Chapter 6

Z-Transform

清大電機系林嘉文
[email protected]
03 5731152
03-5731152
Original PowerPoint slides prepared by S. K. Mitra 4-1-1
© The McGraw-Hill Companies, Inc., 2007
 z Transform
z-Transform
• The DTFT provides a frequency-domain representation of
discrete-time signals and LTI discrete-time systems
• Because of the convergence condition, in many cases, the
DTFT of a sequence may not exist, thereby making it
impossible to make use of such frequency-domain
characterization in these cases
• A generalization of the DTFT defined by

leads to the z-transform

• z-transform may exist for many sequences for which the
DTFT does not exist
• Use of z-transform permits simple algebraic manipulations
4-1-2
Original PowerPoint slides prepared by S. K. Mitra
© The McGraw-Hill Companies, Inc., 2007
 z Transform
z-Transform
• For a given sequence g[n], its z-transform G(z) is defined
as:

where z = Re(z) + j Im(z) is a complex variable

• If we let z = r ejω, then the z-transform reduces to

• The above can be interpreted as the DTFT of the

modified sequence {g[n]r−n}
• For r = 1 (i.e., |z| = 1), zz-transform
transform reduces to its DTFT,
provided the latter exists
• The contour ||z|| = 1 is a circle in the z-plane
p of unityy
radius and is called the unit circle 4-1-3
Original PowerPoint slides prepared by S. K. Mitra
© The McGraw-Hill Companies, Inc., 2007
 z Transform
z-Transform
• Like the DTFT, there are conditions on the convergence
of the infinite series

• For a given sequence, the set R of values of z for which

its z-transform
z transform converges is called the region of
convergence (ROC)
• From our earlier discussion on the uniform convergence
of the DTFT, it follows that the series

• converges if {g[n]r−n} is absolutely summable, i.e., if

Original PowerPoint slides prepared by S. K. Mitra 4-1-4

© The McGraw-Hill Companies, Inc., 2007
 z Transform
z-Transform
• In general, the ROC R of a z-transform of a sequence
g[n] is an annular region of the z-plane:

where
• Note: The z-transform is a form of a Laurent series and
i an analytic
is l ti ffunction
ti att every point
i t in
i th
the ROC
• Example – Determine the z-Transform X(z) of the causal
sequence x[n] = αn μ[n] and its ROC
• Now
• The above power series converges to

• ROC is the annular region |z| > |α|

Original PowerPoint slides prepared by S. K. Mitra 4-1-5
© The McGraw-Hill Companies, Inc., 2007
 z Transform
z-Transform
• Example – Determine the z-Transform μ(z) of the unit
step function μ[n] can be obtained from

by setting α = 1:

• Note: The unit step function μ[n] is not absolutely

summable, and hence its DTFT does not converge
uniformlyy

Original PowerPoint slides prepared by S. K. Mitra 4-1-6

© The McGraw-Hill Companies, Inc., 2007
 z Transform
z-Transform
• Example – Consider the anti-causal sequence
y[n] = −αnμ[−n −1]
• Its zz-transform
transform is given by

• ROC is the annular region |z| < |α|

Original PowerPoint slides prepared by S. K. Mitra 4-1-7
© The McGraw-Hill Companies, Inc., 2007
 z Transform
z-Transform
• Note: the z-Transforms of two sequences αnμ[n] and
−αnμ[−n −1] are identical even though the two parent
sequences are different
• Only way a unique sequence can be associated with a z-
transform is by specifying its ROC
• The
Th DTFT G(e G( jω) off a sequence g[n]
[ ] converges uniformly
if l
if and only if the ROC of the z-transform G(z) of g[n]
includes the unit circle
• The existence of the DTFT does not always imply the
existence of the zz-transform
transform

Original PowerPoint slides prepared by S. K. Mitra 4-1-8

© The McGraw-Hill Companies, Inc., 2007
 z Transform
z-Transform
• Example – the finite energy sequence

has a DTFT given by

which converges in the mean-square sense

• However, hLP[n] does not have a z-transform as it is not
absolutely summable for any value of r

Original PowerPoint slides prepared by S. K. Mitra 4-1-9

© The McGraw-Hill Companies, Inc., 2007
 Commonly Used z-Transform
z Transform Pairs

