Singing Voice Separation by Low-Rank and Sparse Spectrogram Decomposition with Prelearned Dictionaries

Yu S, Zhang H, Duan Z. Singing Voice Separation by Low-Rank and Sparse Spectrogram Decomposition with Prelearned Dictionaries. J Audio Eng Soc. 2017;65(5):377-388. doi:10.17743/jaes.2017.0009

@article{Yu2017_18731,
  author = {Yu, Shiwei and Zhang, Hongjuan and Duan, Zhiyao},
  title = {{Singing Voice Separation by Low-Rank and Sparse Spectrogram Decomposition with Prelearned Dictionaries}},
  journal = {Journal of the Audio Engineering Society},
  volume = {65},
  number = {5},
  pages = {377--388},
  year = {2017},
  month = may,
  publisher = {Audio Engineering Society},
  doi = {10.17743/jaes.2017.0009},
  url = {https://doi.org/10.17743/jaes.2017.0009}
}

TY  - JOUR
TI  - Singing Voice Separation by Low-Rank and Sparse Spectrogram Decomposition with Prelearned Dictionaries
AU  - Yu, Shiwei
AU  - Zhang, Hongjuan
AU  - Duan, Zhiyao
T2  - Journal of the Audio Engineering Society
J2  - J. Audio Eng. Soc.
VL  - 65
IS  - 5
SP  - 377
EP  - 388
PY  - 2017
DA  - 2017/05/06
DO  - 10.17743/jaes.2017.0009
UR  - https://doi.org/10.17743/jaes.2017.0009
PB  - Audio Engineering Society
LA  - en
AB  - Although the human auditory system can easily distinguish the singing voice from the background music in a music recording, it is extremely difficult for computer systems to replicate this ability, especially when the music mixture is a single channel. The challenge arises from the variety of simultaneous sound sources as well from the rich pitch and timbre variations of a singing voice. Unsupervised spectrogram decomposition involves separating the mixture spectrogram into a sparse spectrogram for the singing voice and a low-rank spectrogram for the background music. This approach has two limitations: the unsupervised nature prevents the prelearning of voice and background in music dictionaries; some components of the singing voice and background music may not show the preferred sparse and low-rank properties. In contrast, the authors propose to decompose the mixture spectrogram into three parts: a sparse spectrogram representing the singing voice, a low-rank spectrogram representing the background music, and a residual spectrogram for the components that are not identified by either the sparse or the low-rank spectrogram. Universal dictionaries for the singing voice and background music are prelearned from isolated singing voice and background music training data, through which prior knowledge of the voice and background music is introduced to the separation process. Evaluations on two datasets show that the proposed method is effective and efficient for both the separated singing voice and music accompaniment at various voice-to-music ratios.
ER  -