Efficient linear phase FIR filter structures exploiting recursive subfilters

Abstract

Introduces efficient linear-phase FIR (finite-impulse response) filter structures using novel recursive subfilters. The novel filter structure is based on subfilters whose filter coefficients are related to each other recursively. A well-known structure that uses this method is the recursive running sum. The use of recursive subfilters is extended to cover the simple structures utilizing real and complex poles of radius r. The proposed linear-phase FIR filter structures result in computationally efficient implementations which require fewer general multipliers than conventional filters. This structure is beneficial for filters with narrow transition bands since the use of a single pole pair near the transition region lowers the required numerator order radically. Since the proposed filters have finite impulse responses, they can be realized efficiently using block processing. >