The z-Transform

The z-Transform
 Counterpart of the Laplace transform for discrete-time signals
 Generalization of the Fourier Transform
 Fourier Transform does not converges for all sequences
 The z-Transform is often time more convenient to use
 The Fourier transform X(e jω) of a sequence x[n] is

   xne
X e j   jn

n 
 The z- transform X(e jω) of a sequence x[n] is :

X z   
n 
xn z n

 z is a complex variable that can be represented as z = r e j

 Substituting z = e j will reduce the z-transform to Fourier transform

2 351M Digital Signal Processing Copyright (C) 2005 Güner

Arslan
 The z-transform and the DTFT
 The z-transform is a function of the complex z variable
 Convenient to describe on the complex z-plane
 If we plot z = e j for  = 0 to 2 we get the unit circle

Im  
X e j

Unit
Circle
r =1
 0

Re 0
2 2

3 351M Digital Signal Processing Copyright (C) 2005 Güner

Arslan
 Convergence of the z-Transform

 DTFT does not always converge


   xne
X e j   jn

n 

 Infinite sum not always finite if x[n] no absolute summable

 Example: x[n] = anu[n] for |a| > 1 does not have a DTFT
 Complex variable z can be written as re j so the z-transform
 
    xnre   xn r e
X re j  j n n  jn

n  n 
 DTFT of x[n] multiplied with exponential sequence r -n
 For certain choices of r the sum maybe made finite

n
 x n r -n
4 351M Digital Signal Processing Copyright (C) 2005 Güner
Arslan
 Region of Convergence
 The set of values of z for which the z-transform converges
 Each value of r represents a circle of radius r
 The region of convergence is made of circles

Im  Example: z-transform converges for values of

0.5< r <2
 ROC is shown on the left
2  In this example the ROC includes the unit
0.5
circle, so DTFT exists
Re  Not all sequence have a z-transform
 Example: xn  coson
 Does not converge for any r
 No ROC, No z-transform
 But DTFT exists?!
 Sequence has finite energy
 DTFT converges in the mean-squared sense
5 351M Digital Signal Processing Copyright (C) 2005 Güner
Arslan
 z–Transform zeros and poles

P( z )
X ( z) 
where P(z) and Q(z) are
polynomials in z.
Q( z )

Zeros: The values of z’s such that X(z) = 0

Poles: The values of z’s such that X(z) = 

6 351M Digital Signal Processing Copyright (C) 2005 Güner

Arslan
 Right-Sided Exponential Sequence Example
 
xn  a un  n
X z   
n 
a unz
n n
 az 
n 0
1 n

x[n]
 For Convergence we require



n
az1   ... n
n 0
-8 -7 -6 -5 -4 -3 -2 -1 1 2 3 4 5 6 7 10
 Hence the ROC is defined as 8 9

az1  1  z  a Im

 Inside the ROC series converges to


X z    
n 1 z
az1  
1  az1 z  a a 1
n 0 ox Re
 Geometric series formula
N2 1
N
 N


a a 1 2

an 
n N
1 a
1
 Region outside the circle of radius a is the ROC
7 351M Digital Signal Processing Copyright (C) 2005 Güner
 Right-sided sequence ROCs extend outsideArslan a circle
 Same Example Alternative Way
 
xn  a un  X z  
n

n 
a unz
n n
 az 
n 0
1 n

N2 1
 N  N

1 2

  n

n N
1
1


  az   az  1 0 1 

 az
n 0
1 n
1  az 1

 For the term with infinite exponential to vanish we need

az1  1  a z

 Determines the ROC (same as the previous approach)

 In the ROC the sum converges to 
X z    
1 n 1
az  1
n 0 1  az
8 351M Digital Signal Processing Copyright (C) 2005 Güner
Arslan
 Left-Sided Exponential Sequence Example

-8 -7 -6 -5 -4 -3 -2 -1 1 2 3 4 5 6 7 8 9 10
n
x(n)  a u(n 1) n ...

