Audio Generation on Voice Bank corpus (VCTK)
Metric: Log-Spectral Distance (higher is better)
Metric: Log-Spectral Distance (higher is better)
| # | Model↕ | Log-Spectral Distance▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | U-Net | 3.2 | Yes | Audio Super Resolution using Neural Networks | 2017-08-02 | Code |
| 2 | U-Net + TFiLM | 2.5 | Yes | Temporal FiLM: Capturing Long-Range Sequence Dep... | 2019-09-14 | Code |
| 3 | U-Net + AFiLM | 2.3 | Yes | Self-Attention for Audio Super-Resolution | 2021-08-26 | Code |