Abstract
This paper presents a new algorithm for addressing the permutation ambiguity in convolutive blind source separation. the proposed algorithm seeks to prevent permutations by frequency oversampling, and then exploiting the induced correlation between bins. Any remaining permutation is then corrected by beam pattern estimation. Cascade initialization is shown to improve system performance while decreasing processing time, while frequency oversampling is shown to increase performance with slight increases in computational time. the algorithm is proven robust against isotropic noise, producing an SIR improvement of 12.9 dB when the reverberation time is 260 ms, and the signals at the input are only 5 dB above the noise floor.* © 2012 IEEE.
Recommended Citation
C. Osterwise and S. L. Grant, "Effect of Frequency oversampling and Cascade Initialization on Permutation Control in Frequency Domain BSS," ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings, pp. 285 - 288, article no. 6287873, Institute of Electrical and Electronics Engineers, Oct 2012.
The definitive version is available at https://doi.org/10.1109/ICASSP.2012.6287873
Department(s)
Electrical and Computer Engineering
Keywords and Phrases
Blind source separation (BSS); convolutive mixture; frequency-domain ICA; permutation problem
International Standard Book Number (ISBN)
978-146730046-9
International Standard Serial Number (ISSN)
1520-6149
Document Type
Article - Conference proceedings
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2024 Institute of Electrical and Electronics Engineers, All rights reserved.
Publication Date
23 Oct 2012
