Feb.2008 DISP Lab 1
FIR and IIR Filter Design
Techniques
FIR 與 IIR 濾波器設計技巧
 Speaker: Wen-Fu Wang 王文阜
 Advisor: Jian-Jiun Ding 丁建均 教授
 E-mail: r96942061@ntu.edu.tw
 Graduate Institute of Communication Engineering
 National Taiwan University, Taipei, Taiwan, ROC
Feb.2008 DISP Lab 2
Outline
 Introduction
 IIR Filter Design by Impulse
invariance method
 IIR Filter Design by Bilinear
transformation method
 FIR Filter Design by Window function
technique
Feb.2008 DISP Lab 3
Outline
 FIR Filter Design by Frequency
sampling technique
 FIR Filter Design by MSE
 Conclusions
 References
Feb.2008 DISP Lab 4
Introduction
 Basic filter classification
 We put emphasis on the digital filter
now, and will introduce to the design
method of the FIR filter and IIR filter
respectively.
Filter
Analog Filter
Digital Filter
IIR Filter
FIR Filter
Feb.2008 DISP Lab 5
Introduction
 IIR is the infinite impulse response
abbreviation.
 Digital filters by the accumulator, the
multiplier, and it constitutes IIR filter
the way, generally may divide into
three kinds, respectively is Direct
form, Cascade form, and Parallel
form.
Feb.2008 DISP Lab 6
Introduction
 IIR filter design methods include the
impulse invariance, bilinear
transformation, and step invariance.
 We must emphasize at impulse
invariance and bilinear
transformation.
Feb.2008 DISP Lab 7
Introduction
 IIR filter design methods
Continuous frequency
band transformation
Impulse Invariance
method
Bilinear
transformation
method
Step invariance
method
IIR filter
Normalized analog
lowpass filter
Feb.2008 DISP Lab 8
Introduction
 The structures of IIR filter
Direct
form 1
Direct form2
b0
b1
b2 b2
b1
b0
-a1
-a2
-a1
-a2
x(n) x(n)Y(n) Y(n)
1
z−
1
z−
1
z−
1
z−
1
z−
1
z−
Feb.2008 DISP Lab 9
Introduction
 The structures of IIR filter
Cascade form
x(n) Y(n)
b0
b1
b2
-a1
-a2
-c1
-c2
d1
d2
Parallel form
Y(n)x(n)
b1
b0
d1
d0
E
-c1
-c2
-a1
-a2
1
z−
1
z−
1
z−
1
z−
1
z−
1
z−
1
z−
1
z−
Feb.2008 DISP Lab 10
Introduction
 FIR is the finite impulse response
abbreviation, because its design
construction has not returned to the
part which gives.
 Its construction generally uses Direct
form and Cascade form.
Feb.2008 DISP Lab 11
Introduction
 FIR filter design methods include the
window function, frequency sampling,
minimize the maximal error, and MSE.
 We must emphasize at window
function, frequency sampling, and MSE.
Window function
technique
Frequency
sampling technique
Minimize the
maximal error
FIR filter
Mean square
error
Feb.2008 DISP Lab 12
Introduction
 The structures of FIR filter
x(n) x(n)
b1
b2
b3
b4
b0
Y(n) Y(n)
Direct form Cascade form
b1
b2
d1
d2
b0
1
z−
1
z−
1
z−
1
z−
1
z−
1
z−
1
z−
1
z−
Feb.2008 DISP Lab 13
IIR Filter Design by Impulse
invariance method
 The most straightforward of these is
the impulse invariance transformation
 Let be the impulse response
corresponding to , and define the
continuous to discrete time
transformation by setting
 We sample the continuous time
impulse response to produce the
discrete time filter
( )ch t
( )cH s
( ) ( )ch n h nT=
Feb.2008 DISP Lab 14
IIR Filter Design by Impulse
invariance method
 The frequency response is the
Fourier transform of the continuous
time function
and hence
'( )H ω
*
( ) ( ) ( )c c
n
h t h nT t nTδ
∞
=−∞
= −∑
1 2
'( ) ( )c
k
H H j k
T T
π
ω ω
∞
=−∞
 
= − 
 
∑
Feb.2008 DISP Lab 15
IIR Filter Design by Impulse
invariance method
 The system function is
 It is the many-to-one transformation
from the s plane to the z plane.
1 2
( ) | )sT cz e
k
H z H s jk
T T
π∞
=
=−∞
 
