TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Papers/Music Source Separation Based on a Lightweight Deep Learni...

Music Source Separation Based on a Lightweight Deep Learning Framework (DTTNET: DUAL-PATH TFC-TDF UNET)

Junyu Chen, Susmitha Vekkot, Pancham Shukla

2023-09-15Music Source Separation
PaperPDFCodeCode(official)

Abstract

Music source separation (MSS) aims to extract 'vocals', 'drums', 'bass' and 'other' tracks from a piece of mixed music. While deep learning methods have shown impressive results, there is a trend toward larger models. In our paper, we introduce a novel and lightweight architecture called DTTNet, which is based on Dual-Path Module and Time-Frequency Convolutions Time-Distributed Fully-connected UNet (TFC-TDF UNet). DTTNet achieves 10.12 dB cSDR on 'vocals' compared to 10.01 dB reported for Bandsplit RNN (BSRNN) but with 86.7% fewer parameters. We also assess pattern-specific performance and model generalization for intricate audio patterns.

Results

TaskDatasetMetricValueModel
Music Source SeparationMUSDB18-HQSDR (avg)8.15Dual-Path TFC-TDF UNet (DTTNet)
Music Source SeparationMUSDB18-HQSDR (bass)7.55Dual-Path TFC-TDF UNet (DTTNet)
Music Source SeparationMUSDB18-HQSDR (drums)7.82Dual-Path TFC-TDF UNet (DTTNet)
Music Source SeparationMUSDB18-HQSDR (others)7.02Dual-Path TFC-TDF UNet (DTTNet)
Music Source SeparationMUSDB18-HQSDR (vocals)10.21Dual-Path TFC-TDF UNet (DTTNet)
2D ClassificationMUSDB18-HQSDR (avg)8.15Dual-Path TFC-TDF UNet (DTTNet)
2D ClassificationMUSDB18-HQSDR (bass)7.55Dual-Path TFC-TDF UNet (DTTNet)
2D ClassificationMUSDB18-HQSDR (drums)7.82Dual-Path TFC-TDF UNet (DTTNet)
2D ClassificationMUSDB18-HQSDR (others)7.02Dual-Path TFC-TDF UNet (DTTNet)
2D ClassificationMUSDB18-HQSDR (vocals)10.21Dual-Path TFC-TDF UNet (DTTNet)

Related Papers

Music Source Restoration2025-05-27Training-Free Multi-Step Audio Source Separation2025-05-26Is MixIT Really Unsuitable for Correlated Sources? Exploring MixIT for Unsupervised Pre-training in Music Source Separation2025-05-12Solving Copyright Infringement on Short Video Platforms: Novel Datasets and an Audio Restoration Deep Learning Pipeline2025-04-30Score-informed Music Source Separation: Improving Synthetic-to-real Generalization in Classical Music2025-03-10Separate This, and All of these Things Around It: Music Source Separation via Hyperellipsoidal Queries2025-01-27Sanidha: A Studio Quality Multi-Modal Dataset for Carnatic Music2025-01-12MAJL: A Model-Agnostic Joint Learning Framework for Music Source Separation and Pitch Estimation2025-01-07