𝒁 − Transform

MATH- 232
Instructor: Dr. Naila Amir
Book: Advanced Modern Engineering Mathematics (4th Edition) by Glyn James
▪ Chapter: 6
▪ Sections: 6.2, 6.3, 6.4, 6.5
𝒁 −Transform & Inverse 𝒁 −Transform: Definition and notation
The 𝒁 −transform of a sequence {𝑥𝑛 }∞
−∞ is defined in general as:
∞
𝑥𝑛
𝒵{𝑥𝑛 }∞
−∞ = 𝑋 𝑧 = ෍ 𝑛 , (∗)
𝑧
𝑛=−∞
whenever the sum exists and where 𝑧 is a complex variable. The process of taking the
𝑍 −transform of a sequence thus produces a function of a complex variable 𝑧, whose form
depends upon the sequence itself. The symbol 𝒵 denotes the 𝒁 −transform operator; when it
operates on a sequence {𝑥𝑛 } it transforms the latter into the function 𝑋(𝑧) of the complex
variable 𝑧. It is usual to refer to {𝑥𝑛 }, 𝑋(𝑧) as a 𝒁 −transform pair, which is sometimes written
as {𝑥𝑛 } 𝑋(𝑧). From time domain (or space domain, 𝑛 −domain) to the 𝑧 −domain we have:
∞

{𝑥𝑛 }∞
−∞ = 𝒵 −1
𝑋 𝑧 = ෍ 𝑥𝑘 𝛿𝑛−𝑘 𝒵 𝑥𝑛 ∞
−∞ =𝑋 𝑧 .
𝑘=−∞
For a causal sequence, that is 𝑥𝑛 = 0, (𝑛 < 0) the 𝑍 −transform given in (∗) reduces to:
∞
𝑥𝑛
𝒵{𝑥𝑛 }∞
0 = 𝑋 𝑧 =෍ 𝑛 . (∗∗)
𝑧
𝑛=0
 Bilateral vs. Unilateral
◼ Two sided or bilateral 𝑍 −transform is defined as:
∞
𝑥𝑛
𝒵{𝑥𝑛 }∞
−∞ = 𝑋 𝑧 = ෍ 𝑛. (∗)
𝑧
𝑛=−∞

◼ Unilateral 𝑍 − transform
∞
𝑥𝑛
𝒵{𝑥𝑛 }∞
0 = 𝑋 𝑧 = ෍ 𝑛. ∗∗
𝑧
𝑛=0

For the present course we shall be concerned with causal sequences, and so the
definition given in (∗∗) will be the one that we shall use henceforth. We shall therefore
from now on take {𝑥𝑛 } to denote {𝑥𝑛 }∞ 0 . Non-causal sequences, however, are of
importance, and arise particularly in the field of digital image processing, among others.
Example:
Determine
𝑧
𝒵 −1 .
𝑧−2
Solution:
In our previous example we determined the 𝑍 −transform pair.
𝑛
𝑧
2 .
𝑧−2
Thus, if
∞
𝑛
𝑛
2 𝑧
𝒵 2 =𝑋 𝑧 =෍ 𝑛 = 𝑧 >2 ,
𝑧 𝑧−2
𝑛=0
then
−1
𝑧
𝒵 = 2𝑛 ,
𝑧−2
𝑛 𝑧
so that 2 , represents a 𝑍 −transform pair.
𝑧−2
Example:
Determine the 𝑍 −transform of the sequence: 𝑥𝑛 = 𝑛 , (𝑛 ≥ 0)
Solution: Using 𝑖𝑖 we have:
𝑛
𝑧
𝒵 𝑎 =𝑋 𝑧 = 𝑧 > |𝑎| .
𝑧−𝑎
Differentiating both sides of above equation with respect to 𝑎, which for the moment we
regard as a parameter, we get:

