| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Speech Separation | WSJ0-2Mix (test) | SDRi (dB)25.2 | 141 | |
| Speech Separation | WSJ0-2Mix | SI-SNRi (dB)25.1 | 65 | |
| Voice Separation | WSJ0 3mix | SI-SNRi19.5 | 14 | |
| Monaural Speech Separation | WSJ0 3mix | ΔSI-SDR (dB)23.7 | 13 | |
| Speech Separation | WSJ0-3mix (clean) | Delta SI-SNR (dB)19.5 | 12 | |
| Interpolation | WSJ0 Audio Spectrogram | Interpolation FID (0.0-0.8)7.5 | 10 | |
| Generative Modeling | WSJ0 Audio Spectrogram | Log P(x)1.94 | 10 | |
| Speech Separation | WSJ0-4mix (test) | SI-SNRi12.9 | 10 | |
| Speech separation | WSJ0 open speaker set (test) | SDR Improvement (dB)8.3 | 9 | |
| Source Separation | WSJ0 3Mix | SI-SDRi24.2 | 8 | |
| Speech Enhancement | WSJ0 simulated (test) | PESQ (0dB)2.972 | 8 | |
| Domain Incremental Speech Enhancement | WSJ0 synthetic | SI-SNR (dB) - Alarm18.72 | 7 | |
| Monaural Speech Separation | WSJ0 4mix | Delta SI-SDR (dB)22 | 7 | |
| Source Separation | WSJ0 2mix | Oracle SI-SNR20.12 | 7 | |
| Speaker Separation | WSJ0-3mix 8kHz (test) | Delta SI-SDR17.8 | 7 | |
| Binaural Target Speech Extraction | WSJ0 Reverb | DNSMOS-OVRL3.141 | 6 | |
| Dereverberation | WSJ0-Reverb (test) | WVMOS4.403 | 6 | |
| Monaural Speech Separation | WSJ0-5mix | ΔSI-SDR (dB)21 | 6 | |
| Source Separation | WSJ0-3mix | Oracle SI-SNR16.85 | 6 | |
| 5-speaker speech separation | WSJ0-5mix (evaluation) | SDRi11.14 | 5 | |
| Voice Separation | WSJ0 5mix | SI-SNRi10.6 | 5 | |
| Source Separation | WSJ0 4mix | Oracle SI-SNR12.88 | 4 | |
| Speech Separation | WSJ0-5mix (test) | SI-SNRi11.7 | 4 | |
| Speech separation | WSJ0 closed speaker set (test) | SDR Improvement (dB)6.54 | 4 | |
| Source Separation | WSJ0 3mix (eval) | SI-SDR10.06 | 4 |