In Fourier analysis, a central theorem for the Fourier Transform states:

$$\mathcal F\{(f*g)(t)\}(\omega)=\mathcal F \{f(t)\}(\omega)\cdot \mathcal F\{g(t)\}(\omega)$$

In other words, convolution turns into multiplication.

In turn convolution is defined as $$(f*g)(t)=\int_{-\infty}^\infty f(\tau)g(t-\tau)d\tau$$

A correllation can be defined similarly:

$$corr(f,g) =\int_{-\infty}^\infty f(\tau)g(\tau-t)d\tau$$

How can we figure out which (if any) integral transform which has a similar rule for correllation as the Fourier Transform has for convolution?

  • 1
    $\begingroup$ There is the so called cross correlation theorem archive.lib.msu.edu/crcmath/math/math/c/c778.htm $\endgroup$ – Davide Morgante Jul 29 '18 at 20:08
  • $\begingroup$ @DavideMorgante that is of interest but what does $f^*(t)$ mean there? $\endgroup$ – mathreadler Jul 29 '18 at 20:11
  • 1
    $\begingroup$ The asterisk usually means the complex conjugate! Your definition of cross correlation is valid if $f$ is real, because usually cross correlation is defined as $$(f\star g) =\int_{-\infty}^\infty f^*(\tau)g(\tau+t)d\tau = \int_{-\infty}^\infty f^*(\tau-t)g(\tau)d\tau \overset{\mbox{ }f\mbox{ real }}{=} \int_{-\infty}^\infty f(\tau-t)g(\tau)d\tau$$ $\endgroup$ – Davide Morgante Jul 29 '18 at 20:17
  • $\begingroup$ ok good it is as I suspected, thank you. feel free to add as an answer if you will. $\endgroup$ – mathreadler Jul 29 '18 at 20:17

Surely there is! It's called the cross correlation theorem for Fourer transform. There's also a similar theorem for Laplace transform (the theorem is stated in the table in the link given, just search for "cross correlation").


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.