The Z-Transform

Motivation

The z-transform is an extension of the discretetime Fourier transform to a function that is

defined on regions of the complex plane.
Some sequences that do not have Fourier
transforms will have z-transforms, but the
reverse is true as well.
For our purposes, the special significance of ztransform is as a tool to facilitate designing
systems with the desired attributes that can be
practically implemented

Topics

Concepts from Complex Variables

Laurent Expansions
Definition of the Z-Transform
Computing Inverse Z-Transforms
Properties of Z-Transforms

Complex Functions (1)

Complex functions are complex valued and also have complex
arguments. The domain for the argument usually spans some
open region of the complex plane as shown below,

z0

z1

Complex Functions (2)

f ( z ) is said to be analytic (or holomorphic) in if it can
be represented as a convergent power series

f ( z ) = an ( z z 0 )

n=0

in a neighborhood of each point of .

Each power series will converge to f ( z ) in a disk whose radius
of convergence is at least as large as the shortest distance
from the center point to the boundary of .
If is a disk to begin with then a single power will be able to
represent f ( z ) throughout , but in general, a single power
series won't be enough.

Complex Functions (2)

The natural domains of analytic function can be complicated.
They can include branch points (e.g. log ( z ) ) or essential
singularities (e.g. e 1 z ). Fortunately, we will mostly work with
a better behaved class of functions called rational functions,
which are formed by taking the ratio of two polynomials :
f ( z) = P ( z) Q ( z)
Rational functions are analytic everywhere except at the roots
of the denominator, which are called poles of the function.
A pole has order n if the root of the denominator has
multiplicity n, and a pole of order 1 is called a simple pole.

Complex Functions (4)

Example : In the figure below, is plotted at each pole,
and at each zero of the rational function :
z j )( z + j )
(
z2 +1
=
2
3
z + 2 z 1 ( z 12 ) ( z + 2 )

This type of figure is called a pole - zero plot

Topics

Concepts from Complex Variables

Laurent Expansions
Definition of the Z-Transform
Computing Inverse Z-Transforms
Properties of Z-Transforms

Laurent Expansion (1)

As noted earlier, if f ( z ) is analytic in the disk z < r then it will
be represented, throughout the disk, by a unique power series :

f ( z ) = an z n
n =0

Laurent Expansion (2)

If f ( z ) is analytic in the exterior region z > r then it can be
represented, throughout that region, by a unique expansion :

f (z) =

n =

an z n

Laurent Expansion (3)

And, if f ( z ) is analytic in the annular region r1 > z > r2 then it
can be represented, in that region, by a unique expansion :

f ( z) =

n =

an z n

r2

r1

Laurent Expansion (4)

An expansion of this type,
f (z) =

n =

an z n

which includes both positive and negative powers of z , is

called a Laurent expansion. The annulus r1 < z < r2 is called
the region of convergence which is abbreviated as ROC.
An important theorem from complex variables is that the
an coefficients are given by the (Cauchy Integral) formula:
f ( z)
1
an =
dz


n +1

2 i z

where the integral is around any circle z = c inside the ROC .

Laurent Expansion (5)

Rational functions are analytic in successive annular regions
that have no poles, but start and end at the radii of poles. E.g.,
z j )( z + j )
(
z2 +1
=
2
3
z + 2 z 1 ( z 12 )( z + 2 )
has the three annular regions shown below, and has different
Laurent expansions in each region.

n
a
z
n
n =

f ( z ) = bn z n
n =

n
cn z
n =

Region1

Region 2

Region 3

Region 1
Region 2
Region 3

Laurent Expansion (6)

The upper limit r2 of the ROC r1 < z < r2 is determined by
the positive index an's, while the lower limit r1 is determined
by the negative index an's.
The sum/product of two Laurent expansions will also be a
Laurent expansion provided that the ROC's of the original
functions overlap. The ROC of the sum/product will at
least contain the intersection of the ROC's of the original
functions, but it may be larger.
Exercise : Find a formula for the cn where

n
n
n
cn z = an z bn z

n =
n=
n=

Topics

Concepts from Complex Variables

Laurent Expansions
Definition of the Z-Transform
Computing Inverse Z-Transforms
Properties of Z-transforms

Definition of the Z-Transform (1)

The Z - transform of the discrete time sequence x [ n] is
the complex function given by the Laurent expansion:
Z { x [ n ]} ( z ) =

n =

x [ n] z n

Positive time indices are assigned to negative powers of z.

