0 ratings0% found this document useful (0 votes) 33 views10 pagesSome FIR Filter Designs
Finite impulse response filters design
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content,
claim it here.
Available Formats
Download as PDF or read online on Scribd
7.2. MAGNITUDE RESPONSE AND PHASE RESPONSE
OF DIGITAL FILTERS
The discrete-time Fourier transform of a finite sequence impulse
response A(n) is given by
M-1
H(e/*) = > bine"? = |H(e™) |e *™ ay
a=0
The magnitude and phase responses are given by
M(o) = |H(e/®)| = (RelH(e?*)] * + Im [H(e*1")°*
= tan me)
©(@) = tan RelH(e™) (7.2)
Filters can have a linear or non-linear phase depending upon the
delay function, namely the phase delay and group delay. The phase and
group delays of the filter are given by
and t= -22), respectively.
B= my
_ @)
? o
382_ Digital Signal Processing
The group delay is defined as the delayed response of the filter as a
function of @ to a signal.
Linear phase filters are those filters in which the phase delay and
group delay are constants, i.e. independent of frequency. Linear phase
filters are also called constant time delay filters. Let us obtain the
conditions FIR filters must satisfy in order to have constant phase and
group delays and hence obtain the conditions for having a linear phase.
For the phase response to be linear
@@) _
o
Therefore,
Oo) =-a@t
where tis a constant phase delay expressed in number of samples. Using
Eq. 7.2,
-T, -KS@S+K
tan-! ImH(e*)
%o) = tan” Re Hie™)
=- Ot
or
Met
> A@)sinwn
-1 nwo
wt = tan 3-5
Y Al) cos on
n=O
Scanned with CamScanner M1
LAG)sinon
n=
tan ot = 2-0
h(n) cos wn
ned
Simplifying, we get
M-1
¥ A(n)sin(@t- wn) =0 (7.3)
a=0
and a solution to Eq.7.3 is given by :
(M-)
ae ae (7.4)
and
h(n) = h(M - 1-n) for 0 hn) cos2nn fT aay
and
A= > A(n)sin2xnfT (7.12)
—
From Eq. 7.11 and 7.12 we infer that H, (f) is an even function and
H((f) is an odd function of frequency. If h(nT’) is an even sequence, the
imaginary part of the transfer function, H,(f), will be zero and if h(n’)
is an odd sequence, the real part of the transfer function, H,( f), will be
zero. Thus an even unit impulse response yields a real transfer function
and an odd unit impulse response yields an imaginary transfer function.
A real transfer function has 0 or + x radians phase shift, while an
imaginary transfer function has + n/2 radians phase shift. Therefore, by
making the unit impulse response either even or odd, one can generate
a transfer function that is either real or imaginary.
‘Scanned with CamScanner in the Gesign of digital fliters two interesting situations are olten
sought after.
(i) For filtering applications, the main interest is in the amplitude
response of the filter, where some portion of the input signal spectrum
is to be attenuated and some portion is to be passed to the output with
no attenuation. This should be accomplished without phase distortion.
Thus the amplitude response is realised by using only a real transfer
function. That is
He®)=H(f)
and
H(f)= 0
(ii) For filtering plus quadrature phase shift. the applications include
integrators, differentiators and Hilbert transform devices. For all these
applications the desired transfer function is imaginary. Thus, the
required amplitude response is realised by using only Hf ). That is
He!®) =j H{f)
Hf) =0
and
Design Equations
The term H(e2) is periodic in the sampling frequency and hence both
H,{f) and H{ f) are also periodic in the sampling frequency. Both H,(f)
and Hf) can be expanded in a Fourier series. Since the real part of the
transfer function, H,(f), is an even function of frequency, its Fourier
series will be of the form
Hf) = a9+ 5, a, €08 (2xnfT) (7.13)
nel
Finite Impulse Response (FIR) Filters 387
The Fourier coefficients a, are given by
hl2
a,=2 J H,(f) cos (2nnfT) df, n#0 (7.14)
S -fl2
and the ap term is given by
1 Ae
a= f Hf) af
he iva
In a similar manner, the imaginary part of the transfer function,
which is an odd function of frequency, can be expanded in the Fourier
series
Hf) = > 6, sin (2xnfT) (7.15)
ml
The Fourier coefficients 6, are given by
— tlt
Scanned with CamScanner The Fourier coefficients 5, are given by
2 M2
b= J Hf) sin (2nnfT) df (7.16)
fs fra
Since an odd function has zero average value, by = 0.