= − 
 
∑
Feb.2008 DISP Lab 16
IIR Filter Design by Impulse
invariance method
 The impulse invariance
transformation does map the -axis
and the left-half s plane into the unit
circle and its interior, respectively
jω
Re(Z)
Im(Z)
1
S domain Z domain
sT
e
jω
σ
Feb.2008 DISP Lab 17
IIR Filter Design by Impulse
invariance method
 is an aliased version of
 The stop-band characteristics are
maintained adequately in the discrete time
frequency response only if the aliased tails
of are sufficiently small.
'( )H ω ( )cH jω
0 ω
'( )H ω
/Tπ 2 /Tπ
( )cH jω
Feb.2008 DISP Lab 18
IIR Filter Design by Impulse
invariance method
 The Butterworth and Chebyshev-I
lowpass designs are more appropriate
for impulse invariant transformation
than are the Chebyshev-II and elliptic
designs.
 This transformation cannot be applied
directly to highpass and bandstop
designs.
Feb.2008 DISP Lab 19
IIR Filter Design by Impulse
invariance method
 is expanded a partial fraction
expansion to produce
 We have assumed that there are no
multiple poles
 And thus
( )cH s
1
( )
N
k
c
k k
A
H s
s s=
=
−
∑
1
( ) ( )k
N
s t
c k
k
h t A e u t
=
= ∑
1
( ) ( )k
N
s nT
k
k
h n A e u n
=
= ∑
1
1
( )
1 k
N
k
s T
k
A
H z
e z−
=
=
−
∑
Feb.2008 DISP Lab 20
IIR Filter Design by Impulse
invariance method
 Example:
Expanding in a partial fraction
expansion, it produce
The impulse invariant transformation
yields a discrete time design with the
system function
2 2
( )
( )
c
s a
H s
s a b
+
=
+ +
1/ 2 1/ 2
( )cH s
s a jb s a jb
= +
+ + + −
( ) 1 ( ) 1
1/ 2 1/ 2
( )
1 1a jb T a jb T
H z
e z e z− + − − − −
= +
− −
Feb.2008 DISP Lab 21
IIR Filter Design by Bilinear
transformation method
 The most generally useful is the
bilinear transformation.
 To avoid aliasing of the frequency
response as encountered with the
impulse invariance transformation.
 We need a one-to-one mapping from
the s plane to the z plane.
 The problem with the transformation
is many-to-one.sT
z e=
Feb.2008 DISP Lab 22
IIR Filter Design by Bilinear
transformation method
 We could first use a one-to-one
transformation from to , which
compresses the entire s plane into
the strip
 Then could be transformed to z by
with no effect from aliasing.
s 's
Im( ')s
T T
π π
− ≤ ≤
's
's T
z e=
σ
jω
'σ
jω
/Tπ−
/Tπ
s domain s’ domain
Feb.2008 DISP Lab 23
IIR Filter Design by Bilinear
transformation method
 The transformation from to is
given by
 The characteristic of this
transformation is seen most readily
from its effect on the axis.
 Substituting and , we
obtain
s 's
12
' tanh ( )
2
sT
s
T
−
=
jω
s jω= ' 's jω=
12
' tan ( )
2
T
T
ω
ω −
=
Feb.2008 DISP Lab 24
IIR Filter Design by Bilinear
transformation method
 The axis is compressed into the
interval for in a one-to-
one method
 The relationship between and
is nonlinear, but it is approximately
linear at small .
( , )
T T
π π
− 'ω
ω
ω 'ω
'ω ω≈
-
ω
'ω
/Tπ
/Tπ−
Feb.2008 DISP Lab 25
IIR Filter Design by Bilinear
transformation method
 The desired transformation to is
now obtained by inverting
to produce
 And setting , which yields
12
' tanh ( )
2
sT
s
T
−
=
2 '
tanh( )
2
s T
s
T
=
s z
1
' ( )lns z
T
=
2 ln
tanh( )
2
z
s
T
=
1
1
2 1
( )
1
z
T z
−
−
−
=
+
Re(Z)
Im(Z)
1
S domain Z domain
1
2
1
2
T
s
z
T
s
+
=
−
jω
σ
Feb.2008 DISP Lab 26
IIR Filter Design by Bilinear
transformation method
 The discrete-time filter design is
obtained from the continuous-time
design by means of the bilinear
transformation
 Unlike the impulse invariant
transformation, the bilinear
transformation is one-to-one, and
invertible.
1 1
(2/ )(1 )/(1 )
( ) ( ) |c s T z z
H z H s − −
= − +
=
Feb.2008 DISP Lab 27
FIR Filter Design by Window
function technique
 Simplest FIR the filter design is
window function technique
 A supposition ideal frequency
response may express
where
( ) [ ]j j n
d d
n
H e h n eω ω
∞
−
=−∞
= ∑
1
[ ] ( )
2
j j n
d dh n H e e d
π
ω ω
π
ω
π −
= ∫
Feb.2008 DISP Lab 28
FIR Filter Design by Window
function technique
 To get this kind of systematic causal
FIR to be approximate, the most
direct method intercepts its ideal
impulse response!
[ ] [ ] [ ]dh n w n h n= g
( ) ( ) ( )dH W Hω ω ω= ∗
Feb.2008 DISP Lab 29
FIR Filter Design by Window
function technique
 Truncation of the Fourier series
produces the familiar Gibbs
phenomenon
 It will be manifested in ,
especially if is discontinuous.
( )H ω
( )dH ω
Feb.2008 DISP Lab 30
FIR Filter Design by Window
function technique
 1.Rectangular window
 2.Triangular window (Bartett window)
1, 0
[ ]
0,
n M
w n
otherwise
≤ ≤
= 

