Speech Denoising on VCTK-DEMAND (test)
3.17PESQPERL-AE
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| PERL-AELoss Term=l1 + AcousticEvent2020.10 | 3.17 | 4.43 | 3.53 | 3.83 | 95 | |
| DEMUCS2020.10 | 3.07 | 4.31 | 3.4 | 3.63 | 95 | |
| T-GSA2020.10 | 3.06 | 4.18 | 3.59 | 3.62 | — | |
| Self-Adapt MHSA2020.10 | 2.99 | 4.15 | 3.46 | 3.51 | — | |
| Hi-Fi GAN2020.10 | 2.94 | 4.07 | 3.07 | 3.49 | — | |
| Deep Feature Loss2020.10 | 2.57 | — | — | — | — | |
| Weiner filtering2020.10 | 2.22 | 3.23 | 2.68 | 2.67 | — | |
| Noisy2020.10 | 1.97 | 3.35 | 2.44 | 2.63 | — |