The Laurent series won't necessarily converge, but when it
does the region of convergence (ROC ) is an annulus,
The Z - transform of a discrete - time sequence is an analytic
function on an annular domain, so the inverse transform should
do the opposite : transform a function that is analytic on a certain
annulus into a sequence which has that function and that annulus
as its Z - transform. The Cauchy integral formula does just that.

Definition of the Z-Transform (2)

Connection to the Fourier Transform : Evaluating the
Z - transform at points z = e j on the unit circle gives :
Z { x [ n ]} ( e

j n

)=

n =

x [ n ] e jn = F { x [ n ]} ( )

This shows that the the Fourier transform coincides with

the values of the Z - transform on the unit circle in the
complex plane.
Notation : In these notes the frequency variable has been
used as the argument for Fourier transforms (X ( ) ). However,
the book uses the argument e j so that the same function
can be used for both the Fourier transform and Z - transform.
We will use both notations, distinguishing them by context.

Definition of the Z-Transform (3)

Connection to the Fourier Transform (continued) :
The Z - transform extends the Fourier transform as follows :
If x [ n ] is absolutely summable then it will have an (ordinary)
Fourier transform, and the Z - transform may extend this
transform to a ROC that contains the unit circle.
If x [ n ] is not absolutely summable then it may not have a
Fourier transform, but may still have a Z - transform that
is defined on some ROC that does not contain the unit circle.
However, there are some sequences (e.g. x [ n ] = sin ( n ) n)
which do not have Z - transforms, yet their Fourier transform
can be defined in an extended sense .

Definition of the Z-Transform (4)

Example 1 (Shifted Impulse) : x [ n ] = [ n n0 ]

( ROC : 0 < z < )

= z cos ( n arg ( z ) )

Z { x [ n ]} ( z ) = z n0
Re z n0

n0

F { x [ n ]} ( ) = e jn0

Re e jn0 = cos ( n0 )

Im z n0 = z

n0

sin ( n0 arg ( z ) )

Im e jn0 = sin ( n0 )

Real and Imaginary Parts for n0 = 3 :

1.5

1.5

0.5

0.5

-0.5

-0.5

-1

-1

-1.5
2

-1.5
2

2
1

0
0

-1

-1
-2

-2

2
1

0
0

-1

-1
-2

-2

Definition of the Z-Transform (5)

Example 2 (Right - Sided Exponential) : x [ n] = a nu [ n] , a = ei

Z { x [ n ]} ( z ) = a n z n =
n =0

1
1 az 1

( ROC :

1 Re az 1
1
Re
=
2
1
1 az 1 2 Re az 1 + az 1

F { x [ n ]} ( ) =

1
1 ae j

z >= a)
Im az 1
1
Im
=
2
1
1 az 1 2 Re az 1 + az 1

( if 1 > = a )

1 cos ( )
1
Re
=
2
1
1 az 1 2 cos ( ) + a

sin ( )
1
Im
=
2
1
1 az 1 2 cos ( ) + a

Definition of the Z-Transform (6)

Example 2 (Right - Sided Exponential Continued) :
Real and imaginary parts for a = .5

5
2

4
2

-1

0
2
0
1

-2
2
0
1

-1
-1

-1
-1

-2

-2

-2

-2

Definition of the Z-Transform (7)

Example 3 (Left - Sided Exponential) : x [ n] = a nu [ n 1] , a = ei
1

Z { x [ n ]} ( z ) = a z

n n

n =

= 1 an z n = 1
n =0

1 Re az 1
1
Re
=
2
1
1 az 1 2 Re az 1 + az 1

F { x [ n ]} ( ) =

1
1 ae j

1
1
ROC : z < = a )
=
1
1 (
1 a z 1 az

Im az 1
1
Im
=
2
1
1 az 1 2 Re az 1 + az 1

( if 1 < = a )

1 cos ( )
1
=
Re
2
1
1 az 1 2 cos ( ) + a

sin ( )
1
=
Im
2
1
1 az 1 2 cos ( ) + a

Note : All is the same as the right - sided exponential but with a different ROC .