2 , 0
2
2[ ] 2 ,
2
0,
n Mn
M
n Mw n n M
M
otherwise
 ≤ ≤


= − < ≤


Feb.2008 DISP Lab 31
FIR Filter Design by Window
function technique
 1.Rectangular window
 2.Triangular window (Bartett window)
0 10 20 30 40 50 60
0
0.5
1
sequence (n)
T(n)
Rectangular window
0 10 20 30 40 50 60
0
0.5
1
sequence (n)
T(n)
Bartlett window
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-100
-50
0
50
100
pi units
FrequencyresponseT(jw)(dB)
Rectangular window
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-100
-50
0
50
100
pi units
FrequencyresponseT(jw)(dB)
Bartlett window
Feb.2008 DISP Lab 32
FIR Filter Design by Window
function technique
 3.HANN window
 4.Hamming window
1 2
1 cos , 0
[ ] 2
0,
n
n M
w n M
otherwise
π  
− ≤ ≤  =  


2
0.54 0.46cos , 0
[ ]
0,
n
n M
w n M
otherwise
π
− ≤ ≤
= 

Feb.2008 DISP Lab 33
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-100
-50
0
50
100
pi units
FrequencyresponseT(jw)(dB)
Hanning window
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-100
-50
0
50
100
pi units
FrequencyresponseT(jw)(dB)
Hamming window
0 10 20 30 40 50 60
0
0.5
1
sequence (n)
T(n)
Hanning window
0 10 20 30 40 50 60
0
0.5
1
sequence (n)
T(n)
Hamming window
FIR Filter Design by Window
function technique
 3.HANN window
 4.Hamming window
Feb.2008 DISP Lab 34
FIR Filter Design by Window
function technique
 5.Kaiser’s window
 6.Blackman window
2
0
0
2
[ 1 (1 ) ]
[ ] , 0,1,...,
[ ]
n
I
Mw n n M
I
β
β
− −
= =
2 4
0.42 0.5cos 0.08cos , 0
[ ]
0,
n n
n M
w n M M
otherwise
π π
− + ≤ ≤
= 

