I have two distinct signals and I would like to compute their correlation. However, I think that just computing the crosscorrelation does not work here, since the signals are from different sources — also different amplitude, and length.

I would like to see that, for example, if the variance of the first goes up, the same happens for the second one. So, I plotted the variance, but I didn't see much. Are there any other characteristics I could look at? Energy?

By the way, I do it in Python.

  • $\begingroup$ Hi: in time domain, you can use cross-correlation if you de-mean both series first. I'm not sure what the analog of cross-correlation is in DSP world. I think it may be called coherence but I'm not sure. $\endgroup$ – mark leeds Oct 28 '19 at 1:24
  • $\begingroup$ A plot of those signals would help. $\endgroup$ – Rodrigo de Azevedo Oct 28 '19 at 7:12
  • $\begingroup$ This screams XY Problem. What are you actually trying to achieve? You must think that one signal is affecting the other, or that there's some third actor that's affecting both signals. $\endgroup$ – TimWescott Oct 28 '19 at 20:15
  • $\begingroup$ Yes there should be a third actor affecting both. I have plot the variances of both and at some points in time I can see if the variance of one is increasing, also the variance of the other one is increasing. But I would like, if there are additional signal properties with which I can "prove" (or just show) a correlation. $\endgroup$ – user45882 Oct 28 '19 at 23:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.