Download as pdf or txt
Download as pdf or txt
You are on page 1of 67

Chapter 5

The Discrete-Time Fourier Transform

Prof. Chih-Cheng Tseng

tsengcc@niu.edu.tw

Reproduced from the PowerPoint of ‘OPPENHEIM: Signals and Systems 2/E (PNIE)’ with the permission of Pearson

Wireless Communication Network Lab. Chih-Cheng Tseng 1


Fourier Family

Periodic Fourier Series (Chap. 3)


Continuous
Non-periodic Fourier Transform (Chap. 4)
Fourier
Discrete-Time Fourier Series or
Family
Periodic Discrete Fourier Transform (DFT)
Discrete (Sec. 3.6 & Sec. 3.7)
Discrete-Time Fourier
Non-periodic Transform (DTFT) (Chap. 5)

Wireless Communication Network Lab. Chih-Cheng Tseng 2


Outline
 Sec. 5.0 Introduction
 Sec. 5.1 Representation of Aperiodic Signals: The DTFT Basic
 Sec. 5.2 The Fourier Transform for Periodic Signals
 Sec. 5.3 Properties of the DTFT
 Sec. 5.4 The Convolution Property
 Sec. 5.5 The Multiplication Property Properties
 Sec. 5.6 Tables of Fourier Properties and of Basic FT Pairs
 Sec. 5.7 Duality
 Sec. 5.8 Systems Characterized by Linear Constant-Coefficient
Difference Equations
 Sec. 5.9 Summary

Wireless Communication Network Lab. Chih-Cheng Tseng 3


Introduction
 Chapter 5 is the discrete counterpart of Chapter 4. The two chapters have
similar structure.

x[n] 
Sec. 3.6 ak 

Chap. 3 a  1 x(t )e  jk0t dt
x(t )  ae T0 
jk 0 t
k
 jk (2 / N ) n
 x[n]e jk (2 / N ) n
k
k  T0
k  N 
ak e 1
N k  N 

continuous time discrete frequency discrete time discrete frequency Discrete Fourier transform (DFT)
periodic in time, period=T0 aperiodic in frequency periodic in time, period=N periodic in frequency, period=N

x(t ) Chap. 4 X ( j ) x[ n]  Chap. 5 X (e j ) 



1


X ( j )e jt d 
2 2 
j j n
= 21

  x(t )e  jt dt X (e )e d  , 
x[n]e  jn
n 
continuous time
continuous time continuous frequency discrete time continuous frequency
aperiodic in
aperiodic in time
time aperiodic in frequency aperiodic in time periodic in frequency, period=2

Wireless Communication Network Lab. Chih-Cheng Tseng 4


Sec. 5.1 The Discrete-Time Fourier Transform
 Key Concepts
• learning the definition of the discrete-time Fourier transform;
• the conditions where the discrete-time Fourier transform exists (absolutely summable
or finite energy);
• the output of the discrete-time Fourier transform has the period of 2.

Wireless Communication Network Lab. Chih-Cheng Tseng 5


5.1.1 Development of the Discrete-Time Fourier Transform
(1/2)

Fourier series for x[n]


x[n]  ae
k N
k
jk (2 / N ) n

1
ak 
N
 x[n]e
k N
 jk (2 / N ) n

N2
1

N

n  N1
x[n]e  jk (2 / N ) n


1

N
 x[n]e
n  
 jk (2 / N ) n

Wireless Communication Network Lab. Chih-Cheng Tseng 6


5.1.1 Development of the Discrete-Time Fourier Transform
(2/2)
periodic in frequency
 with period 2
If one defines X (e )  j

n 
x[ n]e  jn d


1 1
ak 
N
 x[n]e
n 
 jk (2 / N ) n
 X (e jk0 )
N
where ω0=2π/N. Moreover, 

1 1
x[n]  
k N
ak e jk (2 / N ) n
 
k N N
jk0
X (e )e jk0 n

2

k N
X (e jk0 )e jk0n0
periodic in frequency with period 2

When N→∞, ω0=2π/N→0. Set =kω0, d=ω0 and   


k  N 
2

1
x[n] 
2 
2
X (e j )e jn d

Wireless Communication Network Lab. Chih-Cheng Tseng 7


Definition 5.1 Discrete-Time Fourier Transform (1/3)
 Synthesis equation:
1
x[ n] 
2 2
X (e j )e jn d

