English

Investigating U-Nets with various Intermediate Blocks for Spectrogram-based Singing Voice Separation

Audio and Speech Processing 2020-10-09 v3 Machine Learning Multimedia Sound Machine Learning

Abstract

Singing Voice Separation (SVS) tries to separate singing voice from a given mixed musical signal. Recently, many U-Net-based models have been proposed for the SVS task, but there were no existing works that evaluate and compare various types of intermediate blocks that can be used in the U-Net architecture. In this paper, we introduce a variety of intermediate spectrogram transformation blocks. We implement U-nets based on these blocks and train them on complex-valued spectrograms to consider both magnitude and phase. These networks are then compared on the SDR metric. When using a particular block composed of convolutional and fully-connected layers, it achieves state-of-the-art SDR on the MUSDB singing voice separation task by a large margin of 0.9 dB. Our code and models are available online.

Keywords

Cite

@article{arxiv.1912.02591,
  title  = {Investigating U-Nets with various Intermediate Blocks for Spectrogram-based Singing Voice Separation},
  author = {Woosung Choi and Minseok Kim and Jaehwa Chung and Daewon Lee and Soonyoung Jung},
  journal= {arXiv preprint arXiv:1912.02591},
  year   = {2020}
}

Comments

8 pages 4 tables 6 figures, accepted to ISMIR 2020

R2 v1 2026-06-23T12:36:55.167Z