dawn@0: % rdiff_inplace(x,y,d,lags,n) - in place running cross-difference function
dawn@0: %
dawn@0: %  x: column vector
dawn@0: %  y: column vector
dawn@0: %  r: result matrix (time X lag)
dawn@0: %  lags: 2-column matrix of lags 
dawn@0: %  n: (samples) frame-rate & window size (default=1)
dawn@0: %
dawn@0: % Vectors x and y are each delayed by amounts specified by lags and subtracted
dawn@0: % sample to sample.  The difference is squared and added over a time window
dawn@0: % of n samples.  This processing is repeated every n samples, as many times
dawn@0: % as there are columns in r.
dawn@0: %
dawn@0: % A positive lag applied to x causes it to be delayed with respect to y.
dawn@0: % A positive lag applied to y causes it to be delayed with respect to x.
dawn@0: % The number of rows of r must match that of lags. 
dawn@0: %
dawn@0: % Mex function.
dawn@0: % Beware: input arguments are assigned to. This is not matlab-kosher!