The Fourier Transform

R. C. Daileda

Trinity University

Partial Differential Equations

April 17, 2014

Daileda Fourier transforms

The Fourier series representation
For periodic functions

Recall: If f is a 2p-periodic (piecewise smooth) function, then f

can be expressed as a sum of sinusoids with frequencies
0, 1/p, 2/p, 3/p, . . .:
∞ 
X    
nπx nπx
f (x) = a0 + an cos + bn sin ,
p p
n=1

where
Z p sine for bn  
1 z}|{ nπt
an = f (t) cos dt.
p −p p
|{z}
2p when n=0

We can obtain a similar representation for arbitrary (non-periodic)

functions, provided we replace the (discrete) sum with a
(continuous) integral.
Daileda Fourier transforms
The1 Fourier integral representation
For L (R) functions

Theorem
R∞
If f is piecewise smooth and −∞ |f (x)| dx < ∞, then
Z ∞
f (x) = (A(ω) cos(ωx) + B(ω) sin(ωx)) dω,
0

where
Z ∞ Z ∞
1 1
A(ω) = f (t) cos(ωt) dt, B(ω) = f (t) sin(ωt) dt
π −∞ π −∞

Remarks:
R∞
When −∞ |f (x)| dx < ∞, we say that f ∈ L1 (R).
f (x+)+f (x−)
The integral actually equals 2 , which is “almost” f .
Daileda Fourier transforms
Example
If a > 0, find the Fourier integral representation for
(
1 if |x| < a,
f (x) =
0 otherwise.

Note that f is piecewise smooth and that

Z ∞ Z a
|f (x)| dx = dx = 2a < ∞,
−∞ −a

so f has an integral representation.

We find that
Z Z
1 ∞ 1 a
A(ω) = f (t) cos(ωt) dt = cos(ωt) dt
π −∞ π −a

1 sin(ωt) t=a 2 sin(ωa)
=  = .
π ω t=−a πω

Daileda Fourier transforms

Likewise
Z ∞ Z a
1 1
B(ω) = f (t) sin(ωt) dt = sin(ωt) dt = 0.
π −∞ π −a | {z }
odd

It follows that the integral representation of f is


Z ∞

1 if |x| < a,
2 sin(ωa) cos(ωx) f (x+) + f (x−)
dω = = 12 if |x| = a,
π 0 ω 2 

0 otherwise.

Remarks:
It is difficult to evaluate the integral on the left directly; we
have indirectly produced a formula for its value.
Expressing a difficult integral as the Fourier representation of
a “familiar” function gives a powerful technique for evaluating
definite integrals.
Daileda Fourier transforms
“Complexifying” the Fourier integral representation
From Euler’s formula e i θ = cos θ + i sin θ one can deduce
e i θ + e −i θ e i θ − e −i θ
cos θ = and sin θ = .
2 2i

If we apply these in the Fourier integral representation we obtain

Z ∞
f (x) = (A(ω) cos(ωx) + B(ω) sin(ωx)) dω
0
Z Z
1 ∞ i ωx 1 ∞
= (A(ω) − iB(ω)) e dω + (A(ω) + iB(ω)) e −i ωx dω
2 0 2 0
| {z }
sub.−ω for ω
Z ∞ Z 0
1 1
= (A(ω) − iB(ω)) e i ωx dω + (A(−ω) + i B(−ω) )e i ωx dω
2 0 2 −∞ | {z } | {z }
even odd
Z ∞
1
= (A(ω) − iB(ω)) e i ωx dω
2 −∞
Daileda Fourier transforms
The Fourier transform
According to their definitions, we have
Z
1 ∞
A(ω) − iB(ω) = f (x) (cos(ωx) − i sin(ωx)) dx
π −∞
Z
1 ∞
= f (x)e −i ωx dx.
π −∞
Definition: The Fourier transform of f ∈ L1 (R) is
r Z ∞
ˆ π 1
f (ω) = F(f )(ω) = (A(ω) − iB(ω)) = √ f (x)e −i ωx dx.
2 2π −∞

Note that the Fourier integral representation now becomes

Z ∞
−1 ˆ 1
f (x) = F (f )(x) = √ fˆ(ω)e i ωx dω,
2π −∞
which gives the inverse Fourier transform.
Daileda Fourier transforms
Example
For a > 0, find the Fourier transform of
(
1 if |x| < a,
f (x) =
0 otherwise.
2 sin(ωa)
We already found that A(ω) = πω and B(ω) = 0, so
r r
π 2 sin(ωa)
fˆ(ω) = (A(ω) − iB(ω)) = .
2 π ω

Remark: As this example demonstrates, if one already knows the

Fourier integral representation of f , finding fˆ is easy.

Daileda Fourier transforms

 Example
For a > 0, find the Fourier transform of


−1 if − a < x < 0,
f (x) = 1 if 0 < x < a,


0 otherwise.

We have
Z ∞  Z 0 Z a 
1 1
fˆ(ω) = √ f (x)e −i ωx
dx = √ − e −i ωx
dx + e −i ωx
dx
2π−∞ 2π −a 0
  !
1 1 −i ωx 0 1 −i ωx a 1 1

=√ e  − e  =√ · 2 − e i ωa − e −i ωa
2π i ω −a iω 0 2π i ω
r
1 2 (cos(ωa) − 1)
= √ (2 − 2 cos (ωa)) = i .
i ω 2π π ω

Daileda Fourier transforms

Example
Find the Fourier transform of f (x) = e −|x| .

We have
Z ∞
ˆ 1
f (ω) = √ f (x)e −i ωx dx
2π −∞
Z 0 Z ∞ 
1
=√ e x e −i ωx dx + e −x e −i ωx dx
2π −∞ 0
Z 0 Z ∞ 
1
=√ e x(1−i ω) dx + e −x(1+i ω) dx
2π −∞ 0
0  !
ω) ∞
1 e x(1−i ω)  e −x(1+i
=√  − 
2π 1 − i ω −∞ 1 + i ω 0
  r
1 1 1 2 1
=√ + = .
2π 1 − i ω 1 + i ω π 1 + ω2

Daileda Fourier transforms

Properties of the Fourier transform
Linearity: If f , g ∈ L1 (R) and a, b ∈ R, then

af\
+ bg = afˆ + bĝ .

Transforms of Derivatives: If f , f ′ , f ′′ , . . . , f (n) ∈ L1 (R) and

f (k) (x) → 0 as |x| → ∞ for k = 1, 2, . . . , n − 1, then

fd
(n) (ω) = (i ω)n fˆ(ω),

i.e. differentiation becomes multiplication by i ω.

Derivatives of Transforms: If f , x n f ∈ L1 (R), then
1 n
fˆ(n) (ω) = n xd f (ω),
i
i.e. multiplication by x (almost) becomes differentiation.
Daileda Fourier transforms
Remarks
Linearity follows immediately from the definition and linearity
of integration.
For the second fact, if f , f ′ ∈ L1 (R) then
Z ∞
b 1
′
f (ω) = √ f ′ (x)e −i ωx dx
2π −∞
| {z }
u=e −i ωx , dv =f ′ (x) dx
 ∞ Z 
1 −i ωx 
 ∞
=√ f (x)e  + iω f (x)e −i ωx
dx = i ω fˆ(ω),
2π −∞ −∞

provided f (±∞) = 0. The general result follows by iteration,

e.g.
d
fb′′ (ω) = (f ′ )′ (ω) = i ω fb′ (ω) = (i ω)2 fˆ(ω).

The final fact follows by “differentiating under the integral.”

Daileda Fourier transforms
Example
Find the Fourier transform of f (x) = (1 − x 2 )e −|x| .

Using properties of the Fourier transform, we have

′′
fb = e[
−|x| − x\ [
2 e −|x| = e [
−|x| − i 2 e −|x|
r   ′′ 
2 1 1
= +
π 1 + ω2 1 + ω2
r   r
2 1 6ω 2 − 2 2 ω 4 + 8ω 2 − 1
= + = .
π 1 + ω 2 (1 + ω 2 )3 π (1 + ω 2 )3

Remark: The Fourier inversion formula tells us that

Z ∞ 4
ω + 8ω 2 − 1 i ωx
2 )3
e dω = π(1 − x 2 )e −|x| .
−∞ (1 + ω

Imagine trying to show this directly!

Daileda Fourier transforms
Example
2
Find the Fourier transform of the Gaussian function f (x) = e −x .

Start by noticing that y = f (x) solves y ′ + 2xy = 0. Taking

Fourier transforms of both sides gives
ω
(i ω)ŷ + 2i ŷ ′ = 0 ⇒ ŷ ′ + ŷ = 0.
2
The solutions of this (separable) differential equation are
2 /4
ŷ = Ce −ω .

We find that
Z ∞
1 2 1 √ 1
C = ŷ (0) = √ e −x e −i 0x dx = √ π=√ ,
2π −∞ 2π 2
2
and hence ŷ = fˆ(ω) = √1 e −ω /4 .
2
Daileda Fourier transforms
 Example (Another useful property)
1 ˆ ω 
Show that for any a 6= 0, \
f (ax) = f .
|a| a

We simply compute
Z ∞ Z ∞
\ 1 −i ωx 1 1 ˆω 
f (ax) = √ f (ax)e dx = √ f (u)e −i ωu/a du = f
2π −∞ |a| 2π −∞ |a| a
| {z }
sub. u=ax

Example
2
Find the Fourier transform of f (x) = e −x /2 .
√ 2 √
Since f (x) = e −(x/ 2) , taking a = 1/ 2 above and using the
Gaussian example gives
√ 1 √ 2 2
fˆ(ω) = 2 √ e −( 2ω) /4 = e −ω /2 .
2
Daileda Fourier transforms
Example
2 /2
Find the general solution to the ODE y ′′ − y = e −x .

The corresponding homogeneous equation is

yh′′ − yh = 0 ⇒ yh = c1 e x + c2 e −x .

It remains to find a particular solution yp .

We assume that an L1 (R) solution exists, and take the Fourier
transform of the original ODE:
2
2 −ω 2 /2 −e −ω /2
(i ω) ŷ − ŷ = e ⇒ ŷ = 2 .
ω +1
Applying the inverse Fourier transform we obtain
Z ∞ 2
1 −e −ω /2 i ωx
yp = √ e dω.
2π −∞ ω 2 + 1

Daileda Fourier transforms

We can eliminate i by a judicious use of symmetry:
Z ∞ 2 Z ∞ 2
1 −e −ω /2 i −e −ω /2
yp = √ 2+1
cos(ωx) dω + √ 2
sin(ωx) dω
2π −∞ | ω {z } 2π −∞ | ω + 1{z }
even odd
r Z 2
2 ∞ −e −ω /2
= cos(ωx) dω.
π 0 ω2 + 1

Hence, the complete solution is

r Z 2
x −x 2 ∞ −e −ω /2
y = yh + yp = c1 e + c2 e + cos(ωx) dω.
π 0 ω2 + 1

Daileda Fourier transforms