Original PowerPoint slides prepared by S. K. Mitra 4-1-10

© The McGraw-Hill Companies, Inc., 2007
 Rational z-Transform
z Transform
• In tthe
e case o
of LTI d
discrete-time
sc ete t e syste
systemss we
eaaree
concerned with in this course, all pertinent z-transforms
are rational functions of z−1
• That is, they are ratios of two polynomials in z−1

• The degree of the numerator polynomial P(z) is M and

the degree of the denominator polynomial D(z) is N
• An alternate representation of a rational z-transform is as
a ratio of two polynomials in z:

Original PowerPoint slides prepared by S. K. Mitra 4-1-11

© The McGraw-Hill Companies, Inc., 2007
 Rational z-Transform
z Transform
• A rational
at o a z-transform
t a s o cacan be a
alternately
te ate y written
tte in
factored form as

• At a root z = ξl of the numerator p

polynomial
y G(ξ
(ξl), and as
a result, these values of z are known as the zeros of
G(z)
• At a root z = λl of the denominator polynomial G(λl) → ∞,
and as a result, these values of z are known as the
l off G(z)
poles G( )
Original PowerPoint slides prepared by S. K. Mitra 4-1-12
© The McGraw-Hill Companies, Inc., 2007
 Rational z-Transform
z Transform
• Consider
Co s de

• Note G(z) has M finite zeros and N finite poles

• If N > M there are additional N − M zeros at z = 0 (the
origin in the z-plane)
• If N < M there are additional M − N p
poles at z = 0 ((the
origin in the z-plane)

Original PowerPoint slides prepared by S. K. Mitra 4-1-13

© The McGraw-Hill Companies, Inc., 2007
 Rational z-Transform
z Transform
• Example
a p e – tthe
e z-transform
ta so

has a zero at z = 0 and a pole at z = 1

Original PowerPoint slides prepared by S. K. Mitra 4-1-14

© The McGraw-Hill Companies, Inc., 2007
 Rational z-Transform
z Transform
• Apphysical
ys ca interpretation
te p etat o o of tthe
e co
concepts
cepts o
of po
poles
es a
and
d zeros
e os
can be given by plotting the log-magnitude 20log10|G(z)| for

• The magnitude plot

exhibits very large peaks
around the poles of G(z)
(z = 0.4 ± j 0.6928)
• It also exhibits very
narrow and deep wells
around the location of
the zeros (z = 1.2
1 2 ± j 1.2)
1 2)
Original PowerPoint slides prepared by S. K. Mitra 4-1-15
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• ROC
OC o of a z-transform
t a s o issa
an important
po ta t co
concept
cept
• Without the knowledge of the ROC, there is no unique
relationshipp between a sequence
q and its z-transform
• ⇒The z-transform must always be specified with its ROC
• Moreover,, if the ROC of a z-transform includes the unit
circle, the DTFT of the sequence is obtained by simply
evaluating the z-transform on the unit circle
• There is a relationship between the ROC of the z-
transform of the impulse response of a causal LTI
discrete time system and its BIBO stability
discrete-time
• The ROC of a rational z-transform is bounded by the
locations of its poles
Original PowerPoint slides prepared by S. K. Mitra 4-1-16
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• Example – the z-transform H(z) of the sequence h[n] =
(−0.6)nμ[n] is given by

|z| < 0.6

• Here the ROC is just outside the circle going through the
point z = −0.6
• A sequence can be one of the following types: finite-
length, right-sided, left-sided and two-sided
• The
Th ROC d dependsd on th the ttype off th
the sequence off iinterest
t t
Original PowerPoint slides prepared by S. K. Mitra 4-1-17
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• Example – Consider a finite-length sequence g[n] defined
for −M ≤ n ≤ N, where M and N are non-negative integers
and |g[n]| < ∞
• Its z-transform is given by

• Note:
o e G(
G(z)) has
as M po
poles at z = ∞ a
es a and
d N po
poles
es a
at z = 0
• As can be seen from the expression for G(z), the z-
transform of a finite-length
g bounded sequence
q converges
g
everywhere in the z-plane except possibly at z = 0 and/or
at z = ∞

