Filter Design and Implementation - Digital Filter Design Solutions - page 39

function [y] = decx2(h,x,varargin);
%
% decx2 returns the filtered or downsampled (decimated) data
% using a compacted low-pass filter h. Downsampling factor is 2.

%
% If h is a compacted half-band low-pass filter and

% the string 'LPF' is added to the input argument, the

% procedure implements the standard time domain low-pass

% filtering operation.

%
%

If h is a compacted half-band band-pass filter and

% the string 'BPF' is added to the input argument, the

% procedure implements the standard time domain band-pass

% filtering operation.

%
%

If h is a compacted half-band band-pass filter and the

% string 'BPx' is added to the input argument, the

% procedure implements decimation by 2 operation.

%
%

The final data length is approximately half that of the

% original length when the decimation option is chosen.

%
% Usage: [y] = decx2(h,x) implements decimation by 2, low-pass filtering
%

[y] = decx2(h,x,'LPF') for low-pass filtering (no decimation)

% [y] = decx2(h,x,'BPF') for band-pass filtering (no decimation)

% [y] = decx2(h,x,'BPx') implements decimation by 2, band-pass filtering

%
% Input:

h, compact half-band filter coefficients

% x, input data

%
% Output: y, data decimated by a factor of 2 or filter output without decimation
%
% Notes: (1) Both h and x should be a single row or column vectors.
% (2) The output y is a single column vector.

% (3) Ensure that extraneous frequency components cannot fold back into the

% pass band when the 'BPx' decimation option is used.

%
N = length(x);
L = length(h)-1;
str = '';
if length(varargin) == 1
str = [varargin{1}(:)]';
end
P = 4*L -1;
if (length(str) == 3)&((str == 'BPF')|(str == 'BPx')) P = 4*L + 1; end
if (length(str) == 3)&((str == 'LPF')|(str == 'BPF')) y(1:N) = 0; else y(1:floor(N/2)) = 0; end
[r,c] = size(h); if (r > 1)&(c == 1) h = h'; end
[r,c] = size(h);
if (r == 1)&(c > 1) Z = zeros(1,P+1);
[r,c] = size(x); if (r > 1)&(c == 1) x = x'; end
[r,c] = size(x);
if (r == 1)&(c > 1) x = [Z x Z];

Figure 2.25 Matlab function for time domain decimation by 2.