Inverse Fourier Transform of the absolute value of H(f) squared

510 Views Asked by At

Given that we have a frequency response $H(f)$,

What would the inverse fourier transform of

$$|H(f)|^2$$

be in terms of a convolution?

I think it would be simply, $$h(t)*h(t)$$

But the answer seems to be $$h(t)*h(-t)$$

Am I missing something? Is it the absolute value that makes the convolution different?

1

There are 1 best solutions below

0
On BEST ANSWER

Note that in general, i.e., for $h(t)\in\mathbb{C}$, the inverse Fourier transform of $|H(f)|^2$ is given by

$$\mathcal{F}^{-1}\big\{|H(f)|^2\big\}=h(t)\star h^*(-t)\tag{1}$$

where $\star$ denotes convolution, because

$$\mathcal{F}^{-1}\big\{H^*(f)\big\}=h^*(-t)\tag{2}$$

and obviously, as pointed out in a comment, we have

$$|H(f)|^2=H(f)\cdot H^*(f)\tag{3}$$