The Fourier coefficients a,, and b, must be related to the unit pulse
response of the filter.
Case (i): For simple filtering applications, consider Hj(/) = 0. Equation
7.13 can be written as
Hee!®) = Hf) = 09+ 5, a, 008 (2nnfT)
=a,+ > + (2" +2"), sear (7.17)
=
The transfer function can be also written as
He’*)= Y hn)2|,_ janet (7.18)
=O) + Y [A(—n) 2" + An) 2°) eer
mt
Comparing Eqs 7.17 and 7.18, the terms of the unit pulse response
h(nT) can be related to the Fourier coefficients of the frequency
response. This yields
(0) = ay
1
hn) =a,
2" 1 n>o0 (7.19)
388 Digital Signal Processing
Case (ii): For filtering with quadrature phase shift. Consider H, (/)= 0.
From Eq, (7.15)
He?) = jH(f)=j Yb, sin (2xnfT)
nl
Comparing this equation with Eq. 7.18 yields
Alen) = 3 b,
(7.20)
n>0O (7.21)
hin)=-+,
Scanned with CamScanner Equations 7.19 and 7.21 are the even and odd unit pulse responses
respectively, and are used for the design of FIR filters.
Design Procedure
The procedure for designing an FIR digital filter using the Fourier series
method is summarised as follows.
1. Decide whether H,(f) or Hf) is to be set equal to zero. For
filtering applications we typically set H(f) = 0. For integrators,
differentiators and Hilbert transformers, we set H,{ f) = 0.
2. Expand H,( f) or H,{ f) in a Fourier series.
3. The unit pulse response is determined from the Fourier
coefficients using Eqs 7.19 and 7.21.
The number of taps, i.e. the value of M may have to be increased in
order to get a satisfactory sinusoidal steady-state response.
There are two problems involved in the implementation of FIR filters
using this technique. The transfer function H(e/’°) represents a non-
causal digital filter of infinite duration. A finite duration causal filter
can be obtained by truncating the infinite duration impulse response
and delaying the resulting finite duration impulse response. This
modification does not effect the amplitude response of the filter;
however, the abrupt truncation of the Fourier series results in
oscillations in the passband and stopband. These oscillations are due to
slow convergence of the Fourier series, particularly near the points of
discontinuity. This effect is known as the Gibbs phenomenon. These
undesirable oscillations can be reduced by multiplying the desired
impulse response coefficients by an appropriate window function.
Scanned with CamScanner TS EHO tO Gesign a low-pass
digital filter ts approximate the ideal specifications given by
Finite Impulse Response (FIR) Filters 389
1, forlfl0
F ie
f
-2 f 1 cos (onnf) af
tp
-2 snanelT
Fl Qn? J,
Replacing T by W/F, and multiplying both numerator and
denominator by f,, we get,
a - (28) sin 2 xnf,/F
"UR JL anf, /F)
Fi2 f 2F,
=i et f rape 2
aon | HNar pit 7
Therefore, from Eq. 7.19,
2f, f, \[ sin 2 xnf,/F
(0) = ag = eo A(-n) = (4) Seer |
_ (fy [sin 2 nnf,/F
and A(n) = (4) Serer |
Scanned with CamScanner Dace A low-pass filter has the desired response as given
low
; e" O<@<%
Hy(e/®)= x 2
0 gsost
Determine the filter coefficients h(n) for M = 7, using type-I frequency
sampling technique.
Solution The samples of the given frequency response is taken
uniformly at @, = 2nk/M. For 0 < w< 3 the values of k = 0,1. For
§ sos, k=2,3,4, 5and for sas 2x, k=6, Thus, the sampled
frequency response is given by
eJ6rA7, = 0,1
H(k)= 40, h=2,3,4,5
eJerhi7, pig
The filter coefficients h,(n) are given by the inverse discrete
Fourier transform,
i's xk _ 1 j2nkniT
han) = — > A(k)e’ =D Abe
M k=O 7 ke0
=i Serioan el2knlT , 9-J6nk/7 el2nhnty, 6
7 Lixo .
= 1 Si elttn-o0v7 +. gman]
kao
= 3 [as enna + eftxin-3y7)
Since H (k) = H"(M — k), we have e/!?*("-37 = g-J2x("- 397
Therefore,
ha(n) = 3fi+ g/min -37 , 9-J2nin |
h An) = 0.1429 + 0.2857 cos[0.898 (n — 3)]
Scanned with CamScanner
