ECE 3331b

Introduction to Signal Processing

Lecture 9

Instructor: Dr. Ilia G. Polushin

Department of Electrical & Computer Engineering

Faculty of Engineering

The University of Western Ontario

Winter 2017
Topic 3: The z-Transform and Its Application to the
Analysis of LTI Systems

The z-Transform

Properties of the z-Transform

Rational z-Transforms

Inversion of the z-Transform

Analysis of LTI Systems in the z-Domain

The One-sided z-Transform

Rational z-Transforms

Rational z-transforms is a wide and practically most important class of

z-transforms

The z-transform is rational if it can be represented as a ratio of two polyno-

mials in z 1 (or z),
1
b0 + b1 z + b2 z 2 + . . . + bM z M
X(z) = 1 + a z 2 + ... + a z N
a0 + a1 z 2 N

Example: all z-transforms in the Common z-Transform Pairs table are

rational.
Rational z-Transforms

1
Rational z-transform as a ratio of polynomials in z :

1 2 M
PM k
b 0 + b1 z + b2 z + . . . + bM z k=0 bk z
X(z) = 1+a z 2 + ... + a z N
= PN (RT 1)
a0 + a1 z 2 N k=0 ak z k

Zeros of X(z) are the values of z for which X(z) = 0.

Poles of X(z) are the values of z for which X(z) = .
Rational z-Transforms
1
Rational z-transform as a ratio of polynomials in z :

1 2 M
PM k
b0 + b1 z + b2 z + . . . + bM z k=0 bk z
X(z) = 1+a z 2 + ... + a z N
= PN (RT 1)
a0 + a1 z 2 N k=0 ak z k

If a0 = 0 and b0 = 0, one can eliminate the negative powers of z and rewrite (RT1) as a ratio
of polynomials in z:
B(z)
M
M M 1 M 2

b0 z z + (b1 /b0 )z + (b2 /b0 )z + . . . + (bM /b0 )
X(z) = N
(RT 2)
a0 z z N + (a1 /a0 )z N 1 + (a /a )z N
2 0
2 + . . . + (a /a )
N 0

A(z)
Both B(z) and A(z) are polynomials in z, therefore they can be rewritten in factored form,
as follows
b0 N M (z z1 ) (z z2 ) . . . (z zM )
X(z) = z (RT 3)
a0 (z p1 ) (z p2 ) . . . (z pN )
Rational z-Transforms

Rational z-transform in factored form:

b0 N M (z z1 ) (z z2 ) . . . (z zM )
X(z) = z (RT 3)
a0 (z p1 ) (z p2 ) . . . (z pN )

X(z) in (RT3) has M finite zeros at z = z1 , z = z2 , . . ., z = zM . These are the roots of

the numerator polynomial.
Similarly, X(z) in (RT3) has N finite poles at z = p1 , z = p2 , . . ., z = pN . These are the
roots of the denominator polynomial.
If N > M , X(z) also has N M (trivial) zeros at z = 0.

If M > N , X(z) has M N (trivial) poles at z = 0.

Relation between Pole Location and Time-Domain
Behaviour for Causal Signals
Rational z-transform:

1 2 M
b0 + b 1 z + b2 z + . . . + bM z b0 N M (z z1 ) (z z2 ) . . . (z zM )
X(z) = 1+a z 2 + ... + a z N
= z
a0 + a1 z 2 N a0 (z p1 ) (z p2 ) . . . (z pN )

Zeros of X(z) are the values of z for which X(z) = 0.

Poles of X(z) are the values of z for which X(z) = .
Time-domain behaviour of a signal depends strongly on the location of the poles
of its z-transform
Location of the zeros of signal z-transform also influence its time-domain behaviour,
although to a lesser extent.
Relation between Pole Location and Time-Domain
Behaviour for Causal Signals

Example: A signal with one (real) pole z = a

n z 1
x(n) = a u(n) ! X(z) = 1
, ROC : |z| > |a|
1 az

If |a| < 1, i.e., a is inside the unit circle, the signal is decaying.

if |a| > 1, the signal is growing without a bound.

If |a| = 1, the signal has a constant magnitude.

In addition, if a < 0 the response alternates in sign.

Relation between Pole Location and Time-Domain
Behaviour for Causal Signals
Time-domain behaviour of a single-real-pole signal as a function
of the pole location (Figure 3.3.5 from Proakis & Manolakis, 2007)
Relation between Pole Location and Time-Domain
Behaviour for Causal Signals

Example: A signal with a double (real) pole at z = a:

1
n z az
x(n) = na u(n) X(z) = , ROC : |z| > |a|
(1 az 1 )2

If |a| < 1, i.e., a is inside the unit circle, the signal is bounded and eventually decaying
to zero.
if |a| > 1, the signal is growing without a bound.
If |a| = 1, the signal is linearly growing without a bound (!)
In addition, if a < 0 the response alternates in sign.
Relation between Pole Location and Time-Domain
Behaviour for Causal Signals
Time-domain behaviour of a double-real-pole (m = 2) signal as a func-
tion of the pole location (Figure 3.3.6 from Proakis & Manolakis, 2007)
Relation between Pole Location and Time-Domain
Behaviour for Causal Signals

