p - . .

tJJllll ', Determine z-lransform offollowing Sl'qumces

'} X (n) ;;;: /1, 2, 3, 4, 5, 0, 7}

\ } (11) == 11,2,3,4,5,0,7}
w 2 1
I
~ : . j) x (n) = (1, 2, 3, 4, S, O, 7}
5ol11tJOl1 • 1
0':: 1, x1(1) = 2, X1(2) = 3, X1(3) = 4, X1(4) = 5, x 1(5) = 0, x (6) = 7
1.e. XJ ( 'I _ 1
X(z) = _LX(n)z-n
Bf definition,
n=O
6
X1(z) - Ix 1(n)z-n
n=O

Pu!tlllg for x (n) = 1+2z-1 +3z-2 +4z-3 +5z- 4 +0z- 5 +7z-6

1

2 3 4 5 7
X1(z) = 1+-+
z
-+-+ -+-
2 2 z3 z4 z6

ltsult: i) X1(z) is as calculated above.

,.
TECHNICAL PUBLICATIONS - An up thrust for knowledge
 3-6
Digital Signal Processing Applications

.. . e x (z) is conv erge nt for all valu es of

11) X1(z) = oo for z = 0, 1• • 1 z, except
... ) H ROC . Entire z-plane except z = O. z : : o.
111 ence • •
ii) x (n) = {1, 2,3, 4, 5, O, 7}
2
i
i.e. x (0) = 4, x2 (t) = S, x2 (2) = 0, x 2 (3) = 7 and
2
X2(- l) = 3, X2(-2) = 2, X2(-3) = 1

...

Putting for x (n), - l•z 3 +2•z 2 +3z 1 +4z O +5z -l +Oz - 2 +7z -3
2

= z 3 +2z 2 +3z +4+ 5 +.2_

z z3

Result: i) Above equation gives X (z).

2
ii) X2 (z) = 00 for z = 0 and z = oo.

iii) Hence ROC : Entire z-plane exce pt z = O and oo.

 '·
j
. 3-3 z-Transform
. Appficat,ons •
proe!_ss,ng
,t<Jf}, . n of con"ergence (ROC) . .. .
oeg10 . n where z-transfor m converges . From def1rution, 1t is clear
"
. ROC JS. the,_t,;..,,te
regio . . .
power series. This senes JS not converge nt all values of z.
!linfl • • an uuuu
~~1"~af'SfortX' JS in mentioning z-transfor m.
~I i'~()C is useful
~ ff~OC:
.u.aoce O b t values of z for which z-transfor m can be calculated .
.:nn~..,, ...., idea a ou
)Ir. ·ves "'·
·) R~ S1 ed to determine causality of the system.
1 aJl be us
ii) R~ c i... used to determine stability of the system.
IV' caJl V'-
ili)R~
Ad properties of the ROC
PJ" ROC for a finite duration sequence ins:Iudes entire z-plane, except
pertY 1 : The
f'!O d/or lzl = 00 •
:•~an ider the finite duration sequence .x(n) = 11 2 1 2)
rroof : cons i
2 0 -1 2
X(z) = 1· z + 2 •z + 1 z + 2 z = z 2 + 2z +1 + z
Here X(z) = oo for z = 0 and 00• This proves first property.
Pioperty 2 : ROC does not contain any poles.
Proof: The z-transform of a" u(n) is calculated as,

1
X(z) = 1- a z - 1

z
=
z-a
ROC: lzl>lal
This function has pole at z = a. Note that ROC is Iz I> Ia I • This means poles do not
fin ROC. Actually X(z) = oo at poles by definition of pole.
Property 3 : ROC is the ring in the z-plane centered about origin.
Pioof:Consider a"u(n) ~ 1
, ROC: lzl>lal
1-a z-1
or
- a" u (- n -1) ~ 1
, ROC : Iz I< Ia I
1-a z-1

Here observe that I z I is always a circular region (ring) centered around origin.
,z;;'rty 4 : ROC of causal sequence (right hand sided sequence) is of the form

TECHNICAL PUBLICATIONS•· An up thrust for knowledge

 3-4
Digital Signal Processing Applications i . .,.,.

