1

I am analyzing some interferometry data and have a time signal that looks like this:

enter image description here

Our temporal resolution is unfortunately not enough to resolve the fringes.

I need to get the upper envelope of the signal to perform some analyses. I have tried everything (in Python) such as getting the Hilbert transform, using a low-pass filter, a Savitzky–Golay filter but no luck so far in getting something I can work with.

Any suggestions for dealing with the data would be very appreciated.

JRE
  • 2,240
  • 1
  • 11
  • 20
Regulus
  • 11
  • 1
  • It might help if you described the limitations of the approaches you've tried so far. – MBaz May 10 '22 at 21:50
  • With the Hilbert transform, its absolute value mostly behaves as the original signal, which does not help at all. With both of the filters, I am unable to smoothen the signal without having the processed data envelope completely losing its ressemblance to that of the original one. – Regulus May 10 '22 at 21:53
  • Try something simple: e.g. absolute value followed by a low-pass filter. Then post the results and describe what you don't like about them – Hilmar May 10 '22 at 22:29
  • What Hilmar said and see this post for an example: https://dsp.stackexchange.com/questions/81944/envelope-detection-using-hilbert-transform/81945#81945 – Dan Boschen May 10 '22 at 23:30
  • Thanks guys, both your comments helped me finally solve the problem. – Regulus May 11 '22 at 03:11
  • 1
    Check out the information at: https://www.dsprelated.com/showarticle/938.php – Richard Lyons May 11 '22 at 09:31
  • You might want to have a look at this post. – A_A May 12 '22 at 10:03
  • It would help if you posted/pointed to the actual data. Then what signal you expect and what the noise source is. – rrogers May 18 '22 at 00:21

0 Answers0