0
$\begingroup$

I am confused trying to understand the Proof of Fourier Transform from Oppenheim book Signals and Systems. I am pasting the equations directly from the book:

$$\widetilde{x}(t)=\sum_{k=-\infty}^{+\infty}a_ke^{jk\omega_0t} \tag{4.3} $$ $$a_k=\frac{1}{T}\int_{-T/2}^{T/2}\widetilde{x}(t)e^{-jk\omega_0t}dt \tag{4.4} $$

where $\omega_0=2\pi/T$. Since $\widetilde{x}(t)=x(t)$ for $|t| < T/2$, and also, since $x(t)=0$ outside this interval, eq. $(4.4)$ can be rewritten as

$$a_k=\frac{1}{T}\int_{-T/2}^{T/2}x(t)e^{-jk\omega_0t}dt=\frac{1}{T}\int_{-\infty}^{+\infty}x(t)e^{-jk\omega_0t}dt$$

  1. My question is In the last equation on the L.H.S, there is no $\widetilde{x}(t)$ but it was mentioned in the eq. $(4.4)$
  2. And then we have: $$Ta_k= \left.\frac{2 \sin(\omega T_1)}{\omega}\right\rvert_{\omega = k\omega_0}$$ That is, with $\omega$ thought of as a continuous variable, the function ($2\sin\omega T_1/ \omega$) represents envelope of $Ta_k$. Here I want to know what is meant by envelope?I want to add more details from the book exactly " Specifically we have plot $T_0a_k$ rather than $a_k$ and we have also modified the horizontal spacing in each plot. here is the plot " enter image description here c) I want to know that how $T_0a_k$ will be drawn in continuous?
$\endgroup$
  • $\begingroup$ $\tilde x(t)=x(t)$, read all. $\endgroup$ – Yves Daoust Oct 18 '15 at 10:18
  • $\begingroup$ Yes $x'(t)=x(t) for |t|< T/2$ what that means $\endgroup$ – Aadnan Farooq A Oct 18 '15 at 10:25
  • 1
    $\begingroup$ can you do the Riemann integral? $\endgroup$ – robert bristow-johnson Oct 18 '15 at 13:07
  • $\begingroup$ @robertbristow-johnson can you please explain exactly the answer? $\endgroup$ – Aadnan Farooq A Oct 18 '15 at 13:47
2
$\begingroup$

@robertbristow-johnson can you please explain exactly the answer?

oh dear. it's a lot of work and it should be in some textbook.

okay, using your O&S notation:

$$ \tilde{x}(t) = \sum\limits_{k=-\infty}^{+\infty} a_k \ e^{j k \omega_0 t} $$

where

$$ \tilde{x}(t) = \tilde{x}(t+T) \quad \forall t $$

$$ \omega_0 \triangleq \frac{2 \pi}{T} $$

and

$$ a_k = \frac{1}{T} \int\limits_{t_0}^{t_0 + T} \tilde{x}(t) \ e^{-j k \omega_0 t} \ dt $$

and where $t_0$ can be any real value: $-\infty < t_0 < +\infty$

we choose $t_0$ to be $t_0 = -\frac{T}{2}$ so that

$$ a_k = \frac{1}{T} \int\limits_{-\frac{T}{2}}^{+\frac{T}{2}} \tilde{x}(t) \ e^{-j k \omega_0 t} \ dt $$

That's Fourier Series.

A Riemann sum is a sum of areas of rectangles that, for decently well-behaved functions, converges to the area of an integral in the limit that the rectangles get thinner and thinner:

$$ \int\limits_{t_1}^{t_2} f(t) \ dt = \lim_{N \to +\infty} \sum\limits_{n=0}^{N-1} f\left(t_1 + n\frac{t_2 - t_1}{N} \right) \ \frac{t_2 - t_1}{N} $$

for $t_1 < t_2$.

this (Riemann integration) works for most ordinary functions that are decently well-behaved. for nastier functions, another method of defining integration may have to be used, but we're not dealing with that here.

so do you understand Fourier series and Riemann integration?

now, define

$$ \tilde{x}(t) = \begin{cases} x(t), & \text{if } |t| < \frac{T}{2} \\ x(t - mT), & \text{if } |t| > \frac{T}{2} \text{ and where } m \triangleq \left\lfloor \frac{t}{T} + \frac{1}{2} \right\rfloor \end{cases} $$

$\lfloor u \rfloor$ is the $\operatorname{floor}(u)$ function that is equal to the largest integer no larger the argument $u$. (it's always rounding down.) $\left\lfloor u + \frac{1}{2} \right\rfloor$ is always rounding to nearest.

you can see that $\tilde{x}(t)$ satisfies periodicity: $ \tilde{x}(t) = \tilde{x}(t+T) \quad \forall t $. and you can also see that

$$ x(t) = \lim_{T\to +\infty} \tilde{x}(t) \quad \forall t $$

so we're taking a finite section of $x(t)$ and periodically extending that to get $\tilde{x}(t)$. but as the period gets longer and longer and goes to $\infty$ the periodic extension becomes the same as $x(t)$.

