Blind source separation of convolutive mixtures of speech in frequency domain

Shoji Makino; Hiroshi Sawada; Ryo Mukai; Shoko Araki

doi:10.1093/ietfec/e88-a.7.1640

Blind source separation of convolutive mixtures of speech in frequency domain

Shoji Makino^*, Hiroshi Sawada, Ryo Mukai, Shoko Araki

^*Corresponding author for this work

Research output: Contribution to journal › Review article › peer-review

57 Citations (Scopus)

Abstract

This paper overviews a total solution for frequency-domain blind source separation (BSS) of convolutive mixtures of audio signals, especially speech. Frequency-domain BSS performs independent component analysis (ICA) in each frequency bin, and this is more efficient than time-domain BSS. We describe a sophisticated total solution for frequency-domain BSS, including permutation, scaling, circularity, and complex activation function solutions. Experimental results of 2 × 2, 3 × 3, 4 × 4, 6 × 8, and 2 × 2 (moving sources), (#sources × #microphones) in a room are promising.

Original language	English
Pages (from-to)	1640-1654
Number of pages	15
Journal	IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
Volume	E88-A
Issue number	7
DOIs	https://doi.org/10.1093/ietfec/e88-a.7.1640
Publication status	Published - 2005 Jul
Externally published	Yes

Keywords

Adaptive beamformer
Blind source separation
Convolutive mixtures
Frequency-domain BSS
Independent component analysis
Microphone array

ASJC Scopus subject areas

Signal Processing
Computer Graphics and Computer-Aided Design
Electrical and Electronic Engineering
Applied Mathematics

Access to Document

10.1093/ietfec/e88-a.7.1640

Cite this

@article{15859135def941f6a55f245059d3b145,

title = "Blind source separation of convolutive mixtures of speech in frequency domain",

abstract = "This paper overviews a total solution for frequency-domain blind source separation (BSS) of convolutive mixtures of audio signals, especially speech. Frequency-domain BSS performs independent component analysis (ICA) in each frequency bin, and this is more efficient than time-domain BSS. We describe a sophisticated total solution for frequency-domain BSS, including permutation, scaling, circularity, and complex activation function solutions. Experimental results of 2 × 2, 3 × 3, 4 × 4, 6 × 8, and 2 × 2 (moving sources), (#sources × #microphones) in a room are promising.",

keywords = "Adaptive beamformer, Blind source separation, Convolutive mixtures, Frequency-domain BSS, Independent component analysis, Microphone array",

author = "Shoji Makino and Hiroshi Sawada and Ryo Mukai and Shoko Araki",

year = "2005",

month = jul,

doi = "10.1093/ietfec/e88-a.7.1640",

language = "English",

volume = "E88-A",

pages = "1640--1654",

journal = "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences",

issn = "0916-8508",

publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",

number = "7",

}

TY - JOUR

T1 - Blind source separation of convolutive mixtures of speech in frequency domain

AU - Makino, Shoji

AU - Sawada, Hiroshi

AU - Mukai, Ryo

AU - Araki, Shoko

PY - 2005/7

Y1 - 2005/7

N2 - This paper overviews a total solution for frequency-domain blind source separation (BSS) of convolutive mixtures of audio signals, especially speech. Frequency-domain BSS performs independent component analysis (ICA) in each frequency bin, and this is more efficient than time-domain BSS. We describe a sophisticated total solution for frequency-domain BSS, including permutation, scaling, circularity, and complex activation function solutions. Experimental results of 2 × 2, 3 × 3, 4 × 4, 6 × 8, and 2 × 2 (moving sources), (#sources × #microphones) in a room are promising.

AB - This paper overviews a total solution for frequency-domain blind source separation (BSS) of convolutive mixtures of audio signals, especially speech. Frequency-domain BSS performs independent component analysis (ICA) in each frequency bin, and this is more efficient than time-domain BSS. We describe a sophisticated total solution for frequency-domain BSS, including permutation, scaling, circularity, and complex activation function solutions. Experimental results of 2 × 2, 3 × 3, 4 × 4, 6 × 8, and 2 × 2 (moving sources), (#sources × #microphones) in a room are promising.

KW - Adaptive beamformer

KW - Blind source separation

KW - Convolutive mixtures

KW - Frequency-domain BSS

KW - Independent component analysis

KW - Microphone array

UR - http://www.scopus.com/inward/record.url?scp=26044473134&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=26044473134&partnerID=8YFLogxK

U2 - 10.1093/ietfec/e88-a.7.1640

DO - 10.1093/ietfec/e88-a.7.1640

M3 - Review article

AN - SCOPUS:26044473134

SN - 0916-8508

VL - E88-A

SP - 1640

EP - 1654

JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

IS - 7

ER -

Blind source separation of convolutive mixtures of speech in frequency domain

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this