Speech Editing on EARS-WHAM Long Replacement (noisy)
0.1WERVoiceCraft
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VoiceCraftCondition=Clean2025.05 | 0.1 | 0.54 | 4.32 | 4.23 | |
| Ground TruthCondition=Noisy2025.05 | 0.12 | 0.51 | 3.97 | 3.82 | |
| FluentSpeechCondition=Noisy2025.05 | 0.22 | 1.84 | 1.5 | 1.49 | |
| SeamlessEditCondition=Noisy, In-Context Learning (ICL)=true2025.05 | 0.22 | 0.75 | 3.65 | 3.11 | |
| VoiceCraftCondition=Noisy2025.05 | 0.23 | 1.16 | 2.93 | 2.6 | |
| SeamlessEditCondition=Noisy, In-Context Learning (ICL)=false2025.05 | 0.24 | 0.86 | 3.42 | 3.13 |