Problem #4

UCI. MAE200B, winter 2006. by nasser Abbasi.

Problem: Prove the convolution theorem for the Fourier transform MATH

Solution

Some definitions first. The Fourier transform is defined as (using textbook definition)

MATH

and the inverse fourier transform is

MATH

And the convolution is defined as

MATH

Start the proof by showing that MATH, this implies that MATH

So, take the fourier transform of the convolution, we obtain

MATH

Let $z=t-\tau $, hence $t=z+\tau $, and so $dt=dz$. When $t=\pm \infty $, $z=\pm \infty $ and the above can be written as

MATH

Exchange the integrals order we obtain

MATH

Break the exponential into product of 2 exponential

MATH

We can now move the marked terms outside the inner integral since they do not depend on $z$

MATH

But by definition MATH and similarly MATH, hence eq (1) can be written as MATH

This results means that MATH

This completes the proof. (We should also proof the other direction since this is an IFF problem, i.e. starting from MATH take its inverse fourier transform we would get the convolution). But I think this proof done here should be sufficient.