0

I have absolutely basic question about sampling rate with regard to this tutorial: http://www.mathworks.com/help/signal/examples/measuring-signal-similarities.html (first section)

Two first signals have FS=4096, while the third is 8192. In simple words, the third signal was sampled two times more frequently. They say that to equalize the sample rates "The safest way to do this is to resample the signal with a lower sampling rate". So, their procedure actually takes every second sample of the two first signals. But wasn't the third signal was sampled two times more frequently? My intuition was exactly the opposite: if something was sampled more frequently, then I would take every second sample of it.

What do I miss? Thanks a lot,

Vadim

user1597969
  • 405
  • 2
  • 5
  • 13

2 Answers2

2

I think the sentence "The safest way to do this is to resample the signal with a lower sampling rate." is perhaps omewhat ambiguous - what it should say is "The safest way to do this is to resample the signal with the lower sampling rate.", i.e. you need to upsample the signal with the lower sample rater to make it the same as the other signal, that way you don't lose any information (as you would if you downsampled the signal with the higher sample rate).

Paul R
  • 208,748
  • 37
  • 389
  • 560
  • 1
    OK. I think I see what I missed. The resample code they wrote for any single sample adds additional value, which is interpolation between two proximate value. So, I artificially get two times higher sampling rate. – user1597969 Dec 16 '13 at 10:52
1

"The safest way to do this is to resample the signal with a lower sampling rate" thus in that case procedure will not take like You stated every second sample (if so the first and second signal will have 4 times less samples than third one), it will simply interpolate samples between two adjacent points for first and second signals.

In my opinion it is easier to resample the signal with lower sampling rate by interpolation than by truncation of samples from signal with higher fs (in that case aliasing may occur).

pezetem
  • 2,503
  • 2
  • 20
  • 38