Definition of the Z-Transform (8)

Example 3 (Left - Sided Exponential Continued) :
Real and imaginary parts for a = 2.0

2
0

-2
-4

-6
-8

0
-5

-10
-2

-1

-1

-2

-1

-1
0

0
1
2

-2

1
2

-2

Definition of the Z-Transform (9)

Example 4 (Two - Sided Exponential) : x [ n ] = 2 n u [ n ] + 2n u [ n 1]
2 u [ n]
n

x [ n]

1 z
1
2

( ROC :

z >

1
2

),

1
ROC : z < 2 )
(

1
Z 1 2z
32 z 1
1
1
1

=
ROC
:
(
2 < z < 2)

2
5
1
Z 1 z
1 2z
1 2 z + z
2

2n u [ n 1]

Definition of the Z-Transform (10)

Sequence

[ n m]
a nu [ n]

a nu [ n 1]
na u [ n ]
n

na u [ n 1]
n

Transform
z m
1 (1 az 1 )
1 (1 az 1 )

az
az

1
1

(1 az )
(1 az )

ROC
z >0
z >a
z <a

1 2

z >a

1 2

z <a

r n cos (0 n ) u [ n ]

1 z 1 cos 0
1 2 z 1 cos 0 + z 2

z >r

r n sin (0 n ) u [ n ]

z 1 sin 0
1 2 z 1 cos 0 + z 2

z >r

Definition of the Z-Transform (11)

ROC Properties
The ROC is must be either a disk, the exterior of a disk,
a ring (annulus) or the entire plane
The ROC cannot contain any poles
If x [ n ] has finite length then there can only be a pole at z = 0
The right side of the sequence determines the upper limit
of the ROC and the left side determines the lower limit
An impulse response h [ n ] will be stable if the ROC of its
Z - transform contains the unit circle
If the impulse response h [ n ] is causal then the ROC of its
Z - transform will extend to

Topics

Concepts from Complex Variables

Laurent Expansions
Definition of the Z-Transform
Computing Inverse Z-Transforms
Properties of the Z-Transform

Computing Inverse Z-Transform (1)

Basic Methods to Compute Inverse Z - Transforms :
1. Apply the Cauchy integral formula using the values on
a circle z = r inside the region of convergence. If r = 1
this is equivalen to the inverse Fourier transform.
2. By inspection using the table and linearity. E.g. the two sided exponential example given earlier
3. Power series expansion:
log (1 + z 1 ) = z 1 z 2 2 + z 3 3 +

1 (1 + z 2 ) = 1 z 2 + z 4

Computing Inverse Z-Transform (2)

Partial Fraction Expansion: An important application of
Z - transforms is to study LCCDE systems. The impulse response
sequences of these systems have Z - transforms that are
rational functions, and important system characteristics are
revealed by the placement of roots and poles of these rational
functions.

Synthesis of LCCDE systems may require finding an impulse

response h [ n ] that is the inverse Z - transform of a rational
function H ( z ) that is specified by a list of roots and poles.

Computing Inverse Z-Transform (3)

Partial Fractions Continued: The rational function H [ z ] is usually
expressed as a ratio of polynomials in z -1 rather than z , thus

(1 c z )
M

bM z + + b0 b0
=
H (z) =
M
aN z + + a0 a0

k =1
N

d
z
1
( k )

1
P
z
(
)
b0
=
a0 Q ( z 1 )

k =1

where c1 , , cM are the roots and d1 , , d N are the poles.

Synthetic division may be necessary to find polynomials
q ( z 1 ) = qL z L + + q0 and r ( z 1 ) = rN 1 z N +1 + + r0 such that
P ( z 1 )

Q ( z 1 )

= q ( z 1 ) +

r ( z 1 )

Q ( z 1 )

where the degree of r ( z 1 ) is less than the degree of Q ( z 1 ) .

