# Fourier of boxcar vs rectangular

1. Jun 12, 2014

### quantumlight

So say I do a fourier transform of a rectangular function with magnitude 1 from (0, NT). The fourier transform of this would be:

$f(jΩ) = \frac{1-e^{-jΩNT}}{jΩ} = NT\cdot{e^{-jΩNT/2}}\cdot{sinc(ΩNT/2)}$

Now say if I sample this rectangle at time T producing N samples, the DTFT of this is:

$f(e^{jw}) = \frac{1-e^{-jwN}}{1-e^{-jw}} = e^{-jw(N-1)/2}\cdot\frac{sin(wN/2)}{sin(w/2)}$

Since DTFT and Fourier Transform is related by Ω = wT where

$f(e^{jw}) = \frac{1}{T}\sum{f(jΩ + j2\pin)}$

Now if I try this method I get to this point:

$f(e^{jw}) = e^{-jw(N-1)/2}\cdot{sin(wN/2)}\cdot{\sum\frac{1}{w+2{\pi}n}}$

This is where I get stuck, because that last summation needs to somehow equal sin(w/2) or 1-e-jw. The summation is from negative infinity to infinity.

Was wondering if there is some math trick that gives the result of that summation. Thanks.

Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Can you help with the solution or looking for help too?
Draft saved Draft deleted