 Analysis equation:

X (e j )  
n 
x[ n]e  jn

where  means the integration from 0 to 0+2 and 0 is any real number.
2

 Specially, if 0=0, synthesis equation:


1 2

j j n
x[ n]  X ( e ) e d
2 0

Wireless Communication Network Lab. Chih-Cheng Tseng 8


Definition 5.1 Discrete-Time Fourier Transform (2/3)
 Note:
• The discrete-time Fourier transform can also be viewed as a special case of the
discrete-time Fourier series where N→∞.
• The output of the discrete-time Fourier transform, i.e., X(ej), always has the period of
2 .

 
X ( e j (   2 ) )  
n 
x[n]e j (  2 ) n  
n 
x[n]e  jn e  j 2 n

 
n 
x[ n]e  jn  X (e j )

Wireless Communication Network Lab. Chih-Cheng Tseng 9


Definition 5.1 Discrete-Time Fourier Transform (3/3)
 The output of the DT Fourier Transform always has the period of 2.

Wireless Communication Network Lab. Chih-Cheng Tseng 10


Example 5.1
x[ n]  a nu[ n], a 1
  
1
X (e )  j
 x[n]e
n 
 j n
  a u[n]e
n 
n  j n
  (ae
n 0
 j n
) 
1  ae  j
.

0< a <1 1< a < 0

Wireless Communication Network Lab. Chih-Cheng Tseng 11


Example 5.2 (1/2)
n
x[ n]  a , a 1
   1
X (e j )  
n 
x[ n]e  jn   a e
n 
 j nn
  e 
a n  j n

n 0
 e
a  n  j n

n 
  j
1 ae
  (ae  j ) n   (ae j )m   j

n0 m 1 1  ae 1  ae j
(1  ae j )  ae j (1  ae j ) 1  ae j +ae j  ae j ae  j
  j j

(1  ae )(1  ae ) 1  ae j  ae  j  ae j ae j
1  a2 1  a2 1  a2
 j  j
 
1  a (e  e )  a 1  2a( 2 )  a 1  2a cos   a 2
2 e j  e j 2

Wireless Communication Network Lab. Chih-Cheng Tseng 12


Example 5.2 (2/2)

Wireless Communication Network Lab. Chih-Cheng Tseng 13


Example 5.3 (1/2)
1, n  N1
x[ n]  
0, n  N1
j N1  j ( N1 1)

e  e N1
X (e j )   x[n]e  jn =  (e  j ) n   j
n  n  N1 1  e
j ( N1 1/2 )
 j /2 j ( N1 1/2 )  j ( N1 1/2 ) e  j ( N1 1/2 )
e (e e ) 2 j( e )
  2j

e  j /2 (e j /2  e  j /2 )
j /2
e  j /2
2 j( e 2j )
sin   N1  12   sin   2 N1  1 2 
 
sin  / 2  sin  / 2 
Wireless Communication Network Lab. Chih-Cheng Tseng 14
Example 5.3 (2/2)

2N1+1
for N1=2

In Examples 5.1, 5.2, and 5.3, all X(ej) have the period of 2.

