Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W.

Schafer

Chapter 3 The Z-Transorm

3.1 z-Transform

Recall the Fourier transform of a sequence x[n] in Chapter 2 was

defined as 

 x[ n] z
X( z ) n
n
The z-transform of a sequence x[n]
is defined as 

X( ej )   x[ n]e 
Two-side or bilateral z-transform
jn
n
X( z )   x[ n] z One-side or unilateral z-transform
n
Def: Z-transform operator n0


Z{ x[ n]}  x[ n] z n  X( z )
n
the
Thenotation
correspondence between a sequence and its z-transform is
indicated by x[ n]  
Z
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

More generally, we can express the complex variable z in polar form as

z  re j
For |z|=1, the z-transorm corresponds to the Fourier transform


Z{ x[ n]}   x[ n] z n  X( z )
n


X( re j )   x[ n ]( re
)
j n


n

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

The equation can be interpreted as the Fourier transform of the product of

the original sequence x[n] and the exponential sequence r-n. Obviously, for
r=1, the z-transform reduces to the Fourier transform of x[n].
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Convergence of z-transform requires


| x[ n] || z | n

n


| x[ n]r n
|
n
Convergence of Fourier transform
requires 

Figure 3.1 The unit circle in the

| x[ n] |, i.e.,
n 
|z|=1 complex z-plane

 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

For any given sequences, the set of values of z for which the z-transform
converges is called the region of convergence, abbreviated as ROC.


| x[ n]r n
|
n

Because of the multiplication of the sequence by the real exponential r-n, it

is possible for the z-transform to converge even if the Fourier transfrom
does not. For example, the sequence x[n] = u[n] is not absolutely
summable, and its Fourier transform does not converge absolutely.
However, r-nu[n] is absolutely summable for r>1.
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Region of Converge (ROC)

Definition : a set of values z for |X(z)| <  , which depends
only on |z| 

| x[ n] || z | n

n
Equation (3.2) is DaisLcraeteu-Triemne StigsnealrPierosce.ssing, 2/E by Alan V. Oppenheim
and Ronald W. Schafer
Chapter 3 The Z-Transorm
Z-transform is a Laurent’s series


X( z ) 


x[ n] z n
n
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

A Laurent series, and therefore the z-transform, represents an analytic
function at every point inside the region of convergence (ROC). In ROC, the
z-transform and all its derivatives must be continuous functions of z within
the ROC. If the ROC includes unit circle, then the Fourier transform and all
its derivatives with respect to  must be continuous functions of . The
sequences also must be absolutely summable, i.e., a stable sequence.
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Z at P(z)=0 are zeros
Closed form: X( z ) 
P( z ) Q( z Z at Q(z)=0 are poles
)
Example 3.1 Right-sided Exponential Sequence
Consider the signal x[n] = anu[n]. It is nonzero only for
n0, which is an example of a right-sided sequence.
Z-transform of x[n]:
 

X( z )   a u[ n] z
n n
  ( az 1 )n
n n0

For X(z) converging, it requiresFor | a | 1, its Fourier transform is


1
X(e j)  by given z  e

n | az

| 
1 n
j.
1  ae j
0 1 z
X( z )   ( az 1 )n   | z || a
n 1 az1
za |
0
For a=1 , x[n] becomes unit step 1
sequence X( z ) | z |
1 z 1
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.2 Left-sided Exponential Sequence

Consider x[n] = -anu[-n-1]. Since the sequence is nonzero for n -1, it is called
left-sided sequence.
 1

X (z)    anu[n 1]z n    a n z n

n n
 

  a  n z n  1  (a  1 z)n
n1 n0

1 z 1
X (z)   , | z || a
1 1 a1 z  z  a 1 az1 |
For | a | 1, its Fourier transform
is 1
j
X(e ) by given z  e
 1  ae j j.
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.3 Sum of Two Exponential Sequences
Consider x[n] 1 1
x[n]  (2 ) n u[n]  ( ) n u[n]

