SOTAVerified

Audio Source Separation

Audio Source Separation is the process of separating a mixture (e.g. a pop band recording) into isolated sounds from individual sources (e.g. just the lead vocals).

Source: Model selection for deep audio source separation via clustering analysis

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 1–10 of 112 papers

TitleStatusHype
Separate Anything You DescribeCode3
Training-Free Multi-Step Audio Source SeparationCode2
A Stem-Agnostic Single-Decoder System for Music Source Separation Beyond Four StemsCode2
Leveraging LLM and Text-Queried Separation for Noise-Robust Sound Event DetectionCode1
Exploring Text-Queried Sound Event Detection with Audio Source SeparationCode1
Unsupervised Composable Representations for AudioCode1
Facing the Music: Tackling Singing Voice Separation in Cinematic Audio Source SeparationCode1
Remastering Divide and Remaster: A Cinematic Audio Source Separation Dataset with Multilingual SupportCode1
Spectral Mapping of Singing Voices: U-Net-Assisted Vocal SegmentationCode1
A Generalized Bandsplit Neural Network for Cinematic Audio Source SeparationCode1
Show:102550
← PrevPage 1 of 12Next →
All datasetsAudioSetMUSIC (multi-source)

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1ST-SED-SEPSDR10.55—Unverified
2Co-SeparationSDR4.26—Unverified
#ModelMetricClaimedVerifiedStatus
1Co-SeparationSAR11.3—Unverified