Should I use the absolute value of fast Fourier transform or the conjugate in signal analysis?

I found this example in MATLAB https://www.mathworks.com/help/matlab/math/fft-for-spectral-analysis.html and I saw that they used the value of conjugate.

Pyy = Y.*conj(Y)/251;

I tried to calculate the magnitude using abs function rather than this line. I get nearly the same output.

Therefore, I am puzzled; should I use the way in this example or the magnitude way?

Answer 1

[Star Strider]

Note that the plot is ‘Power Spectral Density’. Power is the square of magnitude.

The abs value is the square root of the complex value times its conjugate.

[Muhamed Sewidan]

Thank you.

One more question, the example used a second input argument in fft function, which is the number of the time points. Is it necessary to define the number of points inside the Fourier function?

[Star Strider]

As always, my pleasure!

The second argument is not necessary. The fft function will use the row length of the argument to determine the row length of the output. The second argument is only necessary if you want to zero-pad the inplut in order to increase the frequency resolution of the output.

Answer 2

[Bruno Luong]

Y.*conj(Y) is abs(Y)^2

[Muhamed Sewidan]

Thank you.