mastering-audio
Skill7 starsChanged 21 days agoMITView on GitHub
Master audio files (especially Suno/Udio AI music) for streaming release with LUFS normalization, true peak limiting, and optional enhancements. Triggers — "master this", "master the track", "LUFS normalize", "optimize for streaming". Not for a spot's or film's delivery loudness — use video-finish-qc.
In this repository16 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- ad-spot-preprodSkill
- client-roundsSkill
- designed-elementsSkill
- explainer-videoSkill
- film-preprodSkill
- spot-audio-assemblySkill
- video-edit-edlSkill
- video-finish-qcSkill
- video-finishSkill
- video-gen-cost-gateSkill
- video-productionSkill
- video-prompt-dialectsSkill
- video-refs-continuitySkill
- video-take-reviewSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 2 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 30
Run analysis to understand the audio characteristics:
- Very specific: line 48
Use --preset suno (optimized for AI music characteristics)
- Very specific: line 52
Use --preset streaming for -14 LUFS, conservative processing
- Very specific: line 55
Use --preset loud for -9 LUFS with saturation/punch
- Very specific: line 58
Use --preset dynamic for -16 LUFS, minimal processing
- Very specific: line 85
Default output: If no output specified, creates <input>_mastered.wav in same directory.
- Distinctive: line 89
Run analysis on processed file to confirm:
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Audio Mastering Skill
Master audio files for streaming release using DSP analysis and processing.
Tool Location
~/.claude/skills/mastering-audio/scripts/master.mjsWorkflow
Phase 1: Analyzing Audio
Run analysis to understand the audio characteristics:
node ~/.claude/skills/mastering-audio/scripts/master.mjs "<INPUT_FILE>" --analyze --jsonMetrics to evaluate:
- LUFS (loudness) - streaming target is -14 LUFS
- True Peak - should be below -1 dBTP
- Stereo width and correlation
- DC offset presence
- Detected issues (clipping, low headroom, etc.)
Phase 2: Selecting Approach
Based on analysis, determine processing approach:
If Suno/Udio AI music:
- Use
--preset suno(optimized for AI music characteristics) - Addresses: harsh 2-4kHz, clipping, overloud output
If targeting streaming platforms:
- Use
--preset streamingfor -14 LUFS, conservative processing
If maximum loudness needed:
- Use
--preset loudfor -9 LUFS with saturation/punch
If preserving dynamics:
- Use
--preset dynamicfor -16 LUFS, minimal processing
Custom approach (advanced):
--target-lufs N Target LUFS (default: -14)
--ceiling N True peak ceiling dB (default: -1)
--deharsh Reduce harsh frequencies
--add-air High-frequency sparkle
--warmth Tape saturation
--punch Transient enhancement
--stereo-width N Width percentage (100 = unchanged)
--bit-depth N 16 or 24 (default: 24)
--sample-rate N Output rate (default: 48000)Phase 3: Processing Audio
Execute mastering with chosen settings:
# Using preset
node ~/.claude/skills/mastering-audio/scripts/master.mjs "<INPUT_FILE>" "<OUTPUT_FILE>" --preset suno
# Custom settings
node ~/.claude/skills/mastering-audio/scripts/master.mjs "<INPUT_FILE>" "<OUTPUT_FILE>" --target-lufs -14 --ceiling -1 --deharsh --add-airDefault output: If no output specified, creates <input>_mastered.wav in same directory.
Phase 4: Verifying Results
Run analysis on processed file to confirm:
node ~/.claude/skills/mastering-audio/scripts/master.mjs "<OUTPUT_FILE>" --analyzeSuccess criteria:
- LUFS within ±0.5 of target
- True peak at or below ceiling
- No clipping warnings
- Stereo correlation positive (not phase-inverted)
Presets Reference
| Preset | LUFS | Ceiling | De-harsh | Air | Warmth | Punch | Width |
| streaming | -14 | -1 | Yes | No | No | No | 100% |
| loud | -9 | -0.3 | Yes | Yes | Yes | Yes | 105% |
| dynamic | -16 | -1 | No | No | No | No | 100% |
| suno | -14 | -1 | Yes | Yes | No | Yes | 102% |
Processing Chain
- DC offset removal
- Hybrid dynamic processor (de-harsh)
- High-frequency exciter (air)
- Tape warmth saturation
- Multiband transient enhancement (punch)
- Stereo width adjustment
- Bass mono (< 200 Hz)
- Final filters (30 Hz HPF, 18 kHz LPF)
- LUFS normalization
- Lookahead true peak limiter
Dependencies
- Node.js 18+
- ffmpeg (for audio decoding)
- npm dependencies (run
npm installin skill directory once)
Example Session
User: master this suno track for spotify release
Assistant:
1. Analyzing: node master.mjs song.mp3 --analyze --json
→ LUFS: -9.2, True Peak: +0.3 dBTP (clipping detected)
2. Selecting: Using suno preset (AI music, targeting -14 LUFS)
3. Processing: node master.mjs song.mp3 song_mastered.wav --preset suno
→ Applied: de-harsh, air, punch, bass mono, LUFS normalization, limiting
4. Verifying: node master.mjs song_mastered.wav --analyze
→ LUFS: -14.0, True Peak: -1.0 dBTP (no issues)
→ Ready for streaming release- Deeper context — the upstream projects behind the rules here:
video-production/references/CONTEXT-MAP.md§ Where the deeper context lives.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

