I am studying the properties of the Fourier transform. Fourier transform: $$\hat{f}(\xi)=\frac{1}{\sqrt{2\pi}}\int_\limits{-\infty}^{\infty}f(t)e^{it\xi}dt\:\:,\:\: \xi\in\mathbb{R}$$

Notes: If $f\in L_1(\mathbb{R})\\||f||=\int_\limits{-\infty}^{\infty}|f(t)|dt<\infty$

$\hat{f}(\xi)\in C_B(\mathbb{R})\\||g||_{C_B(\mathbb{R})}=\sup_{t\in\mathbb{R}}|g(t)|$

Then $\sup_{\xi\in\mathbb{R}}|\hat{f}(\xi)|\leqslant\frac{1}{\sqrt{2\pi}}||f||_1$

Observation: I tried to understand this notes on Fourier transform in the following way:

I want to prove $\sup_{\xi\in\mathbb{R}}|\hat{f}(\xi)|\leqslant\frac{1}{\sqrt{2\pi}}||f||_1$ and I started using: $\hat{f}(\xi)=\frac{1}{\sqrt{2\pi}}\int_\limits{-\infty}^{\infty}f(t)e^{it\xi}dt\leqslant\frac{1}{\sqrt{2\pi}}\int_\limits{-\infty}^{\infty}f(t)\frac{1}{\sqrt{2\pi}}\int_\limits{-\infty}^{\infty}e^{it\xi}dt=\frac{1}{\sqrt{2\pi}}||f||_1\int_\limits{-\infty}^{\infty}e^{it\xi}dt$

However I guess after computation that $\int_\limits{-\infty}^{\infty}e^{it\xi}dt=0$.


1) What do you think of what I have done so far? How can I prove the property?

2) Fourier transform here presented is $\hat{f}(\xi)=\frac{1}{\sqrt{2\pi}}\int_\limits{-\infty}^{\infty}f(t)e^{it\xi}dt$. However in books I find $2\pi$ instead of $\sqrt{2\pi}$. Does this formula intends to compute the coefficients? How should I derive it with functional analysis tools?

Thanks in advance!

  • $\begingroup$ Any comments will be welcomed $\endgroup$ – Pedro Gomes Apr 18 '18 at 18:02

That is not valid, for $e^{it\xi}\notin L^{1}$.

Rather, the trick is to use absolute value into the integrals:

\begin{align*} |\widehat{f}(\xi)|&=\left|\dfrac{1}{\sqrt{2\pi}}\int_{-\infty}^{\infty}f(t)e^{it\xi}dt\right|\\ &\leq\dfrac{1}{\sqrt{2\pi}}\int_{-\infty}^{\infty}|f(t)e^{it\xi}|dt\\ &=\dfrac{1}{\sqrt{2\pi}}\int_{-\infty}^{\infty}|f(t)|dt\\ &=\dfrac{1}{\sqrt{2\pi}}\|f\|_{L^{1}(-\infty,\infty)}. \end{align*}

  • $\begingroup$ The constant depends on the authors. $\endgroup$ – user284331 Apr 18 '18 at 18:06
  • $\begingroup$ Thanks! What do you mean by the constant? Any thoughts on my second question? $\endgroup$ – Pedro Gomes Apr 18 '18 at 18:07
  • $\begingroup$ The $1/\sqrt{2\pi}$. Yes, that formula is to derive the coefficients of Fourier series in a compact way. $\endgroup$ – user284331 Apr 18 '18 at 18:09
  • $\begingroup$ When I tried to derive the Fourier coefficient I used a projection over $e^{it\xi}$ for the space of periodic functions in the interval $[-\pi,\pi]$. Unfortunately I am not understanding on how the coefficients differ from the transform. Is the transform the coefficient? Thanks! $\endgroup$ – Pedro Gomes Apr 18 '18 at 18:12
  • $\begingroup$ Yes, you are right, I should say it precisely, for the Fourier series coefficients, the interval is $[-\pi,\pi]$, not $(-\infty,\infty)$, and this version is to use for the theory of singular integrals. $\endgroup$ – user284331 Apr 18 '18 at 18:14

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.