1 1
X( z )   {( )n u[ n]  (  3)n u[ n]}z n
n 2 3
 
1 1
 ( )n u[ n] z n   ( )n u[ n] z n
n 2 n 3
 
1 1
  ( z )   (  z 1 )n
1 n

n0 2 3
n0

1 1
1 1 2(1 12z 1 ) 2z( z  12 )
   
1 1 1 1 1 1
1 2z 1 1 3 z 1 (1 2z 1 )(1 3z 1 ) ( z  2 )( z  3 )
( 1 )n u[ n] Z 1 | z | 1 , (  1 )n u[ n] Z 1 | z |1
2 1 2 3 1 3
1 z 1 1 z 1
2 3
( 1 )n u[ n]  (  )n u[ n] Z 1

1 | z | 1
12 3 1 2
1 z 1 1 z 1
1 3
2
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.5 Two-sided Exponential Sequence
Consider the Sequence 1 1
x[ n]  ( 3 )n u[ n]  (2 )n u[ n 1]
1
( 1 )n u[ n] Z 1 | z |
 1 3
1 z 1
3 3
1 1 1
(  )n u[ n 1] Z  | z |
2 1
2
1 2z 1

1 1 1
X( z )  1
| z |,| z |
1 1 1 1
 1 z 1 z
3 3
1 2 1 2 1
2(1 z ) 2z( z  )
 12  12

1 1 1 1
(1 z 1 )(1 z1 ) ( z  )( z 
)
3 2 3
2
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

The aforementioned sequences have infinite lengths, which contribute

terms of the form (1-az-1) at its denominators. It should be noticed that
the term (1-az-1) introduces both a pole and a zero.

For a sequence, x[n], with finite length, the z-transform is

N2
X( z )  x[ n] zn
, N1  n  2
 n N N 1

It has no problem of convergence, because each of the terms |x[n]z-

n| is finite. For example ,x[n]=[n]+[n-5], then X(z)=1+z-5, which is

finite for |z|>0.

 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.6 Finite-length Sequence

x[ n]  a ,0  n  N
n
Consider the signal
1
0,otherwise
N 1  N
Then, 1

X( z )   a n z n   ( az
n n0
1
)n 0 zN aN
1 ( az 1 )N 
 1 az 1 za
N 1
1 z
The ROC is determined by
N 1


n0| az | 
1 n

j( 2k / N )
zk  ae ,k  0,1,..., N
1
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

1.  [n] 1 All z
1
2. u[n] | z |
1 z 1 1
1
3.  u[n 1] | z | 1
1  z 1
4.  [n  m] z -m All z except 0 (if m  0) or  (if m  0)

5. anu[n] 1 | z || a |
1  1az 1
6.  a n u[n 1] | z || a
1  az 1 |
7. na n u[n] az 1 | z || a
|
(1  az 1 )2
8.  na nu[n 1] az 1 | z || a
|
(1  az 1 )2
1 [cos ω 0 ]z 1
9. [cos ω0 n]u[n] | z |
1 [2 cos ω0 ]z 1  z 2 1
[sin ω0 ]z 1
10.[sin ω0 n]u[n] | z |
1 [2 cos ω0 ]z 1  z 2 1
1 [r cos ω0 ]z 1
11.[r n cos ω0 n]u[n] | z | r
1 [2r cos ω0 ]z 1  r 2 z 2
1 [r sin ω0 ]z 1
12.[r n sin ω0 n]u[n] | z | r
1 [2r cos ω0 ]z 1  r 2 z 2

 a n ,0  n  N 1 1  a N z N
| z | 0
13.  1  az 1
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

3.2 Properties of the Region of Convergence for the z-Transform

Property 1: The ROC is a ring or disk in the z-plane centered at the

origin; i.e., 0rR rL .
Property 2: The Fourier transform of x[n] converges absolutely if and
only if the ROC of the z-transform of x[n] includes the unit
circle.
Property 3: The ROC cannot contain any pole.
Property 4: If x[n] is a finite-duration sequence, i.e., a sequence
that is zero except in a finite interval -<N1 n N2<
, then the ROC is the entire z-plane, except possibly
z=0 or z= .
Property 5: If x[n] is a right-sided sequence, i.e., a sequence that is
nonzero for n<N1< , the ROC extends outward from the
outermost (i.e., largest magnitude) finite pole in X(z) to
(and possibly including) z= .
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

