Whirling:A music visualizer that uses audio feature extraction and segmentation to create more representative visuals from create numpy array python Watch Video

⏲ Duration: 3 min 12 sec
✓ Published: 05-Sep-2020

Description:
I built a music visualizer that tries to intelligently understand as much about the song that's playing so it could create a more representative visual. All songs played have to undergo a two step preprocessing algorithm that runs a source separator and a feature extractor on each of the generated sources. I use Spleeter and Librosa to do source separation. Spleeter is able to separate the drums, vocals, bass and other for each track. Librosa can separate the percussion and harmonics. For featur