Feb.2008 DISP Lab 35
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-100
-50
0
50
100
pi units
FrequencyresponseT(jw)(dB)
Blackman window
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-150
-100
-50
0
50
100
pi units
FrequencyresponseT(jw)(dB)
Kaiser window
 5.Kaiser’s window
 6.Blackman window
0 10 20 30 40 50 60
0
0.5
1
sequence (n)
T(n)
Blackman window
0 10 20 30 40 50 60
0
0.5
1
sequence (n)
T(n)
Kaiser window
FIR Filter Design by Window
function technique
Feb.2008 DISP Lab 36
FIR Filter Design by Window
function technique
( / )s Mω
Window Peak sidelobe level
(dB)
Transition
bandwidth
Max. stopband
ripple(dB)
Rectangular -13 0.9 -21
Hann -31 3.1 -44
Hamming -41 3.3 -53
Blackman -57 5.5 -74
Feb.2008 DISP Lab 37
FIR Filter Design by Frequency
sampling technique
 For arbitrary, non-classical
specifications of , the calculation
of ,n=0,1,…,M, via an appropriate
approximation can be a substantial
computation task.
 It may be preferable to employ a
design technique that utilizes
specified values of directly,
without the necessity of determining
' ( )dH ω
( )dh n
' ( )dH ω
( )dh n
Feb.2008 DISP Lab 38
FIR Filter Design by Frequency
sampling technique
 We wish to derive a linear phase IIR
filter with real nonzero . The
impulse response must be symmetric
where are real and denotes
the integer part
( )h n
[ /2]
0
1
2 ( 1/ 2)
( ) 2 cos( )
1
M
k
k
k n
h n A A
M
π
=
+
= +
+
∑
kA [ / 2]M
0,1,...,n M=
Feb.2008 DISP Lab 39
FIR Filter Design by Frequency
sampling technique
 It can be rewritten as
where and
 Therefore, it may write
where
1
/ 2 /
0
/2
( )
N
j k N j kn N
k
k
k N
h n A e eπ π
−
=
≠
= ∑ 0,1,..., 1n N= −
1N M= + k N kA A −=
/ 2 /
( ) j k N j kn N
k kh n A e eπ π
=
1
0
/2
( ) ( )
N
k
k
k N
h n h n
−
=
≠
= ∑
0,1,..., 1n N= −
Feb.2008 DISP Lab 40
FIR Filter Design by Frequency
sampling technique
 with corresponding transform
where
 Hence
which has a linear phase
1
0
/2
( ) ( )
N
k
k
k N
H z H z
−
=
≠
= ∑
/
2 / 1
(1 )
( )
1
j k N N
k
k j k N
A e z
H z
e z
π
π
−
−
−
=
−
' ( 1)/2 sin / 2
( )
sin[( / / 2)]
j T N
k k
TN
H A e
k N T
ω ω
ω
π ω
− −
=
−
Feb.2008 DISP Lab 41
FIR Filter Design by Frequency
sampling technique
 The magnitude response
which has a maximum value
at where
' sin / 2
( )
sin[( / / 2)]
k k
TN
H A
k N T
ω
ω
π ω
=
−
kN A
/k sk Nω ω= 2 /s Tω π=
Feb.2008 DISP Lab 42
FIR Filter Design by Frequency
sampling technique
 The only nonzero contribution to
at is from , and hence
that
 Therefore, by specifying the DFT
samples of the desired magnitude
response at the
frequencies , and setting
'( )H ω
kω ω= '
( )kH ω
'( )k kH N Aω =
'
( )dH ω
kω
'
( ) /k d kA H Nω= ±
Feb.2008 DISP Lab 43
FIR Filter Design by Frequency
sampling technique
 We produce a filter design from
equation (5.1) for which
 The desired and actual magnitude
responses are equal at the N
frequencies
'
'( ) ( )k d kH Hω ω=
kω
Feb.2008 DISP Lab 44
FIR Filter Design by Frequency
sampling technique
 In between these frequencies, is
interpolated as the sum of the
responses , and its magnitude
does not, equal that of
'( )H ω
'
( )kH ω
'
( )dH ω
Feb.2008 DISP Lab 45
FIR Filter Design by Frequency
sampling technique
 Example: For an ideal lowpass filter
from , we would
choose
 The frequency samples are
indeed equal to the desired
' 1, 0,1,...,
( )
0, 1,...,[ / 2]
d k
k P
H
k P M
ω
=
= 
= +
'
( ) /k d kA H Nω= ±
( 1) / ( 1), 0,1,...,
0, 1,...,[ / 2]
k
k
M k P
A
k P M
 − + =
= 
= + '
( )kH ω
'
( )d kH ω
Feb.2008 DISP Lab 46
FIR Filter Design by Frequency
sampling technique
 The response is very similar to the
result form using the rectangular
window, and the stopband is similarly
disappointing.
 We can try to search for the optimum
value of the transition sample would
quickly lead us to a value of
approximately , k p≠0.38( 1) /( 1)p
pA M= − +
Feb.2008 DISP Lab 47
FIR Filter Design by MSE
 : The spectrum of the filter we
obtain
 : The spectrum of the desired
filter
 MSE=
( )H f
( )dH f
( ) ( )∫−
−
−
2/
2/
21 s
s
f
f ds dffHfHf
0 0.1 0.2 0.3 0.4 0.5
-0.5
0
0.5
1
1.5
Feb.2008 DISP Lab 48
FIR Filter Design by MSE
 Larger MSE, but smaller maximal
error
 Smaller MSE, but larger maximal
error
0 0.1 0.2 0.3 0.4
-0.5
0
0.5
1
1.5
0 0.1 0.2 0.3 0.4
-0.5
0
0.5
H(F) H(F) - H (F)d
0 0.1 0.2 0.3 0.4
-0.5
0
0.5
1
1.5
0 0.1 0.2 0.3 0.4
-0.5
0
0.5
H(F) H(F) - H (F)d
Feb.2008 DISP Lab 49
FIR Filter Design by MSE
 1. ( ) ( ) ( ) ( )∫∫ −−
−
−=−=
2/1
2/1
22/
2/
21
dFFHFRdffHfRfMSE d
f
f ds
s
s
( ) ( ) dFFHFnns d
k
n
∫ ∑−
=
−=
2/1
2/1
2
0
|| 2cos][ π
( ) ( ) ( ) ( ) dFFHFnnsFHFnns d
k
n
d
k
n
∫ ∑∑−
==






