AGNEL INSTITUTE OF TECHNOLOGY AND DESIGN

AGNEL TECHNICAL EDUCATIONAL COMPLEX

ASSAGAO, BARDEZ-GOA. 403 507
(Approved by AICTE, Affiliated to Goa University)
DEPARTMENT OF ECE
ECE 4.2 SIGNALS AND SYSTEMS
Module 3- COURSE MATERIALS

Representation of aperiodic signal Continuous Time Fourier transform(CTFT)

Let us consider a aperiodic signal x(t) of finite duration,

x(t) = 0 in the interval |t| > T1, we can construct a periodic signal x(t) for which x(t) is
one period.

if we choose the period T is very large, x(t) is identical to x(t).

The fourier series representation of periodic signal x(t)
k= jkωo t 1
T/2
-jkωo t 2
x(t)=  a e , a k   x(t) e dt ,o = ,
k=- k T T
-T/2

x(t)=x(t) for t < T / 2 so a k can be written as


1 -jkωo t
ak  
T -
x (t) e dt , define the envelope X( ) of To a k


-jkω t
Let X( )   x (t) e
-
dt ---------(1)

Page | 1
From Equation (1) and FS coefficient a k ,
1
ak  X(ko )
T
k= jkωo t k= 1 jkωo t
x(t)=  a e   X(ko )e
k=- k k=- T
2 2
o  , T= as T , x(t)=x(t) ,o  0,
T o
so the summation becomes integral

1 jkω t
x(t)  
2 -
X( )e d --------(2)

The equations 1 & 2 are called F.T pair, the equation 2 is called synthesis equation and 1 is
called analysis equation. X(ω) is called spectrum of x(t). the Fourier transform gives the
frequency components of the time domain signal.

Find the Fourier transform and sketch the spectrum. x(t) = e-at u(t) a > 0
x(t) = e-at u(t) a > 0
  
X(j ) =  x(t)e-j t dt   e-at u(t)e-j t dt   e-at e-j t dt
- - 0

 
 e-(a+j )t  1
=  e-(a+j )t dt =   =
0  -(a+j )  0 a+j
magnitude spectrum phase spectrum
1  
X(j )  X(j ) = -tan -1  
a2  2 a

ω -a 0 a
|X(ω)| 1/√2 a 1/a 1/√2 a
∟X(ω) π/4 0 π/4

Sketching of spectrums
Magnitude spectrum Phase spectrum

Page | 2
Find the fourier transform and sketch the spectrum. x(t) = e- a t u(t) a > 0
-a t
x(t) = e u(t) a > 0
  0  0 
X(j ) =  x(t)e -j t
dt   e e e dt   e e dt =  e e
- a t -j t at -j t -at -j t (a-j )t -(a+j )t
dt  e dt  dt
- -  0 - 0
0 
 e(a-j )t   e-(a+j )t  1 1 2a
=   +  -(a+j )  = a-j  a+j = a 2   2  X(j ) magnitude spectrum
 (a-j )    0

ω -a 0 a
|X(ω)| 1/ a 2/a 1/ a

Find the fourier transform and sketch the spectrum.

x(t) = 1 |t| ≤ T1
=0 |t| > T1

 T1
 e-j t 
T1

X(j ) =  x(t)e -j t

dt   e -j t
dt   
- -T1  -j  -T1
-e-jT1  e jT1 2(e jT1 -e-jT1 ) 2
=   sin  T1
j 2j 
sin( )
si nce sinc( ) =

2 2T1
sin  T1 can be written as sin  T1
  T1
sin  T1  T1
X(j )=2T1 ,  =  T1 ,  =
 T1 
 T1
X(j )=2T1 sin c( ),

 2 3
amplitude 2T1 and X(j ) gives zero values for  =  , , ...
T1 T1 T1

Page | 3
Sinc spectrum

Find the inverse fourier transform and sketch the spectrum.

X(jω) = 1 |ω| ≤ W
=0 |ω| > W

 W
1  e j t 
W
1 1
- X(j )e d  2 
j t j t
e d 
2  jt  -W
x(t)=
2 -W

e jWt  e jWt 1
=  sin Wt
j2 t t
sin( )
si nce sinc( ) =

1 W
sin Wt can be written as sin Wt
t  Wt
W sin Wt Wt
x(t )= ,  = Wt ,  =
 Wt 
W Wt
x(t)= sin c( ),
 
W  2 3
amplitude and x(t ) gives zero values for t =  ,  ,  ...
 W W W

Page | 4
Spectrum

Fourier Transform of Periodic signals

Let us consider a signal x(t) with fourier transform X(ω) which is single impulse
area 2π at ω =ωo
 
X(j )=  2 a k (  ko ) , where a k   x(t )e jko t
This is the relationship between
k=- k 
Fourier transform and fourier series coefficients.

