I have a large data set that was collected by a 5 GHz oscilloscope, so I basically have an array of voltages with each index being a very small time step apart. I am looking for a specific signal in this data set that would only be present for a very small time frame and so most of the data will not contain this signal. The issue is that the signal I am looking for has an amplitude less than or equal to that of the background noise. The signal should also be a fairly wideband.

My question starts then with how to find this signal? My first thoughts are there must be something I can do to take advantage of the fact that my noise and signal probably follow different distributions, but I am not sure.

I know this is a fairly open-ended question so I am just looking for suggestions into general methodologies that help uncover a signal that is not consistent with background noise, but of the same or lesser amplitude.

  • $\begingroup$ Do you know exactly what the signal will look like? $\endgroup$ – Jim Clay Apr 6 '14 at 22:33
  • $\begingroup$ not exactly sure, but it should occur over a relatively brief period of time compared to the sampled data time frame and should be decaying in frequency $\endgroup$ – PumpkinPie Apr 6 '14 at 22:37
  • $\begingroup$ How will you know you've found the signal if you aren't sure what it looks like? (e.g. Why wouldn't a random answer be good enough?) $\endgroup$ – hotpaw2 Apr 6 '14 at 23:23
  • $\begingroup$ I know the exact moment in time when the signal should appear, and I know it should be decaying in frequency. That is all I have to go on though. $\endgroup$ – PumpkinPie Apr 6 '14 at 23:43
  • $\begingroup$ Ah, so you know when it happens, and from the sampled data at that point you want to figure out what it looks like. Is that right? $\endgroup$ – Jim Clay Apr 7 '14 at 0:25

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.