3.2 Properties of the Region of Convergence for the z-Transform

Property 7: A two-sided sequence is an infinite-duration seqeunce that

is neither right sided nor left sided. If x[n] is a two-sided
sequence, the ROC will consist of a ring in the z-plane,
bounded on the interior and exterior by a pole and,
consistent with property 3, not containing any poles.
Property 8: The ROC must be a connected region.
Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim an d Ronald W. Schafer

3.8
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.8 Stability, Causality, and the ROC
Considering a system with impulse response h[n], its z-Transform H(z)
has pole- zero plot shown in Fig. 3.11. Therefore, there are three ROC
possibilities.
(i) | Z | 1 (ii) 1 | Z | 2 (iii)2 | Z
2 |
2

Noncaus Noncaus Causal

al Not- al Not-
stable Stable stable

3.9 3.8
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

3. The Inverse z-Transform

1. Inspection Method
z 1 z 1
(3.35)  ( a ) u[ n 1] 
n
a u n 
n
, z 
1
1 az a . 1 az1 , | z || a
 |
   
 1 
 1  1 1
X (z)  (3.36) X( z )  1 , |z|
 , z 2 2
1 1
 1 2z  >   1 
 1 2 z 

 

When use Table 3.1 to find the inverse z-transform, we should find
its ROC to find its correct z-transform.
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

If M>N, then M-N poles at z=0, zeros at |z|=
3.3.2 Partial Fraction Expansion
If M<N, then N-M zeros
M
at z=0, poles at |z|=
M N M k
b z k
k
z  b
k 0 z
X (z) k
. (3.37) X (z)  k
. (3.38)
N
0 zM N
a z N k (3.40)
 a z k
k

k 0
k
k M
0

b (1 c z k
1

X (z)  0 k 1
N
) (3.39)
(3.41)
. a 1
0
(1 d
k 1 k z
N
Ak
If M<N and poles are all first )
order
X (z) 
k 1 d kz
1
. (3.42)
(3.40)
 1
Obviously, the common denominator of the functions in Eq. (3.42) is the
same as the denominator in Eq. (3.41). Multiplying both sides of Eq. (3.42) by
1
(1 d k z and evaluating for z=dk shows the coefficients Ak,
) A  (1 d z1 ) X (3.43)
(3.41)
zdk
(z)
k k
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.9 Second-Order z-Transform
Consider a sequence x[n] with z-transform
1
X (z)  1 1 , z  2.
1 (3.42)
(1 z1 )(1 z1 )
4 2
Right-sided
Left or Right-sided?
A1 A2
X( z ) 
 1 1
(1 4z1 ) (1 2z 1 )
1 1 1
A1  (1 z 1 z 1/  1 A2  z z 1/ 
)X( z )| 4 4
(1 2 )X( z )| 2 2
1 2
X( z ) 1 1  1 1
 (1 z ) (1 z )
4 2
n n
1 1
x[ n]  2  u[ n]    u[ n]
2 4
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

M

k b z k
M
MN N
Ak
X (z)  Nk 0 (3.37)
(3.39) N X (z)  
r

Br z  k 11 d z 1
. (3.45)
(3.43)
.  ak z  k r 0
k

k (3.42)
0
A more general form of a z-Transform X(z) with M>N contains poles at
z=0, single-order poles and multiple-order poles, which can be written as
follows,
MN N s
Ak Cm
X (z)  r 
0 Br z k 1,k
r
 1 d kz 1
 
m (1 d iz
1 m
)
. (3.46)
(3.44)
i 1

The coefficients Cm can be obtained from the equation

1  d sm  s  1
Cm sm sm  (1 di w) X (w )  (3.47)
(3.45)
 (s  m)!(d )i  dw w 1
i
. d
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

