Room Impulse Response (RIR) Prediction on Matterport3D (Seen environments)
1.1STFTFEW-SHOTRIR
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| FEW-SHOTRIRAcoustic echoes=true, Visual input=true, LD loss=true2022.06 | 1.1 | 0.43 | 106 | 8.66 | |
| Fast-RIR++2022.06 | 1.37 | 1.25 | 137 | 13.7 | |
| FEW-SHOTRIRLD loss=false2022.06 | 1.39 | 1.6 | 347 | 14 | |
| FEW-SHOTRIRAcoustic echoes=false2022.06 | 1.63 | 0.63 | 334 | 19.5 | |
| FEW-SHOTRIRVisual input=false2022.06 | 1.63 | 0.67 | 332 | 19.2 | |
| FEW-SHOTRIRN (Number of observation poses)=12022.06 | 1.69 | 0.74 | 362 | 17.2 | |
| AnalyticalRIR++2022.06 | 2.94 | 0.98 | 463 | 28.1 | |
| Linear Interpolation2022.06 | 4.44 | 1.22 | 393 | 24.3 | |
| Nearest Neighbor2022.06 | 4.65 | 1.15 | 385 | 24.4 |