Wireless Communication Network Lab. Chih-Cheng Tseng 15


5.1.3 Convergence Issues Associated with the Discrete-Time
Fourier Transform
 Convergence
• absolutely summable



n 
x[ n]  

• finite energy



2
x[ n]  
n 

Wireless Communication Network Lab. Chih-Cheng Tseng 16


Sec. 5.2 The Fourier Transform for Discrete-Time Periodic
Signals
 Key Concepts
• the discrete-time Fourier transform pairs of Eq. (1)
• the discrete-time Fourier transform pairs of Eq. (2)
• learning the method to calculate the discrete-time Fourier transform for discrete-time
periodic signals

Wireless Communication Network Lab. Chih-Cheng Tseng 17


Fourier Transform of x[ n]  e j0n

If X (e j )   2 (  
l 
0  2 l )

1 2 1 2 
x[ n] 
2  0
X ( e )e j j n
d 
2 
0
 2
l 
(   0  2 l )e j n
d

1 2

2  0
2 (  0 )e jn d  e j0n discrete-time periodic signal


e j0n 
F
  2 (  0  2 l ) (1)
l 
Wireless Communication Network Lab. Chih-Cheng Tseng 18
Fourier Transform of x[ n]  
k N
ak e jk (2 / N ) n

If x[ n]  
k N
ak e jk (2 / N ) n , then
 
  jk (2 / N ) n  j n 
X (e j )  
n 
x[ n]e  jn  
n  k  N
ak e jk (2 / N ) n e jn   ak   e
k  N   n
e 

   2 k     2 k 
  ak   2     2 l      2 ak     2 l  (2)
k  N   l    N   l  k  N   N 

 2 k 
  2 ak    
k    N   2
N
 4 2  4
N N N

Wireless Communication Network Lab. Chih-Cheng Tseng 19


Procedures to Determine the DT FT for A Periodic DT
Signal x[n]=x[n+N]
Use the discrete-time Fourier series (Sec. 3.6) to express x[n] by
x[ n]  
n N
ak e jk (2 / N ) n

 2 k 

1
Then, X (e )   2 ak   
j
 where ak  N  x[ n]e  jk (2 / N ) n
k   N  n N

Note: The FT of a periodic CT signal



X ( j )   2 a  (  k )
k 
k 0

Wireless Communication Network Lab. Chih-Cheng Tseng 20


Example 5.5
 Consider the periodic signal
1 j0n 1  j0n 2
x[ n]  cos 0 n  e  e , with 0 
2 2 5
x[ n]  x[n  5] a1 = a1 = 1/2
X (e j )     2
5
      , 2
5
    
X (e j )  X (e j (  2 ) ) 要記得寫

Wireless Communication Network Lab. Chih-Cheng Tseng 21


Example 5.6 (1/2)
discrete-time periodic impulse train

 Consider the periodic signal x[ n]    [n  kN ] and x[n]=x[n+N]
k 

1 N 1 1 N 1 
ak   x[ n]e  jk (2 / N ) n     [n  kN ]e  jk (2 / N ) n
N n 0 N n0 k 
1 N 1 1
   [n]e  jk (2 / N ) n 
N n 0 N

2 
 2 k 
X (e ) 
N
j

k 
  
 N


Wireless Communication Network Lab. Chih-Cheng Tseng 22
Example 5.6 (2/2)
x[ n]   k   [ n  kN ]


discrete-time
Fourier transform

If N=1,
 
x[ n]    [
k 
n  k ]  1 
F
 X ( e j
)  2     2 k 
k 

If N,
 
x[ n]    [n  kN ]   [n]  X (e ) 
k 
F j
  [
n 
n ]e  j n
1

Wireless Communication Network Lab. Chih-Cheng Tseng 23


Sec. 5.3 Properties of the Discrete-Time Fourier Transform
 Key Concepts
• Learning the properties of the discrete-time Fourier transform, including
(i) periodicity; (ii) linearity;
(iii) time shifting; (iv) frequency shifting;
(v) conjugation; (vi) conjugate symmetry for real input;
(vii) even-real input to even-real output, even-odd input to imaginary-odd output;
(viii) differencing; (ix) accumulation;
(x) time reversal; (xi) time expansion;
(xii) differentiation in frequency;
(xiii) Parseval’s relation

Wireless Communication Network Lab. Chih-Cheng Tseng 24


5.3.1 Periodicity of the Discrete-Time Fourier Transform

X (e j (  2 ) )  X (e j )

