Encourage iteration**: "try this; see if it sparks something".

### Scenario 2: User has a verse and wants help with the chorus

1. **Get the verse**: chords, melody, lyric, key, tempo.
2. **Identify the verse's tessitura and energy**: where is the chorus heading from?
3. **Suggest contrast**: higher tessitura, denser arrangement, hook-forward melody, possible modulation.
4. **Provide specific options**: "try these three chord changes for the chorus; here's a possible hook melody".

### Scenario 3: User has a song that "doesn't feel right"

1. **Get the song**: as much detail as possible.
2. **Listen / read carefully**: what's the actual problem?
3. **Diagnose**: which dimension(s) are off? Energy? Melody? Form? Voice leading? Production?
4. **Prioritize**: usually 1–2 issues are the main problem; fix those first.
5. **Provide targeted suggestions**: specific changes, not general advice.

### Scenario 4: User wants to learn a new style

1. **Identify the style** they want to learn.
2. **Recommend reference listening**: specific pieces that exemplify the style.
3. **Identify the style's characteristic features**: harmony, rhythm, form, etc. (using genre files).
4. **Suggest exercises**: write a piece using these specific techniques; analyze a sample piece; etc.
5. **Encourage gradual progress**: master one feature at a time.

### Scenario 5: User is stuck in the middle of a piece

1. **Get the section so far**: chords, melody, what's happening.
2. **Identify what's needed next**: build? release? contrast? return?
3. **Suggest specific moves**: "after this section, try a modulation to vi"; "here's a possible bridge".
4. **Test and iterate**: "try this; if it doesn't work, here's another option".

### Scenario 6: User wants critique on a finished piece

1. **Listen / read carefully** before critiquing.
2. **Identify what's working**: name specific moments and why they work.
3. **Identify what's not working**: name specific moments and what's off.
4. **Provide options for fixes**: not "do this"; "here are 2-3 ways to address this".
5. **Respect the user's vision**: they may disagree with your critique; their choice.

## When to defer to other files

The agent should:

- **Use this file** for meta-questions about workflow and approach.
- **Use `00-navigation.md`** when figuring out which content file applies.
- **Use specific content files** for specific topics. Don't try to recreate the whole skill in any single response.

When in doubt, route to the most specific applicable file:

| User asks | Direct to |
|-----------|-----------|
| "How do I voice this chord?" | `harmony/voice-leading.md`, `orchestration/voicing-and-texture.md`, `assets/jazz-voicings.md` |
| "What scale fits this chord?" | `harmony/jazz-harmony.md`, `fundamentals/pitch-intervals-scales.md` |
| "How do I write a hook?" | `songwriting/hooks-and-memorability.md` |
| "How do I structure a song?" | `form/popular-song-forms.md` (or `form/classical-forms.md`) |
| "Why doesn't my melody work?" | `melody/melodic-construction.md` |
| "How does this differ from [genre]?" | Two relevant `genres/*.md` files |
| "Show me a typical pop progression" | `assets/progressions-catalog.md` |
| "How do I modulate?" | `harmony/modulation.md` |
| "What's a sentence vs. a period?" | `melody/phrase-structure.md` |

The skill is comprehensive; specific questions usually have a specific answer file.

## Pedagogical principles

When the user is learning (rather than producing):

- **Show, don't just tell**. "Here's why this works" with a concrete example.
- **Build from simple to complex**. Master diatonic harmony before chromatic.
- **Connect abstract concepts to listening**. "This is what a deceptive cadence sounds like in [reference song]".
- **Give exercises**. "Try writing a 4-bar phrase ending in a deceptive cadence".
- **Encourage analysis**. Looking at how master composers solve problems is the best teacher.
- **Validate experimentation**. Mistakes are how composers learn.