−





−=
2/1
2/1
00
2cos][2cos][ ππ
( ) ( )
1/2
1/2
0 0
[ ]cos 2 [ ]cos 2
k k
n
s n n F s F dF
τ
π τ π τ
−
= =
= ∑ ∑∫
( ) ( ) ( )
1/2 1/2
2
1/2 1/2
0
2 [ ]cos 2
k
d d
n
s n n F H F dF H F dFπ
− −
=
− +∑∫ ∫
Feb.2008 DISP Lab 50
FIR Filter Design by MSE
 2. when n ≠ τ,
when n = τ, n ≠ 0,
when n = τ, n = 0,
 3. The formula can be repressed as:
( ) ( ) 02cos2cos
2/1
2/1
=∫−
dFFFn τππ
( ) ( ) 2/12cos2cos
2/1
2/1
=∫−
dFFFn τππ
( ) ( ) 12cos2cos
2/1
2/1
=∫−
dFFFn τππ
( ) ( ) ( )dFFHdFFHFnnsnssMSE dd
k
n
k
n
∫∫ ∑∑ −−
==
+−+=
2/1
2/1
22/1
2/1
01
22
2cos][22/][]0[ π
Feb.2008 DISP Lab 51
FIR Filter Design by MSE
 4. Doing the partial differentiation:
 5. Minimize MSE: for all n’s
( )∫−
−=
∂
∂ 2/1
2/1
2]0[2
]0[
dFFHs
s
MSE
d ( ) ( )∫−
−=
∂
∂ 2/1
2/1
2cos2][
][
dFFHFnns
ns
MSE
dπ
0
][
=
∂
∂
ns
MSE
( )∫−
=
2/1
2/1
]0[ dFFHs d ( ) ( )∫−
=
2/1
2/1
2cos2][ dFFHFnns dπ
[ ] [0]
[ ] [ ]/ 2 for n=1,2,...,k
[ ] [ ]/ 2 for n=1,2,...,k
[ ] 0 for n<0 and n N
h k s
h k n s n
h k n s n
h n
=
+ =
− =
= ≥
Feb.2008 DISP Lab 52
Conclusions
 FIR advantage:
1. Finite impulse response
2. It is easy to optimalize
3. Linear phase
4. Stable
 FIR disadvantage:
1. It is hard to implementation than IIR
Feb.2008 DISP Lab 53
Conclusions
 IIR advantage:
1. It is easy to design
2. It is easy to implementation
 IIR disadvantage:
1. Infinite impulse response
2. It is hard to optimalize than FIR
3. Non-stable
Feb.2008 DISP Lab 54
References
 [1]B. Jackson, Digital Filters and Signal
Processing, Kluwer Academic Publishers 1986
 [2]Dr. DePiero, Filter Design by Frequency
Sampling, CalPoly State University
 [3]W.James MacLean, FIR Filter Design
Using Frequency Sampling
 [4] 蒙以正 , 數位信號處理 , 旗標 2005
 [5]Maurice G.Bellanger, Adaptive Digital
Filters second edition, Marcel dekker 2001
Feb.2008 DISP Lab 55
References
 [6] Lawrence R. Rabiner, Linear Program
Design of Finite Impulse Response Digital
Filters, IEEE 1972
 [7] Terrence J mc Creary, On Frequency
Sampling Digital Filters, IEEE 1972

