Roughly speaking, the periodogram can be used to determine which frequencies are found in a signal.
However, why is the periodogram defined as the Fourier transform of the signal's autocorrelation? Why don't we take the Fourier transform of the signal itself if we're just looking for frequencies in the signal (I thought this would be more direct and simpler)?
If someone can help me understand why we need to use the autocorrelation, I would appreciate it very much. Thanks.