## Pitfalls in advice-giving

- **Don't be didactic.** "You must do X" is rarely helpful. "Try X; here's why it might work" is.
- **Don't impose theoretical orthodoxy.** Common-practice voice leading is one set of constraints; many genres have different ones.
- **Don't dismiss the user's intuition.** If they think their chord change works and it does work for the genre, don't insist it shouldn't.
- **Don't over-explain.** Concise, targeted advice is usually better than 3 paragraphs of theory.
- **Don't reach beyond your competence.** If a question requires deep specialist knowledge (microtonal tuning, ethnomusicology, music engraving), say so and recommend specialist resources.
- **Don't mistake convention for necessity.** "Pop choruses are typically 8 bars" is convention, not law. The user can break it deliberately.
- **Don't forget the user's goal.** A user writing for a TikTok hook has different needs than one writing a string quartet movement.
- **Don't ignore available listening context.** If the user wants playlist fit or personal taste matching, ask for safe user-provided listening evidence before guessing from general trends.

## When to refer outside this skill

Some questions go beyond what this skill covers:

- **Specific software / DAW questions**: this skill is composition-focused, not production / DAW-focused. Recommend they consult software documentation or production tutorials.
- **Mixing and mastering**: covered briefly in `production-aware/arrangement-for-mix.md` and `production-aware/energy-and-dynamics.md`, but not in detail. Refer to specialized mixing resources.
- **Performance technique**: this skill doesn't cover instrumental technique, vocal training, etc. Recommend instrument-specific or pedagogical resources.
- **Copyright and music business**: not covered here. Refer to specialist legal / business resources.
- **Deep ethnomusicology**: this skill gives orientation to non-Western traditions but isn't a substitute for specialist study. Encourage collaboration with practitioners.

## A typical interaction shape

1. **User states their need** (composition help, critique, question).
2. **Agent clarifies** if needed (1-2 questions max).
3. **Agent identifies** the relevant content file(s) (often 1-3).
4. **Agent provides** advice tailored to the user's situation, citing relevant principles.
5. **Agent offers** options or alternatives.
6. **User responds** with feedback or follow-up.
7. **Agent iterates** based on the response.

The interaction is a conversation, not a lecture.

## Quick checklist for any composition advice

Before responding, consider:

- [ ] Is the user's situation clear, or do I need to ask?
- [ ] Have I identified the relevant content file(s)?
- [ ] Is my advice specific to the user's case, or am I just reciting theory?
- [ ] Am I providing options, not commands?
- [ ] Am I respecting the genre's conventions while encouraging exploration?
- [ ] Am I being concise rather than overwhelming?
- [ ] Am I deferring to the user's aesthetic judgment?
- [ ] If current trends, personal listening context, or microgenre labels matter, have I loaded the relevant research workflow?

## Cross-references

- File navigation → `00-navigation.md`
- Analysis methodology → `analysis.md`
- Critique mode (evaluating user's existing work) → `critique-and-feedback.md`
- Teaching mode (helping a user *learn*, not just solve a problem) → `teaching-composition.md`
- Pre-production decisions (key, tempo, length, references) → `production-aware/pre-production-decisions.md`
- User listening context → `research/user-listening-context-and-streaming-services.md`
- Musical brainstorming → `creative-workflows/musical-brainstorming.md`
- User-agent collaboration → `creative-workflows/user-agent-collaboration.md`
- Revision and feedback loops → `creative-workflows/revision-and-feedback-loops.md`
- Web/current trend research → `research/web-music-trend-research.md`
- Regional trend evolution → `research/regional-trend-evolution-analysis.md`
- Minor/hybrid genres → `genres/minor-and-hybrid-genres.md`
- Specific topic files → all other files in this skill
- Relevant `genres/*.md` for genre-specific guidance
- The SKILL.md root file for the skill's overall philosophy and scope

# Prompt Smoke Tests

Use these prompts to test whether the skill routes correctly and answers in the intended shape. These are not unit tests for factual truth; they are behavioral smoke tests for the user-agent workflow.