Find the fourier transform, x(t) = sin ωot

k= jkωo t
x(t)   a e
k=- k
1 1
k  1 a1  , k  1 a -1  
2j 2j

X( )=  2 a k (  ko )  2 a1 (  o )  2 a -1 (  o )
k=-

2 2  π
X( )=  (  o )   (  o )   (  o )   (  o )
2j 2j j j

Page | 5
Find the fourier transform, x(t) = cos ω ot

k= jkωo t
x(t)   a e
k=- k
1 1
k  1 a1  , k  1 a -1 
2 2

X( )=  2 a k (  ko )  2 a1 (  o )  2 a -1 (  o )
k=-

2 2  π
X( )=  (  o )   (  o )   (  o )   (  o )
2 2 1 1

Differences b/w F.T and F.S

Fourier Series Fourier Transform

1 Applying to periodic signals Applying to aperiodic signals
2 The resultant spectrum is The resultant spectrum is
discrete. continuous.

Find the Fourier Transform of x(t) = e j  o t

Given x(t)=e jot is a complex exponential signal i.e. periodic signal

To find a fourier Transform for a periodic signal,
Let X(j ) =2 ( -o )
The Inverse Fourier Transform of X(j ) is,
 
1 1
 X(j )e d   2 ( - )e
j t j t
x(t)= d
2 2
o
- -

x(t)    ( -o )e j t d  This Integration has a value when  =o because of  ( -o ),
-

x(t)  e jot
From the above Proof, the IFT of 2 ( -o ) is e jot
or in other word FT of e jot is 2 ( -o )

Page | 6
   
Finding of Fourier Transform x(t)= s i n  t  
 3 4 
   
G iv e n s ig n a l x (t)= s in  t   , L et  o 
 3 4  3
F i r s t s t e p E x p r e s s t h e g i v e n s i g n a l i n E u le r f o r m t o f i n d a k v a lu e s ,
             
j t  j t  j  j t   j   j t 
   e  3 4 
 e  3 4 
e  4 
e  3 
e  4 
e  3 
x (t)= s in  t     
 3 4  2 j 2 j 2 j
   
j   j 
 4  j  o t   4   j  o t 
e e e e
x (t)      (1)
2 j 2 j
W K T , F o u rie r S e rie s -
k= jk ω o t -j2 ω o t -jω o t jω o t j2 ω o t
x (t)   ake  a 2 e a 1e  a 0 + a 1e  a 2e    (2)
k = -
-jω o t
E q u a t i n g ( 1 ) a n d ( 2 ) a n d f i n d t h e a k v a lu e s , c o e f f i c i e n t o f e is a 1 ,
   
j   j 
jω o t  4   4 
e e
c o e ffic ie n t o f e is a 1 , s o fro m e q u a tio n a 1  ,a 1  
2 j 2 j

F o r a p e rio d ic s ig n a l F T is X (j )= 
k=-
2  a k  (  k  o )

V a lu e o f k = - 1 a n d 1
   
j   j 
e  4 
 e  4 

X ( j  ) = 2  a 1 (    o )  2  a  1 (    o )    (  )  (  )
j 3 j 3

Find x(t) from X(j )=2 ( )   (  4 )   (  4 )

Given X(j )=2 ( )   (  4 )   (  4 )

1
 X(j )e
j t
Inverse Fourier Transform Equation, x(t)= d
2 -

Substitute X(j ) in the above equation and find x(t)


1
 2 ( )   (  4 )   (  4 )e
j t
x(t)= d
2 -

1   
x(t)=  2   e   e     (  4 )e j t d   e j4 t 
j4 t -j4 t

2  - 
1  e j4 t  e-j4 t  (e j4 t  e-j4 t )
x(t)=  2   e j4 t   e-j4 t   1   1  1   cos(4 t )
2 2 2 2

Representation of aperiodic signal Discrete Time Fourier transform(DTFT)

Let us consider a aperiodic signal x[n] of finite duration from –N1 to N1,

Page | 7
 ~
x[n] = 0 |n| > N1, we can construct a periodic signal x[n] for which x[n]) is one
period.

~
if we choose the period N is very large, x[n] is identical to x[n]. The fourier series
~
representation of periodic signal x[n]
~ k= N jk  2N  n 1 N1 ~ -jk  2N  n 2
x[n]=  a e  
, a k   x[n]e   ,o = ,
k=- N k N N 1
N
~
When N increases x[n]=x[n] so a k can be written as
 2 
1  -jk  
ak  
N n 
x[n]e  N  n , define the envelope X(e j ) of Na k
n 
Let X(e j )   x[n]e
n 
 j n
---------(1)

From Equation (1) and FS coefficient a k ,

1
ak  X(e ko )
N
~ k= N jk  2  k= N 1 jkωon
x[n]=  a e  N  n   X(e ko )e
k=- N k k=- N N
2 2 ~
o  , N= as N , x[n]=x[n] ,o   0,
N o
so the summation becomes integral
1 jkω n
x[n]  
2 2
X(e j ) e d --------(2)

The equations 1 & 2 are called F.T pair, the equation 2 is called synthesis equation and 1 is
called analysis equation. X(ejω) is called spectrum of x[n]. the fourier transform gives the
frequency components of the time domain signal.

