misuba / raw

raw

random audio workstation. basically a command-line program to generate entire songs from short sample loops.

for me, making an album of music means devising a single process of making music and using that process to generate tracks. my past processes were focused on melody and harmony. for generations I found a process around generative chord changes. for 18,000 I used a coding framework to write small snippets for generative audio. for at the place I tried to devise a paper-and-pen method of developing melodies.

in this current process I'm focusing on effects and structure. I plan on only using samples and trying to elevate the samples by restructuring and juxtaposed effects chains - and have its creation based entirely in a terminal. this current project is based off my other projcets dnb.lua, makebreakbeat, and sampswap. this one should be better capable of cross-platformability.

requirements

• python3 numpy+scipy
• sox
• supercollider

you can install python3 anyway you want, and then install the dependencies with:

python3 -m pip install numpy scipy

sox can be downloaded here (or Linux: apt get sox).

SuperCollider is spread across several places (because you'll need to install all the plugins):

• SuperCollider
• SuperCollider sc3-plugins
• SuperCollider mi-ugens
• SuperCollider ported plugins

you can click those links to download the files for your system. its also possible to install them by source (i.e. for Raspberry Pi where binaries aren't available):

git clone https://github.com/schollz/raw && cd raw
sudo apt install cmake
make supercollider
make sc3plugins
make miugens
make portedplugins

usage

> git clone https://github.com/schollz/raw
> cd raw
> go build -v
> ./raw stemstitch --config example/config.toml

confg file

see example/config.toml and its directory for how to structure samples for a song and add configuration. basically, the config file lets you setup the song structure for each track, and setup effects based on probabilities (0-1).

license

MIT