normalize-audio
Installation
SKILL.md
Normalize Audio
Two-pass loudnorm to hit a precise integrated LUFS + true-peak target. Reads the user's default target from preferences.json if present; otherwise accepts an inline target.
This is the apply step that pairs with the read-only audio-analysis skill.
Procedure
1. Inputs
| Field | Default |
|---|---|
| Source | required (video or audio) |
| Target | from preferences.json (loudness.target); else ask. Choices: youtube (-14/-1), spotify (-14/-1), apple-music (-16/-1), podcast (-16/-1), broadcast (-23/-2) |
| Output | sibling <basename>.normalized.<ext> (or project's working/ if inside an index project) |
Resolve target → I (integrated LUFS), TP (true peak), LRA (default 11).