Exploiting Narrowband Efficiency for Broadband Convolutive Blind Source Separation

Based on a recently presented generic broadband blind sourceseparation (BSS) algorithm for convolutive mixtures, we propose inthis paper a novel algorithm combining advantages of broadbandalgorithms with the computational efficiency of narrowbandtechniques. By selective application of the Szego theo...

Full description

Saved in:
Bibliographic Details
Published in:EURASIP journal on advances in signal processing 2007-01, Vol.2007 (1), p.016381, Article 016381
Main Authors: Aichner, Robert, Buchner, Herbert, Kellermann, Walter
Format: Article
Language:English
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:Based on a recently presented generic broadband blind sourceseparation (BSS) algorithm for convolutive mixtures, we propose inthis paper a novel algorithm combining advantages of broadbandalgorithms with the computational efficiency of narrowbandtechniques. By selective application of the Szego theorem whichrelates properties of Toeplitz and circulant matrices, a newnormalization is derived as a special case of the genericbroadband algorithm. This results in a computationally efficientand fast converging algorithm without introducing typicalnarrowband problems such as the internal permutation problem orcircularity effects. Moreover, a novel regularization method forthe generic broadband algorithm is proposed and subsequently alsoderived for the proposed algorithm. Experimental results inrealistic acoustic environments show improved performance of thenovel algorithm compared to previous approximations.
ISSN:1687-6180
1687-6172
1687-6180
DOI:10.1155/2007/16381