Proof : Consider right han d

·
side d seq uen ce a
" ( ) It' R
u n • s OC
•
IS I z I > I a ~
a,iij

• I- l'h11s the~~
of right hand sided sequence is of the form of Iz I> r wh ere 'r'
is the radius
of the Cir~
Property s: ROC of left sided sequence is of the for m Izl<
r.

Proof : Consider left sided sequence - a" u (- n -1) . Its RO C is

Iz I<
of left sided sequence is inside the circle of rad ius 'r'.
IaI- l'hus the~Q:
•
Property 6: ROC of two sided sequence is the con cen tric ringm z-plan
• e.
Proof: We know that ROC of x(n) = a"u (n) +b "u( -n- l) IS I
al< lzl
< lb!, Which.IS~,
concentric ring as shown in Fig. 3.1.4.
I ~
V-:::~Anerties of z-Transform
if.I proP 8
•a•~•ua,;9;1:l:l,lf,I111
.di Linearity
JP"' x1(n) < z >X1(z), ROC : a1 <Iz I< b1
ff x2(n) < z >X2(z), ROC : a2 <Iz I< b2 , then
~
... (3.2.1)

ROC is intersection of ROC's of X1(z) and X2 (z).

nie 00 00

rroof: X(z) = 1',x(n)z-n = L[a 1 x 1(~)+a 2 x 2 (n)]z-n

n=-00 n=-00
00 00

= La1xi(n)z-n + La2x2(n)z-n
n=-00 n=-00

00 00

= a1 L x 1(n) z-n +a 2 Lx2(n) z-n Since a 1 and a2 are constants

n=-00 n=-00

TECHNICAL PUBLICATIONS"'- An up thrust for knowledge

 3- 12
Digital Signal Processing Applications
l-1r,
c!l)llf,
Explanation .. ~
. f is linear, The addition of the sequences . .
Th1s means z-trans onn . . m hrne
adds their respective z-transforms. The ROC is the intersection o cloll'\ail'\ I
r ov aSo
X 1(z) and X 2 (z). erlap Of

Use . be used to evaluate x(n) from X(z). X(z) can b

Tius property can f b e ex
Then inverse z-trans orm ecomes x 1 ( n) + x 2 ( n)+ PresS<!d
X1 (z) +X2 (z)+....... • ....... • as
IfJj Time Shifting or Translation
If l'(n) ~ X(z), ROC: r 1 <lzl< r 2
then

r x(tt-k) ....-4 z-k X(z) I ••• (3.2.2}

00

Proof : Z{x(n-k)}= Lx(n -k) z-n

n=-oo

Let n-k = m. Hence n = k+m and m = - 00

to + 00 • i.e.,
Z{x(n-k)} = I,.x(m)z-Ck+m)
m=-oo
00

= I,x(m)z-k .z-m = z-k Lx(m)z-m = z-k X(z)

m=-oo m=-oo

Explanation
The delay of k samples in time domain is equivalent to multiplying the z-transform
.. by z -k in frequency domain.

Use
-
This property is useful in determining z-transforms of advanced / delayed sequences.

ffII Scaling in z-Domain or M~ltiplication -by Exponential

Let x(n)~X(z), ROC: r 1 <1zl<r2

I then ... (3.23)

Proof:
n=-oo n=-oo

TECHNICAL PUBLICATIONS'". An up thrust for knowledge

 .,~
~
JI
l'

3- 13
. App/ic~at~io~n~s_ _ _ __::___ _ _ _ _ _ _ _ _ _ _ _ _ _z-Transform
_ _ __
~ prc,C!!_sstng
I _JSif"~
l' V__ri
ri,do
r down nthe 'z' plane by
ws that sea mg ' '
the factor 'a' is equivalent to
; ~ ~rtY sho ·n sequence by a . Scaled by a.
1 ..i,is pror- t·me domai
l'v . the l
, 1tipl}ril'S
d'\I