Original PowerPoint slides prepared by S. K. Mitra 4-1-18

© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• Example – A right-sided sequence with nonzero sample
values for n ≥ 0 is sometimes called a causal sequence
• Consider a causal sequence u1[n], with its z-transform
given below

• It can be shown that U1(z) converges exterior to a circle |z|

= R1, including the point z = ∞
• On the other hand, a right-sided sequence u2[n] with
nonzero sample values only for n ≥ − M with M
nonnegati e has a z-transform
nonnegative transform U2(z) ith M poles at z = ∞
( ) with
• The ROC of U2(z) is exterior to a circle |z| = R4, excluding
the point z = ∞
Original PowerPoint slides prepared by S. K. Mitra 4-1-19
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• Example – A left-sided sequence with nonzero sample
values for n ≤ 0 is sometimes called a anti-causal
sequence
• Consider a causal sequence v1[n], with its z-transform
given below

• It can be shown that V1(z) converges interior to a circle |z|

= R3, including
i l di th the point i tz=0
• On the other hand, a right-sided sequence V2[n] with
nonzero sample values only for n ≤ N with N nonnegative
has a z-transform V2(z) with N poles at z = 0
• The ROC of V2(z) is interior to a circle |z| = R4, , excluding
the point
Original PowerPoint z = 0 by S. K. Mitra
slides prepared 4-1-20
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• Example – The z-transform of a two-sided sequence w[n]
can be expressed as

• The first term on the RHS,, can be interpreted

p as the z-
transform of a right-sided sequence and it thus converges
exterior to the circle |z| = R5
• The second term on the RHS, can be interpreted as the z-
transform of a left-sided sequence and it thus converges
interior to the circle |z| = R6
• If R5 < R6, there is an overlapping ROC: R5 < |z| < R6
• If R5 > R6, there
th iis no overlap
l (th
(the z-transform
t f do
d nott exist)
i t)
Original PowerPoint slides prepared by S. K. Mitra 4-1-21
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• Example – The z-transform of a two-sided sequence w[n]
can be expressed as
u[n] = αn
where α can be either real or complex
• Its z-transform is g
given by
y

• The first term on the RHS converges for |z| > |α| , whereas
the second term converges
g ||z|| < |α|
| |
• There is no overlap between these two regions
• Hence,, the z-transform of u[n]
[ ] = αn does not exist
Original PowerPoint slides prepared by S. K. Mitra 4-1-22
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• The ROC of a rational z-transform cannot contain any
poles and is bounded by the poles
• To show that the z-transform is bounded by the poles,
assume that the z-transform X(z) has simple poles at z = α
and z = β
• Assuming
A i th thatt the
th corresponding
di sequence x[n] [ ] is
i a right-
i ht
sided sequence, x[n] has the form
x[n] = (r1αn + r2βn) μ[n − N0],
] |α| < |β|
where N0 is a positive or negative integer
• Now,
Now the z-transform
z transform of the right
right-sided
sided sequence γnμ[n −
N0] exists if
for some z
Original PowerPoint slides prepared by S. K. Mitra 4-1-23
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• The following condition holds for |z| > |γ| but not for |z| ≤ |γ|

• Therefore, the z-transform of

x[n]
[ ] = (r
( 1αn + r2βn) μ[
μ[n − N0], |α|
| | < |β|
has an ROC defined by |β| < |z| ≤ ∞
• Likewise,, the z-transform of a left-sided sequence
q
x[n] = (r1αn + r2βn) μ[− n − N0], |α| < |β|
has an ROC defined by 0 ≤ |z| < |α|
• Finally, for a two-sided sequence, some of the poles
contribute to terms in the p parent sequence
q for n < 0 and
the other poles contribute to terms n ≥ 0 4-1-24
Original PowerPoint slides prepared by S. K. Mitra
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• The ROC is thus bounded on the outside by the pole with
the smallest magnitude that contributes for n < 0 and on
the inside by the pole with the largest magnitude that
contributes
t ib t for
f n≥0
• There are three possible ROCs of a rational z-transform
with poles at z = α and z = β (|α| < |β|)

