I love music. I have a rather large digital collection of it, hoarded over many years, and I intend to grow it indefinitely. There are a wide variety of formats, from FLAC, to Ogg Theora, to MP3, to (very poor) WMA.

Converting all the music at once, preserving the exact directory structure, is more of a challenge than it ought to be – ideally, a command should exist, like rsync, that intelligently decides what to convert and what to just copy (converting 96kbps mp3 files to a different lossy format will not do it many favours).

That’s what shrinkmusic is. It takes an input directory, an output directory, computes a plan – what to convert, what to keep, what to skip due to already existing in the output directory, and what to ignore due to a user flag – and then executes it, in parallel, using as many jobs as you’d like.

Features

• Add new music without re-converting an entire library
• Decides (albeit by pretty dumb heuristics) what is worth converting, and what is OK to just copy
• Splits FILE.flac and FILE.cue pairs automatically
• Parallel execution
• Dry run mode
• Can ignore specific files/directories (e.g., I don’t care to copy hoarded podcasts – readily available on the internet anyway – over)

Install

Install using Stack or Nix.

$ git pull https://github.com/mikeplus64/shrinkmusic.git
$ cd shrinkmusic
$ stack install
$ nix-env -f ./default.nix -i # alternative for Nix

Usage

Usage: shrinkmusic (-i|--input INPUT) (-o|--output OUTPUT)
                   (-b|--bitrate BITRATE) (-j|--jobs JOBS) [-z|--ignore IGNORE]
                   [-d|--dry-run]

Available options:
  -h,--help                Show this help text
  -i,--input INPUT         Input directory
  -o,--output OUTPUT       Output directory
  -b,--bitrate BITRATE     bitrate for audio
  -j,--jobs JOBS           Number of jobs to use
  -z,--ignore IGNORE       Ignore this file
  -d,--dry-run             Do nothing; just output the plan