Fir and iir filter_design

  • 1.
    Feb.2008 DISP Lab1 FIR and IIR Filter Design Techniques FIR 與 IIR 濾波器設計技巧  Speaker: Wen-Fu Wang 王文阜  Advisor: Jian-Jiun Ding 丁建均 教授  E-mail: r96942061@ntu.edu.tw  Graduate Institute of Communication Engineering  National Taiwan University, Taipei, Taiwan, ROC
  • 2.
    Feb.2008 DISP Lab2 Outline  Introduction  IIR Filter Design by Impulse invariance method  IIR Filter Design by Bilinear transformation method  FIR Filter Design by Window function technique
  • 3.
    Feb.2008 DISP Lab3 Outline  FIR Filter Design by Frequency sampling technique  FIR Filter Design by MSE  Conclusions  References
  • 4.
    Feb.2008 DISP Lab4 Introduction  Basic filter classification  We put emphasis on the digital filter now, and will introduce to the design method of the FIR filter and IIR filter respectively. Filter Analog Filter Digital Filter IIR Filter FIR Filter
  • 5.
    Feb.2008 DISP Lab5 Introduction  IIR is the infinite impulse response abbreviation.  Digital filters by the accumulator, the multiplier, and it constitutes IIR filter the way, generally may divide into three kinds, respectively is Direct form, Cascade form, and Parallel form.
  • 6.
    Feb.2008 DISP Lab6 Introduction  IIR filter design methods include the impulse invariance, bilinear transformation, and step invariance.  We must emphasize at impulse invariance and bilinear transformation.
  • 7.
    Feb.2008 DISP Lab7 Introduction  IIR filter design methods Continuous frequency band transformation Impulse Invariance method Bilinear transformation method Step invariance method IIR filter Normalized analog lowpass filter
  • 8.
    Feb.2008 DISP Lab8 Introduction  The structures of IIR filter Direct form 1 Direct form2 b0 b1 b2 b2 b1 b0 -a1 -a2 -a1 -a2 x(n) x(n)Y(n) Y(n) 1 z− 1 z− 1 z− 1 z− 1 z− 1 z−
  • 9.
    Feb.2008 DISP Lab9 Introduction  The structures of IIR filter Cascade form x(n) Y(n) b0 b1 b2 -a1 -a2 -c1 -c2 d1 d2 Parallel form Y(n)x(n) b1 b0 d1 d0 E -c1 -c2 -a1 -a2 1 z− 1 z− 1 z− 1 z− 1 z− 1 z− 1 z− 1 z−
  • 10.
    Feb.2008 DISP Lab10 Introduction  FIR is the finite impulse response abbreviation, because its design construction has not returned to the part which gives.  Its construction generally uses Direct form and Cascade form.
  • 11.
    Feb.2008 DISP Lab11 Introduction  FIR filter design methods include the window function, frequency sampling, minimize the maximal error, and MSE.  We must emphasize at window function, frequency sampling, and MSE. Window function technique Frequency sampling technique Minimize the maximal error FIR filter Mean square error
  • 12.
    Feb.2008 DISP Lab12 Introduction  The structures of FIR filter x(n) x(n) b1 b2 b3 b4 b0 Y(n) Y(n) Direct form Cascade form b1 b2 d1 d2 b0 1 z− 1 z− 1 z− 1 z− 1 z− 1 z− 1 z− 1 z−
  • 13.
    Feb.2008 DISP Lab13 IIR Filter Design by Impulse invariance method  The most straightforward of these is the impulse invariance transformation  Let be the impulse response corresponding to , and define the continuous to discrete time transformation by setting  We sample the continuous time impulse response to produce the discrete time filter ( )ch t ( )cH s ( ) ( )ch n h nT=
  • 14.
    Feb.2008 DISP Lab14 IIR Filter Design by Impulse invariance method  The frequency response is the Fourier transform of the continuous time function and hence '( )H ω * ( ) ( ) ( )c c n h t h nT t nTδ ∞ =−∞ = −∑ 1 2 '( ) ( )c k H H j k T T π ω ω ∞ =−∞   = −    ∑
