Title :
Design of two-channel linear-phase FIR PR filter banks with even length filters using convolution matrices
Author :
Muthuvel, A. ; Makur, Anamitra
Author_Institution :
Dept. of Electr. Commun. Eng., Indian Inst. of Sci., Bangalore, India
fDate :
12/1/2000 12:00:00 AM
Abstract :
In this paper, the design problem of two-channel linear-phase finite-impulse response perfect reconstruction fitter banks in which both the filters are of even length (which are known as type A filter banks) is addressed. The condition on the determinant of the polyphase matrix is translated in terms of well-known convolution matrices. The perfect reconstruction condition is obtained as a system of linear equations. Using the derived condition, an algorithm for the design of type A filter bank by searching the entire class in a sequential approach for both equal and unequal length cases is presented. In this approach, the first analysis filter is chosen such that there exists a filter forming a type A system. The second analysis filter is then chosen using the remaining degrees of freedom. These two steps are iterated, leading to joint optimization of both the filters
Keywords :
FIR filters; channel bank filters; circuit optimisation; convolution; digital filters; linear phase filters; convolution matrices; degrees of freedom; equal length case; even length filters; joint optimization; linear equations; perfect reconstruction condition; polyphase matrix; sequential approach; two-channel linear-phase FIR PR filter banks; type A filter banks; unequal length case; Algorithm design and analysis; Channel bank filters; Convolution; Equations; Filter bank; Finite impulse response filter; Lattices; Linear matrix inequalities; Low pass filters; Matrix decomposition;
Journal_Title :
Circuits and Systems II: Analog and Digital Signal Processing, IEEE Transactions on