mp3glitch

Python tool for glitching MP3s, creating interesting sounds for music production. Includes many options to shape glitching amount, character, and timbre.

The mechanics behind this are described in this article, and I discuss using Python to do the glitching in this article.

Should work for both constant bit rate (CBR) and variable bit rate (VBR) MP3s.

Installing

To install with pipx:

pipx install git+https://codeberg.org/reillypascal/mp3glitch.git

Usage

• Provide the path of an input MP3 file, followed by the desired output file name, OR provide an input and output directory. NOTE: input/output must both be either files or directories.

mp3glitch <input_file_or_directory> <output_file_or_directory>

• Options
  • -h, --help show this help message and exit
  • -p, --prob percent probability of glitching (float)
  • -m, --hexmin decimal representation of minimum hex value to insert (int)
  • -M, --hexmax decimal representation of maximum hex value to insert (int)
  • -f, --framemin minimum position in frame to glitch (float, 0-1)
  • -F, --framemax maximum position in frame to glitch (float, 0-1)
  • -s, --spacingmin minimum spacing between glitched frames (int)
  • -S, --spacingmax maximum spacing between glitched frames (int)
  • -w, --width number of hex digits to insert in each glitch (int)
  • -l, --limit max number of glitches per frame (0 = no limit) (int)

NOTE: while the resulting MP3 is still playable, it will likely have e.g., denormal values, and with greater glitch probability or hexmax values, this can still cause difficulty with playback.

When composing with this tool, I usually use FFmpeg to convert a WAV sample to MP3, glitch it, and then convert back to WAV in order to have the noisy artifacts without the playback difficulty. FFmpeg may report errors on some frames, but it will still produce the expected glitchy output.