5
$\begingroup$

Let $f:\mathbb{R}\to \mathbb{R}$ be a function in $L^2$ satisfying $|f(x)|\ll e^{-a_1 x}$, $a_1>0$, for $x\to \infty$. (Variant: assume as well that $|f(x)|\ll e^{a_2 x}$, $a_2>0$, for $x\to -\infty$.) What is the fastest $\widehat{f}(t)$ can decay (in both directions)?

EDIT: I should emphasize I would be very interested in a simple, concrete example that shows that the bound below (which is correct, as has been confirmed) is tight.


I can show that $\widehat{f}(t)$ cannot decay as fast as $e^{-e^{r |t|}}$ (even in one direction) if $r\cdot \max(a_1 a_2) >\pi^2$. (A one-sided assumption is probably enough, though I haven't checked. ) Can that decay be reached for some $r>0$? (Perhaps even for $r = \pi^2/\max(a_2,a_2)$?)

--

UPDATE: what if $|f(x)|$ decreases faster than any $e^{-A x}$, $A$ arbitrary, as $x\to \infty$? How fast can $\widehat{f}(t)$ decay then (in both directions)?

$\endgroup$
6
  • 1
    $\begingroup$ Maybe something in the question isn't clear to me, but isn't a Gaussian $f(x) = e^{-x^2}$ a counterexample to the conjecture? Both $f(x)$ and its Fourier transform $\hat{f}(t)$ decay faster than any exponential (with linear argument). $\endgroup$ Commented Nov 4, 2022 at 0:27
  • 1
    $\begingroup$ Sure, I had that in mind, and just gave that example below. Notice the double exponential in my question, though: I am asking whether one can find $f$ that decays exponentially such that $\widehat{f}(t)$ decays as fast as $\exp(-\exp(r |t|))$ (double exponential). $\endgroup$ Commented Nov 4, 2022 at 1:24
  • $\begingroup$ I just realized that I asked a very similar question nine years ago: mathoverflow.net/questions/111826/… Back then, my condition on $f$ was two-sided, and what I wished to obtain on $\widehat{f}$ was one-sided, but the desired bounds were otherwise the same. $\endgroup$ Commented Nov 4, 2022 at 1:26
  • $\begingroup$ Of course, sorry, I missed the double exponential in the lower estimate on $\hat{f}(t)$. My eyes just skipped over that and saw it as a single exponential. $\endgroup$ Commented Nov 4, 2022 at 10:56
  • $\begingroup$ For $\sigma>0$, I think the transform of $f(x):=\Gamma(\sigma+iax)$ is $$\widehat {f}(u)=\frac{2\pi}{a}\exp(2\pi \sigma u/a) \exp(-e^{2\pi u /a}).$$ Therefore, $\widehat{f}(u)\widehat{f}(-u)$ is the transform of a convolution of two gamma functions. (I am translating from my writing in 2006, this explain the parameters) $\endgroup$
    – juan
    Commented Nov 18 at 8:47

1 Answer 1

6
$\begingroup$

The simplest result can be stated when both your conditions hold. If they are satisfied, your Fourier transform is analytic and bounded from above in the strip $-a_2<\mathrm{Im}\, s<a_1$, and bounded from above in any smaller strip. Such function has a limit rate of decay $\exp\left(-e^{\pi|s|/a}\right)$, where $a=a_1+a_2$ is the width of your strip, in the sense that if $|\hat{f}(s)|\leq \exp\left(-e^{(\pi/a+\epsilon)|s|}\right)$ when $s\to +\infty$ or if $s\to-\infty$, then $f\equiv 0$. This is a simple consequence of the Phragmen-Lindelof Principle, see for example,

B. Levin, Lectures on entire functions, AMS 1996.

With a bit more labor (using Poisson integral for a strip) one can prove that this remains true when $a_1=0$ or $a_2=0$, that is when we have one-sided decay condition on $f$. More precisely, we have a bounded analytic $\hat{f}$ function in the strip $0<\mathrm{Im}\, s<a$, say continuous at the boundary, and we want to know how fast it can decrease on the real line. Let $u$ be the least harmonic majorant of $\log|\hat{f}|$. Then, as a harmonic function bounded from above, it must be equal to the Poisson integral of its boundary values. In particular, this Poisson integral must converge. This gives the maximum possible rate of decrease (of $u(x)$ to $-\infty$). Since Poisson kernel for the strip of width $a$ is $\mathrm{Im}\coth(\pi z/(2a))$ we obtain the result.

Function $\hat{f}(s)=\exp(-\cosh \pi s/a)$ shows that the result is best possible. This function is in $L^2$ so $f$ is also in $L^2$. Correct exponential decrease of $f$ is easy to prove directly.

Edit. I found a reference, which also contains an example that shows that it is best possible, even in the class of probability densities ($f\geq 0$):

Ju. Linnik and I. Ostrovskii, Decomposition of random variables and vectors, AMS 1977,

Chap. II, section 4, at the very end of this section. Their example has an additional feature that $f\geq 0$, and for this reason it requires some labor.

$\endgroup$
12
  • $\begingroup$ Wait. What if $f(x)$ is the Gaussian $f(x) = e^{-\pi x^2}$? Then $f$ is its own dual, and decays faster than any exponential in both directions. $\endgroup$ Commented Nov 4, 2022 at 1:20
  • 1
    $\begingroup$ @H. A. Helfogt: I corrected. Thanks for your remark. Phragmen-Lindelof is of course applied to $\log|\hat{f}|$. $\endgroup$ Commented Nov 4, 2022 at 1:28
  • $\begingroup$ Ah, sure, that's what I was saying I can show. For a one-sided condition, I'd try the Bargmann transform and then the Phragmen-Lindelof principle. $\endgroup$ Commented Nov 4, 2022 at 1:39
  • $\begingroup$ But is that tight? Can you construct a non-zero example with $|\hat{f}(s)|\ll \exp(-e^{(\pi/a) |s|})$? $\endgroup$ Commented Nov 4, 2022 at 2:49
  • $\begingroup$ @H. A. Helfgott: one can easily construct such an analytic function. More labor is required to construct a positive definite one, so that it is a Fourier transform. I am too lazy to work this out, but I am sure that the result is best possible. $\endgroup$ Commented Nov 4, 2022 at 2:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .