EBU4375: SIGNALS AND SYSTEMS

LECTURE 15
Fourier Transform
Fourier Transform

• The Fourier Series (FS) can only be applied to periodic

signals

• Non-periodic signals cannot be analysed using Fourier

series, the Fourier Transform (FT) is required

The Fourier Transform (FT) is defined as: In this course, the Fourier Transform will
always be denoted by an uppercase letter or
symbol, whereas time signals will be denoted
t  by lowercase letters or symbols.
Xf    x(t )dt
e  j t

t  

X(f) is the frequency signal x(t) is the time signal

Fourier Transform – Rectangular Pulse (also
known as Rect Function)

 T T
1 ,  t 
x(t )   2 2
0 , all other t

t
T T

2 2
 Fourier Transform – Rectangular Pulse (also
known as Rect Function)
x(t) is an even signal
Method 1:
even odd
t  t 
Xf    x(t )dt 
e  jt
 (cos(t )  j sin(t ))  x(t )dt
t   t  

2 t  T / 2 2  T 
T /2
1
X  f   2  (1) cos(t )dt  sin(t )   sin   sin( f T )
0
 t 0    2   f
sin( )
Let’s recall the sinc function sinc( )  . After some mathematical

sin( f T )
Manipulation, we have X  f   T . Therefore,
 fT

 T 
 
X(f)  T sinc f T  T sinc 
 2 
 Fourier Transform – Rectangular Pulse (also
known as Rect Function)
Method 2:
t  t T / 2
  jt
t T /2 
X  f    e  x(t )dt   e 1 dt  
 jt  jt e

t   t  T / 2   j t  T / 2 
Therefore

Xf 
1
 j
 
e  jT / 2  e  jT / 2  
j
e e 
1  jT / 2  jT / 2

1
j 2 f

e  j fT
 e  j fT


Using identity sin  

1  j
2j
 
e  e  j , we get X  f  
1
j 2 f
2 j sin( f T ).

 sin( f T ) 
And after some mathematical manipulation, we get X  f   T   .
  fT 
 sin(  ) 
Using identity sinc     , we get X  f   T sinc( f T ) .
  
Fourier Transform – Rectangular Pulse (also
known as Rect Function)
X(f )
1

0.8

0.6
T
0.4

0.2

-20 -10 10 20
f
-0.2

1
f 
T 1 2 3
X ( f )  0, f  , , ,
T T T
Fourier Transform – Rectangular Pulse (also
known as Rect Function – General Formula)

The Fourier transform of a rectangular pulse is given as

 T 
t
Arect    
AT sinc f T  AT sinc 
T   2 

t
T T

2 2
Fourier Transform – Delta Function

x(t ) X(f )
aT  constant
area = aT
2a in all cases T 0

a
a/2

t
T T T T
T   T
2 4 4 2
Fourier Transform – Delta Function

 (t ) 1


t f
0 0
Fourier Transform Table
Fourier Transform Table
Fourier Transform Table
Trigonometric Identities
Some useful trigonometric identities:
Fourier Transform Properties

Part I:
• Linearity
• Time Shift
• Time Scaling
• Multiplication by a Complex Exponential (Frequency Shift)
• Tutorial

Part II:
• Multiplication by a Sinusoid
• Differentiation in the Time Domain
• Integration in the Time Domain
• Convolution in the Time Domain
• Multiplication in the Time Domain
• Tutorial
Fourier Transform Properties – Linearity
The Fourier transform is a linear operation; that is, if xt   X ( f ) and
vt   V ( f ) , then for any real or complex scalars a, b

ax(t )  bvt   aX ( f )  bV ( f )
The properties of linearity can be proved by computing the Fourier
transform of ax(t )  bvt  : By definition of the Fourier transform,

ax(t )  bvt    ax (t )  bv t e  j t
dt

By linearity of integration,
  

 ax (t )  bv t e  j t
dt  a  x (t ) e  j t
dt  b  v t e  j t
dt
  

and thus ax(t )  bvt   aX ( f )  bV ( f )

