Digital Signal Processing Reference
In-Depth Information
Appendix E: Finite Impulse Response Filter
Design Equations by the Frequency
Sampling Design Method
Recall in Section 7.5 in Chapter 7 on the “Frequency Sampling Design Method” that we obtained
N
1
k ¼
0
HðkÞW
kn
1
N
hðnÞ¼
(E.1)
N
where
hðnÞ
,0
n N
1, is the causal impulse response that approximates the finite impulse
response (FIR) filter,
HðkÞ
,0
k N
1, represents the corresponding coefficients of the discrete
Fourier transform (DFT), and
W
N
¼ e
j
2
p
N
. We further write the DFT coefficients,
HðkÞ
,0
k
N
1, in polar form:
HðkÞ¼H
k
e
j
4
k
;
0
k N
1
(E.2)
where H
k
and
4
k
are the
k
th magnitude and the phase angle, respectively. The frequency response of
the FIR filter is expressed as
N
1
n¼
0
hðnÞe
jn
U
Hðe
j
U
Þ¼
(E.3)
Substituting (E.1) into (E.3) yields
N
1
N
1
k ¼
0
HðkÞW
k
N
e
j
U
n
1
N
Hðe
j
U
Þ¼
(E.4)
n¼
0
Interchanging the order of the summation in Equation
(E.4)
leads to
N
1
k ¼
0
HðkÞ
N
1
n¼
0
ðW
N
e
j
U
Þ
n
1
N
Hðe
j
U
Þ¼
(E.5)
Since
W
N
e
j
U
¼ðe
j
2
p
=N
Þ
k
e
j
U
¼ e
ðj
U
2
p
k=NÞ
and using the identity
P
N
1
n¼
0
r
n
¼
1
þ r þ
1
r
N
1
r
þ
/
þ r
N
1
2
r
¼
, we can write the second summation in Equation
(E.5)
as
Search WWH ::
Custom Search