Video-to-Audio Generation on Kling-Eval (test)
474.56FDPaSSTV-AURA
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| V-AURAGeneration Approach=Autoregressive-based2025.12 | 474.56 | 33.15 | 3.24 | 5.8 | 25 | 0.86 | 0.13 | |
| FoleyCrafterGeneration Approach=Diffusion-based2025.12 | 322.63 | 22.3 | 2.47 | 7.08 | 22 | 1.23 | 0.22 | |
| FrierenGeneration Approach=Diffusion-based2025.12 | 293.57 | 16.86 | 2.95 | 7.32 | 21 | 0.86 | 0.16 | |
| ThinkSound(w/o. CoT)Generation Approach=Diffusion-based2025.12 | 228.68 | 9.92 | 2.39 | 6.86 | 22 | 0.67 | 0.22 | |
| MMAudioGeneration Approach=Diffusion-based2025.12 | 205.85 | 9.01 | 2.17 | 9.59 | 30 | 0.56 | 0.27 | |
| HunyuanVideo-FoleyGeneration Approach=Diffusion-based2025.12 | 202.12 | 6.07 | 1.89 | 8.3 | 38 | 0.54 | 0.24 | |
| DreamFoleyGeneration Approach=Autoregressive-based2025.12 | 127.73 | 6.88 | 1.65 | 8.63 | 31.2 | 0.6 | 0.32 |