Ever wanted to mix 150 drums into a single waveform? Curious to see what patterns emerge across genres? But that's too much work to do in Ableton, right? That's okay because... PYTHON IS MY DAW <3
I used ZFTurbo's Music-Source-Separation-Training model to get drum stems, librosa and some web scraping to get the BPMs, PyRubberband to change each track's tempo, and soundfile and numpy to mix the tracks. View the source code here. What would've taken hours in Ableton took me hours in Python <3
Play with the gate to reveal the loudest (and therefore most frequently occurring) drum sequence.
Here's what 50 Detroit Techno drum sequences sound like together:
Here's what 150 Detroit Techno drum sequences sound like together:
The albums used for this project:
All rights belong to the respective artists and labels. No music was used to train any AI. If you'd like your work removed from this project, please reach out.
This website is a work in progress. Next up... Budots