There is good visual stuff in the Glitch art thread. I would like to try datamoshing by messing with WAV audio files and maybe other formats first with a text/hex editor and then standard Unix tools like sed(1), and play them primarily on norns.
Any experiences or pointers? WAV files are pretty straightforward so I wonder can serious glitching occur? How corrupt files will Softcut and SuperCollider tolerate? Has datamoshing audio files yielded something interesting for you or others?
I remember being curious but not particularly brave about editing compression codecs… like opening the MP3 codec used by VLC in a text editor and messing with variables, moving things around, and attempting to encode audio with that.
Also, I believe that the particular effect of datamoshing has to do with manipulating hidden frames that are interleaved within the video file that reference relative change. I wonder if there’s a way to do this but with the compression methods in say, MP3, which exploits psychoacoustics. I’ve always been curious to hear what would happen if you turned the MP3 algorithm to 11…
M-xhexl-find-fileRET in Emacs, enter a hex value with C-M-x
snatch a valid WAV RIFF header with dd bs=44 count=1 if=a_normal_sample.wav of=header.wav and then cat header.wav any_other_file.png > any_other_file.png.wav
I wrote a command line tool for turning arbitrary files on your computer into audio. It is designed for batch processing as I wanted to process gigabytes of data.
Compiles for win, mac, linux
Let me know if I could help by either providing a binary for your platform or guiding the compile process.
Thank you for the kind feedback I hoped the web-based PhD might put it out into the open more so its great to see people coming across it. Let me know if you have any questions