Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I need prove, that a distribution $$\langle F_f,\phi \rangle= p.v. \int\limits_{-1/2}^{1/2}\frac{\phi(t)}{t\cdot \ln{|t|}}\mathrm{d}t$$ belongs to $S^\prime$ (adjoint to Schwartz space) and I need find, Fourier transform of the distribution.

share|cite|improve this question
Is $F_f$ supposed to be the Fourier Transform of $f$? How is $f$ incorporated on the right-hand side? – robjohn May 23 '12 at 18:14
$f~$ is generating function for $F_f~$, so $$f=\frac{1}{t\cdot \ln{|t|}}$$ – user31497 May 23 '12 at 18:45
So are you looking for $$ \int_{-1/2}^{1/2}\frac{e^{-2\pi i\xi t}}{t\log|t|}\,\mathrm{d}t $$ ? – robjohn May 23 '12 at 18:57
Yes, but I think there are some trick in this situation. And I can't prove, that $F_f\in S~'$. It's a main problem. – user31497 May 23 '12 at 19:18
No, every distribution is infinitely differentiable (in the weak sense). – Vobo May 24 '12 at 7:20
up vote 5 down vote accepted

If $\phi\in\mathcal{S}$, $\psi(t)=\frac{\phi(t)-\phi(0)}{t}\in C^\infty$, and by the Mean Value Theorem, $\|\psi\|_{L^\infty}\le\|\phi'\|_{L^\infty}$ Thus, $$ \begin{align} \lim_{\epsilon\to0}\left(\int_{-1/2}^{-\epsilon}\frac{\phi(t)}{t\log|t|}\,\mathrm{d}t+\int_{\epsilon}^{1/2}\frac{\phi(t)}{t\log|t|}\,\mathrm{d}t\right) &=\lim_{\epsilon\to0}\left(\int_{-1/2}^{-\epsilon}\frac{\psi(t)}{\log|t|}\,\mathrm{d}t+\int_{\epsilon}^{1/2}\frac{\psi(t)}{\log|t|}\,\mathrm{d}t\right)\\ &=\int_{-1/2}^{1/2}\frac{\psi(t)}{\log|t|}\,\mathrm{d}t\tag{1} \end{align} $$ Therefore $$ \begin{align} \left|\text{p.v.}\int_{-1/2}^{1/2}\frac{\phi(t)}{t\log|t|}\,\mathrm{d}t\right|&=\left|\int_{-1/2}^{1/2}\frac{\psi(t)}{\log|t|}\,\mathrm{d}t\right|\\ &\le\left|\int_{-1/2}^{1/2}\frac{1}{\log|t|}\,\mathrm{d}t\right|\;\|\psi\|_{L^\infty}\\ &\le\frac{1}{\log(2)}\|\phi'\|_{L^\infty}\tag{2} \end{align} $$ Inequality $(2)$ says that $\langle F_f,\cdot\rangle\in\mathcal{S}'$.

The Fourier Transform of the distribution would be $$ \begin{align} \text{p.v.}\int_{-1/2}^{1/2}\frac{e^{-2\pi i\xi t}}{t\log|t|}\,\mathrm{d}t &=-i\int_{-1/2}^{1/2}\frac{\sin(2\pi\xi t)}{t\log|t|}\,\mathrm{d}t\\ &=-2i\int_0^{1/2}\frac{\sin(2\pi\xi t)}{t\log|t|}\,\mathrm{d}t\tag{3} \end{align} $$ I can not find a closed form for $(3)$.

share|cite|improve this answer
Thank you very much. But if we try to take the inverse Fourier transform, then how can we justify the irregularity of the resulting function. – user31497 May 23 '12 at 22:15
@user31497: I think what you are asking is that since the Fourier transform of an $L^1$ function is bounded, how do we justify that $\frac{1}{x\log|x|}$ blows up near $0$. The Fourier transform of most $L^1$ functions are not themselves $L^1$. This means that the inverse Fourier transform does not necessarily converge absolutely. However, they converge weakly, that is $$ \int f(x)\overline{\phi(x)}\,\mathrm{d}x=\int \hat{f}(x)\overline{\hat{\phi}(x)}\,\mathrm{d}x $$ for any $\phi\in\mathcal(S)$. – robjohn May 24 '12 at 0:00
Excuse me, but how did you find the Fourier Transform for this function? Do we have to use statement:$<\widehat{f},g>=<f,\widehat{g}>$? – user31497 May 24 '12 at 8:33
Yes. Plug in $\hat{g}(\tau) = \int \exp (-2\pi i \tau t) g(t) \mathrm{d}t$ into $\langle \mathcal{F}_f,\hat{g}\rangle$ from the definition and then interchange the order of integration. – Willie Wong May 24 '12 at 8:43
But why we can change the order of integration? – user31497 May 24 '12 at 15:07

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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