Text-to-Speech Synthesis on Multi-Source Test Set (LibriTTS/LibriVox/YouTube/KIDS)
3.7MOSGround Truth
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Ground Truth2025.06 | 3.7 | 0 | 4.37 | 93.21 | |
| E2-TTS2025.06 | 3.41 | -0.23 | 4.37 | 91.73 | |
| Vevo2025.06 | 3.36 | 0.08 | 4.01 | 90.2 | |
| F5-TTS2025.06 | 3.33 | -0.34 | 4.1 | 91.16 | |
| MaskGCT2025.06 | 3.28 | -0.17 | 4.39 | 91.76 | |
| FishSpeech2025.06 | 3.24 | -0.43 | 3.58 | 89.88 | |
| TorToiSe2025.06 | 3.22 | -0.57 | 2.73 | 88.95 | |
| VoiceCraft2025.06 | 3.15 | -0.44 | 3.66 | 88.3 | |
| WhisperSpeech2025.06 | 3.12 | -0.73 | 2.68 | 87.91 | |
| HierSpeech++2025.06 | 3.08 | -0.86 | 3.48 | 88.63 | |
| StyleTTS22025.06 | 3.01 | -0.66 | 2.93 | 85.87 | |
| Pheme2025.06 | 2.99 | -1 | 3.35 | 88.84 | |
| OpenVoice2025.06 | 2.92 | -1.21 | 2.59 | 83.32 | |
| VALL-E2025.06 | 2.9 | -0.6 | 3.43 | 83.59 | |
| GPTSoVITS2025.06 | 2.81 | -0.57 | 3.82 | 89.22 | |
| XTTS2025.06 | 2.77 | -0.73 | 2.57 | 88.2 | |
| MetaVoice2025.06 | 2.49 | -1.23 | 2.18 | 87.38 | |
| Bark2025.06 | 2.49 | -1.12 | 2.51 | 85.21 | |
| ParlerTTS2025.06 | 2.39 | -1.19 | 2.93 | 84.88 | |
| NaturalSpeech22025.06 | 2.05 | -1.42 | 2.06 | 81.71 | |
| SpeechT52025.06 | 1.98 | -1.56 | 2.63 | 84.84 |