Text-to-Binaural Audio Generation on SpatialTAS (test)
4.93FDTAS
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| TAStext conditions=true, spatial coherence augmentation=true2025.06 | 4.93 | 1.44 | 0.58 | 2.23 | 3.07 | 2.45 | 6.99 | 8.16 | |
| TAS w/o Flippertext conditions=true, spatial coherence augmentation=false2025.06 | 5.08 | 1.72 | 0.61 | 2.15 | 4.14 | 2.89 | 8.63 | 10.03 | |
| TAS w/o texttext conditions=false, spatial coherence augmentation=true2025.06 | 6.77 | 2.54 | 0.63 | 2 | 5.87 | 4.03 | 9.25 | 11.4 | |
| PseudoBinauralre-trained on SpatialTAS=true2025.06 | 7.23 | 2.81 | 0.65 | 1.85 | 6.39 | 4 | 10.36 | 12.91 | |
| Mono-MonoDescription=duplicating the mono audio2025.06 | 9.03 | 3.67 | 0.99 | 1.61 | 19.66 | 18.12 | 12.79 | 15.33 |