Wireless Communication Network Lab. Chih-Cheng Tseng 25


5.3.2 Linearity of the Discrete-Time Fourier Transform

If
F
x1[ n]   X 1 (e j ) and x2 [n] 
F
 X 2 (e j )

then
F
ax1[n]  bx2 [n]   aX 1 (e j )  bX 2 (e j )

Wireless Communication Network Lab. Chih-Cheng Tseng 26


5.3.3 Time Shifting and Frequency Shifting of the Discrete-
Time Fourier Transform
If
x[ n] 
F
 X (e j )

then
F
x[ n  n0 ]   e jn0 X (e j ) (time-shifting property)

and
j0 n j ( 0 ) (frequency shifting property or
e x[ n] 
 X (e
F
) modulation property)

Wireless Communication Network Lab. Chih-Cheng Tseng 27


Example 5.7

H hp (e j )  H lp (e j (  ) )

ideal highpass filter ideal lowpass filter

hhp [n]  e j n hlp [n]  ( 1) n hlp [n]

Wireless Communication Network Lab. Chih-Cheng Tseng 28


Simple Lowpass, Highpass, and Bandpass Filter Conversion
Methods

lowpass filter multiplied by (1)n highpass filter

highpass filter multiplied by (1)n lowpass filter

lowpass filter multiplied by jn bandpass filter


highpass filter multiplied by jn bandpass filter

Wireless Communication Network Lab. Chih-Cheng Tseng 29


5.3.4 Conjugation and Conjugate Symmetry

If x[ n] 
F
 X (e j ) , then x*[ n] 
F
 X * (e  j )
Moreover, if x[n] is real, X (e j )  X * (e  j )

Ev{x[ n]}
F
 Re{ X (e j )} Od {x[ n]}
F
 jIm{ X (e j )}

where

Ev{x[ n]}   x[n]  x[n] / 2 Od {x[n]}   x[n]  x[ n] / 2

Wireless Communication Network Lab. Chih-Cheng Tseng 30


5.3.5 Differencing and Accumulation

If x[ n] 
F
 X (e j ), then

x[ n]  x[n  1] 
F
(1  e j ) X (e j )
n
If y[n]   x[m] , then y[n]y[n1]=x[n]
m 
n 
1

m 
x[ m] 
 F

1 e  j
X (e )   X (e )   (  2 k )
j j0

k 

DC or average value that


t 1
Hint:  x( )d 
F
 X ( j )   X (0) ( ) results from summation
j

Wireless Communication Network Lab. Chih-Cheng Tseng 31


Example 5.8

g[ n]   [n] 
F
 G (e j )  1
If x[n] is a unit step function
n
x[ n]   g[m]  u[n]
m 
then

1
X (e j )   j
G ( e j
)   G ( e j0
)   (  2 k )
(1  e ) k 

1
  j
    (  2 k )
1 e k 

Wireless Communication Network Lab. Chih-Cheng Tseng 32


5.3.6 Time Reversal
x[ n] 
F
 X (e  j )
since
 


n 
x[n]e  jn  
m
x[m]e  j (  ) m  X (e  j )

If x[n] is an even function, X(ej) is even also.


F
x[ n]  x[ n]   X (e j )  X (e  j )

If x[n] is an odd function, X(ej) is odd also.


F
x[ n]   x[n]   X (e j )   X (e  j )
Wireless Communication Network Lab. Chih-Cheng Tseng 33
Summary
 If x[n] is real and even, then X(ejω) is also real and even.
 If x[n] is real and odd, then X(ejω) is purely imaginary and odd.
 If x[n] is purely imaginary and even, then X(ejω) is also purely imaginary and
even.
 If x[n] is purely imaginary and odd, then X(ejω) is real and odd.

