Speech Synthesis on North American English
Metric: Mean Opinion Score (higher is better)
Metric: Mean Opinion Score (higher is better)
| # | Model↕ | Mean Opinion Score▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | Tacotron 2 | 4.526 | No | Natural TTS Synthesis by Conditioning WaveNet on... | 2017-12-16 | Code |
| 2 | WaveNet (Linguistic) | 4.341 | No | Natural TTS Synthesis by Conditioning WaveNet on... | 2017-12-16 | Code |
| 3 | WaveNet (L+F) | 4.21 | No | WaveNet: A Generative Model for Raw Audio | 2016-09-12 | Code |
| 4 | Tacotron | 4.001 | No | Tacotron: Towards End-to-End Speech Synthesis | 2017-03-29 | Code |
| 5 | HMM-driven concatenative | 3.86 | No | WaveNet: A Generative Model for Raw Audio | 2016-09-12 | Code |
| 6 | LSTM-RNN parametric | 3.67 | No | WaveNet: A Generative Model for Raw Audio | 2016-09-12 | Code |
| 7 | means | 0 | No | - | - | - |