1

Calculate fourier transform for function $h$:

$$\hat{h}(\nu)=\int_{-\infty}^\infty e^{-i2\pi\nu t} h(t) dt $$

When $h(t)=1$ and $|t|\le\frac 12$.

And when $h(t)=0$ and $|t|\gt\frac 12$

Also does it hold, that $\int_{-\infty}^\infty |\hat{h}(\nu)|d\nu \lt \infty$? What about $\int_{-\infty}^\infty |\hat{h}(\nu)|^2d\nu \lt \infty$ or $\lim_{|\nu| \to \infty}\hat{h}(\nu)=0$?

ELEC
  • 1,113

2 Answers2

1

The Sinc Function is the fourier transform of a discrete pulse: $$ \begin{align} \hat{h}(\nu) &=\int_{-\infty}^\infty h(t)e^{-2\pi i\nu t}\,\mathrm{d}t\\ &=\int_{-1/2}^{1/2}e^{-2\pi i\nu t}\,\mathrm{d}t\\ &=\frac{e^{\pi i\nu}-e^{-\pi i\nu}}{2\pi i\nu}\\ &=\frac{\sin(\pi\nu)}{\pi\nu}\\[7pt] &=\mathrm{sinc}(\pi\nu) \end{align} $$


Note that $$ \begin{align} \int_k^{k+1}\left|\frac{\sin(\pi\nu)}{\pi\nu}\right|\,\mathrm{d}\nu &\ge\int_k^{k+1}\left|\frac{\sin(\pi\nu)}{\pi(k+1)}\right|\,\mathrm{d}\nu\\ &=\frac2{\pi^2(k+1)} \end{align} $$ Since the Harmonic Series diverges, the integral $$ \int_{-\infty}^\infty|\hat{h}(\nu)|\,\mathrm{d}\nu $$ also diverges.


The Plancherel Theorem says that the $L^2$ norm of a function and its Fourier Transform are equal. Since $$ \int_{-1/2}^{1/2}1^2\,\mathrm{d}t=1 $$ we get $$ \int_{-\infty}^\infty|\hat{h}(\nu)|^2\,\mathrm{d}\nu=1 $$


Since $$ \int_{-1/2}^{1/2}|1|\,\mathrm{d}t=1 $$ The Riemann-Lebesgue Lemma says that $$ \lim_{\nu\to\infty}\hat{h}(\nu)=0 $$ Of course, since we've computed it, we can just see that $$ \lim_{\nu\to\infty}\frac{\sin(\pi\nu)}{\pi\nu}=0 $$

robjohn
  • 345,667
  • The OP's struggling with pretty basic things (see comments) and you already gave it all chewed up... – DonAntonio Jan 08 '14 at 13:11
  • @DonAntonio: I was hoping that the material could be understood, digested, and then they could move on. I don't think that these were homework problems, but I could be wrong. Other than pointing out the Plancherel Theorem and the Riemann-Lebesgue Lemma, the only work I did that was not already done, was to show that $\hat{h}\not\in L^1$. – robjohn Jan 08 '14 at 13:13
  • it in fact looks like a homework problem to me, but probably a basic physics one, not mathematics. – DonAntonio Jan 08 '14 at 13:15
  • @DonAntonio: fair enough. However, ELEC does not have a Physics.SE account (which does not invalidate your point). – robjohn Jan 08 '14 at 13:17
0

Isn't it pretty straightforward, or am I missing something?:

$$\hat h(v)=\int\limits_{-1/2}^{1/2}e^{-2\pi ivt}dt=\left.-\frac1{2\pi iv}e^{-2\pi ivt}\right|_{-1/2}^{1/2}=-\frac1{2\pi iv}\left(e^{-\pi iv}-e^{\pi iv}\right)=\frac1{\pi v}\sin\pi v$$

DonAntonio
  • 211,718
  • 17
  • 136
  • 287
  • Why did the integral gap change from $-\infty \to \infty$ to $\frac{-1}{2} \to \frac 12$ ? I guess I don't understand the properties of the question.. – ELEC Jan 08 '14 at 12:51
  • Because $$h(t)=0;;\forall,|t|>\frac12\implies e^{-2\pi ivt}h(t)=0;;\forall,|t|>\frac12;;\ldots$$ – DonAntonio Jan 08 '14 at 12:52
  • That backwards A means "when" right? how does $h(t)=0$ when $|t|\gt\frac 12$? – ELEC Jan 08 '14 at 13:05
  • No, it means "for all'", and $;h(t)=0;$ for $;|t|>1/2;$ is your definition ! – DonAntonio Jan 08 '14 at 13:10
  • AHHHH!! Thanks!! Got it ! :) – ELEC Jan 08 '14 at 13:11
  • Hehe...a pleasure, @ELEC . Sometimes pretty basic things slip away in front of our own eyes. – DonAntonio Jan 08 '14 at 13:13
  • But still one more question @DonAntonio :) we know $h(t)=1$ when $|t|\le\frac 12$ but why we use $-\frac 12$ as the lower gap and not $-\infty$ ? – ELEC Jan 08 '14 at 13:29
  • Again, @Elec: because the function is zero in $;(-\infty , -1/2);$ so its integral is zero there. Something similar happens with $;(1/2,\infty);$ . We're left only with the integral between $;-1/2;$ and $;1/2;$ because only there the function is non-zero. – DonAntonio Jan 08 '14 at 13:31
  • But if it says the function is 1 when $t$ is less or equal to $1/2$ shouldn't it mean it's 1 for all $t$ values less than $1/2$. shouldn't it say $h(t)=1$ when $-1/2 \le t \le 1/2$ – ELEC Jan 08 '14 at 13:43
  • ELEC, what are you studying? The meaning of $;|t|\le 1/2;$ is exactly $;-1/2\le t \le 1/2;$ , when $;t\in\Bbb R;$ ... – DonAntonio Jan 08 '14 at 13:45
  • Electrical engineering. Thanks! – ELEC Jan 08 '14 at 14:03