Original PowerPoint slides prepared by S. K. Mitra 4-1-25

© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• In general, if the rational z-transform has N poles with R
distinct magnitudes, then it has R + 1 ROCs
• There are distinct sequences with the same z-transform
• Hence, a rational z-transform with a specified ROC has a
unique sequence as its inverse z-transform
• MATLAB [z,p,k] = tf2zp(num,den) determines the zeros,
poles, and the gain constant of a rational z-transform with
the numerator coefficients specified by num and the
denominator coefficients specified by den
• [num,den] zp2tf(z p k) implements the reverse process
[num den] = zp2tf(z,p,k)
• The factored form of the z-transform can be obtained using
sos = zp2sos(z,p,k)
Original PowerPoint slides prepared by S. K. Mitra 4-1-26
© The McGraw-Hill Companies, Inc., 2007
 ROC of a Rational z-Transform
z Transform
• The pole-zero plot is determined using the function zplane
• The z-transform can be either described in terms of its
zeros and poles: zplane(zeros,poles) or, in terms of its
numerator and denominator coefficients zplane(num,den)
• Example – The pole-zero plot of

obtained using MATLAB

Original PowerPoint slides prepared by S. K. Mitra 4-1-27

© The McGraw-Hill Companies, Inc., 2007
 Inverse z-Transform
z Transform
• General Expression: Recall that, for z = re−jω, the z-
transform G(z) given by

is merely the DTFT of the modified sequence g[n]r−n

• Accordingly, the inverse DTFT is thus given by

• B
By making
ki a change
h off variable
i bl z = re−jω
j , the
th previous
i
equation can be converted into a contour integral given by

where C C′ is a counterclockwise contour of integration

defined by |z| = r
Original PowerPoint slides prepared by S. K. Mitra 4-1-28
© The McGraw-Hill Companies, Inc., 2007
 Inverse z-Transform
z Transform
• But the integral remains unchanged when it is replaced with
any contour
t C encircling
i li th the point
i t z = 0 iin th
the ROC off G(z)
G( )
• The contour integral can be evaluated using the Cauchy’s
resid e theorem resulting
residue res lting in

• The above equation needs to be evaluated at all values of n

and is not pursued here
• A rational z-transform G(z) with a causal inverse transform
g[n] has an ROC that is exterior to a circle
• It’s more convenient to express G(z) in a partial-fraction
expansion
p form and then determine g[g[n]] byy summing
g the
inverse transform of the individual terms in the expansion
4-1-29
Original PowerPoint slides prepared by S. K. Mitra
© The McGraw-Hill Companies, Inc., 2007
 Inverse z-Transform by Partial-
Fraction Expansion
• A rational G(z) can be expressed as

• If M ≥ N then G(z) can be re-expressed as

where the degree of P1(z) is less than N

• The rational function P1(z)/D(z) is called a proper fraction
• Example – Consider

• By long division we arrive at

Original PowerPoint slides prepared by S. K. Mitra 4-1-30

© The McGraw-Hill Companies, Inc., 2007
 Inverse z-Transform by Partial-
Fraction Expansion
• Simple Poles: In most practical cases, the rational z-
transform of interest G(z) is a proper fraction with simple
poles
• Let the poles of G(z) be at z = λk 1 ≤ k ≤ N
• A partial-fraction expansion of G(z) is then of the form

• The constants in the partial

partial-fraction
fraction expansion are called
the residues and are given by

• Each term of the sum in partial-fraction expansion has an

ROC given by |z| > |λl| and, thus has an inverse transform
of the form ρ (λ )nμ[n]
l lby S. K. Mitra
Original PowerPoint slides prepared 4-1-31
© The McGraw-Hill Companies, Inc., 2007
 Inverse z-Transform by Partial-
Fraction Expansion
• Therefore,
e e o e, the
t e inverse
e se ttransform
a s o g[g[n]] o
of G(
G(z)) is
sggiven
e by

• N
Note:
t The
Th above
b approachh with
ith a slight
li ht modification
difi ti can
also be used to determine the inverse of a rational z-
transform of a non-causal
non causal sequence
• Example - Let the z-transform H(z) of a causal sequence
h[n]
[ ] be g
given
e by

• A partial-fraction expansion of H(z) is then of the form

Original PowerPoint slides prepared by S. K. Mitra 4-1-32

© The McGraw-Hill Companies, Inc., 2007
 Inverse z-Transform by Partial-
Fraction Expansion
• Now
o

