# Speaker Diarization on VoxCeleb2 (held-out)

Source: [Speaker Diarization on VoxCeleb2 (held-out)](<https://www.sota2.com/research/sota/speaker-diarization-on-voxceleb2-held-out>)

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

Data updated: 2026-06-02T12:27:28.893Z

Evaluation of Speaker Diarization performance on the held-out subset of the VoxCeleb2 dataset, measured using metrics like DER, Predicted K, and Overlap detection.

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 | Speaker Diarization |
| Dataset (including any stated split) | VoxCeleb2 (held-out) |
| Metric | Predicted K |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| 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

- [Predicted K](<https://www.sota2.com/research/sota/speaker-diarization-on-voxceleb2-held-out.md?metric=Predicted%20K>): Higher is better; unit: SCALAR.

- [DER (K unknown)](<https://www.sota2.com/research/sota/speaker-diarization-on-voxceleb2-held-out.md?metric=DER%20(K%20unknown)>): Lower is better; unit: PERCENT.

- [Overlap Detection](<https://www.sota2.com/research/sota/speaker-diarization-on-voxceleb2-held-out.md?metric=Overlap%20Detection>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Predicted K | Result date | Report paper |
| --- | --- | --- | --- |
| Echo (VBx-only) | 3.84 | Not recorded | 6a1e643ddbd7150506f858bf |
| Echo (Full pipeline) | 2 | Not recorded | 6a1e643ddbd7150506f858bf |

### Echo (VBx-only)

Predicted K: 3.84. Result ID: 6a1ecbf1fbc373592c1a728a.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e643ddbd7150506f858bf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1e643ddbd7150506f858bf.md>)

Recorded evaluation configuration: {"Setting":"Stage 6 in isolation","Ground Truth K":"2"}. This record does not assert independent reproduction.

### Echo (Full pipeline)

Predicted K: 2. Result ID: 6a1ecbf1fbc373592c1a728b.

[Reporting paper](<https://www.sota2.com/research/paper/6a1e643ddbd7150506f858bf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1e643ddbd7150506f858bf.md>)

Recorded evaluation configuration: {"Setting":"Stage 7","Ground Truth K":"2","Lead on onset":"1 s"}. This record does not assert independent reproduction.

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