Sie sind auf Seite 1von 30

Fourier Series & The Fourier Transform

What is the Fourier transform?


Fourier cosine and sine series for
even and odd functions
The continuous limit: the Fourier
transform (and its inverse)
Some transform examples
The Fourier transform with respect to
position and the two-dimensional
Fourier transform
( ) ( ) exp( ) F f t i t dt e e

=
}
1
( ) ( ) exp( )
2
f t F i t d e e e
t

=
}
Prof. Rick Trebino, Georgia Tech
What do we hope to achieve with the
Fourier transform?
We desire a measure of the frequencies present in a wave. This will
lead to a definition of the term, the spectrum.
Plane waves have only
one frequency, e
0
.
This light wave has many
frequencies. And the
frequency increases in
time (from red to blue).
L
i
g
h
t

e
l
e
c
t
r
i
c

f
i
e
l
d

Time
Sine wave
#1
Sine wave #2
time
Many waves are sums of sinusoids.
Consider the sum
of two sine
waves of different
frequencies:
Sum
The spectrum of
the sum of the
two sine waves:
Frequency #1 Frequency #2
frequency
Joseph Fourier
Fourier was
obsessed with the
physics of heat
and developed
the Fourier series
and transform to
model heat-flow
problems.
Joseph Fourier 1768 - 1830
Lord Kelvin on Fouriers theorem
Fouriers theorem is not
only one of the most
beautiful results of
modern analysis, but it
may be said to furnish an
indispensable instrument
in the treatment of nearly
every recondite question
in modern physics.

Lord Kelvin
Fourier
decomposing
functions
Here, we write a
square wave as
a sum of sine
waves.
sin(et)
sin(3et)
sin(5et)
Any function can be written as the
sum of an even and an odd function.
( ) [ ( ) ( )] / 2
( ) [ ( ) ( )] / 2
( ) ( ) ( )
E x f x f x
O x f x f x
f x E x O x
+

= +
Fourier cosine series
Because cos(mt) is an even function (for all m), we can write an even
function, f(t), as:






where the set {F
m
; m = 0, 1,

} is a set of coefficients that define the
series.

And where well only worry about the function f(t) over the interval
(,).

f(t) =
1
t
F
m
cos(mt)
m=0

Finding the coefficients F


m

in a Fourier cosine series 0
1
( ) cos( )
m
m
f t F mt
t

=
=

0
1
( ) cos( ) cos( ) cos( )
m
m
f t mt dt F mt mt dt
t t
t t
t

=

' '
=

} }
cos( ) cos( )
0
if m m
mt mt dt
if m m
t
t
t

'
=

'
=

'
=

}
1
( ) cos( )
m
f t mt dt F
t
t
t
t
'

'
=
}
( ) cos( )
m
F f t mt dt
t
t
=
}
To find F
m
: Multiply each side by cos(mt), where m is another
integer, and
integrate:
But:
So:
Dropping the
from the m:
only the m = m
term contributes
yields the
coefficients for
any f(t)!

Fourier sine series
f (t) =
1
t
' F
m
sin(mt)
m=0

Because sin(mt) is an odd function (for all m), we can write


any odd function, f(t), as:
where the set {F ; m = 0, 1,

} is a set of coefficients that define
the series.
m
where well only worry about the function f(t) over the interval (
,).
Finding the coefficients F
in a Fourier sine series
0
1
( ) sin( )
m
m
f t F mt
t

=
'
=

0
1
( ) sin( ) sin( ) sin( )
m
m
f t mt dt F mt mt dt
t t
t t
t

=

' ' '
=

} }
sin( ) sin( )
0
if m m
mt mt dt
if m m
t
t
t

'
=

'
=

'
=

}
1
( ) sin( )
m
f t mt dt F
t
t
t
t
'

' '
=
}
( ) sin( )
m
F f t mt dt
t
t
'
=
}
m
m
To find F: Multiply each side by sin(mt), where m is another integer,
and
integrate:
But:
So:
Dropping the
from the m:
yields the
coefficients for
any f(t)!
only the m = m
term contributes

Fourier series
0 0
1 1
( ) cos( ) sin( )
m m
m m
f t F mt F mt
t t

= =
'
= +

F
m
= f (t) cos(mt) dt
}
' F
m
= f (t) sin(mt) dt
}
So if f(t) is a general function, neither even nor odd, it can be
written:
and
where
even component odd component