. d termining the movement of poles and zeros due to

LIS' • useful in e
15 n
ropertY • domain by sequence a .
1l'i5 P • n in t1J11.e
111oJtiplicatto .
' ..di Time Reversal
i P"" ~{n) ~ X(z), ROC: r1<\zl< '2
J,ll
1 1
[ x <- ~) <-'4 x( z-1) , ROC: -<lzl<- ... (3.2.4)
'2 '1
i
I l'!OOf: Z{x(-n)) =
I n=-oo

: ,dthn=-m,
m=oo m=-oo

• - 1- <- I z I < -1
Roe : r1 < I z -11 < r 2• 1.e.
•• • r2 '1
Explanation

This property shows that inyersion in z-plane is equivalent to folding the time
I dooiain sequence. Note that the scaling of ROC is also inverted.
Use

This property gives the behaviour of the system when poles-zeros are mutually
' inlerchanged.

ID Differentiation in z-Domain or Multiplication by a Ramp

let
x (n) ~ X(z), ROC ~ r 1 <Iz I< r 2 then,

... (3.2.5)

Proof:
L x(n) z-n
00

X(z) ==
n=-oo
d
dz X(z) == f !!.. (x(n) z- 11
] = f x(n) :z z-"
...._____ _»_=_-_oo_d_z_ _ _ _ _,_,=_-_"°_______________,._7_
TECHNICAL PUBLICATIONS... An up thrust lor •now~
 Digital Signal Processing Applications 3-14

• 00

i>,n)·(-n )· z-11-l = - Lnx(n)z- n .z-1

= 11=-00
n=-"°
oo
= _ 2 -1 ~)nx(n)]z -n = -z -1 •Z{ n x( n,
)'
11=-c,O

or Z{n l\n)} = -z :z X(z), ROC: Same as that of x(n)

Explanation

I The differentiation in z-domain is equivalen t to multiplication of th .

sequence by 'n'.
e tuned0
rnaui
Use
Time multiplication of the sequence affects the frequency domain beh .
av1our
derivative in z-plane. This property is used to study the movemen t of poles-
zeros if as
funa
domain sequence is multiplied by n. e

ffII Convolution in Time Domain

Let x 1 (n) ~ X 1 (z), ROC: a 1.< Iz l<b 1

and x 2 (n) ~ X 2 (z), I

ROC: a2 < z l<b 2 then

The ROC is intersection of ROCs of X 1 ( z) and X 2 ( z) .

00

Proof: x 1(n)•x2 (n) = L~1(k)x 2 (n-k)

k=-oo

Z(x1(n)•x2(n)l = .!_[k!_ x1(k)x2 (n-k)}-•

Interchanging orders of summation,

• = J_x1(k ){J_x2 (n-k)z-• }

:'
I Since x 2 (n-k) ~ z-kx 2(z),
00

= L x1(k){z-kx 2 (z)}
k=-oo

= t!_x1 (k)z-k} • X2(z) = X 1(z),X 2(z)

____ ____ ____ _____________________
__:_,_

TECHNICAL PUBLICATIONS.... An up thrust lot lcnowledge

 3-_
·,,gAPP/;cation1ss _ _ _ __.:;_ 15_ _ _ _ _ z-Tran sform
________ __ _
proc!_SSI
j:11'!-
;!,, . equiv alent to conv oluti on of respe ctive time doma in
__.,i,r,,tlO" f z-transforms is
v,· ro£11l't 0
flit p
fl,,
. f·1tering and frequ ency analysis.
Of' ·s used in i
. concept t
'l11i5 f Two Sequ ence s
correlation o
Ad
~ z X( ) ROC : a 1 <Iz I< b1
X1 (1t) ~ 1 Z,
LL"I z X2 ( ) ROC : a 2 < Iz \<b 2 then
X2 (11) ~ Z,
-i X1(n) x2 (n-l )_~ X1(z )X2( z-1) ... (3.2.7)
n,.-oo

.-\ati
rroof: COu-. on of two seque nces is given as,
00

rx1x2 (l) = L X1(n )x2(n -l)

n=-oo
00

= L x 1(n)x 2 [-(l- n)] = X1(l) *X2( -l)

n=-oo

{ t x1(n)x2 (•-Q 1 = Z(x 1(l)*x 2 (-l)}

= Z (x1(l)}·Z (x 2 (-l)} = X 1 (z)-X 2 (z- 1)

