Title :
Precise Filter Design [DSP Tips & Tricks]
Abstract :
In this article, we describe an obscure but simple and powerful method for designing a digital filter that approximates an arbitrary magnitude and phase response. Called frequency-domain least-squares (FDLS) method, it can create a filter roughly comparable in computational burden and latency to that created by the bilinear transform method, with fidelity approaching that of fast convolution. In addition, the FDLS method can be applied to a wide variety of other system identification tasks
Keywords :
digital filters; frequency-domain analysis; least squares approximations; matrix algebra; network synthesis; transforms; arbitrary magnitude; bilinear transform method; digital filter design; frequency-domain least-squares; phase response; precise filter design; Convolution; Delay; Digital filters; Digital signal processing; Equations; Frequency; Linear algebra; Matched filters; Matrices; Transfer functions;
Journal_Title :
Signal Processing Magazine, IEEE
DOI :
10.1109/MSP.2007.273077