Wireless Communication Network Lab. Chih-Cheng Tseng 34


5.3.7 Time Expansion (1/2)
 x[n / k ], if n is a multiple of k
If x( k ) [n]  
0, if n is not a multiple of k
“k”在時間軸上“放” “k”在頻率軸上“縮”
then
x( k ) [n] 
F
 X (e jk )

Proof: let n=rk


 
X ( k ) (e j )  
n 
x( k ) [n]e jn  
r 
x( k ) [rk ]e jrk
 
 
r 
x[rk / k ]e  j ( k ) r  
r 
x[ r ]e  j ( k ) r  X (e jk )

Wireless Communication Network Lab. Chih-Cheng Tseng 35


5.3.7 Time Expansion (2/2)

expansion in time
compression in frequency

Wireless Communication Network Lab. Chih-Cheng Tseng 36


Example 5.9 (1/2)

x[ n]  y(2) [ n]  2 y(2) [ n  1]

 y[n / 2], if n is even


y(2) [n]  
0, if n is odd

 y[(n  1) / 2], if n is odd


y(2) [n  1]  
0, if n is even

Wireless Communication Network Lab. Chih-Cheng Tseng 37


Example 5.9 (2/2)
From Example 5.3 with N1=2 and right time shifted by 2, we have
sin(5 / 2)
Y (e j )  e  j 2
sin( / 2)
 
 j 4 sin(5 )
j
sin  2 N1 1
2
F X (e ) 
y(2) [n]  e sin  / 2 
sin( )
F  j 5 sin(5 )
2 y(2) [n  1]   2e
sin( )
j  j 4  j  sin(5 ) 
X (e )  e (1  2e ) 
 sin( ) 
Wireless Communication Network Lab. Chih-Cheng Tseng 38
5.3.8 Differentiation in Frequency
j
dX ( e )
nx[n] 
F
j
d
Proof:

j
Since X (e )  
n 
x[ n]e  jn
dX (e j ) 
   jnx[n]e  jn
d n 

dX (e j ) 
 j   nx[ n]e  jn
d n 

Wireless Communication Network Lab. Chih-Cheng Tseng 39


5.3.9 Parseval’s Relation

1

2
  X (e
2 j
x[ n]  ) d
n  2 2

Parseval’s relation is also call the energy preservation property.

General form of Parseval’s relation:


1

n 

x[ n] y [n] 
2 
2
X (e j )Y  (e j )d

Wireless Communication Network Lab. Chih-Cheng Tseng 40


Example 5.10
X(ej) is the Fourier transform of x[n].
Then,
1. If x[n] is period, X(ej) is a periodic pulse
train. Hence, x[n] is not period.
2. If x[n] is real, X(ej)=X*(ej). Hence, |X(ej)|
is even and X(ej) is odd. See Slide #11.
3. If x[n] is real and even, |X(ej)| is even and
X(ej) is even. Hence, x[n] is not even.
4. The energy is

1 17

2

2 j
x[ n]  X (e ) d  
n  2 2 4
Wireless Communication Network Lab. Chih-Cheng Tseng 41
Sec. 5.4 The Convolution Property
 Convolution Property

If y[n]  x[n]* h[n]

then Y (e j )  X (e j ) H (e j )

Wireless Communication Network Lab. Chih-Cheng Tseng 42


Example 5.11

If h[ n]   [ n  n0 ] and y[n]  x[n]* h[n]

then 
j
H (e )    [
n 
n  n0 ]e  j n
 e  j n
0

Y (e j )  e  jn0 X (e j )

y[n] = x[n – n0]

Wireless Communication Network Lab. Chih-Cheng Tseng 43


Example 5.12 Ideal Lowpass Filter

1  1 c sin c n c  c n 
  H (e e
j j n j n
h[ n]  )e d  d   sinc  
2  2  c n    

c  c n  F   
sinc     
     2c 
Wireless Communication Network Lab. Chih-Cheng Tseng 44
Example 5.13
If x[ n]   nu[n], h[n]   nu[n], and y[ n]  x[ n]* h[ n]
j 1 j 1
then X (e )   j
, H (e ) 
1 e 1   e  j
j j j 1
Y (e )  H ( e ) X ( e ) 
(1   e  j )(1   e  j )
部分分式法
jA B  
Y (e )   A , B
1e  j
1   e  j    

