In my signal processing pipeline, I normally expect data sampled at 100Hz, on which I then apply a median filter of width 10.

My question is: How should I change the width for data sampled at different rates?

For example, say an incoming data is sampled at 50Hz. Should I then use a median filter of width 5 to be consistent? In other words, should I choose the median filter based on the physical time length (0.1 seconds in our case) instead of number of ticks?

  • 2
    $\begingroup$ That all depends on the purpose of your median filtering – what do you do it for? $\endgroup$ Apr 15 at 11:01
  • $\begingroup$ @MarcusMüller I apply it to smooth the signal to remove potential outliers/noise. In particular, this is wearable sensor data (like Fitbit). $\endgroup$
    – Michael
    Apr 15 at 16:13
  • $\begingroup$ I think the answer is yes. $\endgroup$
    – ZR Han
    Apr 16 at 2:01

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.