For example: X(z) has multiple poles at Z  dk

s
Cm
X(z)  1 m
m (1 d k z )
 1
C1 C2 Cs
 X (z)  L
 (1 d k z 1
) (1 d k z )
1 2
(1 d k z )
1 s

1 s
Multiply X(z) by (1- d k z ) , then we
have
(1- d zz 1))s2k C )  X 1(z)  C 
1 s1 z 1 s
(1
(1dd L s
2 k k C
Let w  zs -1 , the equation becomes
(1- d w)  X (w1 )  C  (1 d w)s1  C  (1 d w)s2  L 
s
C

k 1 k 2 k
s
d sm (1- d w) X (w1
Cm dw sm ) 
(s  m)!(d k )sm
 k
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Inverse Z-transform: Cauchy Integral Method

Cauchy integral
formula: 1, k 
21j 
k 1
z dz  00, k 
C {
 0
Becaus
e  x(n)z  n
X(z)  
Multiply both sides
by 1 zXk1 , then
k 1 1


 (z)  z dz  x(n)  z nk

2πj 2πj C  1dz
 C

 1  X(z) k1zdz 1
 2j z nk1 dz
x(n)

 C C
When – 1 2j k1
n+k=0,  X(z)  z

  x(n) (k 
n)
2j C dz 
1 k1
  X(z)  z 
x(k)2j
C dz
Therefor
1
e, x( n ) 2j C  X( z )
Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer
Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer
Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer
Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer
Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer
Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer
Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.10 Inverse by Partial z-Transform
Consider a sequence x[n] with z-transform
1 2z  1  z2 (1 z1 )2
X (z)  , z 
3 1 1 1
 1 z1  z2 1. (1 z )(1 z1 )
2 2
3.11 3.10
2
Since M=2=N and all poles are all first order, X(z)
can be represented as A1
X( z )  B0 
1 A2 2z
1 1
1 z 1

2
, where B0 =2 is found by long division 12 z 2
 32 z 1 1 z 2  2z 1 1

z 
2 3z 1
2
5z 1 1
1 5z1
X (z)  2 . (3.47)
1 1
 (1 z )(1 z1 )
2
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

X (z)  2
1 5z1
. (3.47)
A A
X( z )   1 1 1
 1 z2 1
 1 B0
1
2 z
(1 2z1 )(1 z1 )
A1 and A2 can be calculated by residual values at z=1/2 and z=1.
(1 1 z1 ) X (z)  [( 1 5z 1 )(1 21 z1 )] 
Re s f (z) 
lim z2 1/
z 1/
2
9 2 z 1/ 2
2 (1 1 z 1 )(1 z 1 )
1 1 5z1 1
Re s f (z)  lim (1 z ) X (z)  )(1 z )]z 1 
[( z1
z (1 1 z )(1 z )
1 1
1
2
8
9 8
Therefore, X (z)  2  . (3.48)
 1 1
1 z 1

1 2z

Check Table 3.1, we get

x[ n]  2[ n]  9 u[ n]  1 n
2

 8u[ n]
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

3.3.3. Power Series Expansions
The defining expression for the z-transform is a Laurent Series where the
sequence values x[n] are the coefficients of z-n. Thus, the z-transform is
given as a power series in the form


X (z)   x n  z 

n
n (3.49)
 L  x2 z 2
+x1 z 
Consider 1 (3.50)
1  z (1 2 z )(1 z )(1 z )
x0  Xx(z)
2 1 1 1

2 z )  z
1 1 1
z +x2 z
1 X( 2
 z 1 z
2
+L,
Example 3.11 Finite- 2
By inspection,
length Sequence
1 1
x[ n]  [ n  2 ]  [ n 1]  [ n]  [ n
1]
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.12 Inverse Transform by Power Series Expansion
Consider the z-transform

X (z)  log(1  az1), | z || a | .

