ELEN E4810: Digital Signal Processing

Topic 4: The Z Transform

After this chapter the student should be know and solve :

1. The Z Transform

2. & the Inverse of Z Transform

1
1. The Z Transform
 Powerful tool for analyzing & designing
DT systems
 Generalization of the DTFT:

G ( z )  z{g[n]}   g[ n] z
n  
n
Z Transform

 z is complex...
j
 ze → DTFT
DTFT of
 z  r.e j
 n
g [ n ]rn
e  j n
r-n·g[n]
2
Region of Convergence (ROC)
 Critical question: 
Does summation G( z)   x[ n
n  
] z n

converge (to a finite value)?

 In general, depends on the value of z
 → Region of Convergence: Im{z}

Portion of complex z-plane

Re{z}
for which a particular G(z) λ
will converge
ROC z-plane
|z| > λ
3
ROC Example
 e.g .x[n]   [n]
n
-2 -1 1 2 3 4
n
“closed form”
1
 X ( z )  n 0  z

 n n
 when
1  z 1 z 1
1 Im{z}

 converges only for z 1  1 Re{z}

i.e. ROC is z  
λ
(previous slide)

   1(e.g. 0.8) - finite energy sequence

   1 (e.g. 1.2) - divergent sequence,
infinite energy, DTFT does not exist
but still has ZT when |z| > 1.2 (in ROC)
M
4
About ROCs
 ROCs always defined in terms of |z|
→ circular regions on z-plane
(inside circles/outside circles/rings)
 If ROC includes Im{z}
unit circle (|z| = 1),
→ g[n] has a DTFT Re{z}

(finite energy
sequence) Unit circle
z-plane
lies in ROC
→ DTFT OK

5
 Another ROC example
 Anticausal (left-sided) sequence:
x[n]   [n  1]
n -5 -4 -3 -2 -1
1 2 3 4 n

X ( z )  n ( [n  1]) z

n n
ROC:
|λ| > |z|
  n    z  m 1  z
1 n n  m m

1
   z
 1

1  z 1
 Same ZT as  [n], different sequence?
n

6
ROC is necessary!
 A closed-form expression for ZT
must specify the ROC:
x[n]  n [n]  X(z)  1 Im
1  z 1

-4 -3 -2 -1 1 2 3 4
n ROC z   Re

x[n]   n
 X(z) 
1 λ

-4 -3 -2 -1
1  z 1
z-plane
1 2 3 4 n ROC z  
 A single G(z) expression can match
DTFTs?

several sequences with different ROCs

7
 Rational Z-transforms
 G(z) expression can be any function;
rational polynomials are important
class:
Pz p0  p1z 1  …  pM 1z (M 1)  pM z M
Gz  
Dz  d0  d1z  …  dN 1z
1 (N1)
 dNz N

 By convention, expressed in terms of z-1

– matches ZT definition
 (Reminiscent of LCCDE expression...)

8
 Factored rational ZTs
 Numerator, denominator can be
factored:

Gz 
M
 
p0  1 1    z 1 z M p 0  1

M
z    
N

d0  1 1    z 1
 z d 0  1 z    
N N

 {  } are roots of numerator

→ G(z) = 0 → {  } are the zeros of G(z)
{ }are roots of denominator


→ G(z) = ∞ → { }are the poles of G(z)

9
Pole-zero diagram
 Can plot poles and zeros on
complex z-plane:
Im{z}
poles 
(cpx conj for real g[n])
o
×
o × o Re{z}
× 
o

zeros  
z-plane

 (Value of) expression determined by roots

10
Z-plane surface
 G(z): cpx function of a cpx variable
 Can calculate value over entire z-plane

ROC
not
shown!!

M
11
 ROCs and sidedness
 Two sequences have: G(z)  1
1  z1
10
ROC z    g[n]   [n]
n

8
RIGHT-SIDED
6
n
λ
1

(| | < 1 )
4

ROC z    g[n]   [n  1]

n
0.5
2

n
0

−1 −0.5
−0.5
0

LEFT-SIDED
−1
0.5

z-plane
1

 Each ZT pole → region in ROC outside

or inside | | for R/L sided term in g[n]
 Overall ROC is intersection of each term’s
12
 ZT is Linear
 G(z)  Zgn   g[n]zn Z Transform
n

y[n]  g[n]  h[n]