𝑑 𝑑𝑎 𝑛 𝑑 𝑧
𝑛
𝒵 𝑎 =𝒵 = ,
𝑑𝑎 𝑑𝑎 𝑑𝑎 𝑧 − 𝑎
leading to
𝑧
𝒵 𝑛𝑎𝑛−1 = 𝑧 > |𝑎| .
(𝑧 − 𝑎)2
In particular, for 𝑎 = 1, the last equation takes the form:
𝑧
𝒵 𝑛 = 2
𝑧 >1 .
(𝑧 − 1)
Example:
Determine the 𝑍 −transform of the sequence: 2𝑛 = 0,2,4,6, … .
Solution: We know that
∞
𝑛 𝑧
𝒵 𝑛 = 𝒵 0,1,2,3, … = ෍ 𝑛 = 𝑧 >1 .
𝑧 (𝑧 − 1)2
𝑛=0
Now
∞ ∞
2𝑛 𝑛
𝒵 2𝑛 = 𝒵 0,2,4,6, … = ෍ 𝑛 = 2 ෍ 𝑛 = 2𝒵 𝑛 .
𝑧 𝑧
𝑛=0 𝑛=0
Thus,
2𝑧
𝒵 2𝑛 = 𝑧 >1 .
(𝑧 − 1)2
Example: unit pulse or impulse sequence
Determine the 𝑍 −transform of the unit pulse or impulse sequence: 𝛿𝑛 = 1,0,0,0, … .
Solution:
The unit impulse sequence is “a sequence of discrete samples having unit magnitude at origin
and zero magnitude at all other sample instants”.
1; 𝑛 = 0
𝛿𝑛 = 𝛿 𝑛 = ቊ
0; 𝑛 ≠ 0
Now
∞
𝛿𝑛 1
𝒵 𝛿𝑛 = 𝒵 1,0,0,0, … = ෍ 𝑛 = 0 = 1
𝑧 𝑧
𝑛=0
The ROC is all 𝑧, where 𝑧 is a complex variable.
Example: unit step function
Determine the 𝑍 −transform of the unit step function.
Solution:
The unit step function is defined as:
1; 𝑛 ≥ 0
𝑢 𝑛 =ቊ
0; 𝑛 < 0
Now
∞ ∞
∞ 𝑢𝑛 1 1 𝑧
𝒵 𝑢 𝑛 −∞ = ෍ 𝑛
=෍ 𝑛 = −1
= valid for 𝑧 > 1
𝑧 𝑧 1−𝑧 𝑧−1
𝑛=−∞ 𝑛=0
The ROC is 𝑧 > 1.
Note: We can generalize above result to determine the 𝑍 −transform of 𝑎𝑛 𝑢 𝑛 , where 𝑎 is a
real or complex constant. Thus,
∞
𝑎 𝑛 1 𝑧
𝑛
𝒵 𝑎 𝑢[𝑛] = 𝑋 𝑧 = ෍ 𝑛 = = valid for 𝑧 > |𝑎| .
𝑧 1 − 𝑎𝑧 −1 𝑧−𝑎
𝑛=0
Example: reverse step function
Determine the 𝑍 −transform of the reverse step function: −𝑢 −𝑛 − 1 .
Solution: The reverse step function is defined as:
−1; 𝑛 ≤ −1
−𝑢 −𝑛 − 1 = ቊ
0; 𝑛 > −1
Now
∞ −1 ∞
∞
−𝑢 −𝑛 − 1 (−1) (−1)
𝒵 −𝑢 −𝑛 − 1 −∞ = ෍ 𝑛 = ෍ 𝑛 = ෍ −𝑡
𝑧 𝑧 𝑧
𝑛=−∞ 𝑛=−∞ 𝑡=1
∞ ∞
−𝑧 𝑧 1
= ෍ −(𝑧)𝑘+1 = −𝑧 ෍ 𝑧𝑘 = = = valid for 𝑧 < 1 .
1 − 𝑧 𝑧 − 1 1 − 𝑧 −1
𝑘=0 𝑘=0
The ROC is 𝑧 < 1.
Note: We can generalize above result to determine the 𝑍 −transform of −𝑎𝑛 𝑢 −𝑛 − 1 ,
where 𝑎 is a real or complex constant. Thus,
𝑛
𝑧 1
𝒵 −𝑎 𝑢 −𝑛 − 1 = = −1
valid for 𝑧 < |𝑎| .
𝑧 − 𝑎 1 − 𝑎𝑧
𝒁 −transform of some
important functions
Sampling:
Sequences are often generated in engineering applications through the sampling of continuous-time
signals, described by functions 𝑓(𝑡) of a continuous-time variable 𝑡. Here we shall not discuss the
means by which a signal is sampled, but merely suppose this to be possible in idealized form.
Figure illustrates the idealized sampling process in which a continuous-
time signal 𝑓(𝑡) is sampled instantaneously and perfectly at uniform
intervals 𝑇, the sampling interval. The idealized sampling process
generates the sequence:
𝑓 𝑛𝑇 = 𝑓 0 , 𝑓 𝑇 , 𝑓 2𝑇 , . . . , 𝑓 𝑘𝑇 , . . . .
We can take the 𝑍 −transform of the sequence as:
∞
𝑓(𝑛𝑇)
𝒵 𝑓(𝑛𝑇) = ෍ .
𝑧𝑛
𝑛=0 Sampling of a continuous-time signal.
whenever the series converges.
Example:
The signal 𝑓(𝑡) = 𝑒 −𝑡 𝑢(𝑡) is sampled at intervals 𝑇. What is the 𝑍 − transform of the resulting
sequence of samples?
Solution: Sampling the causal function 𝑓(𝑡) generates the sequence:

