Forums > MaxMSP

Cross Correlation

May 12, 2014 | 9:52 am

Hi Forum,

looking for a way to cross correlate two audio signals using fft~, I only found older threads that ultimately didn’t lead to any results. Do any of the original posters remember how they ended up solving their problem? Does anybody have any experience with cross correlation here?

Thanks


May 14, 2014 | 5:13 am

bump. Or rather bang.


May 15, 2014 | 2:09 am

correlation is mathematically very similar to convolution and yes, both can be performed in the frequency domain.
so, if you know how to do fast convolution correctly (complex multiplication, correct windowing and overlap) you just have to time reverse one of the input signals before convolution and you should end up with cross correlation.
another thing is handling the result, i.e. extracting the information that you need.


Viewing 3 posts - 1 through 3 (of 3 total)