Page | 8
 Find the Fourier transform and sketch the spectrum. x[n] = a n u[n] a <1
Given x[n] = a n u[n]
n 
Fourier Transform of the signal x[n], X(e j )   x[n]e
n 
 j n

 n  n 1 
n  n  n  n
1
X(e ) =  a u[n]e
j n  j n
 a e n  j n
  ae  j
    a  (1  a ) 
n  n0 n0 1  ae  j  n0 
1
X(e j ) =
1  a(cos   j sin  )
magnitude spectrum phase spectrum
1 1  a sin  
X(e j )  = , X(e j ) = -tan -1  
(1  a cos  ) 2  a 2 sin 2  (1  a cos  ) 2  a 2 sin 2   1  a cos  
a=0.5
ω 0 π/2 π 3π/2 2π
|X(ejω)| 1 1 1 1 1
=2 =1.12 =0.66 =1.12 =2
1 a 1 a 2 1 a 1  a2 1 a
∟ X(ejω) 0 -tan-1a= -0.46 0 tan-1a=0.46 0

Sketching of spectrums
Magnitude spectrum Phase spectrum

Find the fourier transform and sketch the spectrum.x[n]=δ[n-1] + δ[n+1]

Given x  n  =δ  n-1 + δ  n+1

n 
X(e j )    δ  n-1 + δ  n+1 e
n 
 j n
 e  j  e j  2 cos( )

Page | 9
 n
Find the fourier transform and sketch the spectrum. x[n] =a a <1
n
Given x[n] = a
For the Given signal x[n] = a n n  0
a  n n<0
n 
Fourier Transform of the signal x[n], X(e j )   x[n]e
n 
 j n

n  n 1 n  n 1 n 
   ae  j     ae     ae j 
n 1
X(e j ) =  a n e  j n  a  n  j n n j n
e   j
n 0 n  n0 n  1  ae n 1
n  n 
   ae j   1  1     ae j 
1 n 1 n 1 1
  j  j
1   j
 1
1  ae n 1 1  ae n0 1  ae 1  ae j
1 ae j 1  ae j +ae j (1  ae  j ) 1  a2
   
1  ae  j 1  ae j (1  ae  j )(1  ae j ) (1  ae  j  ae j +a 2 )
1  a2
X(e j ) =  Real, only magnitude Spectrum
(1  2acos +a 2 )
a=0.5
ω 0 π/2 π 3π/2 2π
|X(ejω)| 1 a 1  a2 1 a 1  a2 1 a
=3 =0.6 =0.33 =0.6 =3
1 a 1  a2 1 a 1  a2 1 a

Find the fourier transform and sketch the spectrum.

x[n]=1 n=-2,-1,0,1,2
Given x[n] = 1 n=-2,-1,0,1,2
n 
Fourier Transform of the signal x[n], X(e j )   x[n]e
n 
 j n

n2
X(e j ) =  e  j n  e j 2  e j  1  e  j 2  e  j  1  2 cos   2 cos 2  Real , only magnitude Spectrum
n 2

ω 0 π/2 π 3π/2 2π
|X(ejω)| 5 -1 1 -1 5

Page | 10
Find the fourier transform and sketch the spectrum.

x[n] = 1 |n| ≤ M
=0 |t| > M

Given x[n] = 1 -M  n  M
nM
Fourier Transform of the signal x[n], X(e j )   1e
n  M
 j n

To Convert into finite series, Let p=n+M, when n=-M, p=0, n=M, p=2M
p2 M p2 M
1  e  j (2 M 1)   p  2 M p 1  a 2 M 1 
X(e j )  e  j ( p  M )
 e j M e  j p
 e j M   j    a  1  a 
p 0 p 0  1 e   p 0 
 j (2 M2 1)  j (2 M2 1)   
(2 M 1) (2 M 1)
 j  j
j e j M e 2
e e   by taking e
2
outside in NR 
X(e ) =
j
  j

j
  j
 
e 2
 e 2
 e 2
 and e 2 in the DR 
 (2 M  1) 
sin    0
X(e j )   2   when  =0,2 ,4 ,........

  0
sin  
2
 (2 M  1)  (2 M  1)
cos   
lt
Use L Hospital Rule,   0, 2 , 4  2  2  2M  1

  1
cos  
2 2
X(e j )  2 M  1 when  =0,2 ,4 ,........

Page | 11
Spectrum

Find the inverse DT fourier transform and sketch the spectrum.