{𝑓(𝑛𝑇)} = { 𝑓(0), 𝑓(𝑇), 𝑓(2𝑇), . . . , 𝑓(𝑘𝑇), . . . }

= 1, 𝑒 −𝑇 , 𝑒 −2𝑇 , 𝑒 −3𝑇 , . . . , 𝑒 −𝑘𝑇 , . . . .

Thus, the 𝑍 −transform is given as:

∞ ∞ ∞ 𝑛
𝑓(𝑛𝑇) 𝑒 −𝑛𝑇 1 1 𝑧
𝒵 𝑓(𝑛𝑇) = ෍ =෍ 𝑛 =෍ = = , 𝑧 > 𝑒 −𝑇 .
𝑧𝑛 𝑧 𝑧𝑒 𝑇 1 − (𝑧𝑒 𝑇 )−1 𝑧 − 𝑒 −𝑇
𝑛=0 𝑛=0 𝑛=0
Zero and Pole Plot (Pole-Zero Plot)
Recall that the 𝑍 −transform of the sequence: 𝑎𝑛 𝑢 𝑛 , where 𝑎 is a real or complex constant
and 𝑢 𝑛 is the unit step function defined as:
1; 𝑛 ≥ 0
𝑢 𝑛 =ቊ
0; 𝑛 < 0
is given by:
∞
𝑎 𝑛 1 𝑧
𝑛
𝒵 𝑎 𝑢[𝑛] = 𝑋 𝑧 = ෍ 𝑛 = = valid for 𝑧 > |𝑎| . (𝑖)
𝑧 1 − 𝑎𝑧 −1 𝑧−𝑎
𝑛=0

For the present case the ROC is 𝑧 > |𝑎|.

Note that 𝑋(𝑧) is a rational function of 𝑧. Consequently, just as with
rational Laplace transforms, it can be characterized by its zeros (the
roots of the numerator polynomial) and its poles (the roots of the
denominator polynomial). From (𝑖) we see that there is one zero at
𝑧 = 0 and one pole at 𝑧 = 𝑎. The ROC and the pole-zero plot for this
example is shown in the accompanying figure.
Example:
the 𝑍 −transform of −𝑎𝑛 𝑢 −𝑛 − 1 , where 𝑎 is a real or complex constant is:
𝑛
1 𝑧
𝒵 −𝑎 𝑢 −𝑛 − 1 = −1
= valid for 𝑧 < |𝑎|
1 − 𝑎𝑧 𝑧−𝑎

There is one zero at 𝑧 = 0 and one pole at 𝑧 = 𝑎.

