Speech Source Separation Using a Multi-Pitch Harmonic Product Spectrum-Based Algorithm

Ahmed R, Gil-Pita R, Ayllón D, Álvarez L. Speech Source Separation Using a Multi-Pitch Harmonic Product Spectrum-Based Algorithm. In: AES Convention 130. Audio Engineering Society; 2011. Paper 8408. Available from: https://aes.org/publications/elibrary-page/?id=15875

@inproceedings{Ahmed2011_15875,
  author = {Ahmed, Rehan and Gil-Pita, Roberto and Ayllón, David and Álvarez, Lorena},
  title = {{Speech Source Separation Using a Multi-Pitch Harmonic Product Spectrum-Based Algorithm}},
  booktitle = {AES Convention 130},
  note = {Paper 8408},
  year = {2011},
  month = may,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=15875}
}

TY  - CPAPER
TI  - Speech Source Separation Using a Multi-Pitch Harmonic Product Spectrum-Based Algorithm
AU  - Ahmed, Rehan
AU  - Gil-Pita, Roberto
AU  - Ayllón, David
AU  - Álvarez, Lorena
T2  - AES Convention 130
M1  - Paper 8408
PY  - 2011
DA  - 2011/05/06
UR  - https://aes.org/publications/elibrary-page/?id=15875
PB  - Audio Engineering Society
LA  - en
AB  - This paper presents an efficient algorithm for separating speech signals by determining multiple pitches from mixtures of signals and assigning the sources to one of those estimated pitches. The pitch detection algorithm is based on Harmonic Product Spectrum. Since the pitch of speech signals fluctuates readily, a frame-based algorithm is used to extract the multiple pitches in each frame. Then, the fundamental frequency (pitch) for each source is estimated and tracked after comparing all the frames. The estimated fundamental frequency of the sources is then used to generate a set of binary masks that allow separating the signals in the Short Time Fourier Transform domain. Results show a considerable separation of the speech signals, justifying the feasibility of the proposed method.
ER  -