Toward overcoming fundamental limitation in frequency-domain blind source separation for reverberant speech mixtures

Lae Hoon Kim; Mark Hasegawa-Johnson

doi:10.1109/ACSSC.2010.5757618

Toward overcoming fundamental limitation in frequency-domain blind source separation for reverberant speech mixtures

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Blind source separation can be implemented in the frequency domain using one-tap multiplication operation in each frequency bin, but only when the frame length is long enough to disregard temporal aliasing effects. If we take a short-time frequency transformation with a window shorter than a room reverberation time, the justification above does not hold anymore. In this paper, we present an appropriate representation in the short-time frequency domain. The suitability is justified by showing the equivalence with the original time domain approach under the overlap-add context. Experimental validation using a corpus synthesized by convolution with measured sets of room impulse responses is also provided.

Original language	English (US)
Title of host publication	Conference Record of the 44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010
Pages	542-545
Number of pages	4
DOIs	https://doi.org/10.1109/ACSSC.2010.5757618
State	Published - 2010
Event	44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010 - Pacific Grove, CA, United States Duration: Nov 7 2010 → Nov 10 2010

Publication series

Name	Conference Record - Asilomar Conference on Signals, Systems and Computers
ISSN (Print)	1058-6393

Other

Other	44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010
Country/Territory	United States
City	Pacific Grove, CA
Period	11/7/10 → 11/10/10

ASJC Scopus subject areas

Signal Processing
Computer Networks and Communications

Online availability

10.1109/ACSSC.2010.5757618

Library availability

Discover UIUC Full Text

Cite this

Kim, L. H., & Hasegawa-Johnson, M. (2010). Toward overcoming fundamental limitation in frequency-domain blind source separation for reverberant speech mixtures. In Conference Record of the 44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010 (pp. 542-545). Article 5757618 (Conference Record - Asilomar Conference on Signals, Systems and Computers). https://doi.org/10.1109/ACSSC.2010.5757618

Toward overcoming fundamental limitation in frequency-domain blind source separation for reverberant speech mixtures. / Kim, Lae Hoon; Hasegawa-Johnson, Mark.
Conference Record of the 44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010. 2010. p. 542-545 5757618 (Conference Record - Asilomar Conference on Signals, Systems and Computers).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Kim, LH & Hasegawa-Johnson, M 2010, Toward overcoming fundamental limitation in frequency-domain blind source separation for reverberant speech mixtures. in Conference Record of the 44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010., 5757618, Conference Record - Asilomar Conference on Signals, Systems and Computers, pp. 542-545, 44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010, Pacific Grove, CA, United States, 11/7/10. https://doi.org/10.1109/ACSSC.2010.5757618

Kim LH, Hasegawa-Johnson M. Toward overcoming fundamental limitation in frequency-domain blind source separation for reverberant speech mixtures. In Conference Record of the 44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010. 2010. p. 542-545. 5757618. (Conference Record - Asilomar Conference on Signals, Systems and Computers). doi: 10.1109/ACSSC.2010.5757618

@inproceedings{4870d2bd4f40437c87194207d90f08ba,

title = "Toward overcoming fundamental limitation in frequency-domain blind source separation for reverberant speech mixtures",

abstract = "Blind source separation can be implemented in the frequency domain using one-tap multiplication operation in each frequency bin, but only when the frame length is long enough to disregard temporal aliasing effects. If we take a short-time frequency transformation with a window shorter than a room reverberation time, the justification above does not hold anymore. In this paper, we present an appropriate representation in the short-time frequency domain. The suitability is justified by showing the equivalence with the original time domain approach under the overlap-add context. Experimental validation using a corpus synthesized by convolution with measured sets of room impulse responses is also provided.",

author = "Kim, {Lae Hoon} and Mark Hasegawa-Johnson",

year = "2010",

doi = "10.1109/ACSSC.2010.5757618",

language = "English (US)",

isbn = "9781424497218",

series = "Conference Record - Asilomar Conference on Signals, Systems and Computers",

pages = "542--545",

booktitle = "Conference Record of the 44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010",

note = "44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010 ; Conference date: 07-11-2010 Through 10-11-2010",

}

TY - GEN

T1 - Toward overcoming fundamental limitation in frequency-domain blind source separation for reverberant speech mixtures

AU - Kim, Lae Hoon

AU - Hasegawa-Johnson, Mark

PY - 2010

Y1 - 2010

N2 - Blind source separation can be implemented in the frequency domain using one-tap multiplication operation in each frequency bin, but only when the frame length is long enough to disregard temporal aliasing effects. If we take a short-time frequency transformation with a window shorter than a room reverberation time, the justification above does not hold anymore. In this paper, we present an appropriate representation in the short-time frequency domain. The suitability is justified by showing the equivalence with the original time domain approach under the overlap-add context. Experimental validation using a corpus synthesized by convolution with measured sets of room impulse responses is also provided.

AB - Blind source separation can be implemented in the frequency domain using one-tap multiplication operation in each frequency bin, but only when the frame length is long enough to disregard temporal aliasing effects. If we take a short-time frequency transformation with a window shorter than a room reverberation time, the justification above does not hold anymore. In this paper, we present an appropriate representation in the short-time frequency domain. The suitability is justified by showing the equivalence with the original time domain approach under the overlap-add context. Experimental validation using a corpus synthesized by convolution with measured sets of room impulse responses is also provided.

UR - http://www.scopus.com/inward/record.url?scp=79958005036&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79958005036&partnerID=8YFLogxK

U2 - 10.1109/ACSSC.2010.5757618

DO - 10.1109/ACSSC.2010.5757618

M3 - Conference contribution

AN - SCOPUS:79958005036

SN - 9781424497218

T3 - Conference Record - Asilomar Conference on Signals, Systems and Computers

SP - 542

EP - 545

BT - Conference Record of the 44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010

T2 - 44th Asilomar Conference on Signals, Systems and Computers, Asilomar 2010

Y2 - 7 November 2010 through 10 November 2010

ER -

Toward overcoming fundamental limitation in frequency-domain blind source separation for reverberant speech mixtures

Abstract

Publication series

Other

ASJC Scopus subject areas

Online availability

Library availability

Related links

Fingerprint

Cite this