X(ejω) = 1 |ω| ≤ W
=0 |ω| > W

Given X(e j ) = 1 -W    W

1
Inverse Fourier Transform of X(e j ), x[n]=  X(e
j
)e j n d 
2 
W
1 e 
j n
W
1 1
 1e e jWn  e  jWn 
j n
x[n]= d    
2 W
2  jn  W 2 jn 
sin Wn 
x[n]= 
n
sin( )
si nce sinc( ) = , To convert add W in the NR & DR

sin Wn W sin Wn
can be written as
n  Wn
W sin Wn Wn
x[n]= , In the equation  = Wn,  =
 Wn 
W Wn
x[n]= sin c( ),
 
W  2 3
amplitude is and x[ n] gives zero values for n =  ,  ,  ...
 W W W

Page | 12
Spectrum

Fourier Transform of a Discrete Time Periodic Signal

In Continuous time periodic signal, the fourier transform executes at ω=ω o. However, the
DTFT must be periodic in ω with period 2π (because e jo n is periodic with period 2π). The
fourier transform has impulses at ωo, ωo±2π, ωo±4π and so on. So the Fourier Transform of
e jo n is a periodic impulse train and it can be written as,


X(e j )   2 (  
l 
o  2 l )

Consider a periodic sequence x[n] with period N and with FS representation

 2 
jk  n
x[n]  
k  N 
ake  N 

The Fourier Transform of this Equation is,


 2 k 
X  e j    2 a    k 
k  N 
 
Find the Fourier Transform of x[n]  2  cos  n  
6 8
   2 
Given x[n]  2  cos  n   , o    , N  12
6 8 6 N 6
Given Signal is a periodic signal, to find the FT of periodic signal we have to find a k ,
 2   2   2 
jk  n j n  j n
x[n]  
k  N 
ake  N 
 a1e  N 
 a 0  a 1e  N 

             
j n   j n  j n  j   j n   j 
  e 6 8
e 6 8
e 6 
e 8
e 6 
e 8
x[n]  2  cos  n    2    2 
6 8 2 2 2 2

Page | 13
        
j   j  j   j 
j o n  8  j o n  8 8 8
e e e e e e
x[n]  2   , a 1   a 0  2, a1 
2 2 2 2

 2 k  
The Fourier Transform X  e j    2 ak        2 ak    o k 
k   N  k 
X  e j   2 a1   o   2 ao   2 a1   o 
   
 j  j 
8 8
X  e j   2
e e
   o   2 2    2     o 
2 2
   
   
X e    e
 j  j 
j
      4     e  8     
8

 6  6
Find the Inverse Fourier Transform of

      
X  e j     2   2 k   2     2 k   2     2 k  
k    2   2 

      
Given X  e j    2   2 k   2    2  2 k   2    2  2 k  
k 

   
X  e j   2    2      2    
 2  2
 
1 1      
 X(e j )e j n d    2    2    2   2    2  e
j n
x[n]= d
2 
2 

1  
j n j n
 
j n

j n  
x[n]   2  2 e  2 e
2 2
  1 e  e
2 2
 1  2 cos  n 
2   2 

Properties of CT and DT fourier transform

Fourier transform pair CT FT


-jω t
X( )   x (t) e
-
dt ---------(1) analysis equation(FT)


1 jω t
x(t) 
2 
-
X( )e d --------(2)synthesis equation(IFT) used for decomposition

of signal into linear combination of complex exponential signal

Fourier transform pair DT FT


X(e j )   x[n]e
n 
 j n
---------(1) analysis equation(FT)

1 jωn
 X(e
j
x[n]  )e d --------(2)synthesis equation(IFT)
2 

Page | 14
1) Linearity-CTFT
Let x(t) and y(t) are signals with fourier transform of X(jω) and Y(jω)
x(t) F.T X(jω) y(t) F.T Y(jω)

F.T( a x(t)+by(t)) = a X(jω) +bY(jω)

Proof

FT  ax(t) + by(t)     ax(t) + by(t)  e
-jω t
dt
-
 
-jω t -jω t
 a  x (t) e dt  b  y (t) e dt
- -

= aX(j )+bY(j )
1-b)Linearity-DTFT
Let x[n] and y[n] are signals with fourier transform of X(ejω) and Yejω)
x[n] F.T X(ejω) y[n] F.T Yejω)

F.T(a x[n]+by[n]) = a X(ejω) +b Y(ejω)

Proof

FT  ax[n] + by[n]    ax[n] + by[n]e  j n

n 
 
  ax[n]e
n 
 j n
  by[n]e
n 
 j n

j j
= aX(e )+bY(e )

2 a)Time shifting-CTFT
-jωt 0
X(jω) is the fourier transform of x(t) , e X(j ) is the fourier transform of x(t-to).

-jωt 0
x(t) F.T X(jω) , x(t-to) F.T e X(j )
Proof

FT  x(t-t o )  
-jω t
 x(t-t
-
o )e dt

let  =t-t o ,d =dt , t=  t o


-jω(  t o ) -jω t 0
  x( ) e
-
d = e X( )

2-b Time shifting-DTFT

-jωno
X(ejω) is the fourier transform of x[n] , e X(e j ) is the fourier transform of x[n-no]

Page | 15
 -jωno
x[n] F.T X(ejω) , x[n-no) F.T e X(e j )
Proof

FT  x[n-n o ]    x[n-n o ]e  j n
n 

let p =n-n o ,n=p+n o


  x[p]e
p 
 j ( p  no )
=e  j no X(e j )

3 a)Frequency shifting-CTFT
jωot
X(jω) is the fourier transform of x(t) , Fourier transform of e x(t ) is X(j(ω-ωo))

jωot
x(t) F.T X(jω) , e x(t ) F.T X(j(ω-ωo))

