6

Suppose $f$ is a test function and thus compactly supported on some set $S$. Its Fourier transform is

\begin{equation} \widehat{f}(p) = \int_{-\infty}^{\infty} dx \, e^{ipx} f(x) \, . \end{equation}

Is it possible for this Fourier transform $\widehat{f}$ to vanish on some interval $(a,b)$?

I have tried a number of things and believe that this should not be possible, also with regard to this question. More generally, is this possible if $f$ is a Schwartz function?

I would also appreciate if anyone could direct me to a general discussion of the relation between the support of a function and that of its Fourier transform.

2 Answers2

9

Of course not. Since $f$ has compact support, then $\hat{f}$ is the restriction to $R$ of a function which is analytic in the whole complex plane (ie an entire function). It cannot be identically 0 on an interval (principle of isolated zeros) if $f$ is not zero almost everywhere.

6

Concerning the Schwartz class the Fourier transform is an isomorphism from $\cal{S}(\mathbb{R})$ onto itself. As $\cal{S}(\mathbb{R})$ contains functions with compact support, the answer is yes.

  • 1
    I don't quite follow why the isomorphism property (as inner product spaces I assume) implies the result. It could be that all images of compact support functions do not have the desired property. – Zarathustra Sep 10 '24 at 18:23
  • 1
    We have $f\in S$ iff $\widehat{f}\in S.$ I have added a link in my answer. – Ryszard Szwarc Sep 10 '24 at 18:42
  • 3
    The question in OP: "Does there exist a nonzero function in $S$ so that its Fourier transform vanishes on some interval ?" The answer is yes. The answer is no for functions with compact support. – Ryszard Szwarc Sep 11 '24 at 06:09
  • Thank you, that does indeed answer the question. – Zarathustra Sep 13 '24 at 03:12
  • Does not test functions always have compact-support? – Joako Sep 13 '24 at 04:26