F
m
vs. m
m
We really need two such plots, one for the cosine series and another
for the sine series.
Let the integer
m become a
real number
and let the
coefficients F
m

become a
function F(m).
F(m)
1
0 20
10
30
We can plot the coefficients of a Fourier
series.
The Fourier transform
Consider the Fourier coefficients. Lets define a function F(m) that
incorporates both cosine and sine series coefficients, with the sine
series distinguished by making it the imaginary component:
( ) cos( ) f t mt dt
}
( ) sin( ) i f t mt dt
}
F(m) F
m
i F =
( ) ( ) exp( ) F f t i t dt e e

=
}
The Fourier
Transform
m
F(e) is called the Fourier Transform of f(t). It contains
equivalent information to that in f(t). We say that f(t) lives in
the time domain, and F(e) lives in the frequency domain.
F(e) is just another way of looking at a function or wave.
Lets now allow f(t) to range from to , so well have to integrate
from to , and lets redefine m to be the frequency, which well
now call e:
The inverse Fourier transform
0 0
1 1
( ) cos( ) sin( )
m m
m m
f t F mt F mt
t t

= =
'
= +

1
( ) ( ) exp( )
2
f t F i t d e e e
t

=
}
Inverse
Fourier
Transform
Now transform the sums to integrals from to , and again
replace F
m
with F(e). Remembering the fact that we introduced a
factor of i (and including a factor of 2 to account for double
summing in integrating from to , rather than 0 to ), we have:
The Fourier Transform takes us from f(t) to F(e).
How about going back?
Recall our formula for the Fourier Series of f(t) :
The Fourier transform and its inverse
The Fourier Transform and its Inverse:
Inverse Fourier Transform
Fourier Transform
( ) ( ) exp( ) F f t i t dt e e

=
}
1
( ) ( ) exp( )
2
f t F i t d e e e
t

=
}
So we can transform to the frequency domain and back.
Interestingly, these transformations are very similar.
These equations together are Fouriers Theorem. Feel free to plug
one into the other to verify that you get the function you started with!
There are several ways to denote the Fourier transform of a
function.

If the function is labeled by a lower-case letter, such as f,
we can write:
f(t) F(e)

If the function is already labeled by an upper-case letter, such as E,
we can write:
Fourier transform notation
( ) ( ) E t E e ( ) { ( )} E t E t F


Sometimes, this symbol is
used instead of the arrow:
or:
The spectrum
We define the spectrum, S(e), of a wave E(t) to be:
2
( ) { ( )} S E t e F
This is the measure of the frequencies present in a wave.

What about the phase of the Fourier transform, which we lose when
we take its squared magnitude? It yields the phase of a given
frequency, which yields information about when a particular frequency
occurs.
Example: the Fourier transform of a
rectangle function, rect(t)
1/ 2
1/ 2
1/ 2
1/ 2
1
( ) exp( ) [exp( )]
1
[exp( / 2) exp(
F i t dt i t
i
i i
i
e e e
e
e e
e

= =

= / 2)|

}
( sinc( F e e ) = /2)
e
sinc(e/2)
The sinc function is
important in many fields.
rect(t)
t
-.5 0 .5
1
zeros at
e = 2nt
(n 0)
exp( / 2) exp(
2
i i
i
e e
e
1 / 2)
=
( / 2)
sin(e
e
/ 2)
=
( / 2)
The Fourier transform of a Gaussian is a
Gaussian!
f(t) = exp(t
2
/2)
e
t
Fourier transforms can be difficult to do. But heres one you can try
on your own if you like.
f(t) and F(e) look similar, but remember that these functions live in
different spaces (have different arguments)!
2
( ) 2 exp( / 2) F e t e =
Scale Theorem
{ ( )} ( / ) / f at F a a e = F
{ ( )} ( ) exp( ) f at f at i t dt e

=
}
F
{ ( )} ( ) exp( [ / ]) / f at f u i u a du a e

=
}
F
( ) exp( [ / ] ) / f u i a u du a e

=
}
( / ) / F a a e =
If a < 0, the limits flip when we change variables, introducing a
minus sign, hence the absolute value.
Assuming a > 0, change variables: u = at, so dt = du / a
Proof:
The Fourier transform
of a scaled function, f(at):
The Scale
Theorem
in action
f(t) F(e)
Short
pulse
Medium-
length
pulse
Long
pulse
The shorter
the pulse,
the broader
the spectrum!
This is the essence
of the Uncertainty
Principle!
e
e
e
t
t
t
The pulse width
There are many definitions of the
width or length of a wave or pulse.

