Digital Signal Processing Reference
In-Depth Information
Fig. 2.8 Discrete-time
domain representation of a
digital system
Digital System
y
(
n
),
Output
Signal
x
(
n
),
Input
Signal
h
(
n
)
y
(
n
) =
h
(
n
) *
x
(
n
)
y
ð
n
Þ¼
x
ð
n
Þ
h
ð
n
Þ¼
X
1
x
ð
k
Þ
h
ð
n
k
Þ
ð
2
:
6
Þ
k
¼1
Note For causal digital systems h(n) = 0 for n \ 0.
A graphical illustration of a digital system is shown in Fig. (
2.8
).
Eigenfunctions of LTI Digital Systems
If the input signal to an LTI digital system is x
ð
n
Þ¼
e
jxnT
s
¼
e
jnX
;
then the output
is given by:
y
ð
n
Þ¼
x
ð
n
Þ
h
ð
n
Þ¼
h
ð
n
Þ
x
ð
n
Þ¼
X
1
h
ð
k
Þ
e
j
ð
n
k
Þ
X
¼
e
jnX
X
1
h
ð
k
Þ
e
jkX
:
k
¼1
k
¼1
If one defines H
ð
e
jX
Þ¼
P
k
¼1
h
ð
k
Þ
e
jkX
;
then for an input of e
jnX
one obtains the
output y
ð
n
Þ¼
e
jnX
H
ð
e
jX
Þ:
Hence, e
jnX
is an eigenfunction, and the associated
eigenvalue is H
ð
e
jX
Þ
(Compare with the analogous result obtained for continuous-
time systems in
Sect. 1.2.3.4
).
Analyzing the Stability of Digital Systems in the Time Domain
For a digital system to be BIBO stable, the output y(n) should be bounded when
the input x(n) is bounded, i.e.,
j
y
ð
n
Þj
\
1
when
j
x
ð
n
Þj
\A(which is a finite constant)
;8
n
:
Using the inequality |a ? b| B |a| ? |b|, it follows that
j
X
1
x
ð
k
Þ
h
ð
n
k
Þj
X
1
j
x
ð
k
Þjj
h
ð
n
k
Þj:
k
¼1
k
¼1
If |x(k)| \ AVk, then
j
y
ð
n
Þj¼j
X
x
ð
k
Þ
h
ð
n
k
Þj
A
X
j
h
ð
n
k
Þj
A
X
1
1
1
j
h
ð
m
Þj;
k
¼1
k
¼1
m
¼1
where m = n - k. Hence, a digital system is BIBO-stable if
P
k
¼1
j
h
ð
k
Þj
\
1:
(Compare with the condition for stability of analog systems found in
Sect. 1.2.3.5
).
Search WWH ::
Custom Search