Role-Play Text-to-Speech on Role-Play TTS Curated Evaluation Set (Subjective)
4.411MOSGround Truth
Evaluation results
| Method | Links | |
|---|---|---|
| Ground Truthresynthesized_via_Step-Audio-2_tokenizer=true2026.01 | 4.411 | |
| Our Proposed2026.01 | 3.646 | |
| MiMo-Audio-7B-Instructresynthesized_via_Step-Audio-2_tokenizer=true2026.01 | 2.471 | |
| GPT-Audioresynthesized_via_Step-Audio-2_tokenizer=true2026.01 | 1.752 | |
| Step-Audio-2-mini2026.01 | 1.707 |