Gaussian Mixture Model for Singing Voice Separation from Stereophonic Music

Kim M, Beack S, Choi K, Kang K. Gaussian Mixture Model for Singing Voice Separation from Stereophonic Music. In: AES Conference: 43rd International Conference: Audio for Wirelessly Networked Personal Devices. Audio Engineering Society; 2011. Paper 6-2. Available from: https://aes.org/publications/elibrary-page/?id=16121

@inproceedings{Kim2011_16121,
  author = {Kim, Minje and Beack, Seungkwon and Choi, Keunwoo and Kang, Kyeongok},
  title = {{Gaussian Mixture Model for Singing Voice Separation from Stereophonic Music}},
  booktitle = {AES Conference: 43rd International Conference: Audio for Wirelessly Networked Personal Devices},
  note = {Paper 6-2},
  year = {2011},
  month = sep,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=16121}
}

TY  - CPAPER
TI  - Gaussian Mixture Model for Singing Voice Separation from Stereophonic Music
AU  - Kim, Minje
AU  - Beack, Seungkwon
AU  - Choi, Keunwoo
AU  - Kang, Kyeongok
T2  - AES Conference: 43rd International Conference: Audio for Wirelessly Networked Personal Devices
M1  - Paper 6-2
PY  - 2011
DA  - 2011/09/06
UR  - https://aes.org/publications/elibrary-page/?id=16121
PB  - Audio Engineering Society
LA  - en
AB  - This paper presents an adaptive prediction method about source-specific ranges of binaural cues, such as inter-channel level difference (ILD) and inter-channel phase difference (IPD), for centrally positioned singing voice separation. To this end, we employ Gaussian mixture model (GMM) to cluster underlying distributions in the feature domain of mixture signal. By regarding responsibilities to those distinct Gaussians as unmixing coefficients of each mixture spectrogram sample, the proposed method can reduce artificial deformations that previous center channel extraction methods usually suffer, caused by their imprecise or rough decision about ranges of central subspaces. Experiments on commercial music show superiority of the proposed method.
ER  -