Video-Driven Text-to-Speech on GRID standard (test)
7.68LSE-CGROUND-TRUTH
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| GROUND-TRUTHEvaluation Source=Original speech/video numbers reported in VISUALTTS [42]2021.11 | 7.68 | — | 6.87 | — | — | — | — | — | |
| GROUND-TRUTHEvaluation Source=Original speech/video on current study test set2021.11 | 7.24 | 2.68 | 6.73 | 26 | — | — | — | — | |
| VDTTS-GRIDTraining Set=GRID2021.11 | 6.97 | 2.55 | 6.85 | 26 | 7.89 | 0.14 | 0.07 | 0.11 | |
| VDTTS-LSVSRTraining Set=LSVSR2021.11 | 5.85 | 2.1 | 7.93 | 55 | 12.81 | 0.37 | 0.21 | 0.32 | |
| VISUALTTSTraining Set=GRID2021.11 | 5.81 | — | 8.5 | — | — | — | — | — | |
| TTS-TEXTONLYTraining Set=LibriTTS, Task=Zero-shot voice transfer2021.11 | 3.39 | 1.51 | 10.44 | 19 | 15.76 | 0.48 | 0.3 | 0.42 |