Intelligible Audio Generation on AC-Filtered (test)
0.523CLAP ScoreGround Truth
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Ground Truth2025.10 | 0.523 | — | — | — | — | 17.47 | 4.16 | 4.45 | 4.5 | |
| ControlAudio2025.10 | 0.513 | 3.52 | 1.45 | 32.55 | 4.43 | 6.84 | 4.31 | 4.15 | 3.82 | |
| VoiceLDM-Stiming support=LLM-predicted2025.10 | 0.479 | 4.46 | 1.52 | 47.08 | 3.4 | 43.21 | 2.62 | 2.55 | 2.51 | |
| VoiceLDM-Mtiming support=LLM-predicted2025.10 | 0.458 | 5.9 | 1.43 | 46.4 | 3.16 | 8.84 | 4.18 | 3.64 | 3.47 | |
| VoiceDiT2025.10 | 0.22 | 4.6 | — | — | — | 7.09 | — | — | — | |
| AudioLDM 2 Speechtiming support=LLM-predicted2025.10 | 0.078 | 23.55 | 3.58 | 102.84 | 1.52 | 32.74 | 2.85 | 1.92 | 1.6 |