 Properties of the 𝒁 −transform

Note: 𝑥𝑛 = 𝑥[𝑛]
The first shift property (delaying)

The second shift property (advancing)

Solution:
Discrete-Time Systems and Difference Equations
▪ the Laplace transform technique is used to solve differential equations and it provides us
with a way of characterizing a continuous-time system.
▪ In fact, much could be deduced concerning the behaviour of the system and its properties
by examining its transform-domain representation, without looking for specific time-
domain responses at all.
▪ A linear discrete-time system can be modeled by means of a difference equation and
𝑍 − transform plays an analogous role to the Laplace transform for such systems, by
providing a transform-domain representation of the system.
▪ Difference equations arise in a variety of ways, sometimes from the direct modelling of
systems in discrete time or as an approximation to a differential equation describing the
behaviour of a system modelled as a continuous-time system.
▪ The 𝑍 −transform behaves much like the Laplace transform and can be applied to
difference equations to produce frequency and time domain responses. In other words,
the 𝑍 −transform technique can be used to determine solutions of difference equations.
Difference Equation
▪ Suppose that a sequence of observations {𝑥𝑘 } is being recorded and we receive observation 𝑥𝑘
at (time) step or index 𝑘. We might attempt to process this sequence of observations {𝑥𝑘 }
using the discrete-time feedback system as shown in figure.

Discrete time signal

processing system.

▪ At time step 𝑘 the observation 𝑥𝑘 enters the system as an input, and, after combination with
the ‘feedback’ signal at the summing junction 𝑆, proceeds to the block labelled 𝐷. This block is
a unit delay block, and its function is to hold its input signal until the ‘clock’ advances one step,
to step 𝑘 + 1.
▪ At this time, the input signal is passed without alteration to become the signal 𝑦𝑘+1 , which is
the (𝑘 + 1)𝑡ℎ member of the output sequence 𝑦𝑘 .
▪ At the same time this signal is fed back through a scaling block of amplitude 𝛼 to the summing
junction 𝑆.
Difference Equations
▪ This process is instantaneous, and at 𝑆 the feedback signal is subtracted from the next input
observation 𝑥𝑘+1 to provide the next input to the delay block 𝐷.
▪ The process then repeats at each ‘clock’ step.
▪ To analyse the system, let {𝑟𝑘 } denote the sequence of input signals to 𝐷; then, owing to the
delay action of 𝐷, we have:
𝑦𝑘+1 = 𝑟𝑘 .
▪ Also, from the feedback action we get:
𝑟𝑘 = 𝑥𝑘 − 𝛼𝑦𝑘 ,
where 𝛼 is the feedback gain.
▪ Combination of the two expressions gives:
𝑦𝑘+1 = 𝑥𝑘 − 𝛼𝑦𝑘 or 𝑦𝑘+1 + 𝛼𝑦𝑘 = 𝑥𝑘 . (∗)
▪ This is a first-order difference equation, and it relates adjacent members of the sequence {𝑦𝑘 }
to each other and to the input sequence {𝑥𝑘 } .
▪ A solution of the difference equation (∗) is a formula for 𝑦𝑘 , the general term of the output
sequence {𝑦𝑘 }, and this will depend on 𝑘 and the input sequence 𝑥𝑘 . For the present case
solution will depend on feedback gain 𝛼 as well.
Example:
The difference equation that represents the system shown in the accompanying figure, is given as:
𝑦𝑘+2 + 𝑎𝑦𝑘+1 − 𝑏𝑦𝑘 = 𝑥𝑘 𝑘 ≥ 0 ,
where {𝑥𝑘 } and {𝑦𝑘 } are the input and output sequences respectively, 𝐷 is the unit delay block and
𝑎 and 𝑏 are constant feedback gains. If 𝑎 = 1, 𝑏 = 2 and the input sequence {𝑥𝑘 } is the unit step
sequence, then solve the resulting difference equation provided: 𝑦0 = 0 and 𝑦1 = 1.
Solution:
Solution:
Solution:
 Practice Problems:
1. Find the 𝑍 −transform and sketch the pole-zero plot with the ROC for each of the following
sequences:

2. The continuous-time function 𝑓 𝑡 = cos 𝜔𝑡 𝑢(𝑡), where 𝜔 is a constant and 𝑢(𝑡) is unit
step function, is sampled in the idealized sense at intervals 𝑇 to generate the sequence
{cos 𝑛𝜔𝑇}. Determine the 𝑍 − transform of the sequence.

