I only know about this from images and some videos and articles I've read, but I haven't managed to find an explanation. It only says they exist. This is what I mean: enter image description here

Blue is the frequency domain of a given signal, while green are copies that appear for some reason.

And a side question - when I do fft in matlab, these copies don't appear so are they automatically filtered out in matlab fft function?

  • 2
    $\begingroup$ so, what articles did you read? "there's images of the original signal" isn't something anyone tells on their own, it will follow after an introduction to the topic of discrete signals. Could you try to pinpoint where you're stuck in understanding this? This all sounds like you just need to turn back a few pages in your book on the basics of signal theory. You see the $f_s$ at the axes, right? So, do you know what that is? If not, that's the first thing to learn... $\endgroup$ Sep 18 '21 at 19:11
  • $\begingroup$ Asking us where spectral images come from is really asking for us to write the first couple of chapters of any signal theory book as answer - VTC as too broad. $\endgroup$ Sep 18 '21 at 19:13
  • 1
    $\begingroup$ Even though your question is probably still too broad, please edit it to state what articles and videos you're talking about. That looks like the sort of spectral replication that you get from sampling -- which, as pointed out, is explained in signal processing books. $\endgroup$
    – TimWescott
    Sep 18 '21 at 20:57

It's the spectrum of a discrete signal: sampling in time $\Leftrightarrow$ periodizing in frequency - explained in detail here. Overlap means there's aliasing, and we require a higher sampling rate. FFT returns one period of this spectrum since it contains all the information. (More precisely, your image shows DTFT, and DFT (which FFT implements) is a sampling of DTFT).


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