Audio-driven avatar generation on HDTF
0.9192CSIMVidu S1
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Vidu S1Instruction Following=true, Real-Time=true, Resolution=540p (960×540)2026.07 | 0.9192 | 7.847 | 0.566 | 42 | — | |
| HeyGenInstruction Following=false, Real-Time=true2026.07 | 0.9191 | 8.037 | 0.4864 | 25 | — | |
| Kling Avatar 2.0Instruction Following=true, Real-Time=false2026.07 | 0.8688 | 8.158 | 0.5406 | — | — | |
| LemonSliceInstruction Following=false, Real-Time=true, Resolution=368×5602026.07 | 0.8407 | 7.921 | 0.5196 | — | — | |
| StableAvatar-1.3BInstruction Following=false, Real-Time=false, Resolution=480p2026.07 | 0.8358 | 11.18 | 0.556 | — | — | |
| LiveAvatarInstruction Following=false, Real-Time=true2026.07 | 0.8127 | 8.447 | 0.5639 | — | — | |
| OmniAvatarInstruction Following=false, Real-Time=false, Resolution=480p2026.07 | 0.8062 | 9.242 | 0.5476 | — | — | |
| Wan2.2-S2V-14BInstruction Following=false, Real-Time=false, Resolution=480p/720p2026.07 | 0.7936 | 8.255 | 0.551 | — | — | |
| Hallo3Instruction Following=false, Real-Time=false, Resolution=480×7202026.07 | 0.7698 | 8.66 | 0.5313 | — | — |