Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Blind Source Separation of Speech Mixtures using a Simple and Computationally Efficient Time- Frequency Approach

Ballal, Tariq ; Grbic, Nedelko LU and Mohammed, Abbas (2006) Science of Electronic, Technologies of Information and Telecommunications (SETIT 2007)
Abstract
A very simple and extremely computationally efficient algorithm for blind separation of two speech sources from two mixtures is presented in this paper. The algorithm exploits the approximate W-disjoint orthogonality of speech signals and assumes specific sensors (microphones) setting that allows the sources to possess a feature we call cross high-low diversity. Two sources are said to be cross high-low diverse (CH-LD) if the two sources are not both close to the same sensor. A source is said to be close to a sensor, if its energy at that sensor is higher than its energy at the other sensor. With this assumption and the W-disjoint orthogonality, it was found that a speech source can easily be extracted from any of the two mixtures with... (More)
A very simple and extremely computationally efficient algorithm for blind separation of two speech sources from two mixtures is presented in this paper. The algorithm exploits the approximate W-disjoint orthogonality of speech signals and assumes specific sensors (microphones) setting that allows the sources to possess a feature we call cross high-low diversity. Two sources are said to be cross high-low diverse (CH-LD) if the two sources are not both close to the same sensor. A source is said to be close to a sensor, if its energy at that sensor is higher than its energy at the other sensor. With this assumption and the W-disjoint orthogonality, it was found that a speech source can easily be extracted from any of the two mixtures with good SIRs (signal-to-interference ratios) based on simple algorithm that compares the ratios of the magnitudes of the time-frequency representations of the two mixtures. The proposed algorithm was tested using different mixtures and has proved to be efficient with both instantaneous and echoic real mixtures. Finally, performance optimization and future expendability to non-CH-LD sources was found possible. (Less)
Please use this url to cite or link to this publication:
author
; and
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
host publication
Science of Electronic, Technologies of Information and Telecommunications, SETIT 2007
conference name
Science of Electronic, Technologies of Information and Telecommunications (SETIT 2007)
conference location
Hammamet, Tunisia
conference dates
2007-03-25 - 2007-03-29
language
Swedish
LU publication?
no
id
eee124f3-06d6-4ae0-8f86-5476320e4985
alternative location
http://www.setit.rnu.tn/last_edition/setit2007/TS/138.pdf
date added to LUP
2016-06-23 16:40:53
date last changed
2021-03-23 22:33:05
@inproceedings{eee124f3-06d6-4ae0-8f86-5476320e4985,
  abstract     = {{A very simple and extremely computationally efficient algorithm for blind separation of two speech sources from two mixtures is presented in this paper. The algorithm exploits the approximate W-disjoint orthogonality of speech signals and assumes specific sensors (microphones) setting that allows the sources to possess a feature we call cross high-low diversity. Two sources are said to be cross high-low diverse (CH-LD) if the two sources are not both close to the same sensor. A source is said to be close to a sensor, if its energy at that sensor is higher than its energy at the other sensor. With this assumption and the W-disjoint orthogonality, it was found that a speech source can easily be extracted from any of the two mixtures with good SIRs (signal-to-interference ratios) based on simple algorithm that compares the ratios of the magnitudes of the time-frequency representations of the two mixtures. The proposed algorithm was tested using different mixtures and has proved to be efficient with both instantaneous and echoic real mixtures. Finally, performance optimization and future expendability to non-CH-LD sources was found possible.}},
  author       = {{Ballal, Tariq and Grbic, Nedelko and Mohammed, Abbas}},
  booktitle    = {{Science of Electronic, Technologies of Information and Telecommunications, SETIT 2007}},
  language     = {{swe}},
  title        = {{Blind Source Separation of Speech Mixtures using a Simple and Computationally Efficient Time- Frequency Approach}},
  url          = {{http://www.setit.rnu.tn/last_edition/setit2007/TS/138.pdf}},
  year         = {{2006}},
}