 1 
y[n] 
 
 u[n] 
 
 u[n] 
n

 
  n1u[n]   n1u[n]
n

Wireless Communication Network Lab. Chih-Cheng Tseng 45


Example 5.14

W1 (e j )  X (e j (  ) ) W2 (e j )  H lp (e j ) X (e j (  ) )

W3 (e j )  W2 (e j (  ) )  H lp (e j (  ) ) X (e j ( 2 ) )  H lp (e j (  ) ) X (e j )

W4 (e j )  H lp (e j ) X (e j )
HPF LPF
Y (e j )  W3 (e j )  W4 (e j )  [ H lp (e j (  ) )  H lp (e j )] X (e j )

Wireless Communication Network Lab. Chih-Cheng Tseng 46


Sec. 5.5 The Multiplication Property (1/2)
 Multiplication in the time domain is equal to convolution in the frequency
domain

If y[n]  x1[n]x2 [n],


1

j j  j (  )
then Y (e )  X ( e ) X ( e )d
2 2  1 2

Wireless Communication Network Lab. Chih-Cheng Tseng 47


Sec. 5.5 The Multiplication Property (2/2)

Proof:  
Y (e j )  
n 
y[n]e jn   x1[ n]x2 [n]e  jn
n 

 1 
    X 1 (e )e d  x2 [n]e  jn
j j n

n   2 
2

j   j (  ) n 

1
  X 1 (e )   x2 [n]e  d
2 2 
 n 
1

2 
2
X 1 (e j ) X 2 (e  j (  ) )d

Wireless Communication Network Lab. Chih-Cheng Tseng 48


Slide #44
sin Wn W  Wn  F   
 sinc  
Example 5.15 (1/2) n    
  
 2W 

x[ n]  x1[ n]x2 [n]


sin( n / 2) 
x1[n]   X 1 (e )   ( ) and X 1 (e j )  X 1 (e j (  2 ) )
j

n 
sin(3 n / 4) j 
x2 [n]   X 2 (e )   ( ) and X 2 (e j )  X 2 (e j (  2 ) )
n 3 / 2 Xˆ (e ) 1
j

1
j
ˆ j  X ( e ) for      
X 1 (e )   1
2  2 
0  2  2 

0 otherwise 4 4 4 4

X 2 (e j )
1

X (e j )  1
2 
X 1 (e j ) X 2 (e  j (  ) ) d
2 7 6 5  3 2 
4 4 4 4 4 4
0 
4
2 3
4 4
 5 6 7 2
4 4 4

 
 1
2  Xˆ 1 (e j ) X 2 (e  j (  ) ) d  1
2  X 2 (e j ) Xˆ 1 (e  j (  ) )d
 
Wireless Communication Network Lab. Chih-Cheng Tseng 49
Example 5.15 (2/2)
X 2 (e j ) X 2 (e j ) X 2 (e j )
1 1 1

2 7 6 5  3 2  0  2 3  5 6 7 2  2 7 6 5  3 2  0  2 3  5 6 7  2  2 7 6 5  3 2  0  2 3  5 6 7  2 
4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

X 2 (e j ) X 2 (e j ) X 2 (e j )
1 1 1

2 7 6 5  3 2  0  2 3  5 6 7 2  2 7 6 5  3 2  0  2 3  5 6 7 2  2 7 6 5  3 2  0  2 3  5 6 7 2 
4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

X 2 (e j ) X 2 (e j ) X 2 (e j )
1 1 1

2 7 6 5  3 2  0  2 3  5 6 7 2  2 7 6 5  3 2  0  2 3  5 6 7 2  2 7 6 5  3 2  0  2 3  5 6 7 2 
4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

Wireless Communication Network Lab. Chih-Cheng Tseng 50


Sec. 5.6 Tables of FT Properties and Basic FT Pairs --- Table
5.1 (1/2)

(followed by the next slide)

Wireless Communication Network Lab. Chih-Cheng Tseng 51


Sec. 5.6 Tables of FT Properties and Basic FT Pairs --- Table
5.1 (2/2)

