Pitch+Intensity-controlled Audio Generation on Song Describer Dataset (test)
77.86FDopenl3LatCH-F
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| LatCH-FControl(s)=Pitch+Intensity2026.03 | 77.86 | 0.58 | 37.36 | 15.27 | 0.057 | 21.3 | 5.8 | 3.7 | 3.9 | 1.7 | 1.5 | |
| ReadoutControl(s)=Pitch+Intensity2026.03 | 104.45 | 0.53 | 38.34 | 1.59 | 0.047 | 16.1 | 5.69 | 3.5 | 3.8 | 3.1 | 3.5 | |
| LatCH-BControl(s)=Pitch+Intensity2026.03 | 125.7 | 0.49 | 35.46 | 3.35 | 0.03 | 19.5 | 5.69 | 3.1 | 4 | 4.7 | 3.8 | |
| End-to-endControl(s)=Pitch+Intensity2026.03 | 147.64 | 0.65 | 33.95 | 1.76 | 0.033 | 261.1 | 37.23 | 3.8 | 4.1 | 3.5 | 4.3 |