hello
maybe this ?
define a buffer lenth and an overlap (here I took the max possible overlap) and use corrcoef (take the non diagonal term)
% dummy data
n=1000;
t= 1:n;
x = sin(8*pi*t./max(t))+0.1*rand(1,n);
y = sin(8*pi*t./max(t)+0.25)+0.5+0.1*rand(1,n);
mybuffer = 100; % nb of samples in one buffer (buffer size)
overlap = mybuffer-1; % overlap expressed in samples
%%%% main loop %%%%
m = length(x);
shift = mybuffer-overlap; % nb of samples between 2 contiguous buffers
for ci=1:fix((m-mybuffer)/shift +1)
start_index = 1+(ci-1)*shift;
stop_index = min(start_index+ mybuffer-1,m);
time_index(ci) = floor((start_index+stop_index)/2); % time index expressed as sample unit (dt = 1 in this simulation)
M = corrcoef(x(start_index:stop_index),y(start_index:stop_index));
corcoefficient(ci) = M(2,1);
end
t2 = t(time_index);
figure(1),
plot(t,x,t,y,t2,corcoefficient,'-+r');
legend('signal x ','signal y ','cor coefficient');
