The Fourier transform is defined in Press[1] to be
(1.1)
The use of rather than is my notation as is the use of rather than . The use of for the transform from the time domain to the frequency domain rather than +- k is due to Press. Most documents on the Fourier transform start with a brief word or two in which they blame Press for all of their failings. For my part I am very grateful that he has accepted this blame and established a standard to work from.
A first principles derivation of the Fourier transform along with a reference is detailed in CFOUR.doc .htm. The more fundamental Discrete Fourier Sum is described in Discrete Fourier Transform.doc.
Define
(1.2)
So that
(1.3)
Substituting the second defining equation into the first yields
(2.1)
Which establishes the relationship
(2.2)
For a finite frequency region, the inverse of this becomes
(2.3)
The integral of the delta function about a short time interval
(2.4)
Let
(2.5)
The limit as is [2] (2.6)
So that for finite epsilon and infinite F, the delta function relation is satisfied.
(2.7)
This is in the form of a Gabor transform although it is also a standard method of making otherwise difficult integrals tractable. Evaluate the integral by completing the square
(2.8)
Let
(2.9)
Note that the value of F in the first definition is on the order of , so that for not so large values of F the two forms can be written as
(2.10)
It is of interest to plot 1/F times these two forms for values of Ff between -10 and 10
