Evolutionary Methods for Designing Digital Filters.

In this paper we consider the problem of designing digital filters with arbitrary frequency/phase/impulse responses. We introduce and describe in detail an evolutionary (genetic) algorithm for this difficult problem. In particular, we show how the evolutionary algorithm can evolve an always-stable l...

Full description

Bibliographic Details
Published in:	Contemporary Music Review Vol. 22; no. 3; pp. 5 - 20
Main Authors:	Sharman, Ken, Esparcia-Alcázar, Anna
Format:	Article
Published:	Taylor & Francis Ltd 2003
Subjects:	Digital filters (Mathematics) Algorithms Design Transfer functions Automatic control systems
Online Access:	View this record in EBSCOhost

Description
Summary:	In this paper we consider the problem of designing digital filters with arbitrary frequency/phase/impulse responses. We introduce and describe in detail an evolutionary (genetic) algorithm for this difficult problem. In particular, we show how the evolutionary algorithm can evolve an always-stable lattice filter that can yield an arbitrarily complex response. We illustrate the algorithm with an example involving the evolution of a filter that approximates a head-related transfer function for a virtual reality audio application.