Table 2 Quantitative results of Points2Sound and Mono2Binaural. For each method, we report the performance depending on the number of sources (\(N = 1,2,3\)) and the type of 3D point cloud attributes (depth or rgb-depth), based on the average of the evaluation metrics. Average values for any number of sources are given by\(\overline{\mathrm {d}_\mathrm {ENV}}\)and\(\overline{\mathrm {d}_\mathrm {STFT}}\)
From: Points2Sound: from mono to binaural audio using 3D point cloud scenes
Visual features | \(\mathrm {d}_\mathrm {ENV} \downarrow\) | \(\mathrm {d}_\mathrm {STFT} \downarrow\) | \(\overline{\mathrm {d}_\mathrm {ENV}} \downarrow\) | \(\overline{\mathrm {d}_\mathrm {STFT}} \downarrow\) | |||||
|---|---|---|---|---|---|---|---|---|---|
1 | 2 | 3 | 1 | 2 | 3 | ||||
Mono2Binaural [7] | Depth | 0.038 | 0.101 | 0.132 | 0.192 | 1.300 | 2.029 | 0.090 | 1.174 |
rgb-depth | 0.036 | 0.094 | 0.126 | 0.213 | 1.142 | 1.858 | 0.085 | 1.071 | |
Points2Sound | Depth | 0.016 | 0.080 | 0.122 | 0.082 | 0.885 | 1.736 | 0.072 | 0.901 |
rgb-depth | 0.015 | 0.073 | 0.114 | 0.099 | 0.762 | 1.521 | 0.067 | 0.794 | |