Advanced Digital Signal Processing

Lecture 5
Z Transform

Dr. Tahir Zaidi

z-Transform

Background-Definition
∞
jω
- Fourier transform X (e ) = ∑ x[ n
n = −∞
]e − jω n
extracts the
h essence off x[n]

but is limited in the sense that it can handle stable systems

y only.
y
X (e jω ) converges if ∑ | x[n] |< ∞
i.e., stable system → Fourier Transform converges

- So, we want to extend it such that it can be used as a tool to

analyze
y digital
g systems
y in general.
g
 ∞
jω
Let X r (e ) = ∑ ( x[n]r
n = −∞
−n
)e − jωn

then it converges if ∑ | x[ n ]r −n
|< ∞
Th condition
The diti for
f convergence is
i relaxed!
l d!

g ) x[n] = 2 n u[n]
((e.g.)
| X (e jω ) |=| ∑ 2 n e − jωn |→ ∞
1
| X r (e ) |=| ∑ 2 r e
jω n − n − jωn
|< ∑ 2 r
n −n
=
2
1−
|r|

→ converges if | r |> 2
- This implies
p that X r (e jω ) can handle some systems
y that
X (e jω ) cannot due to divergence.
- Therefore we define z-transform to be
∞
jω
X ( z ) = X r (e ) |re jw = z = ∑ x[
n = −∞
∞
n ] z −n

Representing the condition for convergence of X r (e jω )

in terms of region of convergence RoC.

((e.g.) [ ] = 2nu[n]
g ) in case x[n] [ ]
X r ( e jω ) exists for |r|>2.
So RoC is |z| = |rejω|>2.
So, |>2
In general, if x[n] = a nu[n]
RoC is z>a

- In
I terms
t off X (z ) ,
2
X (e jw ) is a special case
Where z = 1 , or r = 1
causal
(e.g.) x[n] = − a nu[− n − 1]
∞ −1
X ( z ) = − ∑ a nu[− n − 1]z −n
= − ∑ a n z −n
n = −∞ n = −∞
∞ ∞
= −∑ a z = 1 − ∑ (a −1 z ) n
−n n

n =1 n =0

1 − a −1 z
= 1− =
1 − a z 1 − a −1 z
−1

1
=
1 − az −1 2=|a|

RoC : a z < 1 , or z < a

−1
(e.g.) Two - sided sequence
1 n 1 n
x[n] = (− ) u[n] − ( ) u[− n − 1]
3 2
∞
1 n − n −1 1 n − n
X ( z ) = ∑ (− ) z − ∑ ( ) z
n =0 3 n = −∞ 2

1 1
= −
1 1
1 + z −1 1 − z −1
3 2
1 1
z> , z< 1/3 1/2
3 2
Some Common zz-Transforms
Transforms
(1) δ [ n ] ↔ 1 [ all z ]
1
(2) u[ n ] ↔ [ z > 1]
1 − z −1
(3) − u [ − n − 1]
1
↔ [ z < 1]
1 − z −1
(4) δ [ n − m ] ↔ z −m [ all z except 0, if m > 0,
p ∞ , iff m < 0 ]
all z except
1
(5) n
a u [n] ↔ [z > a]
1 − az −1
1
(6) − a n u [ − n − 1] ↔ [z < a]
1 − az −1
az −1
(7) na n u[ n ] ↔ [z > a]
(1 − az −1 ) 2
(8)

(9)

(10)

( )
(11)

(12)

(13)
 Properties of RoC

(1) in general 0 ≤ rR < RoC < rL ≤ ∞

(2) X (e jw ) absolutely converges ↔ UC ⊂ Ro C
(3) RoC cannot contain a pole
(4) FIR sequence → entire z plane, may be except for 0 or ∞
Right sided sequence → outward of the outermost pole
(5) Right-sided
(6) Left-sided sequence → inward from the innermost pole
T id d sequence →a ring
(7) Two-sided i ini between
b t two
t adjacent
dj t rings
i
(8) RoC is a connected region
((e.g.)
g ) If x[n]
[ ] is a sum of 3 sequences
q whose poles
p
are a, b, c respectively,
There exist A possible RoCs as shown below

a b c

All right
right-sided
sided All left-sided

a b c

two left-sided
left sided t right-sided
two i ht id d
 zz‐Transform
Transform Properties
(1) Linearity
ax1[n] + bx2 [n] ↔ aX 1 ( z ) + bX 2 ( z )

(2) Time shifting

x[n − no ] ↔ z − no X ( z )
z −1 1 1
↔ z −1 ⋅ → ( )n u[n] and delay -1
1 −1 1 −1 4
(e.g.) 1 − z 1 − z
4 4 1
z>
1 n−1 4
→ ( ) u[n −1], ]
4
 z‐Transform Properties..(cont.)
(3) Multiplication by an Exponential Sequence
z
z x[n] ↔ X ( )
n
o
zo
z = e jw
g ) (e jw ) n x[n] ↔ X (e − jw ⋅ z )
((e.g.) o o
→ X (e j ( w − w ) )
o

n
r
(e.g.) r n cos wo n ⋅ u[n] = [(e jwo ) n + (e − jwo ) n ]u[n]
2
1
= [(re jwo ) n + (re − jwo ) n ]u[n]
2
1 1 1
↔ [ −1
+ − jwo −1
]
2 1 − re ⋅ z
jw
1 − re z
1 − r cos wo z −1
= −1 2 −2 z >r
1 − 2r cos wo z + r z
z‐Transform Properties..(cont.)
(4) Differentiation
iff i i off X(z)
( )
d
nx[n] ↔ − z x[ z ] RoC = Rx
d
dz
−1
(e.g.) X ( z ) = log(1 + az ) z>a
dX ( z ) − az − 2
=
dz 1 + az −1
d az −1 −1 1
− z X ( z) = −1
= a⋅z ⋅
dz 1 + az 1 + az −1
nx[n] = a (− a ) n −1 u[n − 1]
n −1 an
x[n] = (−1) u[n − 1]
n
z‐Transform Properties..(cont.)
(5) Conjugation of Complex Sequence
x * [ n] ↔ X * ( z * ) RoC = RX

(6) Time-Reversal
1 1
x [ − n] ↔ X ( * )
* *
RoC =
z RX
1
x[−n] ↔ X ( )
z

(7) Convolution-Integration
Convolution Integration
x1[n] * x2 [n] ↔ X 1 ( z ) X 2 ( z )
 Inverse zz‐Transform
Transform
4-Ways:
 Inversion by Contour Integration
• Cauchy integral definition of the inverse z‐Transform

• Example: Inverse DTFT

• Implies.. Contour C is chosen as unit circle
 Inversion Method ‐1
1
Concept of Partial Fraction Expansion-
 Inversion Method
Concept of Partial Fraction Expansion Inversion-
1. Find
i d partial
i l fraction
f i expansion
i method
h d in
i third
hi d
equivalent form

2. Invert by expansion
Doing the Partial Fraction Expansion
Doing the Partial Fraction Expansion‐2
Expansion 2
Writing Down x[n]
X[n] depend on knowing the ROC
 Example‐ ROC
If x[n] is a sum of 3 sequences whose poles
are a, b, c respectively,
There exist A possible RoCs as shown below

a b c

All right-sided

a b c

two left-sided
Example‐ Partial Fraction
Example
Long Division
Finding the coefficients of Poles
Writing Down x[n]
Partial Fraction Expansion
p in MATLAB