[Octave-bug-tracker] [bug #46373] [octave forge] (signal) documentation of xcorr misleading

[Juan Pablo Carbajal]

Follow-up Comment #1, bug #46373 (project octave):

Hi The following passes for me, meaning that it gives the same as xcorr


na   = 5;
a    = ones(1,na);
lag  = -(na-1):(na-1); # doc say N-1 is the default
nl   = length (lag);
ac   = prepad (postpad (a,na+floor(nl/2)), na+nl); # doc says pad with zeros
R_xy = zeros (1,nl);

for k=1:length(lags); 
  R_xy(k) = sum (shift(ac,lag(k)) .* ac); # formula in the doc
end; 

assert (xcorr (a, 'none'), R_xy, sqrt (eps));
assert (xcorr (a, 'biased'), R_xy/na, sqrt (eps));
assert (xcorr (a, 'unbiased'), R_xy./(na-abs(lag)), sqrt (eps)); 


What is exactly the misleading part of the documentation?

@devs: please mark as "needs info" or "closed"

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?46373>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/