I am applying wavelet transform on functional Near Infrared Spectroscopy (fNIRS) signal.

My aim is to get the info that which coefficient corresponds to which frequency in the signal. Is it possible to do so, moreover if there can be some other way please mention it.

  • 1
    $\begingroup$ wavelets are kind of a hybrid time-localized / non-localized frequency estimator. Are you sure you want to use wavelets if you actually want to know freqeuncy? That sounds like a job for the DFT. $\endgroup$ Jan 17 '20 at 12:31
  • $\begingroup$ What if I want to do it using wavelet transform only. Is there any way sir?@MarcusMüller $\endgroup$
    – Warrad
    Jan 18 '20 at 12:22
  • $\begingroup$ Why would you want to use a tool that's giving you a different result than what you need? (That's an honest question, when you say "you want", I assume you have a technical reason.) $\endgroup$ Jan 18 '20 at 12:34
  • $\begingroup$ I know I can do this using FT $\endgroup$
    – Warrad
    Jan 18 '20 at 14:29
  • $\begingroup$ yes, but what's the reason not to? You got to have a reason, because when you look at the FT, it's nothing but the actual projection of your signal onto "prototype" frequency signals; i.e. there can't be an integral transform that is any more optimal for harmonic frequency than the FT, so by choosing to use wavelets, you're actively doing something worse in terms of estimating frequencies. So, what's your reason for using wavelets? From that, we can infer what you'd want to do to reach your goal. $\endgroup$ Jan 18 '20 at 14:32

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.