Fourier Transform Properties – Time Shift
If xt   X ( f ) , then for any positive or negative real number c,

xt  c   X ( f )e  jc
Note that if c > 0, then xt  c  is c-second right sift of xt  ;
if c < 0, then xt  c  is c-second left shift of xt  .
Thus the above transform pair is valid for both left and right shifts of xt  .
To verify this prosperity, first apply the definition of the Fourier transform

  

 x(t  c)e
 j t    j  t  c 
x(t  c) 
 
dt x(t  c)   x t e  
dt
   
_ _ _
Let t  t  c , then t  t  c and dt  d t       j t    jc
   x  t e d t  e
   
 X ( f )e  jc
 Fourier Transform Properties – Time Shift
g (t ) g (t )
A A

t t
T T 0 0 T T

2 2
 t T / 2   t T / 2 
g (t )  Arect  g (t )  Arect 
 T   T 

t
Arect   
AT sinc f T  t
Arect   AT sinc f T  
T  T 

xt   X ( f ) xt   X ( f )
xt  t0   X ( f )e  jt0 xt  t0   X ( f )e  jt0

Therefore, Therefore,

   
T T
j 2 f  j 2 f
G ( f )  AT sinc f T e 2
G ( f )  AT sinc f T e 2
Fourier Transform Properties – Time Scaling

1
If xt   X ( f ) , for any positive real number a, xat   X ( )
f
a a
To verify this prosperity, first apply the definition of the Fourier transform



 j t
x(at )  x ( at ) e dt
_ _ _
Let t  at , then t  t / a and d t  adt ,
   
  j t 
   1 
x(at )   x t e  a 
dt
  
a
   
  j t 
1  
  x  t e  a 
dt
a   
1 f 
 X 
a a
 Fourier Transform Properties – Time Scaling
f
If 0<a<1, x(at ) is a time expansion of x(t ) and X ( ) is a frequency compression of X ( f )
a
f
If a>1, x(at ) is a time compression of x(t ) and X ( ) is a frequency expansion of X ( f )
a

x(t )
1
x(t )  2sinc(2 f )
t
1 1
0
1
xat  
f
X( )
x(t )
a a
1
t x(2t )  sinc( f )
1 1

2 0 2

time compression frequency expansion

Fourier Transform Properties – Multiplication by a
Complex Exponential (Frequency Shift)

If xt   X ( f ) , then

  j t
x t e  X ( f  f0 )
0

So, multiplication a complex exponential in the time domain corresponds to a

frequency shift in the frequency domain.

The proof of this properties follows directly from the definition of the Fourier
transform. You can do the verification after the lecture.
Fourier Transform Properties – Tutorial

1) Find the Fourier transform of

i) g(t) = e‐t sin(2πfo t) u(t)
ii) g(t) = 8 rect(t /4) cos(2π106 t)

2) Find the inverse Fourier transform of

G(f) = 12sinc(4f)sin(4πf)
Fourier Transform Properties

Part I:
• Linearity
• Time Shift
• Time Scaling
• Multiplication by a Complex Exponential (Frequency Shift)
• Tutorial

Part II:
• Multiplication by a Sinusoid
• Differentiation in the Time Domain
• Integration in the Time Domain
• Convolution in the Time Domain
• Multiplication in the Time Domain
• Tutorial
Fourier Transform Properties – Multiplication by a
Sinusoid
If xt   X ( f ) , for any real number f 0 , where 0  2 f 0

1
x t sin(  0 t )  X ( f  f 0 )  X ( f  f 0 ) 
2j
1
x t  cos(  0 t )  X ( f  f 0 )  X ( f  f 0 ) 
2
The proof of this property follows directly from the definition of the Fourier
transform and Euler’s identity.
The signals x(t ) sin(0t ) and x(t ) cos(0t ) can be viewed as amplitude-modulated
signals. More precisely, they are called the modulation theorems of the Fourier
transform.
The above relationships show that modulation of a carrier by a signal xt  results
in the frequency translations X ( f  f 0 ) , X ( f  f 0 ) of Fourier transform X ( f ) .
Fourier Transform Properties – Differentiation in
the Time Domain
If x t   X ( f ) , then
xt   jX ( f )
d
dt
It follows from the above equation that differentiation in the time domain
corresponds to multiplication by j in the frequency domain. To prove this
property, observe that the Fourier transform of dx(t ) / dt is

