*Very small* update on MatAsm.

There IS a bug on the MultiplyMatrices macro, to be specific, the _Mat_Mult_ helper procedure has the bug.

It's possible that it IS the one clobbering memory outside of what it should clobber, which means it was the cause of the memory problems I described in previous versions.

Unfortunately I don't really have that much time to work on this thing ha ha ha.

Basically what's new is the 'CorrelateMatrices' macro, which performs correlation of two signals. The length of the new signal is L1 + L2 -1, where L1 and L2 are the lengths of the signals to be correlated.

If the two signals are the same length and start at time 0, their correlation starts at a negative time. Time zero of their resulting correlation is at the exact middle of the correlation. For signals of different length but still starting at time 0, the time zero of their correlateion is the L2'th sample.

Posted on 2003-04-14 00:14:34 by AmkG
Darn, forgot the attachment...

Why o why can't I add an attachment when editing???

Alan Bakkerou!
Posted on 2003-04-14 00:17:20 by AmkG