Filter Design and Implementation - Digital Filter Design Solutions - page 38

function [hpac] = compacLPF(h);
%
% compacLPF returns the compacted half-band low-pass FIR filter

% coefficients. All zero values are removed from

% the filter leaving half its original length. The output is

% then further reduced by half due to the symmetry of the

% filter. Final length is approximately quarter of its

% original length.

%
% Usage:

[hpac] = compacLPF(h)

%
% Input: h, full length half-band filter coefficients

%
% Output: hpac, compacted half-band filter coefficients

%
% Notes:

(1) The input filter h must be full length containing

% symmetry about center coefficient h(L/2+1).

% (2) L must be even.

% (3) The output filter hpac uses the symmetry of h after zero

% removal and therefore is one quarter of the original length.

% (4) h should be either a single column or row vector.

%
%
[r,c] = size(h); if (r > 1)&(c == 1) h = h'; end
[r,c] = size(h); L = length(h)-1;
hpac(1:fix(L/4)) = 0;
if (r == 1)&(c > 1)
R = rem(L,2);
if R == 0
R1 = rem(L,4);
if R1 == 0
P = L/4;
hpac(1:P) = h(2:2:2*P);
hpac(P+1) = h(L/2+1);
elseif R1 ~= 0
P = floor(L/4)+1;
hpac(1:P) = h(1:2:2*P-1);
hpac(P+1) = h(L/2+1);
end
elseif R ~= 0
disp('compac warning: filter length must be odd.');
end
else
disp('compac warning: h must be a row or column vector.');
end
return

function used to compact half-band low-pass filters.

Figure 2.24 Matlab