dx(t )  jt
 dt e dt
and   x(t )
 jt
The integral can be computed “by parts” as follows: with v  e
dv   je  jt and d  dx(t ) / dt  . Then,

dx(t )  jt t  
 dt e dt  v t    dv Then , xt   0 as t  ,

t  
dt  j  x(t )e dt  jX  f 
 jt
e  jt
x(t )   x(t )( j )e  jt

t    
Fourier Transform Properties – Integration in the
Time Domain
Suppose that xt  has the Fourier transform X  f  . Then the integration of a time
function xt  results in the following generalized transform in the frequency
t
1
 x( )d  X  f   X 0   f 

j

where   f  is the impulse function in the frequency domain.

Note that if the signal xt  has no dc component (e.g. X 0  0 ), then the above
equation reduces to
t
1
x( )d  j X  f 
This means the integration of a time function has the effect of dividing its
Fourier transform by the factor of j .
Fourier Transform Properties – Convolution in the
Time Domain
Given two signals xt  and v(t ) with Fourier transforms X  f  and V  f  , the
Fourier transform of the convolution xt   v(t ) is equal to the product X  f V  f 
which results in the transform pair
xt   v(t )  X  f V  f 
This means that convolution in the time domain corresponds to multiplication
in the frequency domain. 

To prove it, first recall that by definition of convolution, xt   v(t )   x( )v(t   )d
Hence, the Fourier transform of xt   v(t ) is given by 


   j t 
 
x( )v(t   )d e dt rewritten x( )v(t   )e dt d
 j t

_
Let t  t   in the second integral,

     

x( )v(t )e  d    v(t )e d t 
 j  j t
 j ( t   )
d t d rewritten  x( )e
    
Fourier Transform Properties – Multiplication in
the Time Domain
If xt   X  f  and vt   V  f  , then


xt v(t )  X  f  V  f    X ( )V ( f   )d


It is seen that multiplication in the time domain corresponds to convolution in

the Fourier transform domain.

The proof of this property follows from the definition of the Fourier transform
and the manipulation of integrals.
Fourier Transform Properties – Tutorial

1) Find the Fourier transform of a triangular pulse

 t
A 1   , t T
g t    T

0, t T
Let’s draw it
t
g (t )  Atri 
T 
A

t
T T
Fourier Transform Properties – Tutorial

Now let’s by defining a doublet pulse which looks like this

g1 t 
A
T
t
T
-A

It is the superposition of 2 shifted rectangular functions, i.e.,

 t T / 2   t T / 2 
g1 (t )  Arect   Arect 
 T   T 
Fourier Transform Properties – Tutorial

 t T / 2   t T / 2 
g1 (t )  Arect   Arect 
 T   T 

What is the Fourier transform of this doublet pulse?


G1 ( f )  ATsinc( f T ) e j fT
 e  j fT

 j 2ATsinc( f T )sin ( f T )
Fourier Transform Properties – Tutorial
Integrating g1 t  , we get
g 2 (t )
AT

t
T T

Therefore, using the integration property, we get the Fourier

Transform of g 2 t 
1
G2  f   G1  f 
j
sin ( f T )
 ATsinc( f T )
 f
 AT sinc ( f T )
2 2
Fourier Transform Properties – Tutorial

Keep in mind that g t   g 2 t  / T , and therefore G  f   G2  f  / T .

The Fourier transform of a triangular pulse is given as

t
Atri   AT sinc 2 ( f T )
T 
A

t
T T
Fourier Transform Properties – Tutorial

2) Find the inverse Fourier transform of

G(f) = 16sinc2(4(f‐106)) + 16sinc2(4(f+106))

3) Find the Fourier transform of

i) g(t) = 10 tri(2t – 1/2)
ii) g(t) = 8 tri(t /2) cos(2π106 t)