Wireless Communication Network Lab. Chih-Cheng Tseng 52


Sec. 5.6 Tables of FT Properties and Basic FT Pairs --- Table
5.2 (1/2)


k  N 
ak e jk (2 / N ) n

(followed by the next slide)


Wireless Communication Network Lab. Chih-Cheng Tseng 53
Sec. 5.6 Tables of FT Properties and Basic FT Pairs --- Table
5.2 (2/2)

1, 0    W   
X ( )     
 0, W      2W 

Wireless Communication Network Lab. Chih-Cheng Tseng 54


Sec. 5.7 Duality
 Key Concepts
• The duality of the discrete-time Fourier series;
• The duality between the discrete-time Fourier transform and the continuous-time
Fourier series;
• Studying the relations among the Fourier-related transforms in Table 5.3
• The duality property of the continuous-time Fourier transform was discussed in
Section 4.3.6.

Wireless Communication Network Lab. Chih-Cheng Tseng 55


x[n] 
Sec. 3.6 ak 


x(t )   ak e jk t Chap. 3 ak 


0 1
 x(t )e  jk0t dt
k 
T0 T0
 k  N 
ak e jk (2 / N ) n 1
N  k  N 
x[n]e jk (2 / N ) n
continuous time discrete frequency discrete time discrete frequency
periodic in time, period=T0 aperiodic in frequency periodic in time, period=N periodic in frequency, period=N

x(t ) Chap. 4 X ( j ) x[ n]  Chap. 5 X (e j ) 



1


X ( j )e jt d 
2 2 
j j n
= 21

  x(t )e  jt dt X (e )e d  , 
x[n]e  jn
n 
continuous time
continuous time continuous frequency discrete time continuous frequency
aperiodic in
aperiodic in time
time aperiodic in frequency aperiodic in time periodic in frequency, period=2

Wireless Communication Network Lab. Chih-Cheng Tseng 56


5.7.1 Duality in the Discrete-Time Fourier Series (1/3)
 Review on the discrete-time Fourier series

1 1
ak 
N

n N
x[ n]e  jk0n 
N

n N
x[n]e  jk (2 / N ) n

x[ n]  
n N
ak e jk0n  
n N
ak e jk (2 / N ) n

Wireless Communication Network Lab. Chih-Cheng Tseng 57


5.7.1 Duality in the Discrete-Time Fourier Series (2/3)
Suppose that f[k] is the discrete-time Fourier series of g[n]
1
f [k ]  
N n N
g [ n ]e  jk (2 / N ) n
g[ n] 
FS
 f [k ]

Let n=k and k=−n, we have


1
f [ n]   g[ k ]e jk (2 / N ) n
k N N

We find that (1/N)g[−k] corresponds to the Fourier series coefficients of f[n].


g[n]  FS
 f [k ] x  t  
F
 X  j 
1
f [ n]  g[ k ]
FS
X  t  
F
 2 x   j 
N
Wireless Communication Network Lab. Chih-Cheng Tseng 58
5.7.1 Duality in the Discrete-Time Fourier Series (3/3)

Other duality properties of the discrete-time Fourier series:

x[ n  n0 ] 
FS
 ak e  jk (2 / N ) n0 e jm (2 / N ) n x[ n] 
FS
 ak m

x[ n] y[ n]   al bk l
FS 
r N
FS
x[r ] y[n  r ]   Nak bk
l N

Wireless Communication Network Lab. Chih-Cheng Tseng 59


Example 5.16
Let g[n] be a rectangular wave with period N=9. Then,
 1 sin(5 k / 9) , k  multiple of 9
1, n  2 FS  9 sin( k / 9)
g[ n]    bk  
0, 2  n  4 5 , k  multiple of 9
9
Let x[n]=bn. Then,
 1 sin(5 n / 9) , n  multiple of 9
 9 sin( n / 9) 1 9, k  2
x[ n]    ak  g   k   
FS 1
9
5 , n  multiple of 9 0, 2  k  4
9
Wireless Communication Network Lab. Chih-Cheng Tseng 60
x[n] 
Sec. 3.6 ak 


x(t )   ak e jk t Chap. 3 ak 


0 1
 x(t )e  jk0t dt
k 
T0 T0
 k  N 
ak e jk (2 / N ) n 1
N  k  N 
x[n]e jk (2 / N ) n
continuous time discrete frequency discrete time discrete frequency
periodic in time, period=T0 aperiodic in frequency periodic in time, period=N periodic in frequency, period=N

x(t ) Chap. 4 X ( j ) x[ n]  Chap. 5 X (e j ) 



