# Speech Enhancement on LRS3 mixed with VGGSound noises (test)

Source: [Speech Enhancement on LRS3 mixed with VGGSound noises (test)](<https://www.sota2.com/research/sota/speech-enhancement-on-lrs3-mixed-with-vggsound-noises-test>)

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

Data updated: 2026-02-26T05:55:01.352Z

Evaluation of Speech Enhancement performance on the test split of the LRS3 dataset augmented with VGGSound noises, measured using metrics like PESQ, CSIG, CBAK, COVL, and STOI.

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 | Speech Enhancement |
| Dataset (including any stated split) | LRS3 mixed with VGGSound noises (test) |
| Metric | PESQ |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 18 of 18 |
| 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

- [PESQ](<https://www.sota2.com/research/sota/speech-enhancement-on-lrs3-mixed-with-vggsound-noises-test.md?metric=PESQ>): Higher is better; unit: SCALAR.

- [CSIG](<https://www.sota2.com/research/sota/speech-enhancement-on-lrs3-mixed-with-vggsound-noises-test.md?metric=CSIG>): Higher is better; unit: SCALAR.

- [CBAK](<https://www.sota2.com/research/sota/speech-enhancement-on-lrs3-mixed-with-vggsound-noises-test.md?metric=CBAK>): Higher is better; unit: SCALAR.

- [COVL](<https://www.sota2.com/research/sota/speech-enhancement-on-lrs3-mixed-with-vggsound-noises-test.md?metric=COVL>): Higher is better; unit: SCALAR.

- [STOI](<https://www.sota2.com/research/sota/speech-enhancement-on-lrs3-mixed-with-vggsound-noises-test.md?metric=STOI>): Higher is better; unit: SCALAR.

## Recorded results

| Method | PESQ | Result date | Report paper |
| --- | --- | --- | --- |
| Visual-stream method \[2\] | 3.25 | Not recorded | 68fa68f21e209131cea1f2df |
| pseudo-visual speech enhancement | 3.19 | Not recorded | 68fa68f21e209131cea1f2df |
| AO (Audio-Only) | 3.12 | Not recorded | 68fa68f21e209131cea1f2df |
| Visual-stream method \[2\] | 3.05 | Not recorded | 68fa68f21e209131cea1f2df |
| pseudo-visual speech enhancement | 2.99 | Not recorded | 68fa68f21e209131cea1f2df |
| Method \[11\] | 2.95 | Not recorded | 68fa68f21e209131cea1f2df |
| AO (Audio-Only) | 2.93 | Not recorded | 68fa68f21e209131cea1f2df |
| Visual-stream method \[2\] | 2.8 | Not recorded | 68fa68f21e209131cea1f2df |
| pseudo-visual speech enhancement | 2.72 | Not recorded | 68fa68f21e209131cea1f2df |
| Noisy | 2.66 | Not recorded | 68fa68f21e209131cea1f2df |
| Method \[19\] | 2.65 | Not recorded | 68fa68f21e209131cea1f2df |
| AO (Audio-Only) | 2.62 | Not recorded | 68fa68f21e209131cea1f2df |
| Method \[11\] | 2.52 | Not recorded | 68fa68f21e209131cea1f2df |
| Noisy | 2.29 | Not recorded | 68fa68f21e209131cea1f2df |
| Method \[19\] | 2.24 | Not recorded | 68fa68f21e209131cea1f2df |
| Method \[11\] | 2.17 | Not recorded | 68fa68f21e209131cea1f2df |
| Noisy | 1.93 | Not recorded | 68fa68f21e209131cea1f2df |
| Method \[19\] | 1.84 | Not recorded | 68fa68f21e209131cea1f2df |

### Visual-stream method \[2\]

PESQ: 3.25. Result ID: 698d8f258b517eb43322591d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa647b1e209131cea04ba1.md>)

Recorded evaluation configuration: {"SNR":"10db"}. This record does not assert independent reproduction.

