lot.m

里面囊括了基于matlab滤波器设计的各种.m文件

 function p0 = lot(icc,N)
% P0 = LOT(icc,N)
% P0 will be a 2N-by-N transform matrix, whose columns
% will be the basis functions of LOT. N is even integer.
% icc is intersample correlation coefficient. The signal
% model is first order Gauss-Markov process.
%
% LOT is based on the paper "The LOT: Transform Coding
% without blocking effects," written by Henrique S.
% Malvar and David H. Staelin, IEEE Trans. on ASSP, April 1989.
% 

J = antieye(N); D = dct2mtx(N);

y = zeros(N,1); y(1) = 1;
De = D*y;
for I = 1:N/2-1
 x = zeros(N,1);
 x(2*I+1) = 1;
 x = D*x;
 De = [De,x];
end

y(1) = 0;
y(2) = 1;
Do = D*y;
for I = 1:N/2-1
 x = zeros(N,1);
 x(2*(I+1)) = 1;
 x = D*x;
 Do = [Do,x];
end

P = 0.5*[De-Do,De-Do;J*(De-Do),-J*(De-Do)];
for I = 0:2*N-1
 x(I+1) = icc^I;
end 
Rxx = toeplitz(x);
A = P'*Rxx*P;
[Z,D] = eig(A);
p0 = P*Z;