When taking the FT of a signal I always get a sharp peak at exactly half the Nyquist frequency. My signal is shown here:
and its FT here:
The Nyquist frequency is 36.7 KHz. As can been seen in the above image, a sharp frequency peak appears at half of this, at 18.38 KHz.
The signal is the a column vector containing the mean of the rows of an image with random white noise. Therefore, I do not expect to see such an exact peak at half the Nyquist frequency. From the looking at the signal, I get why there is a delta function at 0 Hz (which contains most of the power), as it is basically a straight line with random noise, and the noise is then represented by the higher frequencies at lower amplitudes. The other peaks visible in the FT vary each time the data is captured, but the central peak always stays in the same place and has the same amplitude. This leads me to suspect that it may not be part of the signal and that there may be some mathematical reason I do not understand why I would get such a peak (perhaps in some cases it is common to get a peak at half the Nyquist, or maybe it is related to the delta function in some way?). If anyone could shed some light on this or has seen a similar phenomenon with random data please let me know. Thanks.