### pseudo-visual speech enhancement

PESQ: 3.19. Result ID: 698d8f258b517eb43322591c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Implementation](<https://github.com/Sindhu-Hegde/pseudo-visual-speech-denoising>)

Recorded evaluation configuration: {"SNR":"10db"}. This record does not assert independent reproduction.

### AO (Audio-Only)

PESQ: 3.12. Result ID: 698d8f258b517eb43322591b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

Recorded evaluation configuration: {"SNR":"10db"}. This record does not assert independent reproduction.

### Visual-stream method \[2\]

PESQ: 3.05. Result ID: 698d8f258b517eb433225917.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa647b1e209131cea04ba1.md>)

Recorded evaluation configuration: {"SNR":"5db"}. This record does not assert independent reproduction.

### pseudo-visual speech enhancement

PESQ: 2.99. Result ID: 698d8f258b517eb433225916.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Implementation](<https://github.com/Sindhu-Hegde/pseudo-visual-speech-denoising>)

Recorded evaluation configuration: {"SNR":"5db"}. This record does not assert independent reproduction.

### Method \[11\]

PESQ: 2.95. Result ID: 698d8f258b517eb43322591a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64df1e209131cea06ee4.md>)

Recorded evaluation configuration: {"SNR":"10db"}. This record does not assert independent reproduction.

### AO (Audio-Only)

PESQ: 2.93. Result ID: 698d8f258b517eb433225915.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

Recorded evaluation configuration: {"SNR":"5db"}. This record does not assert independent reproduction.

### Visual-stream method \[2\]

PESQ: 2.8. Result ID: 698d8f258b517eb433225911.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa647b1e209131cea04ba1.md>)

Recorded evaluation configuration: {"SNR":"0db"}. This record does not assert independent reproduction.

### pseudo-visual speech enhancement

PESQ: 2.72. Result ID: 698d8f258b517eb433225910.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Implementation](<https://github.com/Sindhu-Hegde/pseudo-visual-speech-denoising>)

Recorded evaluation configuration: {"SNR":"0db"}. This record does not assert independent reproduction.

### Noisy

PESQ: 2.66. Result ID: 698d8f258b517eb433225918.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

Recorded evaluation configuration: {"SNR":"10db"}. This record does not assert independent reproduction.

### Method \[19\]

PESQ: 2.65. Result ID: 698d8f258b517eb433225919.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63651e209131ce9fe5dc.md>)

Recorded evaluation configuration: {"SNR":"10db"}. This record does not assert independent reproduction.

### AO (Audio-Only)

PESQ: 2.62. Result ID: 698d8f258b517eb43322590f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

Recorded evaluation configuration: {"SNR":"0db"}. This record does not assert independent reproduction.

### Method \[11\]

PESQ: 2.52. Result ID: 698d8f258b517eb433225914.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64df1e209131cea06ee4.md>)

Recorded evaluation configuration: {"SNR":"5db"}. This record does not assert independent reproduction.

### Noisy

PESQ: 2.29. Result ID: 698d8f258b517eb433225912.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

Recorded evaluation configuration: {"SNR":"5db"}. This record does not assert independent reproduction.

### Method \[19\]

PESQ: 2.24. Result ID: 698d8f258b517eb433225913.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63651e209131ce9fe5dc.md>)

Recorded evaluation configuration: {"SNR":"5db"}. This record does not assert independent reproduction.

### Method \[11\]

PESQ: 2.17. Result ID: 698d8f258b517eb43322590e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64df1e209131cea06ee4.md>)

Recorded evaluation configuration: {"SNR":"0db"}. This record does not assert independent reproduction.

### Noisy

PESQ: 1.93. Result ID: 698d8f258b517eb43322590c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

Recorded evaluation configuration: {"SNR":"0db"}. This record does not assert independent reproduction.

### Method \[19\]

PESQ: 1.84. Result ID: 698d8f258b517eb43322590d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68f21e209131cea1f2df.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63651e209131ce9fe5dc.md>)

Recorded evaluation configuration: {"SNR":"0db"}. This record does not assert independent reproduction.

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