3. The causal sequence {𝑥𝑛 } is generated by: 𝑥𝑛 = 1/2 𝑛 𝑛 ≥ 0 . Determine the

𝑍 −transform of the shifted sequence {𝑥𝑛−2 }.

𝟐𝒛+𝟏
4. Determine 𝓩−𝟏 .
(𝒛+𝟏)(𝒛−𝟑)
 Practice Questions
▪ Determine the 𝑍 −transform: 𝑿(𝒛) of the sequence:
𝒏
−𝟏
𝒙𝒏 =𝟑 𝒖 𝒏 − 𝟏 − 𝟐 𝟑 𝒏 𝒖 −𝒏 − 𝟏 .
𝟐
Moreover, sketch the pole-zero plot with the ROC (region of convergence) by clearly
identifying the ROC, poles, and zeros of 𝑋 𝑧 .

▪ The dynamics of a discrete-time system are determined by the difference equation:

𝒚𝒌+𝟐 − 𝟔𝒚𝒌+𝟏 + 𝟗𝒚𝒌 = 𝒙𝒌 ; 𝒌≥𝟎 ,
where {𝒙𝒌 } = {𝜹𝒌 }. Determine the output of the system {𝒚𝒌 } provided 𝒚𝟎 = 𝟎 and
𝒚𝟏 = 𝟎. (For the present case {𝛿𝑘 } = 𝛿[𝑘] is the impulse sequence, 𝑦𝑘 = 𝑦[𝑘] and
𝑥𝑘 = 𝑥 𝑘 .
Practice Problems: Find the 𝑍 −transform and sketch the pole-zero plot with the ROC for
each of the following sequences:

Solution: (𝒃) We know that:

𝑧
𝒵 𝑎𝑛 𝑢[𝑛] = valid for 𝑧 > |𝑎| (𝑖)
𝑧−𝑎
and
𝑧
𝒵−𝑏 𝑛 𝑢 −𝑛 − 1 = valid for 𝑧 < |𝑏| (𝑖𝑖)
𝑧−𝑏
Thus, by using linearity property and eqs. (𝑖) and (𝑖𝑖), with 𝑎 = 1/3 and 𝑏 = 1/2, we have:
𝑛 𝑛
1 1
𝒵 𝑥[𝑛] = 𝑋 𝑧 = 𝒵 𝑢 𝑛 −𝒵 − 𝑢 −𝑛 − 1
3 2
𝑧 𝑧 −1 𝑧 1 1
= − = . valid for < 𝑧 <
𝑧 − 1/3 𝑧 − 1 6 𝑧− 1 1 3 2
2 𝑧 −
2 3
Solution: 𝒃
Thus,
−1 𝑧
𝒵 𝑥[𝑛] = 𝑋 𝑧 = .
6 (𝑧 − 1/2)(𝑧 − 1/3)
For the present case 𝑋(𝑧) has one zero at 𝑧 = 0 and two poles at 𝑧 = 1/2 and 𝑧 = 1/3 and
1 1
that the ROC is: < 𝑧 < .
3 2

One zero at 𝑧 = 0 and

two poles at 𝑧 = 1/2 and 𝑧 = 1/3.

1 1
ROC: < 𝑧 <
3 2 ROC and pole-zero plot
 Practice Questions
Book: Advanced Modern Engineering Mathematics (4th Edition) by Glyn James
▪ Exercise: 6.2
Q#1-2
▪ Exercise: 6.3
Q#3–8
▪ Exercise: 6.4
Q # 11 - 13
▪ Exercise: 6.5
Q # 14 - 19