What is Multi-singer and choir separation?

Multi-singer and choir separation Choose model and processing mode Cost, time, formats and parameters depend on the selected model.

Who is Multi-singer and choir separation for?

Separation & stem split Multi-singer and choir separation Choose model and processing mode Cost, time, formats and parameters depend on the selected model.

What tasks can Multi-singer and choir separation help with?

Separation & stem split; WAV; Remix, practice with accompaniment and edit separate parts.; Supports inputs in formats such as mp3 / wav / flac / ogg / m4a / aac / aif / aiff / mp4 / mov / mkv; UNMIXX; Singer Reference · Concat dual λ=0.1

Multi-singer and choir separation

Q: Which models and related terms are associated with Multi-singer and choir separation?

UNMIXX, UNMIXX source and bundled checkpoint (pinned), UNMIXX paper, UNMIXX paper, Table 3, MedleyVox baseline, MedleyVox duet, UNMIXX paper Table 3, TIGER*

Models, parameters, and sources

UNMIXXUNMIXX source and bundled checkpoint (pinned)UNMIXX paperUNMIXX paper, Table 3MedleyVox baselineMedleyVox duetUNMIXX paper Table 3TIGER*MedleyVox unisonUNMIXX: Untangling Highly Correlated Singing Voices MixturesICASSP2026

Model comparison

Model

Role

Why it is used

Output impact

Scores apply only to the cited dataset, metric and model version. Leading this table does not mean a current world record or a result from this service. Capabilities and tests can justify prioritizing models without comparable scores.

Benchmark chart: Multi-Singer Separation · UNMIXX / Reference Singer Separation / Male & Female Vocal Separation / A Cappella Part Separation · SepACap

The table includes only models with public weights or configurations that can run locally.

Multi-Singer Separation · UNMIXX

UNMIXX paper, Table 3

Values come from the linked public source. Row labels identify the test set or checkpoint; models without a matching result are omitted.

Public sources checked: 2026-09-10

Rank	Test setting	Model	SDRi (dB) ↑	Detailed metrics
#3	MedleyVox duet	MedleyVox baselineUNMIXX paper Table 3	15.10	SI-SDRi (dB) ↑14.20H/SSNR (dB) ↑13.33
#2	MedleyVox duet	TIGER*UNMIXX paper Table 3	16.58	SI-SDRi (dB) ↑15.52H/SSNR (dB) ↑15.14
#1	MedleyVox duet	UNMIXXTelkNetUNMIXX paper Table 3	17.52Best	SI-SDRi (dB) ↑16.47BestH/SSNR (dB) ↑15.96Best
#3	MedleyVox unison	MedleyVox baselineUNMIXX paper Table 3	4.90	SI-SDRi (dB) ↑4.40H/SSNR (dB) ↑7.65
#2	MedleyVox unison	TIGER*UNMIXX paper Table 3	5.96	SI-SDRi (dB) ↑5.31H/SSNR (dB) ↑9.86
#1	MedleyVox unison	UNMIXXTelkNetUNMIXX paper Table 3	7.16Best	SI-SDRi (dB) ↑6.58BestH/SSNR (dB) ↑10.50Best

Reference Singer Separation

Singer-reference paper, Table 1

Scores come from the linked base checkpoint, not a new measurement of the model running here.

Public sources checked: 2026-09-10

Rank	Test setting	Model	SI-SDR (dB) ↑	Detailed metrics
#1	Duet target singer	Concat dual λ=0.1Corresponding tool model: Singer Reference · Concat dual λ=0.1Paper Table 1	5.58Best	Both singers · SI-SDR (dB) ↑0.59Background · SI-SDR (dB) ↑-5.91Interfering singer + background · SI-SDR (dB) ↑7.98Best
#2	Duet target singer	FiLM layer-varyingCorresponding tool model: Singer Reference · FiLM layer-varyingPaper Table 1	3.39	Both singers · SI-SDR (dB) ↑0.17Background · SI-SDR (dB) ↑-10.14Interfering singer + background · SI-SDR (dB) ↑-2.99
#3	Duet target singer	Concat layer-varyingPaper Table 1	2.92	Both singers · SI-SDR (dB) ↑-1.61Background · SI-SDR (dB) ↑-6.85Interfering singer + background · SI-SDR (dB) ↑5.04
#4	Duet target singer	BaselinePaper Table 1	0.33	Both singers · SI-SDR (dB) ↑15.88BestBackground · SI-SDR (dB) ↑-1.71BestInterfering singer + background · SI-SDR (dB) ↑1.47

Male & Female Vocal Separation

Author model card · epoch 267

Values come from the linked public source. Row labels identify the test set or checkpoint; models without a matching result are omitted.

Public sources checked: 2026-09-10

Test setting	Model	SDR (dB) ↑
opencpop + m4singer · 50-song validation · epoch 267	Chorus Male/Female BS-RoFormermodel_chorus_bs_roformer_ep_267_sdr_24.1275.ckpt	24.1275

A Cappella Part Separation · SepACap

SepACap paper, Table 3

Values come from the linked public source. Row labels identify the test set or checkpoint; models without a matching result are omitted.