• Hence

• The inverse transform of the above is therefore given by

Original PowerPoint slides prepared by S. K. Mitra 4-1-33

© The McGraw-Hill Companies, Inc., 2007
 Inverse z-Transform by Partial-
Fraction Expansion
• Multiple o es If G(
u t p e Poles: G(z)) has
as multiple
u t p e po
poles,
es, tthe
e pa
partial-
ta
fraction expansion is of slightly different form
• Let the p
pole at z = ν be of multiplicity
p y L and the remaining gN
− L poles be simple and at z = λ, 1 ≤ l ≤ N − L
• Then the partial-fraction expansion of G(z) is of the form

where the constants are computed using

1≤i≤L

• The residues ρl are calculated as before

Original PowerPoint slides prepared by S. K. Mitra 4-1-34
© The McGraw-Hill Companies, Inc., 2007
 Inverse z-Transform via Long
Division
• Thee z-transform
t a s o G( G(z)) of
o a causa
causal sequence
seque ce {g[n]}
{g[ ]} ca
can be
expanded in a power series in z−1
• In the series expansion,
p the coefficient multiplying
p y g the term
z−n is then the n-th sample g[n]
• For a rational z-transform expressed as a ratio of
polynomials in z−1, the power series expansion can be
obtained by long division
• Example
E l - Consider
C id

• Long division of the numerator by the denominator yields

H(z) =1+1.6 z−1 − 0.52 z−2 + 0.4 z−3 − 0.2224 z−4 + …
• As a result, {h[n]} = {1 1.6 − 0.52 0.4 −0.2224 ....}, n ≥4-1-35
0
Original PowerPoint slides prepared by S. K. Mitra
© The McGraw-Hill Companies, Inc., 2007
 z Transform Properties
z-Transform

Original PowerPoint slides prepared by S. K. Mitra 4-1-36

© The McGraw-Hill Companies, Inc., 2007
 z Transform Properties
z-Transform
• Example
a p e - CoConsider
s de tthe
e ttwo-sided
o s ded sequence
seque ce
v[n] = αnμ[n] − βnμ[−n −1]
• Let x[n] = αnμ[n] and y[n] = − βnμ[
μ[−n
n −1]
1] with X(z) and Y(z)
denoting, respectively, their z-transforms
• Now
and

• Using the linearity property we arrive at

• The ROC of V(z) is given by the overlap regions of |z| > |α|
and
d ||z|| < |β|
Original PowerPoint slides prepared by S. K. Mitra 4-1-37
© The McGraw-Hill Companies, Inc., 2007
 z Transform Properties
z-Transform
• Example
a p e - Determine
ete e tthe
e z-transform
ta so a andd its
ts ROC
OC o
of tthe
e
causal sequence
v[n]
[ ] = rn ((cosωon)μ[n]
)μ[ ]
• We can express x[n] = v[n] + v*[n] where

• The z-transform of v[n] is given by

• Using the conjugation property we obtain the zz-transform

transform of
v*[n] as
||z|| > ||α||
Original PowerPoint slides prepared by S. K. Mitra 4-1-38
© The McGraw-Hill Companies, Inc., 2007
 z Transform Properties
z-Transform
• Us
Using
g tthe
e linearity
ea ty pproperty
ope ty we
e get

• or,

• Example - Determine the zz-transform

transform Y(z) and the ROC of
the sequence
y[n] = (n + 1)αnμ[n]
• We can write y[n] = n x[n] + x[n] where x[n] = αnμ[n]

Original PowerPoint slides prepared by S. K. Mitra 4-1-39

© The McGraw-Hill Companies, Inc., 2007
 z Transform Properties
z-Transform
• Now,
o , the
t e z-transform
t a s o X(z)
( )o
of x[n]
[ ] = αnμ[
μ[n]] iss given
g e by

• Using the differentiation property, we arrive at the z-

[ ] as
transform of nx[n]

