Digital Signal Processing Reference
In-Depth Information
recursive procedures. It is possible to find many such procedures but perhaps the
most effective and the simplest result can be obtained using complex notation of a
pair of orthogonal sine, cosine filters. Such representation for the filters given by
Eqs.
6.26
and
6.27
is the following
ð
cos
ð
x
Þþ
j sin
ð
x
Þ¼
exp
ð
jx
ÞÞ
:
X
0
:
y
ð
n
Þ¼
y
c
ð
n
Þþ
jy
s
ð
n
Þ¼
X
N
1
N
1
2
x
ð
n
k
Þ
exp j
k
ð
6
:
35
Þ
k
¼
0
Rearranging (
6.35
) yields:
X
N
1
j
N
1
2
y
ð
n
Þ¼
exp
X
0
x
ð
n
k
Þ
exp
ð
jkX
0
Þ
ð
6
:
36
Þ
k
¼
0
and writing each component of the filters, output at instant n we obtain:
x
ð
n
Þþ
x
ð
n
1
Þ
exp
ð
jX
0
Þþþ
x
ð
n
N
þ
1
Þ
j
N
1
2
y
ð
n
Þ¼
exp
X
0
f
exp
j
ð
N
1
Þ
X
0
½
g:
ð
6
:
37
Þ
Doing the same at instant n - 1 one gets:
x
ð
n
1
Þþ
x
ð
n
2
Þ
exp
ð
jX
0
Þþþ
x
ð
n
N
Þ
j
N
1
2
y
ð
n
1
Þ¼
exp
X
0
f
exp
j
ð
N
1
Þ
X
0
½
g:
ð
6
:
38
Þ
Multiplying the last equation by adequately chosen complex function and
subtracting it from Eq.
6.37
one can reduce many components simplifying the final
equation to the form:
x
ð
n
Þ
x
ð
n
N
Þ
exp
ð
jX
0
Þ
j
N
1
2
y
ð
n
Þ
y
ð
n
1
Þ
exp
ð
jX
0
Þ¼
exp
X
0
½
:
ð
6
:
39
Þ
Comparing real and imaginary parts of the equation one can get finally the
recursive algorithm of a pair of orthogonal sine, cosine FIR filters.
Y
ð
n
Þ
CY
ð
n
1
Þþ
DX
;
ð
6
:
40
Þ
where
; X
¼
; C
¼
;
Y
ð
n
Þ¼
y
c
ð
n
Þ
y
s
ð
n
Þ
x
ð
n
Þ
x
ð
n
N
Þ
cos
ð
X
0
Þ
sin
ð
X
0
Þ
sin
ð
X
0
Þ
cos
ð
X
0
Þ
cos
ð
a
Þ
cos
ð
b
Þ
sin
ð
a
Þ
D
¼
sin
ð
b
Þ
a
¼
0
:
5
ð
N
1
Þ
X
0
;
b
¼
0
:
5
ð
N
þ
1
Þ
X
0
:
Search WWH ::
Custom Search