I had a course in PDE last year where we used fourier transforms extensively; I understand the rules of manipulation and can prove the scaling theorem directly from the definition using a substitution, but I don't really have any good intuitive argument as to why "compressing" a function causes an expansion of its fourier transform, and vice versa. I have been trying to gain solid intuition behind the various properties of the fourier transform; but have not gotten far with this one. If anyone knows of a website / book or a slick argument that covers this; it'd be greatly appreciated. Thanks!

  • 2
    $\begingroup$ What is the scaling theorem? $\endgroup$ Jul 6 '11 at 5:07
  • $\begingroup$ @Qiaochu It is usually called the similarity theorem. If $\mathcal{F}(u)$ is the FT of $f(x)$ then $\frac{1}{a}\mathcal{F}\left(\left|\frac{u}{a}\right|\right)$ is the FT of $f(ax)$ $\endgroup$
    – kuch nahi
    Jul 6 '11 at 5:59

Intuitively, One way to think about it is that compressing a sinusoid, increases its frequency. So compessing a sum of sinusoids will expand the frequency spectrum.

  • $\begingroup$ Ok cool, thanks. I thought about it a bit and I can see that compressing a signal is equivalent to compressing its component frequencies, which would then cause the frequency spectrum to be spread over a larger range. Is there a good way to see why the amplitude of each frequency's contribution is the scaled by the reciprocal of the compression factor? $\endgroup$
    – Steve
    Jul 6 '11 at 6:28
  • $\begingroup$ the amplitude scaling is due to parsevals. compressing the signal reduces the energy. $\endgroup$
    – Mykie
    Jul 6 '11 at 13:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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