, Explanation
The correlation in time doma in is equiv alent to multiplica
tion of z-tran sform of one
• sequence and z-transform of other seque nce with inversion of z-pla
ne (second sequence).
Use

l Correlation between the two sequences can be determined in frequ

I ~ help of this property.
ency doma in with
1
Ill Multiplication of Two Sequences or Convolution in z-Domain
~
ROC : a 1 < Iz j<b1
I tld

1
-__

I,,,
__ __ __ _T_E-CH_N_ICA-LP-U-B-L/C_A_TI_ON_S___-A_n_up_l_hru_s_lfor_kno__
w/edge
___ __ l_/ii
 , , . tl e closed contou r. It enclose s the ongm and lies in the ROC
H ere c 1s '
1 ) . Thus th ROC • • Wh • h
commo n to bo tl' X 1 ( v) and x _ e 1s mterse • of ROCs
chon f IC •
2( v O X (z) IS
1 anc1
X 2 (z).
Proof: Inverse z-trans form is given as, x(n) =
2 ~j f X(v)vn-l dv
C

Let x(n) = xi(11) xi(n)

Putting inverse z-transform of x 1(n) in above equatio n,

x(n) = ~ ! X 1 (v)v"- 1 dv•x2(n )

21tJY
C

X(z) = f {2~. f X
n=- 00 } C
1
1 (v)v"- dv •x2(n)} z-n

Interchanging the order of integration and summa tion,

X(z) =
2
~. f X 1(v) f v" ,v-1 x 2(n)z-n dv
} C n=-oo

• l
I

Explanation
Multipiication of the two sequences in time domain is equiva lent to integrating the
produc t of their z-transforms with approp riate scaling over closed contou r.

Use
Jhis proper ty is used to study the behavi our of cascade d system s in frequency
domain .

ffU Conjugation of a Complex Sequence

Let

TECHNICAL PUBLICATIONS'". An up thrust for knowledge

 ... (3.2.9)

,....u11atiOl1
~- •
of the complex conJugate • eqmv
• alent to takin g comp1
sequence 1s ex
i-uansforll\ . .
·n z-domam.
~ate I •
f!ld z-Transfonn of Real Part of a Sequence
x(n) ~ X(z), ROC: r 1 <Iz I< r 2 then,
Ltl

Re[x(n)] ~ ½[X(z) + x• (z• )] ... (3.2.10)

rn,of: x(n) = Re[x(n)] + jlm [x(n)] and x• (n) = Re[x(n)] - jlm [x(n)]

Re[x(n)] = ½ [x(n) + x• (n)]

Z[Re(x(n)]} = Z {½[x(n) + x• (n)]}

= ½tZ[x(n)]+ Z[x• (n)]}

1
=
2 [X(z)+x• (z.)]
om Z•Transform of Imaginary Part of Sequence
la
x(n) ~X(z), ROC: r 1 <1zl<r2 then,

Im[x(n)] ~ ;j[X(z)-x• (z.)]

... (3.211)

lino£: Im [~n)] = 1 [ •
2 j x(n)-x (n)]

TECHNICAL PfJBLICATIONS~ • An up thtvsl for lcnowt.dge

 ¥PX ¥:

3-18
Digit al Sign al Processing Applications

Z{lm [x(n )]} = z {2\ [x(n )-x• (n)] } = 2\ { Z[x (n)] -Z[x • (n)J}

= ;j{X (z)- x• (z•) }

ff jfj Parseval's Relation

Let X2 (n) ~ X2 (z) then,

~ 1(n)x•2 (n) =• 11t. J'J: X 1 (v) x•2 ( -1.

"'-'x 2 I c
Jv -1 dv
n=- - V ... (3,2.12)

Proof : Inverse z-transform of X 1(z) is, x 1(n) = 2

~j f X 1 (v) vn- l dv
C

00

••• Ix 1(n)x ;(n) =

n=-o o

f X1 (v/ f x2( n)•( J.J- n]* v- dv

1
= 2 ~.
Jc ln=-o o V .