Proof
 
-j(ωωo )t
FT e o x(t )    x(t )e o e t dt   x(t )e
jω t jω t -jω
dt =X(j(  ωo ))
 
- -
3-b Frequency shifting-DTFT
j( -ωo ) 
X(ejω) is the fourier transform of x[n] , e o x[n] fourier transform is X  e
jω n

 
j( -ωo ) 
X  e
jω n
x[n] F.T X(ejω) , e o x[n] F.T 
 
Proof
 
-j(ωωo )n
FT e o x[n]    x[n]e o e   x[n]e
jω n jω n -jωn
 X(e j( o ) )
  n  n 

4. a) Time Reversal CTFT

X(jω) is the fourier transform of x(t) , X(-jω) is the fourier transform of x(-t).

x(t) F.T X(jω) , x(-t) F.T X(-jω)

Proof

FT  x(-t)  
-jω t
 x(-t) e
-
dt

let  =-t ,-d =dt ,

 
-jω(- ) -j(-ω)
 -  x( ) e d =  x( ) e d  X (  j )
 

Page | 16
4. b) Time Reversal DTFT

X(ejω) is the fourier transform of x[n] , X(e-jω) is the fourier transform of x[-n].

x(t) F.T X(ejω) , x(-t) F.T X(e-jω)

Proof


FT  x[ n]   x[n]e
-jωn

n 

let p =-n ,n=-p

 -j 
 
j p -j(- )p
 
p 
x[ p ]e   x[ p]e
p 
=X e

5. Time scaling-CTFT

Let x(t) signal with fourier transform of X(jω), x(at) signal with fourier transform of 1/|a|
X(ω/a)
x(t) F.T X(jω), x(at) F.T 1/|a| X(ω/a)
proof:

FT  x(at)  
-jω t
 x(at) e
-
dt


let  =at ,d =adt , t=
a

1 -jωa 1 
a -
 x(  ) e d = X( )
a a
6.a)Convolution property-CTFT
Convolution of the two signals in the time domain is equal to the multiplication of their
individual fourier transform in the frequency domain.
Proof
Let x(t) and y(t) are signals with fourier transform of X(jω) and Y(jω)

x(t) F.T X(jω) y(t) F.T Y(jω)

F.T( x(t)*y(t)) = X(jω) Y(jω)

Page | 17
 
FT  x(t) * y(t)     x(t) * y(t)  e
-jω t
dt
-
   
-jω t -jω t
 
- -
 x(τ)y(t-τ) e dtdτ=  x( ) d   y(t- ) e
- -
dt

let p=t-τ ,dp=dt , t=τ+p

 
-jω( τ+p)
=  x( ) d   y(p) e
- -
dp

 
=  x ( ) e  d +  y (p) e
-jω -jωp
dp =X(j )Y(j )
- -
6b)Convolution property-DTFT
Convolution of the two signals in the time domain is equal to the multiplication of their
individual fourier transform in the frequency domain.
Proof
Let x[n] and y[n] are signals with fourier transform X(ejω) and Y(ejω)

x(t) F.T X(ejω) y(t) F.T Y(ejω)

F.T( x[n]*y[n]) = X(ejω) Y(ejω)


FT  x[n]* y[n]     x[n]* y[ n]e
-jωn
n 

wkt x[ n]* y[ n]   x[k ] y[n  k ]
k 

Substitute in the above Equation

 

  x[k ] y[n  k ]e
-jωn

n  k 

let p =n-k ,n=p+k

 j  Y  e j 
   

  
-jω(p+k) -jωk -jωp
 x[k ] y[ p ]e = x[k ]e y[ p ]e X e
k  p  k  p 

7.a)Multiplication property CTFT

Multiplications of the two signals in the time domain is equal to the convolution of
their individual fourier transform in the frequency domain.

x(t) F.T X(jω) y(t) F.T Y(jω)

F.T( x(t)y(t)) = 1/2π(X(jω) * Y(jω) )

Page | 18
Proof

1 1 1 jω
IFT(
2
(X(jω)*Y(jω)) 
2 
-
2
(X(jω)*Y(jω)e t d

2  
 1  jω t
= 
 2 

-
e d 
-
X(ju)Y(jω-ju)du

2  
 1  jω t
=   X(ju) du  Y(jω-ju)e d
 2  - -

let jp=jω-ju ,dp=d , jp+ju=jω

2  
 1 
 X(ju) du  Y(jp)e dp
(jp+ju) t
= 
 2  - -
 
1 ju 1 jp
= 
2 -
X(ju)e t du 
2 -
Y(jp)e t dp =x(t)y(t)

7.b)Multiplication property DTFT

Mulitplication of the two signals in the time domain is equal to the convolution of
their individual fourier transform in the frequency domain.

x[n] F.T X (ejω) y[n] F.T Y (ejω)

F.T(x[n]y[n]) = 1/2π(X (ejω) * Y (ejω))

Proof:

IFT  X(e j ) * Y (e j )   x[n] y[n]


wkt,IFT  X(e )  =x[n] 
1 jωn
 X(e
j j
)e d
2 

 1  1 1 jωn
 X(e j ) * Y (e j )     2  X(e ) * Y (e j ) e
j
IFT  d
 2  2 


 X(jk )Y (j  jk )dk

j j
wkt X(e ) * Y (e ) 


Substitute in the above Equation

 
1 1 jωn

2 

2  X(jk )Y (j  jk )dke

d

let p = -k , =p+k,d  dp
 
1 jkn 1 jpn

2  X(jk )e

dk
2  Y (jp)e

dp  x[n] y[n]

Page | 19
 8. a) Differentiation and Integration in CTFT

Time domain
x(t) is a signal with inverse fourier transform

1 jω t
x(t) 
2 
-
X(  )e d  differentiate with respect to t


dx(t) 1 jω t
dt

2 
-
j  X(  )e d

dx(t) FT
 j  X(  )
dt
Freq domain

-jω t
X(  )   x (t) e
-
dt diff erentiate with respect to 

dX(  ) IFT FT dX(  )

  jtx(t) or tx(t)  j
d d
Integration
t
1
 x( )d 
-
j
X(  ) +  X(0)  (  )
=

8b) Difference in DTFT

X(ejω) is the fourier transform of x[n] , x[n]-x[n-1] fourier transform is 1  e j  X e  j 
Proof:

  
FT  x[n]-x[n-1]     x[n]-x[n-1]e  j n
  x[n]e  j n
   x[n-1]e  j n
n  n  n 

 X(e )  e X(e )= 1  e  X(e )

j j j j j

8c Differentiation in frequency in DTFT

Fourier Transform of nx[n] is j

dX e
j
 
d

X(e j )   x[n]e
n 
 j n

differentiation wrt to 
dX(e j ) 
=   -jnx[n]e  j n
d n 
FT dX(e j )
 nx[n]  j
d

Page | 20
 9. Conjugation and conjugate symmetry

Let x(t) signal with fourier transform of X(jω), fourier transform of x*(t) is X*(-ω)

x(t) F.T X(jω), x*(t) F.T X*(-ω)


X( )   x (t) e-jωt dt
-

X*( )   x* (t) e jωt dt ,  is replaced by -
-

X*(- )   x* (t) e-jωt dt i.e F.T(x*(t)) = X*(- )
-
case 1
If x(t) is real, x(t) =x*(t), so X*(- ) = X( ) or X(- ) = X*( ),so conjugation property
shows that if x(t) is real then X(j ) has conjugate symmetry
X(j ) in the rectangular form,X(j ) = real X(j )+j imagX(j ),
X*(-j ) in the rectangular form,X*(-j ) = real X(-j )-j imagX(-j )
so by conjugate symmetry Re(X(j ))=Re(X(-j )) and Im(X(j ))=-Im(X(-j ))
i.e real part of F.T is an even function and imaginary part is an odd function.
X(j ) in the polar form,X(j ) = X(j ) X(j )
X*(-j ) in the polar form,X*(-j ) = X(-j )   X(-j )
X(j ) X(j )= X(-j )   X(-j )
X(j ) = X(-j ) and X(j )=-X(-j )
case2
If x(t) is real and even then X(j ) is also real and even , the condition is x(t) =x*(t) = x(-t)
 
X(- )   x (t) e jωt dt   x (-t) e jωt dt
- -
replace -t by 

  x (τ) e-jτωdτ=X(ω)
-
case 3
If x(t) is real and odd then X(j ) is also real and odd ,thee condition is x(t) =x*(t) = -x(-t)
 
X(- )   x (t) e jωt dt    x (-t) e jωt dt
- -
replace -t by 

   x (τ) e-jτωdτ=-X(ω)
-
10. Parsevals relationship in CTFT

Parsevals relationship states that the energy or power in the time domain of a given
signal is equal to the energy or power in the frequency domain.

Page | 21
 Proof
 2 2
The energy of the signal E   x (t) dt , x (t) = x(t)x*(t)
-
1  -jω t
x*(t)   X*( )e d
2 -
  1

E   x(t)x*(t)dt =  x(t)  X*( )e-jωt d dt
- - 2 -
1 
 
-jω t 1
 
2 -
x(t)e dt 
-
X*( )d =
2 
-
X( )X*( )d


1
 X( ) d
2
E=
2 -
The energy in the time domain is equal to the energy in the frequency domain
normalized by 2π. The quantity X( ) is the energy spectrum of the signal. This
2

Equation is also referred as Rayleigh’s Energy theorem.