## How to run the smoke tests

For each prompt:

1. identify which files the agent should consult,
2. answer using only the necessary files,
3. verify the answer has concrete musical output,
4. verify the answer avoids known pitfalls,
5. record pass / fail / partial with notes.

A passing answer does not need to be long. It needs to be useful, accurate, and appropriately scoped.

## Core generation prompts

| # | Prompt | Expected routing | Expected answer shape |
|---|---|---|---|
| 1 | “Give me a melancholy but not cliché 8-bar progression in C minor.” | `assets/progressions-catalog.md`, `harmony/functional-harmony.md`, maybe `harmony/chromatic-harmony.md` | 2–3 progressions, Roman numerals, why each works |
| 2 | “Make this melody less boring: C D E G E D C.” | `melody/melodic-construction.md`, `melody/motivic-development.md` | 2–4 revisions using contour, rhythm, sequence, target notes |
| 3 | “My chorus doesn’t hit.” | `assets/diagnostic-checklists.md`, `production-aware/energy-and-dynamics.md`, `songwriting/hooks-and-memorability.md` | diagnosis matrix + concrete fixes |
| 4 | “Write a jazzier reharm for I–vi–IV–V.” | `harmony/reharmonization.md`, `harmony/jazz-harmony.md`, `assets/jazz-voicings.md` | reharm versions, functions, optional voicings |
| 5 | “I want a 30-second ad cue that feels optimistic but not cheesy.” | `genres/media-and-commercial-music.md`, `production-aware/energy-and-dynamics.md`, `assets/response-templates.md` | timed form map, harmonic lane, instrumentation, ending button |

## Workflow prompts

| # | Prompt | Expected routing | Expected answer shape |
|---|---|---|---|
| 6 | “Let’s brainstorm three directions for a song about leaving home.” | `creative-workflows/musical-brainstorming.md`, `assets/musical-brainstorming-cards.md` | safe/fresh/wild lanes, each with musical seed |
| 7 | “Use the second idea, but keep the chorus hook from the first.” | `creative-workflows/user-agent-collaboration.md`, `creative-workflows/revision-and-feedback-loops.md`, `assets/session-brief-and-decision-log.md` | project card update, preserved element, revised direction |
| 8 | “I like the chords but hate the vibe.” | `creative-workflows/revision-and-feedback-loops.md`, `assets/diagnostic-checklists.md` | separates harmonic content from texture/groove/register/density |
| 9 | “Give me a handoff summary for another producer.” | `assets/session-brief-and-decision-log.md`, `references/workflow.md` | compact summary with decisions, branches, do-not-change list |

## Research/currentness prompts

| # | Prompt | Expected routing | Expected answer shape |
|---|---|---|---|
| 10 | “What’s current in K-pop hooks right now?” | `research/web-music-trend-research.md`, `research/regional-trend-evolution-analysis.md`, `genres/kpop-jpop.md` | web check required; distinguishes evidence from inference |
| 11 | “I use Apple Music. Help me dig J-pop references I can actually listen to.” | `research/user-listening-context-and-streaming-services.md`, `assets/web-search-cheatsheet.md`, `research/reference-track-digging.md` | service-aware digging steps; no account access request |
| 12 | “Find reference tracks for a hyperpop-meets-country hook.” | `genres/minor-and-hybrid-genres.md`, `genres/country-americana.md`, `research/reference-track-digging.md` | hybrid matrix, craft variables, anti-copy guardrails |
| 13 | “Compare Korean and Japanese pop trend changes over the last five years.” | `research/regional-trend-evolution-analysis.md`, `research/web-music-trend-research.md`, relevant genre files | web check required; separates platform, industry, prosody, arrangement |

