Computing the autocorrelation Matrix in Matlab

Hello everyone.

Iam reading a course in signal processing and we talked about autocorrelation and the autocorrelation Matrix. As far as I have understood, Matlab can compute the sample autocorrelation sequence by using the xcorr() function.

The result would be according to formula: rx(k) = 1/N*Sum(x(n)*x(n+k)), giving the autocorrelations rxx(0,0), rxx(0,1), rxx(0,2)...

From that I could easily create the autocorrelation Matrix Rxx by i.ex. using the teoplitz() function to get:

 Rxx = rxx(0) rxx*(1) rxx*(2) ...  rxx*(n-1)
       rxx(1) rxx(0)  rxx*(1) ...  rxx*(n-2)
       rxx(2) rxx(1)  rxx(0)  .... rxx*(n-2)

Now Iam wondering what happens if a have an autocorrrelation like this:

rx(k,l) = 1/N*Sum(x(n-k)*x(n+l))

where the autocorrelation depends on both time k and time l.

I would like to compute an autocorrelation of the form:

   Rxx = rxx(0,0) rxx(0,1) rxx(0,2) ...   rxx(0,n-1)
         rxx(1,0) rxx(1,1)  rxx(1,2) ...  rxx(1,n-2)
         rxx(2,0) rxx(2,1)  rxx(2,2) ...  rxx(2,n-2)
         ....     ....      ....     ...  ...

How would can do that in Matlab?

Thanks in advance,

kind regards,

Mike