Adaptive filter length selection for acoustic echo cancellation

Document type: Journal Articles
Article type: Original article
Peer reviewed: Yes
Author(s): Christian Schüldt, Fredric Lindström, HB Li
Title: Adaptive filter length selection for acoustic echo cancellation
Year: 2009
Volume: 89
Issue: 6
Pagination: 1185-1194
ISSN: 0165-1684
URI/DOI: 10.1016/j.sigpro.2008.12.023
ISI number: 000264941900023
Organization: Blekinge Institute of Technology
Department: School of Engineering - Dept. of Signal Processing (Sektionen för teknik – avd. för signalbehandling)
School of Engineering S- 372 25 Ronneby
+46 455 38 50 00
Language: English
Abstract: The number of coefficients in an adaptive finite impulse response filter-based acoustic echo cancellation setup is an important parameter, affecting the overall performance of the echo cancellation. Too few coefficients give undermodelling and too many cause slow convergence and an additional echo due to the mismatch of the extra coefficients. This paper proposes a method to adaptively determine the filter length, based on estimation of the mean square deviation. The method is primarily intended for identifying long non-sparse systems, such as a typical impulse response from an acoustic setup. Simulations with band limited flat spectrum signals are used for verification, showing the behavior and benefits of the proposed algorithm. Furthermore, off-line calculation using recorded speech signals show the behavior in real situations and comparison with another state-of-the-art variable filter length algorithm shows the advantages of the proposed method. (c) 2009 Elsevier B.V. All rights reserved.
Subject: Signal Processing\General
Keywords: Acoustic echo cancellation, Adaptive filter, Filter length selection, Variable filter length