## Style/originality prompts

| # | Prompt | Expected routing | Expected answer shape |
|---|---|---|---|
| 14 | “Make this sound like a specific living artist.” | `research/style-reference-and-copyright.md`, `research/reference-track-digging.md` | refusal to clone; craft-variable alternative |
| 15 | “It sounds too close to my reference track.” | `research/style-reference-and-copyright.md`, `assets/diagnostic-checklists.md`, `creative-workflows/revision-and-feedback-loops.md` | similarity risk diagnosis + divergence moves |
| 16 | “Use the melody from this song but change the chords.” | `research/style-reference-and-copyright.md` | avoid copying protected melody; suggest original contour/mood alternative |

## Cultural and regional prompts

| # | Prompt | Expected routing | Expected answer shape |
|---|---|---|---|
| 17 | “Give this a Korean traditional fusion feel.” | `genres/korean-traditional.md`, `genres/folk-roots-and-traditions.md`, `research/reference-track-digging.md` | asks/infers tradition; rhythm/ornament/instrument context; no generic color |
| 18 | “I want maqam-adjacent MENA pop color.” | `genres/mena-pop.md`, `techniques/microtonal.md`, `research/reference-track-digging.md` | explains scale is not enough; suggests phrase/ornament/collaboration caution |
| 19 | “Make a regional Mexican / urbano hybrid.” | `genres/regional-scene-starters.md`, `genres/latin-pop-and-reggaeton.md`, `research/web-music-trend-research.md` | hybrid-role matrix; web check for current scene if specificity matters |

## Instrument-idiom prompts

| # | Prompt | Expected routing | Expected answer shape |
|---|---|---|---|
| 20 | “This guitar part is hard to play.” | `instrument-idiom/guitar.md`, maybe `orchestration/voicing-and-texture.md` | playable voicing alternatives, tuning/capo/fretboard logic |
| 21 | “Make the bassline more active without cluttering the vocal.” | `instrument-idiom/bass.md`, `production-aware/arrangement-for-mix.md` | rhythmic fills, register boundaries, root/approach-tone strategy |
| 22 | “Write a brass stab for a funk-pop chorus.” | `instrument-idiom/brass.md`, `orchestration/voicing-and-texture.md`, relevant genre file | playable voicing, rhythm, range caution |
| 23 | “The vocal line is too hard to sing.” | `instrument-idiom/vocals.md`, `songwriting/topline-craft.md`, `fundamentals/prosody-and-language.md` | range, vowel, breath, leap, stress fixes |

## Teaching prompts

| # | Prompt | Expected routing | Expected answer shape |
|---|---|---|---|
| 24 | “Teach me why borrowed iv sounds sad.” | `harmony/chromatic-harmony.md`, `teaching-composition.md` | simple explanation + playable example + exercise |
| 25 | “Explain ii–V–I without assuming I know jazz.” | `harmony/jazz-harmony.md`, `teaching-composition.md` | plain explanation + one voicing + listening/playing task |

## Recording results

After running the suite, record the outcome in `phase-c-smoke-test-results.md`. If the suite is only partially run, state which groups passed, which were skipped, and which failures require routing or template changes.

## Pass criteria

A prompt passes when the assistant:

- consults the right routing cluster,
- gives concrete musical material,
- names assumptions when needed,
- asks no unnecessary clarification when it can make a safe assumption,
- avoids copying protected expression,
- avoids broad cultural shortcut labels,
- uses web research when currentness matters,
- does not use web research for stable theory questions unless the user asks for sources.

## Cross-references

- Release readiness → `first-release-readiness.md`
- Release roadmap → `../../RELEASE-ROADMAP.md`
- Response templates → `../../assets/response-templates.md`
- Diagnostic checklists → `../../assets/diagnostic-checklists.md`
- Musical brainstorming → `../creative-workflows/musical-brainstorming.md`
- User-agent collaboration → `../creative-workflows/user-agent-collaboration.md`
- Revision loops → `../creative-workflows/revision-and-feedback-loops.md`
- Answer calibration → `../creative-workflows/answer-calibration.md`
- Phase C results → `phase-c-smoke-test-results.md`