Browse State-of-the-Art › Speech Separation

Speech Separation

120 papers with code · 19 benchmarks · 16 datasets archive 2025-07-28

Speech

The task of extracting all overlapping speech sources in a given mixed speech signal refers to the Speech Separation. Speech Separation is a special scenario of source separation problem, where the focus is only on the overlapping speech signal sources and other interferences such as music or noise signals are not the main concern of the study. A recent representative Github project can be referred to ClearerVoice-Studio.

Source: A Unified Framework for Speech Separation

Image credit: Speech Separation of A Target Speaker Based on Deep Neural Networks

Description from the archive archive 2025-07-28.

Benchmarks archive 2025-07-28

19 leaderboard tables shown for this task, 19 with rows (a “benchmark” on this site is a table with at least one row, as on /sota), ordered by row count. “Best model” is the first row in the archive's own order at snapshot; nothing is re-ranked here and metric direction is not recorded in the archive. PwC's Trend sparklines are not in the archive, so that column is omitted. 10 shown of 19 until expanded.

DatasetBest model (first row in archive order)PaperCodeSyntologyCompare
WSJ0-2mix (40 rows) TF-Locoformer (L) + DM TF-Locoformer: Transformer with Local Modeling by Convolution for... code — Compare
WHAMR! (18 rows) TF-Locoformer (M) TF-Locoformer: Transformer with Local Modeling by Convolution for... code — Compare
Libri2Mix (10 rows) MossFormer2 (w speed perturb) MossFormer2: Combining Transformer and RNN-Free Recurrent Network... code Syntology official (archive's flag): 1 ran · 4 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 1 honoured, 2 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified Compare
WSJ0-3mix (9 rows) SepTDA Boosting Unknown-number Speaker Separation with Transformer... — — Compare
LRS2 (8 rows) IIANet IIANet: An Intra- and Inter-Modality Attention Network for... code Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified Compare
WHAM! (6 rows) SepReformer-L + DM Separate and Reconstruct: Asymmetric Encoder-Decoder for Speech Separation code — Compare
WSJ0-5mix (6 rows) SepTDA Boosting Unknown-number Speaker Separation with Transformer... — — Compare
LRS3 (5 rows) IIANet IIANet: An Intra- and Inter-Modality Attention Network for... code Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified Compare
VoxCeleb2 (5 rows) IIANet IIANet: An Intra- and Inter-Modality Attention Network for... code Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified Compare
WSJ0-4mix (5 rows) SepTDA Boosting Unknown-number Speaker Separation with Transformer... — — Compare
Libri5Mix (4 rows) Separate And Diffuse Separate And Diffuse: Using a Pretrained Diffusion Model for... — — Compare
Libri10Mix (3 rows) Separate And Diffuse Separate And Diffuse: Using a Pretrained Diffusion Model for... — — Compare
GRID corpus (mixed-speech) (2 rows) (unnamed in the archive) MIDI: Multi-Instance Diffusion for Single Image to 3D Scene Generation — — Compare
Libri20Mix (2 rows) Separate And Diffuse Separate And Diffuse: Using a Pretrained Diffusion Model for... — — Compare
LibriCSS (2 rows) Conformer (large) Continuous Speech Separation with Conformer code — Compare
iKala (1 row) U-Net Singing Voice Separation with Deep U-Net Convolutional Networks code — Compare
Libri15Mix (1 row) Hungarian PIT Many-Speakers Single Channel Speech Separation with Optimal... code — Compare
TCD-TIMIT corpus (mixed-speech) (1 row) Audio-Visual concat-ref Face Landmark-based Speaker-Independent Audio-Visual Speech... code — Compare
WSJ0-2mix-16k (1 row) MossFormer2 MossFormer: Pushing the Performance Limit of Monaural Speech... code Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified Compare

Syntology column: samples harvested from the paper's repositories and executed on synthesized fixtures; “ran” is not a correctness claim and does not order the table. Each line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the instrument figure counts failures of Syntology's instrument, not of the code. When the archive marks a repository official for the row's paper, the line starts with that repository's state (“official” is the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from. A dash means no Syntology record for that paper, not a recorded non-run. Each “ran” line links to that paper's sample list. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. For agents: get_harvested_code_for_paper(arxiv_id) lists each paper's samples; how to connect.

Libraries

Not in the archive: the export carries no per-task library table, so there is nothing to show at snapshot 2025-07-28.

Datasets archive 2025-07-28

16 datasets whose archive record lists this task, ordered by the archive's paper count.

Subtasks archive 2025-07-28

1 subtask in the archive's task tree.

Papers archive 2025-07-28

archive papers tagged: 359 · with a code link: 120 · where Syntology ran a sample: 27 (21 with a run with no instrument failure, 6 where every run was a failure of Syntology's instrument) Syntology

Show: all tagged papersonly where code ran (27 of 359 tagged: 21 with a run with no instrument failure, 6 where every run was a failure of Syntology's instrument)

Page 1 of 4: papers 1 to 100 of 359, in archive order: by repositories listed in the archive (most first), then newest first, not by stars (the archive holds no stars, so PwC's “Social” and “Latest” sorts cannot be reproduced). Papers that list no repository come after every paper that lists one.

Papers without a page here are shown as plain text. A Syntology line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the figure “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. When the archive marks a repository official for the paper, the line starts with that repository's state (the archive's flag, not a verdict on who wrote the code); hover it for the repositories the samples that ran came from. The first sentence of the abstract is shown for the first 30 papers; every abstract is on its paper's page.

Syntology lines on 26 of the papers shown; no Syntology record for the others (a paper without an arXiv id cannot be joined to the graph, and absence from the graph layer is not a recorded non-run). “Ran” means the sample executed on a synthesized fixture, not that the paper's result was reproduced; each line links to that paper's sample list. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. For agents: get_harvested_code_for_paper(arxiv_id) lists each paper's samples; how to connect.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections