Neural Vocoding on VCTK 100 audio clips (unseen)
0.0925MAEBigVGAN
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| BigVGANunseen_dataset=true2022.06 | 0.0925 | 0.7684 | 4.001 | 0.3557 | 0.0833 | 0.9672 | |
| BigVGAN-baseunseen_dataset=true2022.06 | 0.1527 | 0.8677 | 3.443 | 0.4526 | 0.1047 | 0.9586 | |
| BigVGAN-baseunseen_dataset=true, ablation=w/o filter2022.06 | 0.1739 | 0.8935 | 3.379 | 0.4924 | 0.1128 | 0.9528 | |
| BigVGAN-baseunseen_dataset=true, ablation=w/o filter & snake2022.06 | 0.1885 | 0.9184 | 3.316 | 0.5159 | 0.1208 | 0.9481 | |
| HiFi-GAN (V1)unseen_dataset=true2022.06 | 0.2541 | 0.9859 | 3.029 | 0.6795 | 0.1336 | 0.9375 | |
| UnivNet-c32+unseen_dataset=true, training_subset=train-clean-360, pretrained_source=open-source repository2022.06 | 0.2753 | 0.9476 | 3.235 | 0.718 | 0.1131 | 0.9535 | |
| UnivNet-c32unseen_dataset=true2022.06 | 0.282 | 0.9586 | 3.184 | 0.7403 | 0.1198 | 0.9434 | |
| WaveFlow-128unseen_dataset=true2022.06 | 0.2982 | 1.0825 | 2.953 | 1.1127 | 0.1213 | 0.9518 | |
| WaveGlow-256unseen_dataset=true2022.06 | 0.5454 | 1.3324 | 3.149 | 2.4512 | 0.1218 | 0.9513 | |
| SC-WaveRNNunseen_dataset=true2022.06 | 0.6619 | 2.7344 | 1.445 | 1.7597 | 0.2528 | 0.8611 |