Music Source Separation Based on a Lightweight Deep Learning Framework (DTTNET: DUAL-PATH TFC-TDF UNET)
arXiv:2309.08684 · doi:10.1109/ICASSP48485.2024.10448020
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.
arXiv commentAccepted for ICASSP 2024. Additional experiments can be found in the published version on IEEE Xplore
References in corpus (6)
- Gaussian Error Linear Units (GELUs)
- Instance Normalization: The Missing Ingredient for Fast Stylization
- Music Source Separation in the Waveform Domain
- Hybrid Spectrogram and Waveform Source Separation
- Music Demixing Challenge 2021
- Sound Demixing Challenge 2023 Music Demixing Track Technical Report: TFC-TDF-UNet v3