10. b) Parsevals relationship in DTFT

Parsevals relationship states that the energy or power in the time domain of a given
discrete time signal is equal to the energy or power in the frequency domain.
Proof

2 2
The energy of the signal E   x [n] , x [n] = x[n]x*[n]
n 

1 
j jωn
x[n]   X(e )e d
2 -
Take Conjugate
1  j -jωn
x *[n]   X (e )e
*
d
2 -
   
1
  x[n]x [n]   x[n] 2 -
-jωn
E x[n]  X* (e j )e d
2 *

n  n  n 

Change the order of summation and integration


1  1  1 

j -jωn j j j 2
=  X *
(e )d  x[ n ]e   X *
(e )X(e )d  =  X(e ) d
2 - n  2 - 2 -
1  j 2
E=  X(e ) d
2 -

11. Duality Property

The fourier transform and inverse fourier transform is quite symmetry. This results in
a duality between time domain and frequency domain. Let x(t) signal with fourier
transform of X(jω).
FT(X(t))= 2πx(-ω).

Page | 22
 x(t) F.T X(jω), X(t) F.T 2πx(-ω)
Proof:
1  jω t
we know that, IFT, x(t)   X( )e d
2 -
 jω
2 x(t)   X( )e t d
-
Replace t by -t, to bring the equation into FT equation
 -jω
2 x(-t)   X( )e t d ,
-
2 x(-t)  X( ), replace t by and viceversa
2 x(- )  X(t )
Note: Example for duality property is FT of square signal gives sinc function, and IFT of
square function gives sinc signal.

The above figure shows the symmetry between two examples and this can be extend to
other fourier transforms. Because of symmetry there is a dual pair with time and
frequency variables interchanged. The above is the best example of duality.

Page | 23
Find the fourier transform using duality property x(t)=1/(2-jt)
1
Given x(t)=
2  jt
1 1
let X(j )   This transform is for the signal x(t)=e 2t u (t ) 
2  j 2  j
1
X ( jt ) 
2  jt
By duality property X(jt )  2 x(- ) =2 e 2 u ( )
Find the fourier transform using duality property x(t)=1/(1+jt)
1
Given x(t)=
1  jt
1 1
let X(j )   This transform is for the signal x(t)=e  t u (t ) 
1  j 1  j
1
X ( jt ) 
1  jt
By duality property X(jt )  2 x(- ) =2 e u (  )
Find the fourier transform using duality property x(t)=1/(9+t2)
1
Given x(t)=
9  t2
1 a t 2a
let X(j )   This transform is for the signal x(t)=e  2
9 2
a  2
3 t 6 1
x(t )  e  , but given expression is ,so multiply and divide by 6
9 2
9  2
1 3 t
x(t )  e
6
1
X ( jt ) 
9  t2
2 3   3 
By duality property X(jt )  2 x(- ) = e  e
6 3
Find the fourier transform using duality property x(t)=1/(1+t2)
1 1 a t 2a
Given x(t)= let X(j )   This transform is for the signal x(t)=e  2
1 t 2
1  2
a  2
t 2 1
x(t )  e  , but given expression is ,so multiply and divide by 2
1  2
1 2
1 t 1
x(t )  e and X ( jt ) 
2 1 t2
2   
By duality property X(jt )  2 x(- ) = e  e
2

Page | 24
Consider a causal LTI system with frequency response H(jω)=1/3+jω. For a given
input the system is produce an output y(t)=e-3tu(t) - e-4tu(t). Find the input x(t)
By convolution property,
Y(jω) = X(jω) H(jω), From this X(jω)= Y(jω)/ H(jω), Take IFT to find x(t)
1
Given y  t   e 3t u  t   e 4t u  t  , H (j )=
3  j
1 1 1
Take FT Y(j )=  
3  j 4  j (3  j )(4  j )
Y(j ) 1 1 1
X(j )   / 
H(j ) (3  j )(4  j ) 3  j 4  j
1
IFT of is e 4t u  t  ,
4  j
The input signal x(t)= e 4t u  t 
Find the IFT
 j
X(j )=
( j )  3 j  2
2

By using partial fraction we can find the inverse FT, let j =x,
x x A B
= 2   
x  3 x  2 ( x  2)( x  1) ( x  2) ( x  1)
A  2, B  1
 j 2 1
so X(j )=  
( j )  3 j  2 ( j  2) ( j  1)
2

 FT 1 
IFT = x(t) = -2e 2t u (t )  e  t u (t )  e  at u (t ) 
 ( j  a ) 
Sampling and aliasing
Theorem
‘ A band limited signal of finite energy which has no frequency components higher
than the ‘Fh’ Hz is completely described by the specifying the values of signal at instant of
time separated by 1/ 2Fh Sec’ (or)
‘ A band limited signal of finite energy which has no frequency components higher
than ‘Fh’ Hz is completely recovered from the knowledge of the samples taken at the rate
of 2 Fh sec’ i.e Fs≥2Fh. Where Fh is the highest frequency of the signal.
‘The sampling rate of 2 Fh samples per sec for a bandwidth of Fh is called as
Nyguist rate’.
The process of converting analog signal into digital signal is called sampling
process. If we choose the sampling rate properly the sequence of sample obtained by
sampling defines the original analog signal. This is the essence of sampling theorem.
Let x(t) is the continuous time to be sampled. δ(t) is the periodic impulse train
which has a sampling period Ts and sampling rate Fs = 1/ Ts.
As a result of this sampling process we get an infinite sequence spaced Ts sec and
denoted by xδ(t).
xδ(t)
Page | 25
 x(t) δ(t)
0
× =

