# Obtain points in signals which cause correlation

I've got two large .wav files (tens or hundreds of minutes each). I know that at some points these files contain very similar areas - actually, these areas are the recordings of the same event from different cameras, so some noise, etc. might be present, but they contain essentially the same things. I want to obtain the timestamps of these areas in each of two arrays. I've been using fftconvolve() in order to determine the offset between these arrays,. When I apply this function, I get the correct offset - if you stack one array under the other and apply the offset, then the desired areas with same recordings will be stacked right on top of each other.

But what I want to get are the exact areas with similar recordings which made this offset to be the most likely one. Is it possibly to do without digging into fftconvolve() source code and customizing it? Maybe the general idea should be to look for similar pieces in arrays after applying the offset, but I'm not sure how to do it and whether it's the easiest way.