= 2 ~j ! X1 (vi[ x 2( v~ Jr v- dv
1

= 2~j ! X1 (v)X i( v~ Jv- dv 1

ffl FI Initial Value Theorem

Let x(n) ~ X(z ), then,

x(O) = lim X(z) ... (3.2.13)

I %➔-

edge
TECHNICAL PUBLICATIONS"'- An up thrust for knowl
 ' -• I __ I I . •

ssing Applications 3 - 19 z-Transform

s;gnslp ,oce~~ :...----- --------- -------

~transform of a causal sequence is given as,

rroof: z-
oo

X(z) = Lx(n)z-n, since x(n) =0 for n < O

n=O

1irn X(z) = lim x(0)+ lim x(l)z-1 + lim x(2)z-2 +· ..

z ➔ oo z ➔ oo z ➔ oo
z ➔ oo

= x(0)+0+0+0+Q• ..
l'(O) = lim X(z)
z ➔ oo
1

explanation
. .

The initial value of the sequence is given by taking limiting value of X(z) as z ➔ 00 •

Use
This property is used to evaluate initial value of the sequence from its z-transfonn.
 l)tttl'lfUnt invmt z-trans_
Example 3.3 1
form of the following :
i) X(z)..: - 1 - ,
ROC: lzl>lal -
1 -• Z -l
sa1urion :
X(z) = l , ROC: \z\>\a\
1-a z-1
1 2 2 3 3
1+a 2 - +a z- +a z- t-Negative power of' z'
1-az-1f1
1-az -1
- +

a3 z-3
a 3 z-3 - a 4 z-4
+

a 4 z-4

Thuswehave,X(z)= 1 =l+a z-1 +a 2 z-2 +a 3 z-3 +· ..

1-az -1
Taking inverse z-transform, x(n)
= {l, a, a2 , a 3 , .. ,}
= a" u(n)
 fxample 3.3.3 Determine inverse z-transform of
1
X(z) = , ROC: Iz I> 1
(1 +z- )(1-z -1) 2
1
PU : May-12. Marks 8
Solution : Step 1 : Converting X(z) to positive power s of z,
z3
X(z) = - - - -
(z+l)( z-1) 2
2
X(z) = z
z (z+l)( z-1) 2

Step 2 : Here there is multiple pole at z = 1. Therefore the partial fraction expan
sion
will be,
X(z) ·= ~ + ~ + A3
z z+l (z-1) (z-1)2 ... (3.3.8)

X(z) z2 1
Ai = (z+ l)·- = --
z Z=-1
(z-1)2 Z=-1 4

A3 = (z-1)2 X(z) --
z2 1
z z=l z+l 2
z=l

A2 = !_{(z
dz
-l)2. X(z)}
z ••• By equation (3.3.3)
Z=l

= !_,( ~) 2
= (z+l) 2z-z 3
dz z+l z=l (z+l)2 4
Z=l
Putting values in equation (3.3.8),
X(z) = 1/4 + 3/4 + 1/2
z z+l z-1 (z-l)2

~3: X(z) = -+-

1/4 z 3/4 z 1/2z
-+- -'--
z+l z-1 (z-l)2

TECHNICAL PUBLICATIONS'"- An up thrust for knowledge

 ~
Dig ita l SignaI r'f'O 3- 34
eessing Applications

4 : RO C is I z I> 1. Le t us us e fol low ing rel ati on s :

St ep
1 _ ,R OC : /zl >/ pk I
p~ u( n) t z +
1- p k Z 1
ms of X( z) will be ,
He nc e inverse z-transform of first tw o ter
!
IZT { l/~1} = (- lt u(n ) an d IZT { 1- z 1
l+ z
34
/ _ } = : (l)n u(n )

For 3
rd
term of X( z) let us use,
p z- l
n p ~ u ( n) < z > k _1 , ROC : I z I > IP k I
2
(1 -p k z )

i.e.