Using the Taylor series
 expansion for log(1+x) with |x|<1,
we obtain
X (z)   (1)
n1 n n a z
n
.
n1
Therfor
e, 
n1 an , n 
x[n] (1) n 1
 0 ,n
0


* Note : log(1  x)


(1) n1 xn , for - 1  x 
n 1
n
1
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.13 Power Series Expansion by Long Division
Consider the z-transform Make the long-division in descending order
1
X (z) , z> (3.53)
 1 az a .
1

Use Long-division, we get

1 az 1  a 2 z
2
1
Since X(z) approaches a finite 1 az 1
constant as z approaches infinity, 1 az 1
the sequences is causal. az 1  a 2 z 2

a 2 z 2
1
 1 az 1
 a 2
z 2

1 az 1  L
Inverse z-
transform
x[ n] 
a n u[ n]
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.13 Power Series Expansion for a Left-sided Sequence
Make the long-division in ascending order
X (z)  1 , z<a. (3.54)
1 az1

Use Long-division, we get  a1 z  a2 z 2

az z
Since X(z) approaches a finite z  a1 z 2
constant as z approaches zero,
the sequences is causal. a1 z 2

a1 z 2L a 2 z 3

x[ n]  a n u[ n
1]
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

3.4 z-Transform Properties
3.4.1 Linearity x1 [ n]  X( z ),ROC  x1
xR2 [ n]  X( z ),ROC  Rx
2

ax1 [ n]  bx2 [ n]  aX1 ( z )  bX 2 ( z ),ROC : xR1 2

 Rx z
x[ n  n0 ]  zn 0 , ROC=Rx (except
3.4.2 Time Shifting for the possible
X( z ) addition or deletion
 of z=0 or z=)
Y( z )  x[ n  n0 ] n

 n z
 

 z n0  x[ n  n 0 ] ( nn 0)

z
n 0


n z
 z n 0 x[ m] z ( m )
m
X( z )
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.14 Shifted Exponential Sequence

Consider the z-transform
1
X( z ) ,| z | 4

z  14
1

From the ROC, we identify this as corresponding to a right-

sided sequence.
4
X  z z 1 X z  4
1 1
 1 z 1  1  z 1
4 n
4
1
xn  4n4 
4 
, or u n
 
1
X z  z 1
xn  1  un 1
n1
 ,
1 1 4
z  1 1 
 4z   4
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

3.4.3 Multiplication by an Exponential Sequence

z n x[ n] z  X( z / z ),ROC | z |
x
R
0 0 0

The notation ROC=|z0|Rx denotes that ROC is Rx scaled by |z0|; i.e., if

Rx is the set of values of z such that rR<|z|<rL, then |z0|Rx is the set of
values of z such that
|z0|rR<|z|<|z0|rL. This in turn can be interpreted as a frequency shift or
translation, associated with the modulation in the time domain by the
complex potential sequence ej0n. That is

e j n x[ n] z  X( e j(  ) )
0 0
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.15 Exponential Multiplication

Calculate the z-Transform of xn  r n cos

nun
0
Rewrite is as
1 1
x[ n] 2 ( re j0
) u[ n] 2 ( re j )n
n 0

u[ n] 1
1 0
) u[ n] 
n z 2
,| z |
2 ( re
j

 1 re z rj 1 0

1 1
( re 0 ) u[ n] 
 j n z
2
,| z |
 1 re z r
 j 1 0

2 1
1
X z  2
1 
2 , z 
1  re z
j 0
1  re z 1
 j 0

r 1  r cos z 1
 z r

1  2r cos0 10 r 2 z 2 ,
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

3.4.4 Differentiation of X(z)

dX( z )
nx[ n] z  z dz ,ROC  Rx

Because
X( z )   x[ n] z
n
n

dX( z )
Consider
z  z  ( n )x[ n] z n1
dz n


  nx[ n] z n 
Z{ nx[ n]}
n
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.16 Inverse of Non-rational z-transform

Find the inverse Z - transform of

X(z)  log(1  az-1), for | z || a

|.
Ans :
dz
dX (z)
Consider nx[n] z 2
 z ,
then dX(z)  az  -z dX(z) az1
dz  dz
1  az1 .
 1 az1
z
 nx[n]   az1 z-1 a(a) n1 u[n 1].