• Using the linearity property we finally obtain

Original PowerPoint slides prepared by S. K. Mitra 4-1-40

© The McGraw-Hill Companies, Inc., 2007
 LTI Discrete-Time Systems in the
Transform Domain
• An LTI ddiscrete-time
sc ete t e system
syste iss co
completely
p ete y ccharacterized
a acte ed in
the time-domain by its impulse response sequence {h[n]}
• Thus, the transform-domain representation
p of a discrete-
time signal can also be equally applied to the transform-
domain representation of an LTI discrete-time system
• Besides providing additional insight into the behavior of LTI
systems, it is easier to design and implement these
systems in the transform
transform-domain
domain for certain applications
• We consider now the use of the DTFT and the z-transform
in developing the transform-domain representations of an
LTI system

Original PowerPoint slides prepared by S. K. Mitra 4-1-41

© The McGraw-Hill Companies, Inc., 2007
 LTI Discrete-Time Systems in the
Transform Domain
• Consider
Co s de LTI ddiscrete-time
sc ete t e syste
systems
s ccharacterized
a acte ed by linear
ea
constant coefficient difference equations of the form

• Applying the z-transform to both sides of the difference

equation and making use of the linearity and the time
time-
invariance properties we arrive at

• A more convenient form of the z-domain representation of

the difference equation is given by

Original PowerPoint slides prepared by S. K. Mitra 4-1-42

© The McGraw-Hill Companies, Inc., 2007
 The Transfer Function
• A ge
generalization
e a at o o of tthe
e frequency
eque cy response
espo se function
u ct o
• The convolution sum description of an LTI discrete-time
y
system with an impulse
p response
p h[n]
[ ] is given
g by
y

• Taking
T ki the
th z-transforms
t f off both
b th sides
id we gett

Original PowerPoint slides prepared by S. K. Mitra 4-1-43

© The McGraw-Hill Companies, Inc., 2007
 The Transfer Function
• Or,
O,

X (z)
• Therefore,

• Hence, H(z) = Y(z)/X(z)

• The function H(z), which is the z-transform of the impulse
response h[n] of the LTI system, is called the transfer
f ti or the
function th system
t f
function
ti
• The inverse z-transform of the transfer function H(z) yields
the impulse response h[n]
Original PowerPoint slides prepared by S. K. Mitra 4-1-44
© The McGraw-Hill Companies, Inc., 2007
 The Transfer Function
• Consider an LTI discrete-time system characterized by a
difference equation

• Its transfer function is obtained by taking the z-transform of

both sides of the above equation

• Or, equivalently as

• An alternate form of the transfer function is given by

Original PowerPoint slides prepared by S. K. Mitra 4-1-45

© The McGraw-Hill Companies, Inc., 2007
 The Transfer Function
• Or, equivalently as

• ξ1, ξ2, …, ξM are the finite zeros, and λ1, λ2, …, λN are the
finite poles of H(z)
• If N > M, there are additional (N − M) zeros at z = 0
• If M > N, there are additional (M − N) poles at z = 0
• For a causal IIR digital filter, the impulse response is a
causal sequence
• The ROC of the causal transfer function is thus exterior to a
circle going through the pole furthest from the origin
• Thus the ROC is given by
Original PowerPoint slides prepared by S. K. Mitra 4-1-46
© The McGraw-Hill Companies, Inc., 2007
 The Transfer Function
• Example - Consider the M-point moving-average FIR filter
with an impulse response

• Its transfer function is then given by

• The transfer function has M zeros on the unit circle at

z = e j2πk /M, 0 ≤ k ≤ M −1
• There are poles at z = 0 and a single pole at z = 1
• The pole at z = 1 exactly cancels the zero at z = 1
• The ROC is the entire z-plane except z = 0 4-1-47
Original PowerPoint slides prepared by S. K. Mitra
© The McGraw-Hill Companies, Inc., 2007
 The Transfer Function
• Example – A causal LTI IIR filter is described by a constant
coefficient difference equation given by
• y[n] = x[n −1] −1.2 x[n − 2] + x[n − 3] +1.3 y[n −1]
−1.04 y[n − 2] + 0.222 y[n − 3]
• Its transfer function is therefore given by

• Alternate forms:

• Note: Poles farthest from z = 0 have a magnitude

• ROC: |z| > 4-1-48
Original PowerPoint slides prepared by S. K. Mitra
© The McGraw-Hill Companies, Inc., 2007
 Frequency Response from Transfer
Function
• If tthe
e ROC
OC oof tthe
e ttransfer
a s e function
u ct o H(z)
( ) includes
c udes tthe
euunitt
circle, then the frequency response H(ejω) of the LTI digital
filter can be obtained simply as follows:

• For a real coefficient transfer function H(z) it can be shown

th t
that

• For
F a stable
t bl rational
ti l transfer
t f function
f ti ini the
th form
f

• the factored form of the frequency response is given

Original PowerPoint slides prepared by S. K. Mitra 4-1-49

© The McGraw-Hill Companies, Inc., 2007
 Geometric Interpretation of
Frequency Response Computation
• Itt iss convenient
co e e t to visualize
sua e tthe
e co
contributions
t but o s o of tthe
e zero
eo
factor (z − ξk) and the pole factor (z − λk) from the factored
form of the frequency response
• The magnitude function is given by

which reduces to

• The phase response for a rational transfer function is of the

form

Original PowerPoint slides prepared by S. K. Mitra 4-1-50

© The McGraw-Hill Companies, Inc., 2007
 Geometric Interpretation of
Frequency Response Computation
• Thee magnitude-squared
ag tude squa ed function
u ct o o
of a real-coefficient
ea coe c e t
transfer function can be computed using

• The factored form of the frequency response

• is convenient to develop a geometric interpretation of the

frequency
q y response
p computation
p from the p
pole-zero p
plot as
ω varies from 0 to 2π on the unit circle
• The geometric interpretation can be used to obtain a sketch
of the response as a function of the frequency
Original PowerPoint slides prepared by S. K. Mitra 4-1-51
© The McGraw-Hill Companies, Inc., 2007
 Geometric Interpretation of
Frequency Response Computation
• A typ
typical
ca factor
acto in tthee factored
acto ed form
o o of tthe
e frequency
eque cy
response is given by
((ejω − ρ
ρejϕ)
where ρejϕ is a zero (pole) if it is zero (pole) factor
• As shown below in the z-plane the factor (ejω − ρejϕ)
represents a vector starting at the point z = ρejϕ and ending
on the unit circle at z = ejω
• As ω is varied from 0 to 2π, the
tip of the vector moves counter-
clockwise
l k i ffrom th the point
i tz=1
tracing the unit circle and back to
the point z = 1
Original PowerPoint slides prepared by S. K. Mitra 4-1-52
© The McGraw-Hill Companies, Inc., 2007
 Geometric Interpretation of
Frequency Response Computation
• Ass indicated
d cated by

the magnitude response |H(ejω)|at a specific value of ω is

given by the product of the magnitudes of all zero vectors
divided by the product of the magnitudes of all pole vectors
• Likewise, from

• we observe that the phase response at a specific value of

ω is obtained by adding the phase of the term p0/d0 and the
linear-phase
linear phase term ω(N − M) to the sum of the angles of the
zero vectors minus the angles of the pole vectors 4-1-53
Original PowerPoint slides prepared by S. K. Mitra
© The McGraw-Hill Companies, Inc., 2007
 Geometric Interpretation of
Frequency Response Computation
• Thus,
us, a
an app
approximate
o ate pplot
ot o
of tthe
e magnitude
ag tude aand
dpphase
ase
responses of the transfer function of an LTI digital filter can
be developed by examining the pole and zero locations
• Now, a zero (pole) vector has the smallest magnitude when
ω=ϕ
• To highly attenuate signal components in a specified
frequency range, we need to place zeros very close to or
on the unit circle in this range
• Likewise, to highly emphasize signal components in a
specified frequency range,
range we need to place poles very
close to or on the unit circle in this range

Original PowerPoint slides prepared by S. K. Mitra 4-1-54

© The McGraw-Hill Companies, Inc., 2007
 Stability Condition in Terms of the
Pole Locations
• In add
addition,
t o , for
o a stab
stableeaandd causa
causal d
digital
g ta filter
te for
o which
c
h[n] is a right-sided sequence, the ROC will include the unit
circle and entire z-plane including the point z = ∞
• An FIR digital filter with bounded impulse response is
always stable
• On the other hand, an IIR filter may be unstable if not
designed properly
• In
I addition,
dditi an originally
i i ll stable
t bl IIR filter
filt characterized
h t i d by
b
infinite precision coefficients may become unstable when
coefficients get quantized due to implementation

