Fourier coefficients relation to Power Spectral Density

1 reply · 4K views · last reply
Skaiserollz89
Messages
9
Reaction score
0
TL;DR
Help deriving a result found in "Numerical Simulation of Optical Wave Propagation" by Jason Schmidt. I'm trying to work out by hand an equation stating that the ensemble average of the squared fourier coefficients of a 2D phase function equals the Power Spectral Density( Phi(fx,fy) multiplied by 1/A, where A is the domain area ( either delta_fx*delta_fy in frequency space, or 1/(L_x*L_y) in real space). I am having trouble seeing how to get this result. Please assist in the derivation.
Typeset page of equations: a double Fourier series phi(x,y) = sum over n and m of c_{n,m} exp(i2pi(f_{x_n} x + f_{y_m} y)), the claim <|c_{n,m}|^2> = Phi(f_{x_n}, f_{y_m}) Delta f_{x_n} Delta f_{y_m}, and Parseval's theorem relating the double integral of |phi|^2 to the integral of Phi.

Typeset equations (4) and (5): |c_{n,m}|^2 equals a double integral over x and y of Phi(f_{x_n}, f_{y_m}) dfx dfy, and |c_{n,m}|^2 is approximately a double sum of Phi(f_{x_n}, f_{y_m}) times Delta f_x Delta f_y, with surrounding explanatory text.

Typeset equations from a LaTeX document: approximation of squared Fourier coefficient |c_{n,m}|^2 as Phi(f_{x_n}, f_{y_m}) times a double sum of delta f_x delta f_y, followed by the ensemble average result <|c_{n,m}|^2> = Phi(f_{x_n}, f_{y_m}) delta f_{x_n} delta f_{y_m}, with surrounding explanatory text.