1  az 1

 x[n]  (1) n1 An u[n 1]

n
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.17 Second-Order Pole

x[ n]  na n u[ n]  n( a n u[ n ])
Then, we
have d az 1
X( z )  ( ) ,| z |
1
dz 1 az
1
(1 az 1
)
2
z  a
az1
na n u[ n] z
 (1 az 1 )2
3.4.5 Conjugation of a Complex Sequence


x[ n]   X( z )
z
X( z )   x[ n] z
x* [ n] z  X *  n
n


( z* ) x *
[ n] z n  ( 
x[ n ]( z* )n )*
ROC=Rx n n

 X * ( z* )
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

3.4.6 Time Reversal


x[ n] z  X( z ) X( z )   x[ n] z
 n 
n

x* [ n] z  X * (
1
*
)  x [ n] z
 * n
 x [ m] z * m
  x[ m ](( z *
)1 )m )*
z n  m ( m

ROC=1/R x   x[ m ]( 1 )( m ) )*  X *
( ) z* z*
1
( m

The notation ROC= 1/Rx implies that Rx is inverted; i.e., if Rx is the set of
values of z such that rR<|z|<rL, then the ROC is the set of values of z such
that 1/rL<|z|<1/rR.


x[ n] z 
X( z ) 1 
 x[ n] z n
X( z ) n
 
x[ n] z  X( )  x[ m] z  x[ m ]( z
 x[ n] z  ( )m
n m 1

z n m
)
m
ROC=1/Rx 
1
  x[ m ]( )m 
1
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.18 Time-Reversed Exponential Sequence
Consider
x[ n]  a  n u[ n]
, which is a time-reversed version of anu[n]. From the time-reverse
property, it follows
X( z )  1  a 1 z 1 
1 az
1 a1 z 1
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

3.4.7 Convolution of Sequence
According to the convolution property,
x [ n]* x
[ n] z  X ( z )X
y[ n]   ( zx)1 [ k ] x2 [ n  k
1
] 
2
Y( z ) k 1{  x1 [ k ] x2 [ n 
 n

 n 2 k 
k ] }z
  
  x1  x 2 [ n  k ]z n

x1 [ k ]{  x2 [ n  ( nk )
}z k
k  [ k ] n  k  k ]z n
 
  x1 [ k ]  x 2 [ m]z m
}z k
 X1 ( z )X 2 ( z
k  { m
)
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

Example 3.19 Convolution of Finite-Length Sequence

Please find the convolution y[n] of two finite - length sequences x1[n] and x2 [n],
in which
x1[n]  [n]  2[n 1]  [n  2],
and
x2[n]  [n]  [n 1].

Answer :

X1(z)  1  2z-1  z - 2
and

X2(z)  1 - z -1

Y(z)  X1(z) X2(z)  (1  2z-1  z - 2 )  ( 1 - z -1)

 1  z -1 - z - 2 - z -3.
 y[n]  [n]  [n 1]  [n  2]  [n  3].
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

3.4.10 Initial Value Theorem

If x[n] is zero for n<0 (i.e., if x[n] is causal), then
x[ 0 ]  lim X( z )
z
causal

X( z )|z   x[ n] z
 n
|z   x[ n] zn |z  x[ 0
n n0 ]
 Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm

3.5 z-Transforms and LTI systems
Example 3.20 Evaluating a Convolution Using the z-Transform
Let h[n]=anu[n] and x[n]=Au[n]. The corresponding z-Transforms are

H (z)

a n z n ,| z || a
n  |
0 1
X (z)



n 
Az n
A
1  z1 1
,| z |
1
1  az
0 z-transform of the convolution of h[n] and x[n] is
If |a|<1, then the
A

Y (z)  X 1 (z) X 2 (z)  Az2  ,| z |

1 1
1
(1  az )(1  z
)
(z  a)(z
1)
Discrete-Time Signal Processing, 2/E by Alan V. Oppenheim and Ronald W. Schafer

Chapter 3 The Z-Transorm