1


X ( j )e jt d 
2 2 
j j n
= 21

  x(t )e  jt dt X (e )e d  , 
x[n]e  jn
n 
continuous time
continuous time continuous frequency discrete time continuous frequency
aperiodic in
aperiodic in time
time aperiodic in frequency aperiodic in time periodic in frequency, period=2

Wireless Communication Network Lab. Chih-Cheng Tseng 61


5.7.2 Duality between the DT Fourier Transform and the CT
Fourier Series

CT Fourier series

FS 1
x(t )  ae
k 
k
jk0t
ak   x (t )e  jk0t dt
T T
0  2 / T

k  n, 0t   t   , k  n
T  2 0  1

DT Fourier transform

1 DTFT
x[ n] 
2 2
X ( e j
)e j n
d  X (e j )  
n 
x[ n]e  jn

Wireless Communication Network Lab. Chih-Cheng Tseng 62


Example 5.17
Given
1, t  T1 FS sin( kT1 )
g (t )  g (t  2 )   ak 
0, T1  t   k

If T1 = /2,

sin( n / 2) DTFT j
1,    / 2
x[ n]  X (e )  
n 0,  / 2    
Slide #44
c  n F   
sinc  c    
     2c 
Wireless Communication Network Lab. Chih-Cheng Tseng 63
Sec. 5.8 Systems Characterized by Linear Constant-
Coefficient Difference Equations
 Using the discrete-time Fourier series to analyze the linear constant-
coefficient difference equation for an LTI system with input x[n] and output
y[n].
N M

 a y[n  k ]   b x[n  k ]
k 0
k
k 0
k

N M

 k
a
k 0
e  jk
Y ( e j
)  k
b e  jk

k 0
X ( e j
)

 k 0 k
M  jk
j
jY (e ) b e
H (e )  
 k 0 k
j N  jk
X (e ) a e

Wireless Communication Network Lab. Chih-Cheng Tseng 64


Example 5.19
Consider a causal LTI system that is characterized by the difference equation
below
3 1
y[n]  y[n  1]  y[n  2]  2 x[n].
4 8
Then,
j 2 2 4 2
H (e )  3  j 1  j 2
 1  j 1  j
 1  j
 1  j
.
1 4 e  8 e (1  2 e )(1  4 e ) 1  2 e 1 4 e

Thus, n n
1 1
h[ n]  4   u[ n]  2   u[ n].
2 4
Wireless Communication Network Lab. Chih-Cheng Tseng 65
Example 5.20
2j
Consider the LTI system of Example 5.19 H (e )  1  j 1  j
.
(1  2 e )(1  4 e )
Let the input to the system be x[ n]   
1 n
4 u[n].
j 2  1  2
Y (e )   1  j 1  j   1  j 
 1  j 1  j 2
 (1  2 e )(1  4 e ) 1
 4  e  (1  2 e )(1  4e )
B 11 B 12 B 21
 1  j
 1  j 2
 1  j
B11  4, B12  2, B21  8
1 4 e (1  4 e ) 1  2 e
  1 n 1
n
1 
n

y[n]   4    2(n  1)    8    u[n]


  4  4  2  
Wireless Communication Network Lab. Chih-Cheng Tseng 66
Sec. 5.9 Summary
 There are a great many similarities between continuous-time and discrete-time
Fourier analysis.
 The discrete-time Fourier transform of an aperiodic signal is always periodic
with period 2.

Wireless Communication Network Lab. Chih-Cheng Tseng 67

You might also like