After reading this question: PSD (Power spectral density) explanation I am still a little confused as to what extra information the PSD gives us over simply taking the magnitude of the fourier transform.

So my understanding is that the PSD is equal to:


but what does that tell me that simply


does not. I am assuming that squaring the fourier transform's magnitude somehow makes the number more physically (or mathematically) relevant? So how do I interpret the values of the power spectral density, especially when the signal is not a physical one and hence has no actual energy (say for example a financial signal or an image)?

Also, does the power spectral density contain information about the probability of the frequencies in the signal? As in is there a way to use the power spectral density to make a statement about the significance level of making a claim about a harmonic frequency in a stochastic signal? So something kind of like checking the significance of the coefficients from an AR model?


1 Answer 1


it doesn't tell you new information, but because the abs value function is non-analytic (i.e. not all of its derivatives are continuous), and the magnitude-squared is, then the latter can be manipulated mathematically in ways that the former cannot. one important property of the power spectrum is that it is the Fourier transform of the autocorrelation function in the time domain.

the PSD gives you the energy of each frequency component for a deterministic signal and the expectation value of the energy for each frequency component for a random (or "stochastic" if you like jargon) signal.

  • $\begingroup$ Can you elaborate on what "expectation value of the energy" means please? Also when you say the energy for each frequency, does that mean if I have a voltage signal made up of two sine waves, the PSD tell me what portion of the energy of the total signal is due to each frequency component? I don't really understand how that translates to say a financial signal. $\endgroup$
    – Dan
    Commented Mar 6, 2014 at 15:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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