1. Not finding help here? Sign up for a free 30min tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Integral computation over Fourier/Convolution

  1. Jan 27, 2013 #1
    Hi,

    I just tried to solve the following integral:

    [tex]
    \int_{-\infty}^{\infty} \frac{2}{1+(2\pi t)^2} \mathrm{sinc}(2t) dt
    [/tex]

    My approach is: Convert both to Fourier domain and the multiplication becomes a convolution. Because of Parsevals theorem, I can either integrate in time or frequency domain. Because of linearity, I can then put the outer integral inside:

    [tex]
    \int_{f=-\infty}^{\infty} \int_{\nu=-\infty}^{\infty} e^{-|\nu|} \frac{1}{2} \Pi\left(\frac{f-\nu}{2}\right) d\nu df \\
    = \int_{\nu=-\infty}^{\infty} e^{-|\nu|} \frac{1}{2} \underbrace{\int_{f=-\infty}^{\infty}\Pi\left(\frac{f-\nu}{2}\right) df}_{2} d\nu
    [/tex]

    It can be clearly seen, that only the double-sided exponential stays whose integral is 2.
    So my total solution is 2.

    However, Wolfram alpha gives me (e-1)/e:

    http://www.wolframalpha.com/input/?i=integral+2%2F%281%2B%282*pi*t%29^2%29+*+sin%28pi*2*t%29%2F%28pi*2*t%29%2Ct%2C-infinity%2Cinfinity

    Can anyone explain where I did wrong?

    Thanks,
    div

    PS: The sinc is defined as normalized, i.e. [itex]\mathrm{sinc}(t)=\sin(\pi t)/\pi t[/itex] ...
     
  2. jcsd
  3. Jan 28, 2013 #2

    jbunniii

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    This is your fundamental problem. Parseval's theorem would apply if you were integrating the magnitude squared of your function, but you are not doing that.

    In general, if your original function is [itex]x(t)[/itex], then
    $$\int_{-\infty}^{\infty} x(t) dt = \left. \int_{-\infty}^{\infty} x(t) e^{-ift} dt \right|_{f = 0} = \hat{x}(0)$$
    where I have assumed the following definition for the Fourier transform:
    $$\hat{x}(f) = \int_{-\infty}^{\infty} x(t) e^{-ift} dt$$
    Thus, integration in the time domain does not correspond to integration in the frequency domain, but rather to evaluation of the Fourier transform at [itex]f = 0[/itex].
     
  4. Jan 28, 2013 #3
    Oh, of course, thanks!!

    Do you have a pointer on how to calculate this integral with Fourier transform theorems?

    Thanks!
    divB
     
  5. Jan 28, 2013 #4

    vela

    User Avatar
    Staff Emeritus
    Science Advisor
    Homework Helper
    Education Advisor

    As jbunniii implied, you want to evaluate
    $$ F\left[\frac{2}{1+(2\pi t)^2} \text{sinc }2t\right]_{f = 0}$$ where F[ ] denotes the Fourier transform. What's the Fourier transform of that product?
     
  6. Jan 28, 2013 #5
    Oh, right, didn't realize that!
    Thanks a lot, got it!
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Integral computation over Fourier/Convolution
  1. Convolution integral (Replies: 1)

  2. Convolution Integral (Replies: 8)

Loading...