2
$\begingroup$

I have two audio signals, both with supposedly matching frequencies at each time window. What I would like to do is to verify that the frequencies are identical. However, one was created at a higher power than the other, so simply checking for a match at each sample won't work. I've already cross-correlated the signals, so there's no time lag between them.

Is there any way to compare the "shapes" of two waveforms of differing amplitudes? If there's some way to drop the amplitude domain from the signal, that would work - I don't know nearly enough about DSP to know if it's easy or not.

$\endgroup$
  • $\begingroup$ try analyzing in frequency domain $\endgroup$ – Sufiyan Ghori Mar 25 '13 at 16:26
  • 1
    $\begingroup$ Normalized cross-correlation should be sufficient, as long as you are sure that at least one of the signals is "correct". $\endgroup$ – Jim Clay Mar 25 '13 at 17:07
  • $\begingroup$ "both with matching frequencies at each time window" meaning the frequency changes from one window to the next? $\endgroup$ – endolith Mar 26 '13 at 20:32
0
$\begingroup$

Sounds to me like you need to look at the signals in the frequency domain. Take the Fourier transform of each signal (use an FFT), and look at the spectrum. I'm not sure if you've got signals with just a single frequency (a tone) or with many frequencies, but in either case if they have the same frequencies in them then the spectra should look the same (except for one being larger than the other).

If you're unfamiliar with frequency domain concepts, I'd suggest you read up on them first. The wikipedia page on Fourier transforms is pretty large, so I'm not sure how good that would be as an introduction, but there are a large number of books on the topic and many websites too.

$\endgroup$

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.