Digital Signal Processing Reference
In-Depth Information
Denoting the impulse response by h(n), i.e. h
ð
n
Þ¼
F
f
d
ð
n
Þg;
one obtains the
discrete convolution equation allowing to calculate the output signal as a function
of the input signal and the system's impulse response:
y
ð
n
Þ¼
X
1
x
ð
k
Þ
h
ð
n
k
Þ
ð
4
:
32
Þ
k
¼
0
or symbolically
y
ð
n
Þ¼
x
ð
n
Þ
h
ð
n
Þ:
ð
4
:
33
Þ
If both the input signal and the system follow the causal conditions, i.e.
x(n) = 0 and h(n) = 0 for n \ 0, then the discrete convolution can be written in
the form:
y
ð
n
Þ¼
X
x
ð
k
Þ
h
ð
n
k
Þ¼
X
n
n
x
ð
n
k
Þ
h
ð
k
Þ
ð
4
:
34
Þ
k
¼
0
k
¼
0
Example
4.12
Applying
the
method
of
discrete
convolution
calculate
step
response of the system with the impulse response h
ð
n
Þ¼
exp
ð
n
Þ:
Solution In accordance with (
4.32
) one can derive:
y
ð
n
Þ¼
X
n
exp
ð
k
Þ¼
1
exp
ð
n
þ
1
Þ
1
exp
ð
1
Þ
:
k
¼
0
Difference Equation
Another way of description of digital transferring of an input series into an output
one is the linear difference equation with constant coefficients:
y
ð
n
Þ¼
X
a
k
y
ð
n
k
Þþ
X
M
N
1
b
k
x
ð
n
k
Þ;
ð
4
:
35
Þ
k
¼
1
k
¼
0
where a
k
,b
k
are constant factors.
To solve Eq.
4.35
M initial conditions are needed. The solution, i.e. the function
of y(n), can be reached multifold. One of the ways is application of the Z trans-
form, the other is an iterative approach, that is easy when a few output samples are
sought only, or when digital means are applied.
Example 4.13
Let there be given a difference equation of the form:
y
ð
n
Þ¼
x
ð
n
Þþ
ay
ð
n
1
Þ:
Search WWH ::
Custom Search