How can I perform Fourier Transformation for non periodic signals

[AndrazS]

I have a signal that looks like this:

It is clear that the signal is not periodic. So how can I perform Fourier Transformation to extract harmonics from the signal?

[wiebe@CARYA]

---

@AndrazS wrote:

So how can I perform Fourier Transformation...

---

Exactly the same as a periodic signal. The difference is that the results are usually useless 🤗.

---

@AndrazS wrote:

So how can I perform Fourier Transformation to extract harmonics from the signal?

---

Use a Short Time Fourier Transform, that takes a joint time\frequency approach.

Take a look at the STFT Spectrogram.vi and STFT Spectrogram PtBtPt.vi.

[Henrik_Volkers]

If you say harmonics, .. harmonic to what?

I assume a basic frequency, so first you find that frequency (tone detection? extract single tone?)  , cut the data to the longest multiple periode of that frequency , do a DFT

more enhanced: find the basic frequency, do a linear least RSME fit with that freq and n harmonics...

nearly the same, but you define the exact location of the now reduced frequency bins, independent of samplerate and other frequency content

if you only have a guess of the basic frequency you can do an extendet 4-parameter sine fit including the harmonics BTDT but can't share the code

[altenbach]

You can Fourier transform any signal and it will tell you if there are dominant global harmonics or not.

There is definitely a dominant high frequency component of about 10 oscillations per mm.

There also seem to be a few slower oscillations at the very end. To get these, you could do a jtfa as has been mentioned. Look into wavelets. (Toolkit needed).

The x-axis is distance (?), so what are the units of the fft? What does the data represent? What do you want to get out of it at the end? Could you attach the actual data?