Foley Generation on VGGSound (test)
11.73FIDMel-QCD
Evaluation Results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mel-QCD2025.03 | 11.73 | 2.96 | 0.4591 | 0.33 | 0.11 | 0.2542 | 0.52 | 0.31 | — | — | — | — | — | — | — | |
| VTA-LDM2025.03 | 11.77 | 4.72 | 0.2772 | 0.37 | 0.11 | 0.2683 | 0.44 | 0.28 | — | — | — | — | — | — | — | |
| FoleyCrafter2025.03 | 13.11 | 4.14 | 0.3154 | 0.43 | 0.13 | 0.2433 | 0.48 | 0.29 | — | — | — | — | — | — | — | |
| DiffFoley2025.03 | 15.15 | 6.47 | 0.2327 | 0.49 | 0.14 | 0.1602 | 0.32 | 0.23 | — | — | — | — | — | — | — | |
| Im2Wav2025.03 | 16.16 | 5.66 | 0.167 | 0.45 | 0.13 | 0.221 | 0.31 | 0.2 | — | — | — | — | — | — | — | |
| SpecVQGAN2025.03 | 19.31 | 6.47 | 0.0564 | 0.45 | 0.1 | 0.2434 | 0.18 | 0.13 | — | — | — | — | — | — | — | |
| Seeing-and-Hearing2025.03 | 20.32 | 6.08 | 0.1056 | 0.68 | 0.17 | 0.2049 | 0.43 | 0.38 | — | — | — | — | — | — | — | |
| Auffusion2025.03 | 22.26 | 5.54 | 0.1836 | 0.44 | 0.12 | 0.223 | 0.35 | 0.22 | — | — | — | — | — | — | — | |
| AC-FoleyComparison Baseline=MMAudio-L-V22026.03 | — | — | — | — | — | — | — | — | — | — | — | — | 61.1 | 21.8 | 83.5 | |
| MultiFoleyConditioning=text control, Evaluation protocol=Human evaluation (2AFC) win rate against FoleyCrafter2024.11 | — | — | — | — | — | — | — | — | 85.8 | 94.5 | 86.5 | 90.2 | — | — | — |