Fourier Transform Properties

From Class Wiki
Revision as of 17:31, 8 November 2009 by Nicholas.Christman (talk | contribs)
Jump to navigation Jump to search

Some properties to choose from if you are having difficulty....

Max Woesner

1. Find [cos(w0t)g(t)]

Recall w0=2πf0, so [cos(w0t)g(t)]=[cos(2πf0t)g(t)]=cos(2πf0t)g(t)ej2πftdt

Also recall cos(θ)=12(ejθ+ejθ),so cos(2πf0t)g(t)ej2πftdt=12[ej2πf0t+ej2πf0t]g(t)ej2πftdt

Now 12[ej2πf0t+ej2πf0t]g(t)ej2πftdt=12ej2π(ff0)tg(t)dt+12ej2π(f+f0)tg(t)dt=12G(ff0)+12G(f+f0)

So [cos(w0t)g(t)]=12[G(ff0)+G(f+f0)]


reviewed by Joshua Sarris


2. Find [g(t)h*(t)dt]

Recall g(t)=1[G(f)]=G(f)ej2πftdf
Similarly, h(t)=1[H(f)]=H(f)ej2πftdf
So [g(t)h*(t)dt]=G(f')ej2πf'tdf'(H(f')ej2πf'tdf')*dt
Now G(f')ej2πf'tdf'(H(f')ej2πf'tdf')*dt=G(f')H*(f')ej2π(f'f')tdtdf'df'

Note that ej2π(f'f')tdt=δ(f'f')

Added step per Nick's suggestion

Substituting gives us G(f')H*(f')ej2π(f'f')tdtdf'df'=G(f')H*(f')δ(f'f')df'df'

And G(f')H*(f')δ(f'f')df'df'=G(f')H*(f')df'

Since f' is a simply a dummy variable, we can conclude that:

[g(t)h*(t)dt]=G(f)H*(f)df

"I was going to make a comment on the delta identity, but after looking at it closer I think it is fine. One comment I have is that you might consider adding one more step, showing the delta function in the integral and pulling the integrands together to make it look like a double integral -- it isn't necessary and I understood the transition, but it helps the proof/identity look a little more complete. Good job!"

Example: a1a2b1b2X(s)Y(s)δ(ss)dsds=a1a2X(s)Y(s)ds

Reviewed by Nick Christman


Nick Christman

Note: After scratching my head for a couple of hours, I decided that I would try a different Fourier Property. In fact, I chose a property that would need to be defined in order to show my second property.

1. Find [g(t)ej2πf0t]

This is a fairly straightforward property and is known as complex modulation

[g(t)ej2πf0t]=[g(t)ej2πf0t]ej2πftdt

Combining terms, we get:

[g(t)ej2πf0t]ej2πftdt=g(t)ej2π(ff0)tdt

Now let's make the following substitution θ=ff0

This now gives us a surprisingly familiar function:

g(t)ej2π(ff0)tdt=g(t)ej2πθtdt

This looks just like G(θ)!

We can now conclude that:

[g(t)ej2πf0t]=G(θ)=G(ff0)

PLEASE ENTER PEER REVIEW HERE



2. Find [g(tt0)ej2πf0t]

-- Using the above definition of complex modulation and the definition from class of a time delay (a.k.a "the slacker function"), I will attempt to show a hybrid of the two...

By definition we know that:

[g(tt0)ej2πf0t]=[g(tt0)ej2πf0t]ej2πftdt

Rearranging terms we get:

[g(tt0)ej2πf0t]ej2πftdt=g(tt0)ej2π(ff0)tdt

Now lets make the substitution λ=tt0t=λ+t0.
This leads us to:

g(tt0)ej2π(ff0)tdt=g(λ)ej2π(ff0)(λ+t0)dλ

After some simplification and rearranging terms, we get:

g(λ)ej2π(ff0)(λ+t0)dλ=g(λ)ej2π(ff0)λej2π(ff0)t0dλ

Rearranging the terms yet again, we get:

g(λ)ej2π(ff0)λej2π(ff0)t0dλ=ej2π(ff0)t0[g(λ)ej2π(ff0)λdλ]

We know that the exponential in terms of t0 is simply a constant and because of the Fourier Property of complex modualtion, we finally get:

[g(t)ej2πf0t]=G(ff0)ej2π(ff0)t0


Reviewed by Kevin Starkey --> add dλ above... other than that it looks good.




Joshua Sarris

Find [sin(w0t)g(t)]


Recall w0=2πf0,

so expanding we have,

[sin(w0t)g(t)]=[sin(2πf0t)g(t)]=sin(2πf0t)g(t)ej2πftdt

Also recall sin(θ)=1j2(ejθejθ),

so we can convert to exponentials.

sin(2πf0t)g(t)ej2πftdt=1j2[ej2πf0tej2πf0t]g(t)ej2πftdt

Now integrating gives us,

1j2[ej2πf0tej2πf0t]g(t)ej2πftdt=1j2ej2π(ff0)tg(t)dt1j2ej2π(f+f0)tg(t)dt=1j2G(ff0)1j2G(f+f0)


So we now have the identity,

[sin(w0t)g(t)]=1j2[G(ff0)G(f+f0)]

or rather

[sin(w0t)g(t)]=12j[G(f+f0)+G(ff0)]

Reviewed by Max Woesner

Also reviewed by Nick Christman -- Looks good.


Kevin Starkey

1. Find [s(t)dt]
First we know that [s(t)dt]=(s(t)dt)ej2πftdt
We also know that [s(t)]=S(f) and ej2πftdt=δ(f)
(+) Which gives us (s(t)dt)ej2πftdt=S(f)δ(f)df
Since δ(f)df is only non-zero at f = 0 this yeilds
S(f)δ(f)df=S(0)
So [s(t)dt]=S(0)

Reviewed by Nick Christman -- I fixed one typo (needed a minus sign in the exponential). I'm not sure about the step (+). I would like to believe it, but I'm just not sure that it works... if you are sure it works, maybe add a little comment to explain it a little better. Other than that, it looks good!


2. Find [ej2πf0ts(t)]
First [ej2πf0ts(t)]=ej2πf0ts(t)ej2πft
or rearranging we get ej2πf0ts(t)ej2πftdt=s(t)ej2πt(ff0)dt
Which leads to s(t)ej2πt(ff0)dt=S(ff0)
So [ej2πf0ts(t)]=S(ff0)


PLEASE ENTER PEER REVIEW HERE