xcorr zero lag

Dear all, How can I use xcorr for 0 lags? In the test I'm running I don't want to 'slide' one sequence over another but keep them fixed. Thanks in advance. Happy new year! Diego

[Diego]

This doesn't work

X1=xcorr(P,Q,0, 'coeff');

I need the signals to be superimposed but avoiding the sliding. Does this makes sense at all?

[Diego]

I'm correlating sequences of length 8.

Is this right?

Thanks,

Diego

[the cyclist]

I don't have the Signal Processing Toolbox, so I can't check myself, but the syntax in your first comment looks right to me. Can you give a more complete example of the code you have written, the result, and why specifically you believe it to be wrong (or not what you expect)?

Answer 1

[Wayne King]

Hi Diego, Why doesn't this work?

[c,lags] = xcorr(randn(1000,1),rand(1000,1),0,'coeff');

This gives you exactly what you have described; the cross correlation between two input vectors at zero lag.

c is the value of the normalized cross correlation sequence at zero lag.

[Diego]

Hi Wayne,

Thanks for your reply.

When I use 0 lags the plotted figures get side by side, not superimposed as when I don't declare the lags.

This is the code of the plot:

delay=d-max(length(P),length(Q));

figure,plot(P);

hold,plot([delay+1:length(P)+delay],Q,'r');

Happy new year!

Diego

[Wayne King]

you're not telling me what d means in the above. are P and Q different lengths? If they are, you can't use the 'coeff' option.

[Diego]

ughh I'm so sorry for omitting the information... the code just grows a lot and is a little messy at this time.

P and Q are the same length.

Thanks for your patience!

P=[M(l,i:i+rep-1)];

Q=[M(z,j:j+rep-1)];

[m,d]=max(X1);

[Diego]

It seems the problem was using the 'delay' for plotting.

I just removed it since at 0 lags it seems to be useless and the plot works as expected.

Many thanks,

Diego