The effective width is the width of a rectangle whose height and
area are the same as those of the pulse.

Effective width Area / height:
We can similarly define widths of F(e), the spectrum, or any other
quantity.
1
( )
(0)
eff
t f t dt
f

A
}
t
f(0)
0
At
eff

t
At
The Uncertainty Principle
The Uncertainty Principle says that the product of a function's widths
in the time domain (At) and the frequency domain (Ae) is 1.
1 1 (0)
( ) ( ) exp( [0] )
(0) (0) (0)
F
t f t dt f t i t dt
f f f


A > = =
} }
(0) (0)
2
(0) (0)
f F
t
F f
e t A A > 2 t e t A A > 1 t v A A >
Combining results:
or:
Use effective widths
assuming f(t) and
F(e) peak at 0:
1 1
( ) ( )
(0) (0)
t f t dt F d
f F
e e e


A A
} }
1 1 2 (0)
( ) ( ) exp(
(0) (0) (0)
f
F d F i d
F F F
t
e e e e e e


A > = |0|) =
} }
1
Av At = Time-Bandwidth Product
Shift Theorem
( ) { }
( ) :
( ) exp( )

( ) exp( [ ])
exp( ) ( ) exp( )

f t a
f t a f t a i t dt
u t a du dt
f u i u a du
i a f u i u du
e
e
e e

=
= =
+
=
}
}
}
The Fourier transform of a shifted function,
Proof :
Change variables: so
F
exp( ) ( ) i a F e e =
{ }
( ) exp( ) ( ) f t a i a F e e = F
The Fourier
transform of a
double pulse
f(t) = rect[(t+a)/w] + rect[(ta)/w]
( ) sinc( / 2) exp( )
sinc( / 2) exp( )
F w w ia
w w ia
e e e
e e
= + +

0
t
a -a
w w
f(t)
e
F(e)
Use the scale and shift theorems
simultaneously:
( ) 2 sinc( / 2) cos( ) F w w a e e e =
Simplifying:
sinc(we/2)
Shift Theorem for the inverse Fourier transform
{ }
1
0 0
( ) exp( ) ( ) F i t f t e e e

= + F
Because the Fourier transform and inverse Fourier transform are
so similar, any theorem for one also holds for the other, subject
to a change in minus sign (and perhaps a factor of 2t).
{ }
0 0
exp( ) ( ) ( ) i t f t F e e e + = F
Taking the Fourier transform of both sides:
2
0
2
0
exp( ) exp( / 2)
exp ( ) / 2
i t t e
e e
+
(


This means that shifting the frequency of a wave moves its spectrum:
e
0
e
0

F(e)
Fourier Transform with respect to space
F {f(x)} = F(k)
( ) ( ) exp( ) F k f x ikx dx

=
}
If f(x) is a function of position,
We refer to k as the spatial frequency.

Everything weve said about Fourier transforms between the t and e
domains also applies to the x and k domains.
k
x
F(k)
f(x)
F
(2)
{f(x,y)} = F(k
x
,k
y
)

= f(x,y) exp[i(k
x
x+k
y
y)] dx dy

If f(x,y) = f
x
(x) f
y
(y),

then the 2D FT splits into two 1D
FT's.

But this doesnt always happen.
The 2D Fourier Transform
}}
F
(2)
{f(x,y)}
x
y
f(x,y)
Suppose we have a function of x and y.
Then:
A 2D Fourier Transform: a square function
x
y
f(x,y)
|F
(2)
{f(x,y)}|
2

Consider a square function in the xy
plane:

f(x,y) = rect(x) rect(y)

The 2D Fourier Transform splits into
the product of two 1D Fourier
Transforms:

F {f(x,y)} = sinc(k
x
/2) sinc(k
y
/2)


This picture is an optical determination
of the mag-squared Fourier Transform
of the square function!