A New Initialization Method for Frequency-Domain Blind Source Separation Algorithms

Frequency-domain blind source separation (BSS) techniques have been proposed with the objective of increasing the speed and/or reducing the computational complexity of conventional algorithms, mainly for applications that involve large convolutive mixture filters. In this letter, we present a new in...

Full description

Saved in:
Bibliographic Details
Published in:IEEE signal processing letters 2011-06, Vol.18 (6), p.343-346
Main Authors: Clark, F S P, Petraglia, M R, Haddad, D B
Format: Article
Language:English
Subjects:
Blind source separation (BSS)
clustering
convolutive mixtures
direction of arrival
Microphones
Noise
Noise measurement
Source separation
Speech
Time frequency analysis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Frequency-domain blind source separation (BSS) techniques have been proposed with the objective of increasing the speed and/or reducing the computational complexity of conventional algorithms, mainly for applications that involve large convolutive mixture filters. In this letter, we present a new initialization method for frequency-domain BSS algorithms employing estimates of the directions of arrival and time-frequency masking, that outperforms the classical pre-whitening initialization technique. Performance and convergence time results of the proposed approach are presented using a frequency-domain BSS algorithm that exploits higher-order frequency dependencies, employing real conference room recordings and simulated data.
ISSN:1070-9908
1558-2361
DOI:10.1109/LSP.2011.2135851