Fourier Transform of a constant

Discussion in 'Physics & Math' started by neelakash, Feb 24, 2007.

  1. neelakash Registered Senior Member

    Messages:
    491
    I am tryung to prove that Fourier Transform of a constant is a Dirac delta function.I have fed f(x)=1 in the formula of Forward Fourier transform and got F(k)=int{exp[-ik*pi*x]}dx
    I know that this is delta function with arguement k.But cannot prove it.It's sure to give infinity at k=0.But,for other values of k I am not getting the integral=0.So,an attempt through the way of definition fails.Can anyone help me?
     
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Tom2 Registered Senior Member

    Messages:
    726
    It's been a while since I've studied this, but one thing I do recall is that rigorously you cannot speak of the values of the delta function, because it really isn't a function at all. It's a distribution, and that distribution has various representations. One representation of the delta function is, as you have discovered,

    \(\delta(x)=\frac{1}{2\pi}\int_{-\infty}^{\infty}e^{ikx}dx\)

    The representation on the right hand side isn't required to vanish for x not equal to zero. (Indeed, the complex exponential isn't even integrable on \((-\infty,\infty)\)!)

    What we do require is that:

    \(\int_{-\epsilon}^{\epsilon}f(x)\delta(x)dx=f(0)\), for any \(\epsilon>0\).
     
    Last edited: Feb 24, 2007
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. Physics Monkey Snow Monkey and Physicist Registered Senior Member

    Messages:
    869
    Hi neelakash,

    The idea here is that for k different than zero the oscillations of the integrand should cancel in some sense. Nevertheless, you should be clear that the integral doesn't really exist as Tom2 said. To see things more clearly, insert a convergence factor \( e^{-\epsilon |x| } \) into your integral. You can now easily compute the integral and you will find a well known representation of the delta function as \( \epsilon \) goes to zero.
     
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. neelakash Registered Senior Member

    Messages:
    491
    Let f(x)=delta(x)
    Then,Fourier Transform: F(k)=int{delta(x)exp[-ik*pi*x]dx}=exp[0]=1
    Thus,Fourier transform of a delta function is a constant.
    Now,let me carry out the inverse transform:
    int{(1)exp[ik*pi*x]dk
    Can you please tell me why is this delta(x)?
     
  8. Physics Monkey Snow Monkey and Physicist Registered Senior Member

    Messages:
    869
    Did you perform the calculation I suggested?
     
  9. neelakash Registered Senior Member

    Messages:
    491
    Here is a method how to do it.It took some time,however.

    First, you find the F.T. of delta(x) and it turns out as a constant F(k)=1
    Then,you did inverse transform of 1 which yields a delta function.
    Like this: int{exp[ik'x']dk'}=2*pi*delta(x')
    Putting k'=x and x'=-k,we have
    int{exp[-ikx]dx}=2*pi*delta(-k)

    Now consider the F.T. of a constant function g(x)=c
    G(k)=int{c exp[-ikx]dx}=c[2*pi*delta(-k)]=2*pi*c delta(k)
     

Share This Page