We study the problem of source separation for music using deep learning with four known sources: drums, bass, vocals and other accompaniments. State-of-the-art approaches predict soft masks over mixture spectrograms while methods working on the waveform are lagging behind as measured on the standard MusDB benchmark... (read more)
PDF
#2 best model for
Music Source Separation
on MUSDB18
TASK | DATASET | MODEL | METRIC NAME | METRIC VALUE | GLOBAL RANK | COMPARE |
---|---|---|---|---|---|---|
Music Source Separation | MUSDB18 | DEMUCS | SDR (vocals) | 6.29 | # 2 | |
Music Source Separation | MUSDB18 | DEMUCS | SDR (drums) | 6.08 | # 1 | |
Music Source Separation | MUSDB18 | DEMUCS | SDR (other) | 4.12 | # 1 | |
Music Source Separation | MUSDB18 | DEMUCS | SDR (bass) | 5.83 | # 1 |