0
$\begingroup$

Here is an image of a filter bank spectrograph, where the main signal is at 2.5kHz and the high signal is at 21.5khz,

What is the high frequency signal? I doubt that birds can make frequencies above 18kHz, so it is probably an effect from the recording equipment? Is it an issue from generic FIR BPF at high frequencies?

enter image description here

enter image description here

$\endgroup$
  • 2
    $\begingroup$ That might well be an artifact, but the question may be impossible to answer without knowing more about the signal, and about your capture and analysis systems. $\endgroup$ – MBaz Oct 24 '19 at 17:37
  • $\begingroup$ and the problem is that we're not ornithologists. I'd strongly recommend you start to "look behind the pictures": You can't ask us about things in pictures that just represent a reduced version of some other signal that we don't know. Maybe these are just artifacts, maybe they're actually there, maybe you're just misinterpreting things - we can't know. So, describe the original signal, and how it comes to be. $\endgroup$ – Marcus Müller Oct 24 '19 at 17:38
  • $\begingroup$ Hi, it's a recording by RainSamba from Cologne, Germany, using a Vivanco EM35 with preamp into Sony DAT-recorder. I had read that birds didn't sing at high freqencies, although I have just found a page that states their call goes to 45khz... storianaturale.comune.fe.it/modules/core/lib/d.php?c=oqXZ3 $\endgroup$ – com.prehensible Oct 24 '19 at 18:25
  • $\begingroup$ @com.prehensible thanks! That's about half the info we'd need: sampling rate, and how this visualization works would also be necessary. $\endgroup$ – Marcus Müller Oct 24 '19 at 19:14
3
$\begingroup$

If the sample rate of data fed to a DFT or FFT for spectrum analysis was around 24 KHz, then what you are seeing is just the complex conjugate negative frequency image of the low (2.5 kHz) frequency audio data. Basically, one should ignore any spectral information in a spectrogram for any frequency around or above half the sample rate, as it is redundant, just a (log distorted?) mirror image of the magnitude spectrum below.

If you want valid spectrographs for higher frequencies, use a higher sample rate, both for acquisition and analysis, with a higher anti-aliasing low pass filter cut-off.

But make sure to low pass filter any data to below half the sample rate before acquiring or decimating for any spectrum analysis, or you will see aliasing artifacts.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.