ferraus master

Automatic master of a finished mix to a delivery target

Synopsis

ferraus master <input> [options] [-o <output>]

Runs the automatic mastering chain and writes a release-ready master. Pre-flight: dc-remove, head/tail silence trim at -70 dB, analyze problems (clipping means declip runs first). Corrective EQ: match-eq toward --reference or the genre curve at --match-amount (1/3-octave bands, at most 4 dB per band), static cuts on persistent narrow resonances, 20 Hz high-pass. Dynamics: four-band multiband compression at 1.5:1 to 2:1 with tempo-dependent release, plus glue compression (1.5:1, 30 ms) when the crest factor is above the genre target. Stereo: widen with mono bass below --bass-mono, never letting broadband correlation drop below 0.3. Color: tape saturation and an exciter layer at --warmth. Loudness: gain to the target, then a 4x-oversampled true-peak limiter whose gain reduction is bounded by --intensity; if the target cannot be reached within the bound the master lands as close as possible and the report says so. Output: TPDF dither when writing 16-bit, master verify, the before/after report (--report), and a loudness-matched before/after pair (--ab). Every stage is an ordinary ferraus command; 'master plan' shows the exact chain and 'master --plan' runs an edited one. --stems (best-effort) separates the mix first and treats vocals, drums, bass, and other separately.

Examples

Master for streaming (-14 LUFS, -1 dBTP) with every default

ferraus master mix.wav -o master.wav

Run an edited plan and write the before/after report

ferraus master mix.wav --plan plan.json -o master_v1.wav --report r1.json

Move toward a reference's tonal balance and write an A/B pair

ferraus master mix.wav --reference ref.wav --match-amount 75% --ab -o master.wav

A loud club master

ferraus master mix.wav --target club --genre edm --intensity heavy -o club.wav

Options

Flag	Type	Default	Meaning
`--target`	streaming \| apple \| youtube \| club \| cd \| broadcast \| podcast \| custom	`streaming`	Delivery target: streaming -14 LUFS/-1 dBTP, apple -16/-1, youtube -14/-1, club -8/-0.3, cd -10/-0.3 (16-bit 44.1 kHz), broadcast -23/-1 (EBU R128), podcast -16 stereo or -19 mono/-1, custom (--lufs, --true-peak)
`--lufs`	gain	—	Integrated loudness for --target custom (e.g. -12LUFS); overrides a named target's loudness
`--true-peak`	gain	—	True-peak ceiling for --target custom (e.g. -1dBTP, default -1dBTP); overrides a named target's ceiling
`--standard`	ebu \| atsc	`ebu`	Broadcast standard for --target broadcast: ebu (-23 LUFS, -1 dBTP) or atsc (-24 LKFS, -2 dBTP)
`--reference`	path	—	Reference track whose tonal balance the corrective EQ moves toward (replaces the genre curve)
`--match-amount`	percent	`50%`	How much of the measured tonal difference the corrective EQ corrects (each band is capped at 4 dB)
`--genre`	auto \| pop \| hiphop \| edm \| rock \| acoustic \| classical \| jazz \| podcast	`auto`	Genre for the tonal curve and dynamics targets; auto guesses from tempo, spectral tilt, low-end share, and crest factor
`--intensity`	light \| medium \| heavy	`medium`	How hard to push: bounds limiter gain reduction (light 2 dB, medium 4 dB, heavy 6 dB) and sets compression ratios; an input that already looks mastered defaults to light
`--width`	percent	`100%`	Stereo width of the side signal (100% unchanged); never narrower than mono bass or wider than broadband correlation 0.3 allows
`--bass-mono`	frequency	`120Hz`	Side signal below this frequency is collapsed to mono
`--warmth`	percent	`15%`	Tape saturation and exciter amount (0% skips the color stage)
`--no-eq`	flag	—	Skip the corrective EQ
`--no-widen`	flag	—	Skip the stereo stage
`--trim-silence`	text	`-70dB`	Trim head and tail silence below this level (dBFS), or 'off'
`--report`	path	—	Write the before/after report (JSON) here
`--ab`	flag	—	Also write a loudness-matched before/after pair (within 0.2 LU) for listening: <output>.ab-before.wav and <output>.ab-after.wav
`--plan`	path	—	Run this (edited) plan instead of planning: the JSON from 'master plan'
`--stems`	flag	—	Best-effort stem-aware mastering: separate (best verified tier), per-stem chains, then sum and master

Global options

Every command also accepts the global options (-o, --json, --dry-run, --overwrite, --progress, ...). See Global options.

Exit codes

Code	Meaning
0	Success
1	Processing failed
2	Usage or parameter error
3	Input missing or unreadable
4	Unsupported format or codec
5	Model not cached and downloads disabled
6	Model download or verification failed
8	Cancelled
10	Network error
70	Internal error (bug)

Common mistakes

Forgetting that outputs are never overwritten: pass --overwrite or a new -o path when re-running.
Percentages accept 80% or the fraction 0.8; a bare 80 means 80x, not 80%.
Gains need a unit when not in dB: 0.5x is a factor, -6 and -6dB are decibels.

For models

MCP tool: ferraus_master
Schema: ferraus schema master
Pipeline: { op: master, args: { ... } } (flag names with _ instead of -)
fx stage: no

{"command": "master", "defaults": {"target": "streaming", "standard": "ebu", "match_amount": "50%", "genre": "auto", "intensity": "medium", "width": "100%", "bass_mono": "120Hz", "warmth": "15%", "trim_silence": "-70dB"}}