3 ms·

There's maybe something useful for you here? https://github.com/facebookresearch/demucs https://github.com/facebookresearch/demucs https://github.com/sigsep/o

There's maybe something useful for you here? https://github.com/facebookresearch/demucs https://github.com/facebookresearch/demucs https://github.com/sigsep/open-unmix-pytorch https://github.com/sigsep/open-unmix-pytorch https://github.com/bytedance/music_source_separation https://github.com/bytedance/music_source_separation https://github.com/deezer/spleeter https://github.com/deezer/spleeter

sdenton4 5y ago
Looks like demucs is using a learned filter bank (so no handling of fft phase), and open umix and spleeter are using spectrogram magnitude masking (just reuse phase from original fft in all channels). Can't immediately tell what bytedance is doing.