# Role-Play Text-to-Speech on Role-Play TTS Curated Evaluation Set (With Audio History)

Source: [Role-Play Text-to-Speech on Role-Play TTS Curated Evaluation Set (With Audio History)](<https://www.sota2.com/research/sota/role-play-text-to-speech-on-role-play-tts-curated-evaluation-set-with-audio-history>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:38:16.657Z

Evaluation of Role-Play Text-to-Speech synthesis quality using character error rate (CER), Pinyin WER, and MCLP metrics on the curated evaluation set that incorporates audio history context.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Role-Play Text-to-Speech |
| Dataset (including any stated split) | Role-Play TTS Curated Evaluation Set (With Audio History) |
| Metric | CER |
| Unit | PERCENT |
| Direction | Lower is better |
| Results loaded | 4 of 4 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [CER](<https://www.sota2.com/research/sota/role-play-text-to-speech-on-role-play-tts-curated-evaluation-set-with-audio-history.md?metric=CER>): Lower is better; unit: PERCENT.

- [Pinyin WER](<https://www.sota2.com/research/sota/role-play-text-to-speech-on-role-play-tts-curated-evaluation-set-with-audio-history.md?metric=Pinyin%20WER>): Lower is better; unit: PERCENT.

- [MCLP](<https://www.sota2.com/research/sota/role-play-text-to-speech-on-role-play-tts-curated-evaluation-set-with-audio-history.md?metric=MCLP>): Higher is better; unit: SCALAR.

## Recorded results

| Method | CER | Result date | Report paper |
| --- | --- | --- | --- |
| Our Proposed | 1.13 | Not recorded | 69802f69a8b01e76cc8e14e8 |
| Step-Audio-2-mini | 3.276 | Not recorded | 69802f69a8b01e76cc8e14e8 |
| MiMo-Audio-7B-Instruct | 10.605 | Not recorded | 69802f69a8b01e76cc8e14e8 |
| GPT-Audio | 11.974 | Not recorded | 69802f69a8b01e76cc8e14e8 |

### Our Proposed

CER: 1.13. Result ID: 699a006d5698665ec48272fd.

[Reporting paper](<https://www.sota2.com/research/paper/69802f69a8b01e76cc8e14e8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69802f69a8b01e76cc8e14e8.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Step-Audio-2-mini

CER: 3.276. Result ID: 699a006d5698665ec48272fc.

[Reporting paper](<https://www.sota2.com/research/paper/69802f69a8b01e76cc8e14e8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b021e209131cea8635d.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### MiMo-Audio-7B-Instruct

CER: 10.605. Result ID: 699a006d5698665ec48272fb.

[Reporting paper](<https://www.sota2.com/research/paper/69802f69a8b01e76cc8e14e8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6955ff7e53fdd5c36173df40.md>)

Recorded evaluation configuration: {"resynthesized\_via\_Step-Audio-2\_tokenizer":true}. This record does not assert independent reproduction.

### GPT-Audio

CER: 11.974. Result ID: 699a006d5698665ec48272fa.

[Reporting paper](<https://www.sota2.com/research/paper/69802f69a8b01e76cc8e14e8.md>)

Recorded evaluation configuration: {"resynthesized\_via\_Step-Audio-2\_tokenizer":true}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.