# 5-speaker speech separation on WSJ0-5mix (evaluation)

Source: [5-speaker speech separation on WSJ0-5mix (evaluation)](<https://www.sota2.com/research/sota/5-speaker-speech-separation-on-wsj0-5mix-evaluation>)

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

Data updated: 2026-02-26T06:02:51.535Z

Evaluation of 5-speaker speech separation performance on the evaluation split of the WSJ0-5mix dataset, measured using SDRi.

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 | 5-speaker speech separation |
| Dataset (including any stated split) | WSJ0-5mix (evaluation) |
| Metric | SDRi |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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

- [SDRi](<https://www.sota2.com/research/sota/5-speaker-speech-separation-on-wsj0-5mix-evaluation.md?metric=SDRi>): Higher is better; unit: SCALAR.

## Recorded results

| Method | SDRi | Result date | Report paper |
| --- | --- | --- | --- |
| TasTas | 11.14 | Not recorded | 68fa6fe31e209131cea47ee9 |
| Nachmani's | 10.5 | Not recorded | 68fa6fe31e209131cea47ee9 |
| TasTas | 10.41 | Not recorded | 68fa6fe31e209131cea47ee9 |
| IRM | 9.6 | Not recorded | 68fa6fe31e209131cea47ee9 |
| DPRNN-TasNet | 8.4 | Not recorded | 68fa6fe31e209131cea47ee9 |

### TasTas

SDRi: 11.14. Result ID: 6990f0583a6c314849047044.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fe31e209131cea47ee9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fe31e209131cea47ee9.md>)

[Implementation](<https://github.com/ShiZiqiang/dual-path-RNNs-DPRNNs-based-speech-separation>)

Recorded evaluation configuration: {"Identity-aware":"true","blocks\_stage\_1":"8","blocks\_stage\_2":"9","online remixing":"true"}. This record does not assert independent reproduction.

### Nachmani's

SDRi: 10.5. Result ID: 6990f0583a6c314849047042.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fe31e209131cea47ee9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa684b1e209131cea1b2cb.md>)

[Implementation](<https://enk100.github.io/speaker_separation>)

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

### TasTas

SDRi: 10.41. Result ID: 6990f0583a6c314849047043.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fe31e209131cea47ee9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fe31e209131cea47ee9.md>)

[Implementation](<https://github.com/ShiZiqiang/dual-path-RNNs-DPRNNs-based-speech-separation>)

Recorded evaluation configuration: {"Identity-aware":"true","blocks\_stage\_1":"6","blocks\_stage\_2":"6"}. This record does not assert independent reproduction.

### IRM

SDRi: 9.6. Result ID: 6990f0583a6c314849047040.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fe31e209131cea47ee9.md>)

Recorded evaluation configuration: {"Description":"Ideal Ratio Mask (Upper bound for STFT based methods)"}. This record does not assert independent reproduction.

### DPRNN-TasNet

SDRi: 8.4. Result ID: 6990f0583a6c314849047041.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6fe31e209131cea47ee9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67421e209131cea15525.md>)

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

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