# Dysarthric Speech Naturalness Assessment on SAP (Speech Accessibility Project)

Source: [Dysarthric Speech Naturalness Assessment on SAP (Speech Accessibility Project)](<https://www.sota2.com/research/sota/dysarthric-speech-naturalness-assessment-on-sap-speech-accessibility-project>)

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

Data updated: 2026-06-18T07:42:21.685Z

Evaluation of models assessing the naturalness of dysarthric speech using the SAP (Speech Accessibility Project) dataset, measured by Mean Squared Error (MSE), Linear Correlation Coefficient (LCC), and Spearman's Rank Correlation Coefficient (SRCC).

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 | Dysarthric Speech Naturalness Assessment |
| Dataset (including any stated split) | SAP (Speech Accessibility Project) |
| Metric | MSE |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 25 of 25 |
| 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

- [MSE](<https://www.sota2.com/research/sota/dysarthric-speech-naturalness-assessment-on-sap-speech-accessibility-project.md?metric=MSE>): Lower is better; unit: SCALAR.

- [LCC](<https://www.sota2.com/research/sota/dysarthric-speech-naturalness-assessment-on-sap-speech-accessibility-project.md?metric=LCC>): Higher is better; unit: SCALAR.

- [SRCC](<https://www.sota2.com/research/sota/dysarthric-speech-naturalness-assessment-on-sap-speech-accessibility-project.md?metric=SRCC>): Higher is better; unit: SCALAR.

## Recorded results

| Method | MSE | Result date | Report paper |
| --- | --- | --- | --- |
| Fine-Tuning (FT) | 0.717 | Not recorded | 6a337bbc206542251d7c5184 |
| Fine-Tuning (FT) | 0.8 | Not recorded | 6a337bbc206542251d7c5184 |
| Fine-Tuning (FT) | 0.819 | Not recorded | 6a337bbc206542251d7c5184 |
| Fine-Tuning (FT) | 0.823 | Not recorded | 6a337bbc206542251d7c5184 |
| Fine-Tuning (FT) | 0.847 | Not recorded | 6a337bbc206542251d7c5184 |
| Fine-Tuning (FT) | 0.88 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 0.909 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 0.93 | Not recorded | 6a337bbc206542251d7c5184 |
| IDT (In-Domain Training) | 0.941 | Not recorded | 6a337bbc206542251d7c5184 |
| Fine-Tuning (FT) | 1 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 1.022 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 1.022 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 1.027 | Not recorded | 6a337bbc206542251d7c5184 |
| Fine-Tuning (FT) | 1.033 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 1.036 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 1.05 | Not recorded | 6a337bbc206542251d7c5184 |
| Fine-Tuning (FT) | 1.053 | Not recorded | 6a337bbc206542251d7c5184 |
| Fine-Tuning (FT) | 1.075 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 1.085 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 1.106 | Not recorded | 6a337bbc206542251d7c5184 |
| IDT (In-Domain Training) | 1.119 | Not recorded | 6a337bbc206542251d7c5184 |
| Joint Training (JT) | 1.121 | Not recorded | 6a337bbc206542251d7c5184 |
| IDT (In-Domain Training) | 1.127 | Not recorded | 6a337bbc206542251d7c5184 |
| IDT (In-Domain Training) | 1.169 | Not recorded | 6a337bbc206542251d7c5184 |
| IDT (In-Domain Training) | 1.354 | Not recorded | 6a337bbc206542251d7c5184 |

### Fine-Tuning (FT)

MSE: 0.717. Result ID: 6a33a10dfbc373592c309b98.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Base","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### Fine-Tuning (FT)

MSE: 0.8. Result ID: 6a33a10dfbc373592c309b99.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large\*","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### Fine-Tuning (FT)

MSE: 0.819. Result ID: 6a33a10dfbc373592c309b92.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Large","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### Fine-Tuning (FT)

MSE: 0.823. Result ID: 6a33a10dfbc373592c309b9c.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Large","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### Fine-Tuning (FT)

MSE: 0.847. Result ID: 6a33a10dfbc373592c309b91.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Base","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### Fine-Tuning (FT)

MSE: 0.88. Result ID: 6a33a10dfbc373592c309b9a.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large+","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 0.909. Result ID: 6a33a10dfbc373592c309b93.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Base","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 0.93. Result ID: 6a33a10dfbc373592c309b97.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Large","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### IDT (In-Domain Training)

MSE: 0.941. Result ID: 6a33a10dfbc373592c309b8d.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Large","QualiSpeech Auxiliary Supervision":"–"}. This record does not assert independent reproduction.

### Fine-Tuning (FT)

MSE: 1. Result ID: 6a33a10dfbc373592c309b90.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large+","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 1.022. Result ID: 6a33a10dfbc373592c309b9d.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Base","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 1.022. Result ID: 6a33a10dfbc373592c309b9f.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large+","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 1.027. Result ID: 6a33a10dfbc373592c309b94.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large\*","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### Fine-Tuning (FT)

MSE: 1.033. Result ID: 6a33a10dfbc373592c309b8f.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large\*","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 1.036. Result ID: 6a33a10dfbc373592c309ba1.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Large","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 1.05. Result ID: 6a33a10dfbc373592c309b96.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Base","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### Fine-Tuning (FT)

MSE: 1.053. Result ID: 6a33a10dfbc373592c309b8e.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Base","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### Fine-Tuning (FT)

MSE: 1.075. Result ID: 6a33a10dfbc373592c309b9b.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Base","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 1.085. Result ID: 6a33a10dfbc373592c309b9e.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large\*","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 1.106. Result ID: 6a33a10dfbc373592c309b95.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large+","QualiSpeech Auxiliary Supervision":"Overall"}. This record does not assert independent reproduction.

### IDT (In-Domain Training)

MSE: 1.119. Result ID: 6a33a10dfbc373592c309b8b.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large+","QualiSpeech Auxiliary Supervision":"–"}. This record does not assert independent reproduction.

### Joint Training (JT)

MSE: 1.121. Result ID: 6a33a10dfbc373592c309ba0.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Base","QualiSpeech Auxiliary Supervision":"Naturalness"}. This record does not assert independent reproduction.

### IDT (In-Domain Training)

MSE: 1.127. Result ID: 6a33a10dfbc373592c309b89.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Base","QualiSpeech Auxiliary Supervision":"–"}. This record does not assert independent reproduction.

### IDT (In-Domain Training)

MSE: 1.169. Result ID: 6a33a10dfbc373592c309b8c.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"HuBERT Base","QualiSpeech Auxiliary Supervision":"–"}. This record does not assert independent reproduction.

### IDT (In-Domain Training)

MSE: 1.354. Result ID: 6a33a10dfbc373592c309b8a.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bbc206542251d7c5184.md>)

Recorded evaluation configuration: {"SSL Encoder":"wav2vec 2.0 Large\*","QualiSpeech Auxiliary Supervision":"–"}. This record does not assert independent reproduction.

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