now if you plug in $a_k$ into the Fourier sum:

$$ \tilde{x}(t) = \sum\limits_{k=-\infty}^{+\infty} \frac{1}{T} \int\limits_{-\frac{T}{2}}^{+\frac{T}{2}} \tilde{x}(u) \ e^{-j k \omega_0 u} \ du \ e^{j k \omega_0 t} $$

and let $T \to +\infty$ ...

okay, i got back to this. (sorry for the delay.)

so we've been saying

$$ \begin{align} x(t) & = \lim_{T \to \infty} \tilde{x}(t) \\ & = \lim_{T \to \infty} \sum\limits_{k=-\infty}^{+\infty} \frac{1}{T} \int\limits_{-\frac{T}{2}}^{+\frac{T}{2}} \tilde{x}(u) \ e^{-j k \omega_0 u} \ du \ e^{j k \omega_0 t} \\ & = \lim_{T \to \infty} \sum\limits_{k=-\infty}^{+\infty} \frac{1}{T} \int\limits_{-\frac{T}{2}}^{+\frac{T}{2}} x(u) \ e^{-j k \omega_0 u} \ du \ e^{j k \omega_0 t} \\ \end{align} $$

(note the subtle difference between the last two lines. why can we do that?)

and we're saying

$$ \begin{align} \int\limits_{-\frac{T}{2}}^{+\frac{T}{2}} x(u) \ e^{-j k \omega_0 u} \ du & = \lim_{N \to +\infty} \sum\limits_{n=0}^{N-1} x\left(-\frac{T}{2} + n\frac{+\frac{T}{2} - (-\frac{T}{2})}{N} \right) \ e^{-j k \omega_0 \left(-\frac{T}{2} + n\frac{+\frac{T}{2} - (-\frac{T}{2})}{N} \right)} \ \frac{+\frac{T}{2} - (-\frac{T}{2})}{N} \\ & = \lim_{N \to +\infty} \sum\limits_{n=0}^{N-1} x\left(-\frac{T}{2} + n\frac{T}{N} \right) \ e^{-j k \omega_0 \left(-\frac{T}{2} + n\frac{T}{N} \right)} \ \frac{T}{N} \\ \end{align} $$

putting the two limits together:

$$ \begin{align} x(t) & = \lim_{T \to \infty} \sum\limits_{k=-\infty}^{+\infty} \frac{1}{T} \int\limits_{-\frac{T}{2}}^{+\frac{T}{2}} x(u) \ e^{-j k \omega_0 u} \ du \ e^{j k \omega_0 t} \\ & = \lim_{T \to \infty} \sum\limits_{k=-\infty}^{+\infty} \frac{1}{T} \lim_{N \to +\infty} \sum\limits_{n=0}^{N-1} x\left(-\frac{T}{2} + n\frac{T}{N} \right) \ e^{-j k \omega_0 \left(-\frac{T}{2} + n\frac{T}{N} \right)} \ \frac{T}{N} \ e^{j k \omega_0 t} \\ & = \lim_{T \to \infty} \lim_{N \to +\infty} \sum\limits_{k=-\infty}^{+\infty} \sum\limits_{n=0}^{N-1} x\left(-\frac{T}{2} + n\frac{T}{N} \right) \ e^{-j k \omega_0 \left(-\frac{T}{2} + n\frac{T}{N} \right)} \ \frac{T}{N} \ e^{j k \omega_0 t} \frac{1}{T} \\ & = \lim_{N \to +\infty} \sum\limits_{n=0}^{N-1} \lim_{T \to \infty} \sum\limits_{k=-\infty}^{+\infty} x\left(-\frac{T}{2} + n\frac{T}{N} \right) \ e^{-j k \omega_0 \left(-\frac{T}{2} + n\frac{T}{N} \right)} \ \frac{1}{N} \ e^{j k \omega_0 t} \\ \end{align} $$

(answer not done yet, still being developed.)

$\endgroup$
  • $\begingroup$ i have to quit now. i'll return to this later. but you should be able to see where i am going with it. $\endgroup$ – robert bristow-johnson Oct 18 '15 at 19:01
  • $\begingroup$ i'd also be happy if someone else finishes this. bye for now. $\endgroup$ – robert bristow-johnson Oct 18 '15 at 19:04
  • $\begingroup$ I understand the part of Fourier series, but I am confused that why we are doing Riemann Sum? is it related to the second part of my question, Evelope? $\endgroup$ – Aadnan Farooq A Oct 19 '15 at 6:34
  • 1
    $\begingroup$ @robertbristow-johnson, grab a + for this effort of re-writing book chapter ;) $\endgroup$ – jojek Oct 19 '15 at 9:28
  • 1
    $\begingroup$ as you can see i still have a problem with $\frac{T}{N}$ which is indeterminate. so i still have to fix it. where i am trying to get to is: $$ x(t) = \frac{1}{2 \pi} \int\limits_{-\infty}^{+\infty} \int\limits_{-\infty}^{+\infty} x(u) e^{-j \omega u} du \ e^{j \omega t} \ d\omega $$ and identifying the inner integral as $X(j \omega)$ . i'll get back to this tomorrow, i think. (anyone else is free to fix and finish it.) $\endgroup$ – robert bristow-johnson Oct 20 '15 at 23:49

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.