Public sources checked: 2026-09-10

Rank	Test setting	Model	SDRi (dB) ↑
#3	Alto	X-UMXSepACap paper Table 3 · finger snap is not evaluated	13.5
#4	Alto	DPTNetSepACap paper Table 3 · finger snap is not evaluated	11.9
#1	Alto	MRDLASepACap paper Table 3 · finger snap is not evaluated	14.7Best
#5	Alto	MBRSepACap paper Table 3 · finger snap is not evaluated	6.3
#2	Alto	SepACapSepACap paper Table 3 · finger snap is not evaluated	14.6
#5	Bass	X-UMXSepACap paper Table 3 · finger snap is not evaluated	9.1
#2	Bass	DPTNetSepACap paper Table 3 · finger snap is not evaluated	19.7
#4	Bass	MRDLASepACap paper Table 3 · finger snap is not evaluated	10.2
#3	Bass	MBRSepACap paper Table 3 · finger snap is not evaluated	17.8
#1	Bass	SepACapSepACap paper Table 3 · finger snap is not evaluated	23.2Best
#4	Lead vocal	X-UMXSepACap paper Table 3 · finger snap is not evaluated	7.5
#2	Lead vocal	DPTNetSepACap paper Table 3 · finger snap is not evaluated	8.9
#3	Lead vocal	MRDLASepACap paper Table 3 · finger snap is not evaluated	8.7
#5	Lead vocal	MBRSepACap paper Table 3 · finger snap is not evaluated	0.7
#1	Lead vocal	SepACapSepACap paper Table 3 · finger snap is not evaluated	13.0Best
#3	Soprano	X-UMXSepACap paper Table 3 · finger snap is not evaluated	10.7
#4	Soprano	DPTNetSepACap paper Table 3 · finger snap is not evaluated	8.5
#2	Soprano	MRDLASepACap paper Table 3 · finger snap is not evaluated	11.8
#5	Soprano	MBRSepACap paper Table 3 · finger snap is not evaluated	4.5
#1	Soprano	SepACapSepACap paper Table 3 · finger snap is not evaluated	13.1Best
#5	Tenor	X-UMXSepACap paper Table 3 · finger snap is not evaluated	10.2
#2	Tenor	DPTNetSepACap paper Table 3 · finger snap is not evaluated	14.9
#3	Tenor	MRDLASepACap paper Table 3 · finger snap is not evaluated	11.3
#4	Tenor	MBRSepACap paper Table 3 · finger snap is not evaluated	10.3
#1	Tenor	SepACapSepACap paper Table 3 · finger snap is not evaluated	17.0Best
#4	Vocal percussion	X-UMXSepACap paper Table 3 · finger snap is not evaluated	21.0
#3	Vocal percussion	DPTNetSepACap paper Table 3 · finger snap is not evaluated	21.9
#2	Vocal percussion	MRDLASepACap paper Table 3 · finger snap is not evaluated	22.1
#5	Vocal percussion	MBRSepACap paper Table 3 · finger snap is not evaluated	19.3
#1	Vocal percussion	SepACapSepACap paper Table 3 · finger snap is not evaluated	22.5Best

Paper and benchmark notes

UNMIXX: Untangling Highly Correlated Singing Voices Mixtures

ICASSP - 2026

Paper

Singer-Informed Vocal Source Separation for Multi-Singer Music Mixtures

IWAENC - 2026

Extract the singer matching a separate enrollment recording with the official Concat dual λ=0.1 or FiLM layer-varying checkpoint.

Paper

Source Separation for A Cappella Music

arXiv - 2025

Separate alto, bass, finger snap, lead vocal, soprano, tenor and vocal percussion with SepACap.

Paper

Official / repositories

UNMIXX source and bundled checkpoint (pinned)Singer-reference source and official checkpoint links (pinned)MSST inference source (pinned)Official chorus gender checkpoint SepACap source (pinned)SepACap checkpoint (pinned)

Papers / technical notes

UNMIXX paper Singer-reference separation paper SepACap paper

Parameter guide

Optional start of the processed segment in seconds.
Optional end of the processed segment in seconds; it must be later than the start.

Multi-singer and choir separation

Models, parameters, and sources

UNMIXX

Singer Reference · Concat dual λ=0.1

Singer Reference · FiLM layer-varying

Chorus Male/Female BS-RoFormer

SepACap

Model comparison

Benchmark chart: Multi-Singer Separation · UNMIXX / Reference Singer Separation / Male & Female Vocal Separation / A Cappella Part Separation · SepACap

Multi-Singer Separation · UNMIXX

Reference Singer Separation

Male & Female Vocal Separation

A Cappella Part Separation · SepACap

Paper and benchmark notes

UNMIXX: Untangling Highly Correlated Singing Voices Mixtures

Singer-Informed Vocal Source Separation for Multi-Singer Music Mixtures

Source Separation for A Cappella Music

Official / repositories

Papers / technical notes

Parameter guide

Workflow

Before you submit

Quick feedback