X ( z )    a nu(n  1)z n x(n)
n  
1
   a n z n
n   For convergence of X(z), we require that


 | a 1z | 1

  a z n n | a 1 z |  
n 1 n 0
Im

 1   a n z n | z || a |
n 0
1 
a ax X ( z )  1   (a 1 z ) n  1 
1

z
Re n 0 1  a 1 z z  a

9
| z || a |
3 5 1 M D i g i t a l S i g n a l P r o c e s s i n g Copyright (C) 2005 Güner
Arslan
 Two-Sided Exponential Sequence Example
n n
 1 1
xn     un -   u- n - 1
 3 2 ROC : 
1 1
z 1
3
0 
 1 1   1 1  1
 n  z   z   z
 1 1   3   3 

1 3
 z   
n 0  
3 1 1
1  z 1 1  z 1 1 1
3 3 ROC : z 1
2
 0 1
 1 1   1 1   z
1 n  z    z 
 1    2   1 2

1 2
 z   Im
n    
2 1 1
1  z 1 1  z 1
2 2

1 1
 1 
2 z z   3x 2
 12  oo x Re
X z  
1 1
  1
1 1
1  z 1 1  z 1  1  1 12
 z   z  
3 2  3  2

10 351M Digital Signal Processing Copyright (C) 2005 Güner

Arslan
 Example: A Finite Sequence

x(n)  an , 0  n  N 1
1 1
1  (az 1 ) N 1 zN  aN
X ( z)   a z   ( az )
N N

 1
n n n

  N 1
n 0 n 0 1  az 1 z za

Im ROC: 0 < z < 

N-1 zeros
ROC does not include any pole.

N-1 poles Always Stable

Re

11 351M Digital Signal Processing Copyright (C) 2005 Güner

Arslan
 Properties of The ROC of Z-Transform

 The ROC is a ring or disk centered at the origin

 DTFT exists if and only if the ROC includes the unit circle
 The ROC cannot contain any poles
 The ROC for finite-length sequence is the entire z-plane
 except possibly z = 0 and z = 
 The ROC for a right-handed sequence extends outward from the
outermost pole possibly including z = 
 The ROC for a left-handed sequence extends inward from the
innermost pole possibly including z = 0
 The ROC of a two-sided sequence is a ring bounded by poles
 The ROC must be a connected region
 A z-transform does not uniquely determine a sequence without
specifying the ROC
12 351M Digital Signal Processing Copyright (C) 2005 Güner
Arslan
 Stability, Causality, and the ROC

 Consider a system with impulse response h[n]

 The z-transform H(z) and the pole-zero plot shown below
 Without any other information h[n] is not uniquely determined
 |z| > 2 or |z| < ½ or ½ < |z| <2
 If system stable ROC must include unit-circle: ½ < |z| <2
 If system is causal must be right sided: |z| >2

13 351M Digital Signal Processing Copyright (C) 2005 Güner

Arslan
 Z-Transform Pairs

Sequence z-Transform ROC

All z
(n) 1
All z except 0 (if m>0)
(n  m) z m or  (if m<0)
1
| z | 1
u (n ) 1  z 1
1
 u (n  1) | z | 1
1  z 1
1
anu(n) | z || a |
1  az1
1
 a u(n 1)
n | z || a |
1  az 1
14 351M Digital Signal Processing Copyright (C) 2005 Güner
Arslan
 Z-Transform Pairs

Sequence z-Transform ROC

1  [cos 0 ] z 1
[cos0n]u(n) | z | 1
1  [2 cos 0 ]z 1  z  2
[sin 0 ]z 1
[sin0n]u(n) | z | 1
1  [2 cos 0 ]z 1  z  2
1
1  [ r cos  ] z
[r cos 0 n]u (n)
n 0
| z | r
1  [2r cos 0 ]z 1  r 2 z  2
1
[ r sin  ] z
[r sin 0 n]u ( n)
n 0
| z | r
1  [2r cos 0 ]z 1  r 2 z  2
a n 0  n  N  1 1 a N zN
 1
| z | 0
0 otherwise 1  az
15 351M Digital Signal Processing Copyright (C) 2005 Güner
Arslan
16 351M Digital Signal Processing Copyright (C) 2005 Güner
Arslan
 Example1: Second-order z-transform

Partial Fraction Expansion

M
bo Π (1  ck z 1 )
k 1
X ( z)  N
ao Π (1  d k z 1 )
k 1
N

 1 d
Ak
X ( z)  1
k 1 kz

Ak  (1  d k z 1 ) X ( z ) | z  d k

17 351M Digital Signal Processing Copyright (C) 2005 Güner

Arslan
 Example2: Finite-Length Sequence

18
⇔
351M Digital Signal Processing Copyright (C) 2005 Güner
(n  m) z m
Arslan
Example3: Evaluating a Convolution Using the z-
Transform

19 351M Digital Signal Processing Copyright (C) 2005 Güner

Arslan
Example4: Exponential Multiplication

20 351M Digital Signal Processing Copyright (C) 2005 Güner

Arslan
Example5: Inverse of Non-Rational z-Transform

⇔
21 351M Digital Signal Processing Copyright (C) 2005 Güner
Arslan