Text-to-Sounding-Video Generation on AVSync15
765.74FVDBridgeDiT
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| BridgeDiTProcess=T → VA2025.10 | 765.74 | 21.33 | 5.34 | 1.3 | 28.52 | 35.95 | 34.59 | 27.5 | |
| Wan + SDAProcess=T → V ∥ T → A2025.10 | 828.33 | 22.56 | 11.9 | 3.17 | 28.12 | 30.78 | 26.22 | 20.5 | |
| Wan + MMAudioProcess=T → V + V → A2025.10 | 828.33 | 22.56 | 7.98 | 1.4 | 28.12 | 34.64 | 33.5 | 24.3 | |
| Wan + SeeingHearingProcess=T → V + V → A2025.10 | 828.33 | 22.56 | 14.21 | 2.89 | 28.12 | 25.52 | 35.87 | 20.8 | |
| Wan + ThinkSoundProcess=T → V + V → A2025.10 | 828.33 | 22.56 | 17.31 | 5.58 | 28.12 | 30.92 | 20.7 | 16.9 | |
| Wan + PrismAudioProcess=T → V + V → A2025.10 | 828.33 | 22.56 | 9.79 | 2.34 | 28.12 | 33.53 | 25.25 | 23.3 | |
| JavisDiTProcess=T → VA2025.10 | 878.7 | 23.23 | 13.48 | 3.5 | 28.05 | 22.99 | 20.75 | 15.8 | |
| MTVProcess=T → VA2025.10 | 982.09 | 24.6 | 16.46 | 3.53 | 27.66 | 21.22 | 15.84 | 14.9 | |
| JointDiTProcess=T → I + I → VA2025.10 | 992.71 | 25.2 | 6.51 | 1.77 | 29.94 | 30.34 | 34.17 | 15.6 | |
| SSVGProcess=T → VA2025.10 | 1,028.78 | 31.44 | 14.35 | 4.35 | 25.78 | 23.67 | 22.45 | 12.6 | |
| CoDiProcess=T → VA2025.10 | 1,387.14 | 39.24 | 16.56 | 5.24 | 24.96 | 17.94 | 10.79 | 8.1 | |
| SDA + TempoTokenProcess=T → A + A → V2025.10 | 1,516.53 | 42.3 | 11.9 | 3.17 | 20.56 | 30.78 | 17.63 | 21.5 | |
| SDA + TPosProcess=T → A + A → V2025.10 | 1,975.22 | 92.95 | 11.9 | 3.17 | 18.64 | 30.78 | 25.45 | 17.6 |