Computing Inverse Z-Transform (4)

Partial Fractions Continued : The simplest case is when the
denominator has no repeated roots. In that case, let
r ( z 1 )

A1
AN
=
++
.
1
1
1
Q ( z ) (1 d1 z )
(1 d N z )

To find the constant Ak , multiply both sides by (1 d k z 1 ) :

r ( z 1 )(1 d k z 1 )
Q(z

A1 (1 d k z 1 )

(1 d1 z

+ + Ak + +

AN (1 d k z 1 )
1
1

d
z
( N )

and then evaluate at z = d k to get

Ak =

(1 d d )(1 d
1
1 k

r ( d k1 )

1
1
1
d
1

d
d

d
d
)(
)
(
k 1 k
k +1 k
N k )

Computing Inverse Z-Transform (4)

Partial Fractions Continued : The partial fraction expansion
results from substituting the Ak 's into
P ( z 1 )

Q(z

= qL z

A1
AN
+ + q0 +
+ +
1
1
1
1

d
z

d
z
( 1 )
( N )

The inverse Z - transform can then be computed one term at

a time using the linearity property and the table results, giving
b0
h [ n ] = ( q0 [ n ] + + qL [ n L ])
a0
n
b0 d1 u [ n 1] : z < d1
+ A1
+ + AN
n

: z > d1
a0 d1 u [ n ]

d Nn u [ n 1] : z < d N

n
: z > d N
d N u [ n ]

This expression gives the expansions for all possible ROCs

Computing Inverse Z-Transform (5)

1 jz )(1 + jz )
(
z +1
Partial Fractions Example :
=
z z 1
(1 z )(1 + 2 z )
2

3
2

1
2

Step 1: (Synthetic division)

1

q
z
(
) =1
z +2
z +1

= 1+

2
3 1
3 1
2
3 1
z 2 z 1
z 2 z 1
r ( z ) = 2 z + 2
Step 2 : (Find the Partial Fraction Coefficients)
2

r ( z 1 )

3
2
2

3 1
z 1 + 2
+2
A1
A2
2 z
=
=
=
+
1
1
1
1
3 1
1
1
Q ( z ) z 2 z 1 (1 2 z )(1 2 z ) (1 2 z ) (1 2 z 1 )

A1 =

A2 =

3
2
2

r ( z 1 )(1 12 z 1 )

(1

1
2

z 1 )(1 2 z 1 )

r ( z ) ( z 2 z 1 )

1
1 1
1

z
1

2
z
( 2 )(
)

2) + 2
(
=
= 53
(1 2 ( 2 ) )
3
2

z = 12

=
z =2

32 ( 12 ) + 2

(1 ( ) )
1
2

1
2

5
3

Computing Inverse Z-Transform (5)

Partial Fractions Example Continued:
Step 3: (Find the inverse - Transform)
h [ n] = [ n]
( 1 )n u [ n 1] : z < 1 2n u [ n 1] : z < 2
2
2
5
5
+3

n
1 n
1
: z > 2
: z > 2 2 u [ n ]
( 2 ) u [ n ]

Topics

Concepts from Complex Variables

Laurent Expansions
Definition of the Z-Transform
Computing Inverse Z-Transforms
Properties of Z-transforms

Properties of Z-Transform (1)

Linearity:

x [ n] X ( z ) , y [ n] Y ( z )

x [ n] + y [ n] X ( z ) + Y ( z )

( ROC ROCX ROCY )

Z { x [ n ] + y [ n ]} ( z ) = Z { x [ n ]} ( z ) + Z { y [ n ]} ( z )

Time Shift:

x [ n ] X ( z ) x [ n n0 ] z n0 X ( z )

Z { x [ n n0 ]} ( z ) =
=

n =

n =

=z

n0

( ROC = ROCX )

x [ n n0 ] z n
x [n] z

( n + n0 )

x [ n] z

n =

= z n0 Z { x1 [ n ]} ( z )

Summary

The z-transform of a sequence is the Laurent

expansion formed by using the terms of the
sequence as the coefficients.
The z-transform extends the Fourier transform to
regions of the complex plane
Elementary properties of the z-transform, have
been derived
Some simple z-transforms have been evaluated
Evaluation by partial fraction expansion has been
described for evaluating z-transforms of rational
functions.