Original PowerPoint slides prepared by S. K. Mitra 4-1-55

© The McGraw-Hill Companies, Inc., 2007
 Stability Condition in Terms of the
Pole Locations
• A causa
causal LTI d
digital
g ta filter
te is
s BIBO
O stab
stable
e if a
and
doonly
y if its
ts
impulse response h[n] is absolutely summable, i.e.,

• We now develop p a stability

y condition in terms of the p
pole
locations of the transfer function H(z)
• The ROC of the z-transform H(z) of the impulse response
sequence h[n] is defined by values of |z| = r for which
h[n]r−n is absolutely summable
• Thus, if the ROC includes the unit circle |z| = 1, then the
digital filter is stable, and vice versa

Original PowerPoint slides prepared by S. K. Mitra 4-1-56

© The McGraw-Hill Companies, Inc., 2007
 Stability Condition in Terms of the
Pole Locations
• Example
a p e – Co
Consider
s de tthe
e causa
causal LTI IIR ttransfer
a s e function:
u ct o

• The plot of the impulse response is shown below

• As can be seen from the above plot, plot the impulse response
coefficient h[n] decays rapidly to zero value as n increases
• The absolute summability condition of h[n] is satisfied satisfied,
⇒ H(z) is a stable transfer function
Original PowerPoint slides prepared by S. K. Mitra 4-1-57
© The McGraw-Hill Companies, Inc., 2007
 Stability Condition in Terms of the
Pole Locations
• Now,
o , consider
co s de tthe
e case when
e tthee ttransfer
a s e function
u ct o coe
coef. a
are
e
rounded to values with 2 digits after the decimal point:

• A plot of the impulse response of is shown below

• In this case, the impulse response coefficient increases

p y to a constant value as n increases
rapidly
• Hence, is an unstable transfer function
Original PowerPoint slides prepared by S. K. Mitra 4-1-58
© The McGraw-Hill Companies, Inc., 2007
 Stability Condition in Terms of the
Pole Locations
• Thee stab
stability
ty test
testing
goof a IIR ttransfer
a s e function
u ct o iss therefore
t eeoea an
important problem
• In most cases it is difficult to compute
p the infinite sum

• For a causal IIR transfer function

function, the sum S can be
computed approximately as
• The p partial sum is computed
p for increasing
g values of K until
the difference between a series of consecutive values of SK
is smaller than some arbitrarily chosen small number, which
is typically 10−66
• For a transfer function of very high order this approach may
nott be
b satisfactory
ti f t
Original PowerPoint slides prepared by S. K. Mitra 4-1-59
© The McGraw-Hill Companies, Inc., 2007
 Stability Condition in Terms of the
Pole Locations
• Consider
Co s de tthee causa
causal IIR d
digital
g ta filter
te with
t a rational
at o a ttransfer
a se
function H(z) given by

• Its impulse
p response
p {{h[n]}
[ ]} is a right-sided
g sequence
q
• The ROC of H(z) is exterior to a circle going through the
pole furthest from z = 0
• But stability requires that {h[n]} be absolutely summable
• This in turn implies that the DTFT of {h[n]} exists
• Now, if the ROC of the z-transform H(z) includes the unit
circle, then

Original PowerPoint slides prepared by S. K. Mitra 4-1-60

© The McGraw-Hill Companies, Inc., 2007
 Stability Condition in Terms of the
Pole Locations
• Conclusion:
Co c us o All po poles
es of
o a causa
causal stab
stable
e ttransfer
a s e function
u ct o
H(z) must be strictly inside the unit circle
• The stabilityy region
g ((shown shaded)) in the z-plane
p is shown
below

Original PowerPoint slides prepared by S. K. Mitra 4-1-61

© The McGraw-Hill Companies, Inc., 2007
 Stability Condition in Terms of the
Pole Locations
• Example:
a p e The
e factored
acto ed form
o o of

is

which has a real pole at z = 0.902 and a pole at z = 0.943

• Since both poles are inside the unit circle H(z) is BIBO stable
• Example: The factored form of

is

which has a pole at z = 1 and the other inside the unit circle
• Since one pole is not inside the unit circle, H(z) is not BIBO
stable
Original PowerPoint slides prepared by S. K. Mitra 4-1-62
© The McGraw-Hill Companies, Inc., 2007