The original question was posted here.
I have a signal, which I'd like to treat as a non-continuous function now, let it be $signal(t)$. It looks like this:
Zoomed in a bit:
I create a Lomb-Scargle Periodogram using $signal(t)$, which I understand to be very similar to non-uniform FFT (that is to say, the $t$s for which $signal(t)$ is defined, are not always separated by equal distance - as it is the case now).
The computed plot, frequency vs power:
I am happy with the large peak on the left side of the plot, but I am confused about the peaks centered around 48. Zoomed in on those:
They seem to be a reflection of the peaks on the left. I suspect that it is something to do with the Nyquist frequency. The difference between largest and smallest $t$ is around $27.8$, and I have around $1280$ values of $t$ (nearly but not exactly equidistance from each other), so the sampling rate is $1280/27.8 \approx 46$. This is closed to the previously observed $48$, but also clearly distinct from it (ie the previously observed 48 was an approximation, not an exact value, but it was definitely not $46$).
How can I explain this peculiarity of the frequency-power plot of this function?