Y ( z )   (g[n]  h[n]) z n

Linear 
  [(g[n]z  h[n]z )]  G ( z )  H ( z )
n n

 Thus, if y[n]  1 n1[n]  2 n2 [n]

then 1 2 ROC:
Y (z)  z>λ 1 ,λ 2
1  1z 1
1  2z 1

13
 ROC intersections
1 1
 Consider G(z)  
1  1z 1
1  2z 1

with |1| < 1 , | 2 | > 1 ... no ROC specified

 Two possible sequences for 1 term...
n n1[n]
  1[n  1]
n

n
 Similarly for  2 ...
 n 2 [n  1] n
or n 2 [n]
n
→ 4 possible g[n] seq’s and ROCs ...
14
 ROC intersections: Case 1
G(z) 
1

1 g[ n]   [ n]   [ n]
n
1
n
2
1  1z 1
1   2 z 1
both right-sided:
n
ROC: |z| > |1| and |z| > |2|
Im

Im
Re
λ λ

15
 ROC intersections: Case 2
1 1 g[n]  1n [ n  1]  n2 [ n  1]
G(z)  
1  1z 1
1   2 z 1
both left-sided: n

ROC: |z| < |1| and |z| < |2|

Im

Re
λ λ

16
 ROC intersections: Case 3
G(z) 
1

1 g[n]   [n]   [ n  1]
n
1
n
2
1  1z 1
1   2 z 1
two-sided:
n

ROC: |z| > |1| and |z| < |2|

Im

Re
λ λ

17
 ROC intersections: Case 4
G(z) 
1
1

1
1
g[n]   [ n  1]   [n]
n
1
n
2
1  1z 1  2z
two-sided:
n
ROC: |z| < |1| and |z| > |2| ?
Im

Re
no ROC
λ   ‹ ...
λ 


18
ROC intersections
 Note: Two-sided exponential
gn   n   n  
n
  n   n 1
n n

ROC ROC Im
|z| > |  | |z| < |  |
 No overlap in ROCs Re

→ ZT does not exist
(does not converge for any z)
19
ZT of LCCDEs
 LCCDEs have solutions of form:
yc[n]  i  i n  ...
n
(same
i s)
 Hence ZT Yc z  1 
1  iz
 Each term  in in g[n] corresponds to a
pole  i of G(z) ... and vice versa
 LCCDE sol’ns are right-sided
‹ ROCs are |z| > |  i | outside
circles
20
Z-plane and DTFT
 Slice between surface and unit cylinder
( z  1  z  e j )isG (e j ), theDTFT

G ( e j )
z  e j

0 / rad/samp º

21
 Some common Z transforms
g[n] G(z) ROC
 [n] 1 ∀z
1
µ[n] 1z 1 |z| > 1
 [n]
n 1
1z 1 |z| > ||
1rcos0 z 1
rncos(0 n)µ[n] |z| > r sum of
r n e j 0 n  r n e  j 0 n
12r cos0 z r z
1 2 2

rsin0 z 1
rnsin(0 n)µ[n] |z| > r
12r cos0 z r z
1 2 2

 j 0
×
“conjugate pole
poles at z  re ×
pair”
22
 Z Transform properties
g[n] — G(z) w/ROC Rg
Conjugation g*[n] G*(z*) Rg

Time reversal g[-n] G(1/z) 1/Rg

Time shift g[n-n0] z-n0G(z) Rg (0/∞?)

Exp. scaling n g[n] G(z/)  Rg

dG(z)
Diff. wrt z ng[n] z Rg (0/∞?)
dz
23
 Z Transform properties
g[n] G(z) ROC
at least
Convolution g[n] ∗h[n] G(z)H(z)
Rg∩Rh
Modulation g[n]h[n]
1
2 j C G v H z
 v v atdvleast
1

RgRh


Parseval:  gnh n  * 1

2 j C G v H
 
* 1 v 1dv
v
n

24
 ZT Example x[n]  r cos( 0 n) [n]
n

x[n]  r n cos( 0 n) [n]; Can express as

1
2
 
[n] (re j0 ) n  (re  j0 ) n  v[n]  v * [n]

 v[n]  1 [n] n ;   re j0

2
 V ( z )  1 j 0 1 and the ROC z  r
(2(1  re z ))

 Hence, X z   V z   V * z *  
 12 1

j 0 1 
1re z
1
1re  j 0 z 1 
1rcos0 z 1
 12r cos z 1 r 2z2
0

25
 Another ZT example
yn  n 1 nn
 x[n] nx[n] where x[n] =  nµ[n]
dX z
X z  1  z
1  z 1 dz
1
( |z| > ||)
 z d  1 
 z

dz 1  z1  (1 z1)2
1
 Y z 
1

 z

1 repeated
ROC |z| > || 1  z (1 z ) (1 z )
1 1 2 1 2 root - IZT

26
2. Inverse Z Transform (IZT)
 Forward z transform was defined as:

G(z)  Zgn   g[n]z n

n
 3 approaches to inverting G(z) to g[n]:
 Generalization of inverse DTFT
 Power series in z (long division)
 Manipulate into recognizable
pieces (partial fractions) the useful
one

27
 IZT #1: Generalize IDTFT
j
 If z  re
j
G ( z )  G ( re )   g[ n]r  n e  jn  DTFT {g[ n]r  n }
1 
 d , IDTFT
n j jn
 so, g[ n]r  G ( re ) e
2 

dz 1
z  re  d  
j n 1  n
 G ( z ) z r dz
jz 2j C Im

Re
Counterclockwise
closed contour at |z| = r
within ROC
 Any closed contour around origin will do
 Cauchy: g[n] = [residues of G(z)zn-1]
28
IZT #2: Long division
 Since G(z)  

g[n]zn
n
if we could express G(z) as a simple
power series G(z) = a + bz-1 + cz-2 ...
then can just read off g[n] = {a, b, c, ...}
 Typically G(z) is right-sided (causal)
P(z)
and a rational polynomial Gz 
D(z)
 Can expand as power series through
long division of polynomials
29
IZT #2: Long division
 Procedure:
 Express numerator, denominator in
descending powers of z (for a causal fn)
 Find constant to cancel highest term
→ first term in result
 Subtract & repeat → lower terms in result
 Just like long division for base-10
numbers

30
 IZT #2: Long division
1
 e.g. H z 
1 2z
1 0.4z1  0.12z2 Result
11.6z1  0.52z2  0.4z3...
1 0.4z1  0.12z2 ) 1 2z1
1 0.4z1  0.12z2
1.6z1 0.12z2
1.6z1  0.64z2  0.192z3
0.52z2  0.192z3
...
31
IZT#3: Partial Fractions
 Basic idea: Rearrange G(z) as sum of
terms recognized as simple ZTs
1
 especially   n  n 
1   z 1
or sin/cos forms
P(z)
i.e. given products
1  z 1  z 

1 1

B A
rearrange to sums  
1  z1 1  z1
32
 Partial Fractions
 Note that: order 2 polynomial
A B C
   u + vz-1 + wz-2
1  z1 1  z1 1  z 1

       
A 1  z 1 1  z 1  B 1  z 1 1  z 1  C 1  z 1 1  z 1 
order 3 polynomial 1  z 1  z 1  z 
1 1 1

 Can do the reverse i.e.

P(z) 
go from to 1N
N
 1 (1   z 1
) 1    z 1
else cancel
 if order of P(z) is less than D(z) w/ long div.
33
Partial Fractions
 Procedure: order N-1

P(z) 

N
F(z)  N 
 1 (1   z )
1 11   z 1


 f n  
N
no repeated   
n
 n
 
poles! 1

  
where   1    z F z z 
1

i.e. evaluate F(z) at the pole (cancels term in

denominator)
but multiplied by the pole term
→ dominates = residue of pole
34
 Partial Fractions Example
1
 Given H z 
1 2z
(again)
1 2
1 0.4z  0.12z
factor:
1 2z1 1 2
  
 1
1 0.6z 1  0.2z1 1 0.6z1
1  0.2z 1

 where:
1
1  1 0.6z 1
H z z0.6 
1 2z
1
1  0.2z z0.6
 1.75
1
2  1 2z  2.75
1 0.6z1 z0.2
35
Partial Fractions Example
Hence H z  1.75 2.75
 
1 0.6z1 1  0.2z1
 If we know ROC |z| > || | i.e. h[n] causal:
 hn  1.750.6 n  2.750.2 n
n n

= –1.75{ 1 -0.6 0.36 -0.216 ...}

+2.75{ 1 0.2 0.04 0.008 ...}
same as
= {1 1.6 -0.52 0.4 ...} long division!

36
 Problems:
Note: you must be know what is the purpose of unit step and how determine the
term (interval) of summation.
Q1:

Q2:

7  3.6 z 1
X b ( z) 
1  0.9 z 1  0.18 z 2

Solve by using long division and partial fractions and compare between your
answers.