Using Fourier Properties or standard Integral?

AI Thread Summary
The discussion revolves around the discrepancies observed when calculating the Fourier Transform of a signal using either Fourier Transform properties or the FT integral. The signal in question is a combination of shifted rectangular functions, leading to different expressions for G(f) depending on the method used. After some algebraic manipulation and applying trigonometric identities, one participant claims to have shown that the two results are indeed equivalent. However, there are concerns about potential errors in calculations, particularly regarding the correct definition of the sinc function and the use of radians. Ultimately, the participants reach a consensus that both methods yield identical results when properly evaluated.
frenzal_dude
Messages
76
Reaction score
0
Hi, sometimes when I'm trying to work out the Fourier Transform of a signal, I get different answers depending on whether I use Fourier Transform properites (such as rect(t) will go to sinc(f) etc) or whether I use the FT integral. Here's an example where I'm not sure which 1 is the correct answer:

The signal is: g(t)=-Arect(\frac{t+T}{T})+Arect(\frac{t}{T})-Arect(\frac{t-T}{T})

Using FT properties:
<br /> G(f)=-ATsinc(Tf)e^{j2\pi fT}+ATsinc(fT)-ATsinc(Tf)e^{-j2\pi fT}

Therefore:

G(f)=-2ATsinc(Tf)cos(2\pi fT)+ATsinc(Tf)

Using the FT Integral:
G(f)=-A\int_{\frac{-3T}{2}}^{\frac{-T}{2}}e^{-j2\pi ft}dt+A\int_{\frac{-T}{2}}^{\frac{T}{2}}e^{-j2\pi ft}dt-A\int_{\frac{T}{2}}^{\frac{3T}{2}}e^{-j2\pi ft}dt

Therefore:

G(f)=-A[\frac{e^{j\pi fT}-e^{j\pi f3T}}{-j2\pi f}]+\frac{A}{\pi f}[\frac{e^{j\pi fT}-e^{-j\pi fT}}{2j}]-A[\frac{e^{-j\pi f3T}-e^{-j\pi fT}}{-j2\pi f}]

Therefore:

G(f)=2TAsinc(fT) -3TAsinc(3fT)
 
Last edited:
Engineering news on Phys.org
I think that both answers are equivalent. After a bit of fiddling around with the usual trig identities I was able to show that :

\sin(x) - \sin(3x) = -2 \cos(2x) \sin(x)

If you use that in your final equation then it reduces to the earlier one that you got "Using FT properties"
 
Hey uart thanks for the reply.
I've tried subbing in different values for A, T and f but I get different answers from the 2 equations, so they're not the same unfortunately.
 
frenzal_dude said:
Hey uart thanks for the reply.
I've tried subbing in different values for A, T and f but I get different answers from the 2 equations, so they're not the same unfortunately.

Yeah I'm pretty sure I can show (algebraically) that they're identical. No time right now but I'll have a look at numerical values later.

BTW Are you sure you're using the correct "sinc" definition : sinc(x) = sin(pi x)/(pi x) is the correct one to use in this case.
 
OK I'm calling "calculator error" on you frenzal_dude. :) Double check that you're using radians and also double check you have the correct sinc definition.

I just checked and I get them both algebraically and numerically identical!
 
Last edited:
Here's the algebraic solution.

Starting with your last resut (I took out the common factor of "fT" as it's obviously the same in both solutions) :

G(f)=2sinc(fT) -3sinc(3fT) = sinc(fT) + sinc(fT) - 3 sinc(3fT)

= sinc(fT) + \frac{1}{\pi f T} \left( \sin(\pi f T) - \sin(3 \pi f T) \right)

Now sub in the trig identity I posted previosly and you get :

G(f)=sinc(fT) -2 sinc(fT) \cos(2 \pi f T)

Which is identical to your "using FT properties" result.
 
uart, thankyou so much for your help!
I'm relieved to know I'm on the right track now.
 
Back
Top