-1

I saw this theorem in Wikipedia, and didn't see it in other place. I tried to proof it, but failed.

How to prove that for $2\pi-$periodic functions one has The convolution theorem for fourier series.: $$ \widehat{f*g}(x) =2π\hat{g}(x)\cdot\hat{f}(x) $$ Note: this is convolution theorem for fourier series, not transforms

see here

Guy Fsone
  • 23,903
Mr.O
  • 584
  • Is there a question you wanted us to answer? – John Hughes Jan 16 '18 at 18:48
  • A proof of this is in nearly every book / lecture notes on fourier analysis – TheWaveLad Jan 16 '18 at 18:49
  • note, this theorem is for fourier series, not fourier transform. I tried to find a proof on the web, and I didn't find even the theorem itself on other sources. – Mr.O Jan 16 '18 at 19:22
  • @JohnHughes It will be nice to see a proof to this theorem. – Mr.O Jan 16 '18 at 19:23
  • Of what theorem? You've shown us a proof of a theorem that isn't the one you care about (apparently). If you could carefully state the theorem whose proof you'd like to see, perhaps we could help you. As @TheWaveLad says, it's almost certainly in any book you pick up, but we could actually tell you the theorem NUMBER in a book or two ... – John Hughes Jan 16 '18 at 19:44
  • @JohnHughes, I don't know what you are talking about. have you seen the photo I uploaded? it's a copy from wikipedia. it states a theorem for fourier series of a convolution. it's not a proof. – Mr.O Jan 16 '18 at 19:52
  • Hint: $$4 \pi^2 \hat f(n) \hat g(n) = \int_{-\pi}^{\pi} f(x) e^{-inx} \mathrm dx \int_{-\pi}^{\pi} g(y)e^{-iny} \mathrm dy = \int_{-\pi}^{\pi} e^{-int} \left( \int_{-\pi}^{\pi} f(x)g(t-x) \mathrm dx \right)\mathrm dt $$ – TheWaveLad Jan 16 '18 at 20:16
  • @ClementC. this is about Fourier series not Fourrier transform. have a close look to see that it is not that obvious and similar. so this is clearly not a duplicate – Guy Fsone Jan 16 '18 at 23:15
  • @GuyFsone My bad, I got the suggestion from the review queue and failed to look closely enough. Thank you. – Clement C. Jan 16 '18 at 23:16
  • @ClementC. me too I got it from the review queues buit.... – Guy Fsone Jan 16 '18 at 23:17
  • @ClementC. you may have look the answer below – Guy Fsone Jan 16 '18 at 23:18

1 Answers1

3

The proof use the same token as in the case of Fourier transform. Indeed, $f$ and $g$ be $2\pi$-periodic functions. But since the function $y\mapsto e^{-iyx}g(y)$ is $2π-$periodic, then using this we have , $$\int_{0}^{2\pi} e^{-ix(t-u)}g(t-u)dt\overset{y=t-u}{=} \color{blue}{\int_{-u}^{2\pi-u} e^{-ixy}g(y)dy =\int_{0}^{2\pi} e^{-ixy}g(y)dy}= 2π\hat{g}(x) $$

Inserting this in the following relation we get

$$ \color{blue}{\widehat{f*g}(x) =} \frac{1}{2\pi}\int_{0}^{2\pi}e^{-ixt}f*g(t)dt = \frac{1}{2\pi}\int_{0}^{2\pi}e^{-ixt}\int_{0}^{2\pi} f(u)g(t-u)dudt\\= \frac{1}{2\pi}\int_{0}^{2\pi}\left(e^{-iux}f(u)\int_{0}^{2\pi} e^{-ix(t-u)}g(t-u)dt\right)du\\= 2π\hat{g}(x) \frac{1}{2\pi}\int_{0}^{2\pi}e^{-iux}f(u)du= \color{red}{2π\widehat{g}(x)\cdot\widehat{f}(x)} $$

we used this An integrable and periodic function $f(x)$ satisfies $\int_{0}^{T}f(x)dx=\int_{a}^{a+T}f(x)dx$.

Guy Fsone
  • 23,903