Talking Head Generation on Verse-Bench
1.62LSE-CUniverse
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| UniverseGeneration Mode=Joint Video-Audio Generation2025.12 | 1.62 | 0.18 | 1.25 | 2.7 | 0.16 | 3.53 | 4.61 | 2.49 | 5.2 | 0.2 | 0.47 | 0.89 | |
| FantasyTalkingGeneration Mode=Audio-driven Generation2025.12 | 2.68 | — | — | — | — | — | — | — | — | 0.07 | 0.42 | 0.87 | |
| Wan-S2VGeneration Mode=Audio-driven Generation2025.12 | 6.49 | — | — | — | — | — | — | — | — | 0.17 | 0.46 | 0.89 | |
| JoVAGeneration Mode=Joint Video-Audio Generation2025.12 | 6.51 | 0.11 | 1.1 | 2.16 | 0.23 | 4 | 5.24 | 2.23 | 5.61 | 0.8 | 0.48 | 0.86 | |
| OVIGeneration Mode=Joint Video-Audio Generation2025.12 | 6.61 | 0.12 | 0.97 | 2.23 | 0.2 | 3.95 | 5.6 | 2.09 | 6.2 | 0.55 | 0.47 | 0.86 |