Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Audio Source Separation research benchmarks · SOTA2 Research

Audio Source Separation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Slakh2100 (test)

20.09SDR Bass

16

Apr 28, 2026

11.61SDR

16

Apr 14, 2026

MUSDB18 SiSEC 2018 (test)

D3Net (proposed)

7.24Vocals Score

10

Feb 26, 2026

MoXaRt XR Scenarios (human speech sessions)

0.382WER

7

Mar 12, 2026

SISEC live-recorded stereo mixtures Three males T60 = 130 ms

Proposed (Inform)

9.1SDR

5

Apr 15, 2026

SISEC live-recorded stereo mixtures Three females T60 = 130 ms

Proposed (Inform)

10.07SDR

5

Apr 15, 2026

SISEC live-recorded stereo mixtures T60 = 250 ms (Three males)

Proposed (Informed)

7.82SDR

5

Apr 15, 2026

SISEC live-recorded stereo mixtures (T60 = 250 ms) (Three females)

Proposed (Informed)

9.4SDR

5

Apr 15, 2026

AudioSet SingleSource (test)

4.26SDR

5

Feb 26, 2026

9.5SI-SNRi (dB)

3

Feb 26, 2026

0.68SDR (Drums)

3

Feb 26, 2026

68Human Preference Score

1

Feb 26, 2026

63Human Preference Score

1

Feb 26, 2026

Popular datasets

SISEC AudioSet MUSIC MUSDB LRS2

Follow for updates

Join SOTA2 Contact