Hello!
You have stumbled on my procedural music project. The project should be in somewhat working order. So far it is very basic and produces pretty crappy tunes... but hey I can call them tunes, so that's already pretty cool... There's some audio demos in the audio_tests
folder.
If you want to try it out yourself, I invite you to clone the repo and run the meta.py
file script on your own machine.
Everything uses the numpy
, scipy
, matplotlib
, os
and pydub
libraries, so if you've got those, you should be good to go. pydub
is only used to convert the .wav
files to .mp3
, to make things lighter, but if you don't convert stuff, you won't need it. Just make sure to turn off any settings converting stuff to MP3.
The idea here is to make a bunch of interelated scripts which will generate a musical album. In this repo, I'll make some sound engines to make noise. I'll also make some automated composition tools which will randomly be able to select from a set of chord and melody generating algorithms.
If everything goes well, the script should be able to produce a complete album that is at least interesting to listen to. Using some random parameters, I think it'd be cool if you could generate basically a entirely new album everytime you run the script.
I haven't decided yet if I'll build the system with a central seed so as to be able to reproduce an album.
So far the examples are very basic. This is mainly due to very rudimentary sound design and rythms. Here are some sound demos: