I feel problem in understanding the proof of Fourier series properties

  1. Time scaling
    \begin{align} b_k &= \frac{1}{T}\int_{T}x(t)e^{jk\omega_0t}dt\\ & = \frac{a}{T}\int_{T/a}x(at)e^{jk(\omega_0a)t}dt\tag{$\scriptstyle{a - \text{scaling factor}}$}\\ & = \frac{a}{T}\int_{T}x(L)e^{jk\omega_0L}\frac{dL}{a}\tag{$\scriptstyle{L=at}$}\\ &= \frac{1}{T}\int_{T}x(L)e^{jk\omega_0L}dL = a_k\\ \end{align} For this I want to ask that how $T/a$ term in the integral changes to $T$, and $dt$ changes to $dL/a$?

  2. Multiplication \begin{align} x(t)y(t)&= \sum^{+\infty}_{k\ =\ -\infty} a_k e^{jk\omega_0t}\sum^{+\infty}_{k\ =\ -\infty} b_k e^{jl\omega_0t}\\ &=\sum^{+\infty}_{k\ =\ -\infty} \sum^{+\infty}_{l\ =\ -\infty} a_k b_le^{j(k+l)\omega_0t}\\ &=\sum^{+\infty}_{m\ =\ -\infty} \left[\sum^{+\infty}_{l\ =\ -\infty} a_{m-l} b_l\right]e^{jm\omega_0t}\tag{$\scriptstyle{m=k+l}$} \end{align} I couldn't understand this last step, why the summation changes from $k$ to only $m$? Shouldn't it be $m-l$ and also is it done here to prove the multiplication property? or there are more steps?

  • $\begingroup$ please write in comment the reason before giving down $\endgroup$ – Aadnan Farooq A Oct 12 '15 at 8:07
  • 1
    $\begingroup$ There is a mistake in your last expression in time scaling. The integrating variable is dL and not dt. In your multiplication steps all the steps are correct. He is taking the substitution m=k+l. so now k becomes (m-l).But the limits of integration of k are expressed in terms of m. m=K+l. now if k=-infinity, m=-infinity and if k=+infinity,m=+infinity. All this is being done to express fourier coefficients of product of x(t)y(t) as discrete convolution of individual fourier series coefficients ak and bk. $\endgroup$ – Karan Talasila Oct 12 '15 at 9:45
  • 2
    $\begingroup$ see when you are making a substitution of L=at in integral, here you are doing definite integrals , so you have to change the limits of integration also. So if L=at ,then you have to change everything with respect to L. so now the integration limits are t=0 and t=T/a before making substitution L=at. After L=at substitution, you have to change limits with respect to L. So when t=0 L=a.0=0 and when t=T/a, L=a.(T/a)=T.That's how limits are changing to T. with regard to dt/a, it is because L=at.differentiate b.s. you will get dL=dt.a. so dt=dL/a. change that dt/a to dL/a. That's a mistake. $\endgroup$ – Karan Talasila Oct 12 '15 at 10:04
  • $\begingroup$ yes i have change it to dL, can you please help me to understand how $T/a$ term changes to$T$ in the integral changes and $dt$ to $dL/a$ $\endgroup$ – Aadnan Farooq A Oct 12 '15 at 10:04
  • 2
    $\begingroup$ @AadnanFarooqA There shouldn't be those $dt$'s in your second question. Can you double-check the $x(t)y(t)$ equation ? $\endgroup$ – Gilles Oct 12 '15 at 14:18
  • Looking at the change of terms in the equality: $$\frac{a}{T}\int_{T/a}x(at)e^{jk(\omega_0a)t}dt = \frac{a}{T}\int_{T}x(L)e^{jk\omega_0L}\frac{dL}{a}$$ $$at = L\implies \begin{cases}\text{if }t=T/a\implies L=T\\adt=dL\implies dt=\frac{dL}{a}\end{cases}$$

  • Looking at the equality again: $$ \sum^{+\infty}_{k\ =\ -\infty} \sum^{+\infty}_{l\ =\ -\infty} a_k b_le^{j(k+l)\omega_0t}=\sum^{+\infty}_{m\ =\ -\infty} \left[\sum^{+\infty}_{l\ =\ -\infty} a_{m-l} b_l\right]e^{jm\omega_0t} $$ $$m=k+l\implies \begin{cases}k=m-l\quad({\scriptstyle{\rm indeed}})\\\text{and } m-l=\pm \infty \implies m=\pm\infty+l=\pm\infty\end{cases}$$


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.