MUSDB
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Music Source Separation | MUSDB18 (test) | SDR (Bass)7.12 | 69 | |
| Music Source Separation | MUSDB18 HQ (test) | SDR (Drums)12.91 | 61 | |
| Music Source Separation | MUSDB18 HQ | Vocals Score10.83 | 21 | |
| Waveform Generation | MUSDB18 out-of-distribution vocal samples HQ (test) | M-STFT0.8623 | 19 | |
| Music Reconstruction | MUSDB | Mel-Loss0.63 | 18 | |
| Audio Super-Resolution | MUSDB18-HQ Out-of-domain | LSD0.97 | 16 | |
| Music Reconstruction | MUSDB18 | Mel-16k Score1.273 | 16 | |
| Vocal Source Separation | MUSDB HQ 18 | cSDR12.72 | 14 | |
| Music Source Separation | MUSDB18 | SDR (Bass)11.31 | 11 | |
| Singing Voice Source Separation | MUSDB (test) | Vocals SDR4.6 | 11 | |
| Audio Source Separation | MUSDB18 SiSEC 2018 (test) | Vocals Score7.24 | 10 | |
| Music Source Separation | MusDB | SDR (All)8.22 | 9 | |
| Timbre Transfer | MUSDB18-HQ (test) | CLAP0.342 | 8 | |
| Vocoding | MUSDB18 (test) | ViSQOL (Bass)4.842 | 8 | |
| Neural Audio Compression | MUSDB18 (test) | FAD130.783 | 8 | |
| Multi-instrument source separation | MUSDB18 SiSEC (test) | Vocals SDR3.46 | 8 | |
| Singing Voice Separation (Accompaniment) | MUSDB SiSec 2018 (test) | Median SDR10.69 | 8 | |
| Singing Voice Separation (Vocals) | MUSDB SiSec 2018 (test) | Median SDR4.58 | 8 | |
| Audio Reconstruction | MUSDB18-HQ (test) | Mel Loss0.5035 | 7 | |
| Audio Super-Resolution | MUSDB18-HQ (test) | MOS4.8 | 6 | |
| Audio Super-Resolution | MusDB 11.025-44.1 kHz | LSD1.12 | 6 | |
| Source Separation | MUSDB vocal accompaniment (test) | SDR6.5 | 6 | |
| Lossless Audio Compression | MusDB18 Mixes 16-bit | Compression Rate2.08 | 5 | |
| Lossless Audio Compression | MusDB18 All 16-bit | Compression Rate2.82 | 5 | |
| Multi-track Music Generation | MUSDB18 | CBS0.574 | 5 |
Showing 25 of 34 rows