Example: A signal with a pair of complex conjugate poles at z = rej :

n z 1 rz 1 cos
x(n) = (r cos n) u(n) X(z) = , ROC : |z| > |r|
1 2rz 1 cos + r2 z 2

Complex conjugate poles at z = rej result in exponentially weighted sinusoidal

signal with frequency ( , ]

if |r| < 1, the amplitude of oscillations is exponentially decaying

If |r| = 1, the amplitude of oscillations remains constant
If |r| > 1, the amplitude of oscillations is exponentially growing
Relation between Pole Location and Time-Domain
Behaviour for Causal Signals
Time-domain behaviour of a signal with a pair of complex conjugate
poles at z = rej!0 (Figure 3.3.7 from Proakis & Manolakis, 2007)
Relation between Pole Location and Time-Domain
Behaviour for Causal Signals
Double pair of complex conjugate poles on the unit circle (i.e., at
z = ej! ) result in sinusoidal signal with linearly growing amplitude
of oscillations.
The Transfer Function of an LTI System
We know that the response of an initially relaxed LTI system to an input sequence
x(n) is given by the convolution formula:

y(n) = h(n) x(n), (1)

where h(n) is the impulse response function of the system.

Taking z-transform of both sides of (1) and using convolution property of the
z-transform, we have
Y (z) = H(n)X(z)

where Y (z) = Z [y(n)], X(z) = Z [x(n)], H(z) = Z [h(n)].

The z-transform of the impulse response function h(n),

+1
X
n
H(z) = Z [h(n)] := h(n)z ,
n= 1

is called the system function (or transfer function) of the LTI system.
The Transfer Function of an LTI System

Determining the response of an (initially relaxed) LTI system based

on the system function H(z):

1. Determine z-transform of x(n):

X(z) = Z [x(n)] .

2. Obtain z-transform of the response Y (z) according to the formula

Y (z) = H(z)X(z)

3. Determine the response y(n) by evaluating the inverse z-transform of Y (z),

1
y(n) = Z [Y (z)]
The Transfer Function of a System Described by
Constant-Coefficient Difference Equation
System:
N
X M
X
ak y(n k) = bk x(n k), a0 = 1. (SY S1)
k=0 k=0

Taking z-transform of both sides of (SYS1) and using the time-shifting

property, we have
XN M
X
ak Y (z)z k = bk X(z)z k ,
k=0 k=0

or, equivalently,

N
! M
!
X X
k k
Y (z) ak z = X(z) bk z . (SY S2)
k=0 k=0

From (SYS2), the system function H(z) of (SYS1) is

P
M
k
bk z
Y (z)
H(z) = = k=0
X(z) PN
k
ak z
k=0
The Transfer Function of a System Described by
Constant-Coefficient Difference Equation

Example 3.3.4 (from Proakis & Manolakis, 2007): Determine the system function and the
impulse response of the system described by the dierence equation
1
y(n) = y(n 1) + 2x(n) (SY S1)
2

Solution:

Computing z-transform of both sides of (SYS1), we have

1 1
Y (z) = z Y (z) + 2X(z)
2
The system function:
Y (z) 2
H(z) = = 1 1
(SF )
X(z) 1 2 z
(SF) has a pole at z = 1/2 and a zero at the origin.
Using the z-transform table, we see that
n
1
h(n) = Z 1 [H(z)] = 2 u(n).
2
Some Common z-Transform Pairs
The System Function of a System Described by
Constant-Coefficient Difference Equation

Example 3.3.4 (from Proakis & Manolakis, 2007): Determine the system function and the
impulse response of the system described by the dierence equation
1
y(n) = y(n 1) + 2x(n) (SY S1)
2

Solution:

Computing z-transform of both sides of (SYS1), we have

1 1
Y (z) = z Y (z) + 2X(z)
2
The system function:
Y (z) 2
H(z) = = 1 1
(SF )
X(z) 1 2 z
(SF) has a pole at z = 1/2 and a zero at the origin.
Using the z-transform table, we see that
n
1
h(n) = Z 1 [H(z)] = 2 u(n).
2
The System Function of a System Described by
Constant-Coefficient Difference Equation
Difference equation vs. system function:

P
M
k
N
X M
X bk z
k=0
ak y(n k) = bk x(n k) H(z) =
PN
k
k=0 k=0 ak z
k=0

Special Case 1: All-Zero System: ak = 0 for all k = 1, . . . , N .

M
X M
X M
k 1 X
y(n) = bk x(n k) H(z) = bk z = M bk z M k
.
z
k=0 k=0 k=0

All zero system has only trivial poles (at z = 0). This is FIR system.
Special Case 2: All-Pole System: bk = 0 for all k = 1, . . . , M .

N
X b0 b0 z N
ak y(n k) = b0 x(n) H(z) = =
P
N
k
P
N
k=0 ak z ak z N k
k=0 k=0

All-pole system has only trivial zeros (at z = 0). This is IIR system.