dsp n06 chapter6
TRANSCRIPT
-
8/10/2019 Dsp n06 Chapter6
1/10
-
8/10/2019 Dsp n06 Chapter6
2/10
-
8/10/2019 Dsp n06 Chapter6
3/10
-
8/10/2019 Dsp n06 Chapter6
4/10
66 Chapter 6.
Figure 6.3: Pole-zero plot and region of convergence for Example 6.3.
Property 1 The ROC is a ring or disk in the Z-plane centered at the origin, i.e. 0 rR |z|< rL< .
Property 2 The DTFT exists if and only if the ROC includes the unit circle.
Property 3 The ROC cannot contain any poles.
Property 4 Ifx[n] is a finite-length sequence, then the ROC is the entire Z-plane withthe possibleexception ofz = 0 or z = .
Property 5 If x[n] is a right-sided sequence, i.e. x[n] = 0 for n < N1 < , for someN1 Z, then the ROC extends outwards from the outermostfinite pole in X(z).
Property 6 If x[n] is a left-sided sequence, i.e. x[n] = 0 for n > N2 > , for someN2 Z, then the ROC extends inwards from the innermostfinite pole in X(z).
Property 7 A two-sided sequence is an infinite-duration sequence that is neither right-sided nor left-sided. If x[n] is a two-sided sequence, then the ROC will consist ofa ring in the Z-plane bounded on the interior and exterior by a pole and does not
contain any poles.Property 8 The ROC must be a connected region.
6.2 The inverse Z-transform
By considering
X
rej
= F
x[n]rn
,
-
8/10/2019 Dsp n06 Chapter6
5/10
6.3. Partial fraction expansion 67
one can obtain
x[n]rn =F1
X
rej
,
or
x[n]rn = 1
2
X
rej
ejn d,
or
x[n] = 12
Xrej rejn d. (6.8)That is, we can recoverx[n] from integrating its Z-transform along a contour z = rej
in its ROC, by fixing r and varying from to . Withz = rej , for fixed r we seethat dz= jrej d= jz d or d= dz
jz.
Since the integral in (6.8) is over a 2 interval of, it corresponds to a traversal aroundthe circle |z|= r. Consequently
x[n] = 1
2j
X(z)zn1 dz, (6.9)
where the symbol denotes integration around a counter-clockwise closed circular contourcentered at the origin and with radius ofr. The value ofr can be chosen such that |z|= ris in the ROC.
6.3 Partial fraction expansion
X(z) =
M1k=0 bkz
kN1k=0 akz
k=
b0M1
k=1
1 ckz
1
a0N1
k=1 (1 dkz1)
. (6.10)
If M < N and all poles are simple, i.e. unique dks, we can do a partial fraction
expansion to get
X(z) =N1k=0
Ak1 dkz1
, (6.11)
where
Ak =
1 dkz1
X(z)
z=dk
. (6.12)
-
8/10/2019 Dsp n06 Chapter6
6/10
68 Chapter 6.
IfM > N, we can obtain the right form by first doing a long division of the numerator bythe denumerator. This gives gives a general form
X(z) =MNr=0
Brzr +
N1k=0
Ak1 dkz1
, (6.13)
where the Brs are obtained from the long division and the Aks are as before.IfX(Z) has a pole of order s atdi and all other poles are simple we get
X(z) =
MNr=0
Brzr +
N1k=0,k=i
Ak1 dkz1 +
sm=1
Cm(1 diz1)m , (6.14)
where
Cm= 1
(s m)!(di)sm
dsm
dwsm
(1 diw)s X
w1
w=d1
i
(6.15)
and the Aks andBrs are obtained as before.Using (6.14) and the ROC, we can invert the Z-transform using the inspection method,
i.e.
anu[n] Z
1
1 az1
, |z|> |a|. (6.16)
anu[n 1] Z
1
1 az1 , |z|< |a|. (6.17)
6.4 Z-transform properties
1. Linearity:
a1x1[n] + a2x2[n] Z a1X1(z) + a2X2(z), ROC containsRx1 Rx2
2. Time-shifting:
x[n n0] Zzn0X(z), ROC =Rx
(except for the possible addition or deletion ofz = 0 or z = )
3. Multiplication by exponential sequence:
zn0
x[n] Z X(z/z0) , ROC =|z0|Rx
-
8/10/2019 Dsp n06 Chapter6
7/10
6.4. Z-transform properties 69
4. Differentiation ofX(z):
nx[n] Z z
dX(z)
dz , ROC =Rx
Example: Inverse of a non-rationalZ-transform
X(z) = log
1 + az1
, |z|> |a|.
dX(z)
dz =
1
1 + az1.
az2
Hence
zdX(z)
dz =
az1
1 + az1Z1
a(a)n1u[n 1] |z|> |a|.
Hencenx[n] =a(a)n1u[n 1], or x[n] = an
(a)n1u[n 1]. Hence
(1)
n1an
nu[n 1]
Z
log 1 + az15. Conjugation of a complex sequence:
x[n] Z X(z) ROC =Rx
6. Time-reversal:
x[n] Z X
1
z
, ROC =
1
Rx
7. Convolution of sequences:
x1[n] x2[n] Z X1(z)X2(z), ROC containsRx1 Rx2
8. Initial value theorem: Ifx[n] is zero for n
-
8/10/2019 Dsp n06 Chapter6
8/10
70 Chapter 6.
6.5 Analysis and characterization of LTI systems using Z-transform
The Z-transform plays a particularly important role in the analysis and representation ofdiscrete-time LTI systems. From the convolution property, for an LTI system with impulse
response h[n] Z H(z) and an inputx[n]
Z X(z), we have,
y[n] = (h x)[n] =h[n] x[n] Z H(z)X(z), (6.18)
where X(z), Y(z) and H(z) are the Z-transforms and the ROC ofY(z) is
Ry =Rx Rh. (6.19)
H(z) is referred to as the system function or transfer function of the system.
6.5.1 Causality
A causal LTI system has an impulse response that is h[n] = 0 for n
-
8/10/2019 Dsp n06 Chapter6
9/10
6.5. Analysis and characterization of LTI systems using Z-transform 71
6.5.3 LTI systems and linear constant-coefficient difference equations
If
y[n] =N1k=1
aky[n k] +M1k=0
bkx[n k],
then
Y(z) =N1
k=1
akzkY(z) +
M1
k=0
bkzkX(z).
Hence
Y(z)
1
N1k=1
akzk
= X(z)
M1k=0
bkzk
or
Y(z)
X(z)=
M1k=0 bkz
k
1 N1
k=1 akzk
=H(z).
Therefore the system function for such systems are easy to write. This form ofH(z) =Y(z)X(z) is also called the transfer function of a linear constant-coefficient difference equationsystem.
-
8/10/2019 Dsp n06 Chapter6
10/10