Ts t Ts t
t
xδ(t) is the signal obtained by multiplying the sequence of x(nTs) by corresponding
sequence of delta function spaced ‘Ts’ sec aparts and summing the individual.

x (t ) =  x(nT ) (t  nT )
n=-
s s

where xδ(t) is the ideally sampled signal, the xδ(t) can also expressed as original signal and
an Ideal sampling function δ(t) i.e
x (t ) = x(t)   (t )
The fourier transform of the sampled signal xδ(t) obtained by convolving the Fourier
transform of x(t) with fourier transform of δ(t).
1  n 
FT( (t))= 
Ts n=-
 ( f 
Ts
)  F s   ( f  nFs )
n=-

FT of x(t) = X(f)

X (f)=X(f)*Fs   ( f  nF )
n=-
s

Interchanging the order of summation

 
X (f)=Fs  X(f)* ( f  nFs )  Fs
n=-
 X(f-nF )
n=-
s

The spectrum Xδ(f) is a continuous spectrum which is periodic with period 1/Ts. It is called
periodic frequency. So the process of uniformly sampling the signal in the time domain
results periodic spectrum in frequency domain with the period equal to sampling
rate.
Xδ(f) Xδ(f)
X(f)

- Fh Fh f - Fs - Fh Fh Fs - Fs
f - Fh Fh Fs

Fs = 2Fh
Fs >2Fh

Page | 26
 If the signal is strictly bandlimited which has no frequency components higher than
the Fh Hz then the Fourier transform of x(t) i.e X(f) has the property that X(f) is zero for all
|f|>Fh

1 
X(f)=Fs  x(nTs)e  j 2 fnTs   x(n/2Fh )e  j 2 fn / 2 Fh
n=- 2 Fh n=-
1 
X(f)= 
2 Fh n=-
x(n/2Fh )e  j fn / Fh

The sequence x(n/2Fh) contains all the information of x(t).

Reconstruction using Interpolation Technique:

Reconstruction of signal x(t) from the sequence of sample values x(n/2Fh)


1
X(f)=
2 Fh
 x(n/2F )e
n=-
h
 j fn / Fh
-Fh <f<Fh

Take inverse fourier transform,

 Fh 
1
x(t)=  X(f)e j 2 ft df= 
2 Fh n=-
 x(n/2F )e h
 j fn / Fh
e j 2 ft df
- -Fh

Interchanging the order of integration and summation

 Fh
1  j fn / Fh j 2 ft
=  x(n/2Fh )  e e df
n=- -Fh
2 Fh

 Fh
1
=  x(n/2Fh )  e j 2 f (t  n / 2 Fh ) df
n=- -Fh
2 Fh

=  x(n/2F )sinc(2F t-n)
n=-
h h - <t< 

This equation termed as interpolation equation used for reconstruction of signal x(t)
from its sample. Here each sample is multiplied by delayed version of interpolation
function and adding the resulting waveforms are to obtain x(t).

Aliasing and its drawbacks and avoiding of aliasing

When the sampling rate ‘Fs’ exceeds the nyquist rate ‘Fs>2Fh’ all replica of X(f)
involved in the construction of Xδ(f) move further apart and there is no problem in
recovering the original signal x(t) from its sampled version x δ(t). This is known as over
sampling.
When the sampling rate ‘Fs’ is less than the nyquist rate ‘Fs<2Fh’ the sampled
signal the spectrum X(f) overlap each other. This is due to under sampling. In this case, the
frequencies X(f) reflected into the lower frequencies in Xδ(f). The phenomenon of a High
frequency component in the spectrum of original signal x(t) to the lower frequency in the
spectrum of its sampled version xδ(t) is called aliasing effect or aliasing error.

Page | 27
Drawback of Aliasing

The original signal x(t) can’t be recovered exactly from the sampled version xδ(t) and
the information lost in sampling process itself.

To avoid Aliasing

Prior to sampling a Lowpass Pre alias filter is used to attenuate the high frequency
components the signal that lie outside the band. Once the filter it, the filtere signal is
sampled at a rate slightly higher than the nyquist rate.
Find the Nyquist rate of the following signal
i)x(t)=1+cos(200πt)+sin(400πt) ii) x(t) = sin2(200πt)

Solution:
Nyquist theorem says that the sampling frequency should be greater than twice of the max
frequency. i.e. Fs≥2Fmax.
i) two frequencies ω1=200π, 2πF1=200π, F1=100, ω2=400π, 2πF2=400π, F2=200
Maximum frequency is 200 so Minimum Nyquist rate is 2xFmax=400 Hz, Fs≥400
ii) Given signal sin2(ωt)=(1-cos2 ωt)/2
Given signal x(t) = sin2(200πt)=(1- cos400π t)/2
ω=400π, 2πF=400π, F=200, Minimum Nyquist Rate is 2xF=400 Hz, Fs≥400
--------------------------------------------------------------------------------------------------------------

Page | 28