sed -i 's/\bmp3gain\b/mp3rgain/g' your_script.sh

See docs/migrating-from-mp3gain.md for the full flag equivalence table, Dockerfile/CI substitution patterns, tag interop notes, and the small set of intentional behaviour differences. Bit-level verification lives in docs/compatibility-report.md.

GUI Application

A native GUI application (mp3rgui) is available for users who prefer a graphical interface.

Features:

• Drag-and-drop file / folder loading (recurses subfolders)
• Track and Album ReplayGain analysis (parallel, with Cancel)
• Apply Track / Album Gain — shares the same applywithoptions pipeline as the CLI
• Options panel: Prevent clipping (-k), Preserve mtime (-p), Wrap mode (-w), Use ID3v2 (-s i), Dry run (-n)
• Modify Gain menu: Apply Track / Album / Manual (-g) / Channel (-l) Gain, Undo (-u), Delete Stored Tags (-s d)
• Analysis menu: Track / Album Analysis, Find Max Amplitude (-x), Check Stored Tags (-s c)
• Stored RG table column shows existing ReplayGain / undo tags (APE / ID3v2 / MP4 freeform) with a per-tag breakdown on hover
• Responsive UI: all batch work runs on a worker thread with per-file progress and a Cancel button

• mp3rgui-*-macos-universal.dmg (macOS)
• mp3rgui--linux-x86_64.tar.gz / mp3rgui--linux-arm64.tar.gz (Linux)
• mp3rgui--windows-x86_64.zip / mp3rgui--windows-arm64.zip (Windows)
• mp3rguiamd64.deb / mp3rguiarm64.deb (Debian/Ubuntu)

macOS manual download: If you see "mp3rgui cannot be opened" warning, run:

> xattr -cr /path/to/mp3rgui.app

This is not needed when installing via Homebrew.

Command-Line Options

| Option | Description | |--------|-------------| | -r | Apply Track gain (ReplayGain) | | -a | Apply Album gain (ReplayGain) | | -g <i> | Apply gain of i steps (1 step = 1.5 dB) | | -d <n> | Modify suggested dB gain by n (mp3gain-compatible; applied with -r / -a) | | -u | Undo gain changes | | -k | Prevent clipping | | -R | Process directories recursively | | --skip-errors | Keep album analysis (-a) going past unreadable files | | -n | Dry-run mode | | -j <n> / --threads <n> | Worker threads for analysis (default: auto, 0=auto, 1=serial) | | -o [fmt] | Output format: text, json, tsv (default: tsv if no argument) |

Run mp3rgain -h for the full list of options.

ReplayGain analysis runs in parallel by default (std::thread::available_parallelism() worker threads). Use -j 1 or MP3RGAIN_THREADS=1 for the legacy serial path. See docs/perf-parallel.md for the design and real-corpus benchmark numbers.

Documentation

• Migration Guide - Drop-in replacement for mp3gain: flag equivalence, sed/Dockerfile/CI substitution patterns, beets config
• Parallel Performance - -j / --threads design and real-corpus benchmark numbers
• Roadmap - Development plans and upcoming features
• Security - Memory safety and CVE analysis
• Compatibility Report - Verification against original mp3gain
• Technical Comparison - Comparison with similar tools
• Use Cases - Integration examples (beets, headroom, etc.)
• Download Stats - Weekly download trends across all platforms

Why mp3rgain?

The original mp3gain has been unmaintained upstream since ~2015 (though distribution maintainers continue to apply security patches). aacgain, its AAC counterpart, has been unmaintained since ~2009 and is effectively unbuildable on modern 64-bit systems. mp3rgain is a modern, memory-safe replacement written in Rust that covers both.

AAC/M4A on the CLI is the differentiator. Among CLIs, rsgain / loudgain / FFmpeg either only write ReplayGain tags (which non-compliant players ignore) or re-encode the audio. foobar2000 has a comparable "Apply ReplayGain to file content" feature for AAC in MP4/MKA, but it is Windows GUI only, has no undo, and is unsuited for batch, headless, or container workflows. mp3rgain fills the cross-platform, scriptable, reversible niche.

mp3rgain implements the ReplayGain 1.0 algorithm (89 dB reference level) for full compatibility with the original mp3gain / aacgain. Loudness values will differ from EBU R128/LUFS-based tools (foobar2000, loudgain, ffmpeg loudnorm).

Use mp3rgain in Docker / CI

Official multi-arch images (linux/amd64, linux/arm64) are published to GHCR:

ghcr.io/m-igashi/mp3rgain:latest
ghcr.io/m-igashi/mp3rgain:v2          # latest 2.x
ghcr.io/m-igashi/mp3rgain:v2.8.0      # exact version

The image is built FROM scratch with a fully static (musl) binary — no shell, no runtime deps, ~2 MB. Drop-in replacement for mp3gain in containerized batch / cron pipelines (e.g. Plex maintenance windows):

# Normalize a music library by mounting it into the container
docker run --rm \
  -v /path/to/music:/music \
  ghcr.io/m-igashi/mp3rgain:latest -r -R /music
Run as your own user so written files keep correct ownership
docker run --rm \
  --user "$(id -u):$(id -g)" \
  -v /path/to/music:/music \
  ghcr.io/m-igashi/mp3rgain:latest -r -R /music

Because the entrypoint is the binary itself, all mp3rgain flags work exactly the same as the host CLI (-r, -a, -R, -k, -u, …).

Library Usage

use mp3rgain::{apply_gain, analyze};
use std::path::Path;
let frames = apply_gain(Path::new("song.mp3"), 2)?;  // +3.0 dB let info = analyze(Path::new("song.mp3"))?;

Contributing

Contributions welcome! See CONTRIBUTING.md.

License

MIT License - see LICENSE.

See Also

• Original mp3gain
• headroom - DJ audio loudness optimizer