Pu tti ng all the sequences together,

1 3
x(n) = (- Itu (n )+ (1 tu (n )+ .!. n( l)n u( n)
4 4 2

= [{ (-1 )" +: + ~ n}< n)

 uence using z-transform
~ ( ( ! ) u(n), x 2 (n) = cos(nn)u(n).
.\'1(11) - 2
PU : May-10, Marks 6

' x (n) == -1 )" u(n) Hence X 1 (z) = - -1 -

.... t1ere 1 (2 1 -1
50111110•• • 1 - 2 z

1- Z-l COS7t
n) == cos ( 1tn) u ( n) Hence X 2 (z) =
and x2 ( 1- 2z-1 cos n + z-2

=
X2 (z) can be rearranged as,
Here
z and X1(z) = z(z+l) = 2(z+ 1) = _z_
1 2 (z+1)2
z +2z+l z+l
z-2

X(z) = X1 (z) • X2 (z)

z z
= 1 z+l
z-2

(I
X(z) = z
z

X(z) =
1{ 1-½z-1 \+!- } 1 ... (3.3.11)

By convolution in time domainI

l'i(n) * x2(n) ~ X1 (z)·X (z) = X(z)

2
TECHNICAL PUBLICATIONS..• An up th111$1 for knowledge
p

Digital Signal Processing Applications 3 - 38

putting for X( z)= X 1 (z)• x 2 (z)

From equation (33

' ,lJ)
of ab ov e eq ua tio n,
Taking inverse z-transfrom of right ha nd sid e

x1(n)•x2 (n) = ½[( ½J +2(-1)"] u(n)

3.6.1 Power Series Expansion Method ___
The z-transform X(z) of a discrete-time signal x(n) is represented as prope r rational transfer
function. .
• f

The transfer function of X(z) is represented as :

N(z) a0 + a1 r 1 + a2 r 2 + ... + am·z-m
X(z) = D(z) = b + b1 r 1 + b2 r 2 + ... + b n '".,.-n
0

The transfer functi on is called proper and rational when all the coefficients of denominator is
non-z ero and n > m.
The powe r series expansion is obtained by long division. For this method, we should have
X(z) with its R.O.C. Now, X(z) is expanded into a power series of the form
00

X(z) = r. Cn rt'
n ■ -00

Equat ion (2) converges in the given R.O.C. Inverse z-transform of X(z) is given by:
x(n) = r 1 [x(z)] = Cn for all n •

This metho d is demonstrated example 3.18.

 DIGITAL SIGNAL PROCESSING E E&TC Z-T

3.6.2 Partial Fraction Expansion Method

nal function as: \
The z-transform of x(n) is represented in ratio
N(z) ao + a1 rl + a2 r2 + ... + am z-m •
X(z) .= D(z) = b0 + b1 z-1 + b2 r 2 + ... + bn
z-n

where, N(z) Numerator poly~omial

D(z) _ Denominator polynomial
ao, a1, a2, ..., am - Coefficients of numerator
b o, b 1, b2, ..., bn• _ Coefficients of den omi nat or

m - Degree of numerator
n - Degree of denominator
Steps for Partial Fraction Expansion :
m, fu_nction is proper.
~- Check the function is proper or not. If n >
vert the. function in
erator and denominator by z". This will con
2. Mul tiply the numz.
positive power of

3. Obtain the equation X(z).

z
- o
e the den omi nato r and obta in root s. The den ominator will be in the form
4. Factoriz
, Pn are called poles. ·
(z - P1}, (z - P2), ... , (z - Pn), where p 1, p2, .... '

expansi_on form as follo ws:

5. Wri te do'h'.n the equ atio n ir, partial fraction
X(z~ C1 C2 Cn
Z = + + ... +
z - P1 z - P2 . . z - Pn
c is calculated as :
c1, c2, •••, Cn are coefficients. The coefficient 0
X(zl
Cn = -(z- Pn)
• z z. = Pn
s to obta in inverse z-transform.
6. Now , we will use standard z-transforrn pair
z z·
an u(n) +-➔ - for R.0.C. lzl . > lal
z-a .
z z
and -a" u(-n - 1) +-➔ ;=- ; for R.0.C. lzl < lal
u The proc edu re is illustrated by example 3.19
.