# Audio Source Separation on MoXaRt XR Scenarios (human speech sessions)

Source: [Audio Source Separation on MoXaRt XR Scenarios (human speech sessions)](<https://www.sota2.com/research/sota/audio-source-separation-on-moxart-xr-scenarios-human-speech-sessions>)

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

Data updated: 2026-03-12T08:44:00.207Z

Evaluation of Audio Source Separation performance on the MoXaRt XR Scenarios dataset, focusing on human speech sessions, measured using WER and DNSMOS quality metrics.

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 | Audio Source Separation |
| Dataset (including any stated split) | MoXaRt XR Scenarios (human speech sessions) |
| Metric | WER |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 7 of 7 |
| 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

- [WER](<https://www.sota2.com/research/sota/audio-source-separation-on-moxart-xr-scenarios-human-speech-sessions.md?metric=WER>): Lower is better; unit: SCALAR.

- [DNSMOS-SIG](<https://www.sota2.com/research/sota/audio-source-separation-on-moxart-xr-scenarios-human-speech-sessions.md?metric=DNSMOS-SIG>): Higher is better; unit: SCALAR.

- [DNSMOS-BAK](<https://www.sota2.com/research/sota/audio-source-separation-on-moxart-xr-scenarios-human-speech-sessions.md?metric=DNSMOS-BAK>): Higher is better; unit: SCALAR.

- [DNSMOS-OVRL](<https://www.sota2.com/research/sota/audio-source-separation-on-moxart-xr-scenarios-human-speech-sessions.md?metric=DNSMOS-OVRL>): Higher is better; unit: SCALAR.

## Recorded results

| Method | WER | Result date | Report paper |
| --- | --- | --- | --- |
| MoXaRt | 0.3824 | Not recorded | 69b248a8d030084c22e2f9d9 |
| AV-Mossformer2 | 0.3956 | Not recorded | 69b248a8d030084c22e2f9d9 |
| MoXaRt | 0.499 | Not recorded | 69b248a8d030084c22e2f9d9 |
| AudioScopeV2 | 0.5263 | Not recorded | 69b248a8d030084c22e2f9d9 |
| Raw Audio | 0.6297 | Not recorded | 69b248a8d030084c22e2f9d9 |
| DAVIS | 0.7259 | Not recorded | 69b248a8d030084c22e2f9d9 |
| Sound of Pixels | 0.8636 | Not recorded | 69b248a8d030084c22e2f9d9 |

### MoXaRt

WER: 0.3824. Result ID: 69b27bf66dbb7075b95d241b.

[Reporting paper](<https://www.sota2.com/research/paper/69b248a8d030084c22e2f9d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b248a8d030084c22e2f9d9.md>)

Recorded evaluation configuration: {"Processing Mode":"Offline"}. This record does not assert independent reproduction.

### AV-Mossformer2

WER: 0.3956. Result ID: 69b27bf66dbb7075b95d241a.

[Reporting paper](<https://www.sota2.com/research/paper/69b248a8d030084c22e2f9d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75eb1e209131cea6a9f5.md>)

[Implementation](<https://github.com/alibabasglab/MossFormer2>)

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

### MoXaRt

WER: 0.499. Result ID: 69b27bf66dbb7075b95d241c.

[Reporting paper](<https://www.sota2.com/research/paper/69b248a8d030084c22e2f9d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b248a8d030084c22e2f9d9.md>)

Recorded evaluation configuration: {"Processing Mode":"Real-Time"}. This record does not assert independent reproduction.

### AudioScopeV2

WER: 0.5263. Result ID: 69b27bf66dbb7075b95d2419.

[Reporting paper](<https://www.sota2.com/research/paper/69b248a8d030084c22e2f9d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d1a1e209131cea37c58.md>)

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

### Raw Audio

WER: 0.6297. Result ID: 69b27bf66dbb7075b95d2416.

[Reporting paper](<https://www.sota2.com/research/paper/69b248a8d030084c22e2f9d9.md>)

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

### DAVIS

WER: 0.7259. Result ID: 69b27bf66dbb7075b95d2418.

[Reporting paper](<https://www.sota2.com/research/paper/69b248a8d030084c22e2f9d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75351e209131cea66710.md>)

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

### Sound of Pixels

WER: 0.8636. Result ID: 69b27bf66dbb7075b95d2417.

[Reporting paper](<https://www.sota2.com/research/paper/69b248a8d030084c22e2f9d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64f91e209131cea077c6.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.
