Repository to keep track of my various ideas using ML to analyse and generate audio.

Some of the things I'm working on:

• ADSR envelope characteristics detection (attack/decay time)
• Finding drum samples which are similar to a given sample based on audio characteristics
• implementing a WaveGAN model in PyTorch to generate drum samples based on sample packs
• Synth parameter inference using Ableton's Wavetable (trying to predict which parameters were used on wavetable given an audio sample)
• using Spotify API to recommend lesser-known artists based on purely audio features (as opposed to popularity/whether similar users like them)