  • 15.
    Feb.2008 DISP Lab15 IIR Filter Design by Impulse invariance method  The system function is  It is the many-to-one transformation from the s plane to the z plane. 1 2 ( ) | )sT cz e k H z H s jk T T π∞ = =−∞   = −    ∑
  • 16.
    Feb.2008 DISP Lab16 IIR Filter Design by Impulse invariance method  The impulse invariance transformation does map the -axis and the left-half s plane into the unit circle and its interior, respectively jω Re(Z) Im(Z) 1 S domain Z domain sT e jω σ
  • 17.
    Feb.2008 DISP Lab17 IIR Filter Design by Impulse invariance method  is an aliased version of  The stop-band characteristics are maintained adequately in the discrete time frequency response only if the aliased tails of are sufficiently small. '( )H ω ( )cH jω 0 ω '( )H ω /Tπ 2 /Tπ ( )cH jω
  • 18.
    Feb.2008 DISP Lab18 IIR Filter Design by Impulse invariance method  The Butterworth and Chebyshev-I lowpass designs are more appropriate for impulse invariant transformation than are the Chebyshev-II and elliptic designs.  This transformation cannot be applied directly to highpass and bandstop designs.
  • 19.
    Feb.2008 DISP Lab19 IIR Filter Design by Impulse invariance method  is expanded a partial fraction expansion to produce  We have assumed that there are no multiple poles  And thus ( )cH s 1 ( ) N k c k k A H s s s= = − ∑ 1 ( ) ( )k N s t c k k h t A e u t = = ∑ 1 ( ) ( )k N s nT k k h n A e u n = = ∑ 1 1 ( ) 1 k N k s T k A H z e z− = = − ∑
  • 20.
    Feb.2008 DISP Lab20 IIR Filter Design by Impulse invariance method  Example: Expanding in a partial fraction expansion, it produce The impulse invariant transformation yields a discrete time design with the system function 2 2 ( ) ( ) c s a H s s a b + = + + 1/ 2 1/ 2 ( )cH s s a jb s a jb = + + + + − ( ) 1 ( ) 1 1/ 2 1/ 2 ( ) 1 1a jb T a jb T H z e z e z− + − − − − = + − −
  • 21.
    Feb.2008 DISP Lab21 IIR Filter Design by Bilinear transformation method  The most generally useful is the bilinear transformation.  To avoid aliasing of the frequency response as encountered with the impulse invariance transformation.  We need a one-to-one mapping from the s plane to the z plane.  The problem with the transformation is many-to-one.sT z e=
  • 22.
    Feb.2008 DISP Lab22 IIR Filter Design by Bilinear transformation method  We could first use a one-to-one transformation from to , which compresses the entire s plane into the strip  Then could be transformed to z by with no effect from aliasing. s 's Im( ')s T T π π − ≤ ≤ 's 's T z e= σ jω 'σ jω /Tπ− /Tπ s domain s’ domain
  • 23.
    Feb.2008 DISP Lab23 IIR Filter Design by Bilinear transformation method  The transformation from to is given by  The characteristic of this transformation is seen most readily from its effect on the axis.  Substituting and , we obtain s 's 12 ' tanh ( ) 2 sT s T − = jω s jω= ' 's jω= 12 ' tan ( ) 2 T T ω ω − =
  • 24.
    Feb.2008 DISP Lab24 IIR Filter Design by Bilinear transformation method  The axis is compressed into the interval for in a one-to- one method  The relationship between and is nonlinear, but it is approximately linear at small . ( , ) T T π π − 'ω ω ω 'ω 'ω ω≈ - ω 'ω /Tπ /Tπ−
  • 25.
    Feb.2008 DISP Lab25 IIR Filter Design by Bilinear transformation method  The desired transformation to is now obtained by inverting to produce  And setting , which yields 12 ' tanh ( ) 2 sT s T − = 2 ' tanh( ) 2 s T s T = s z 1 ' ( )lns z T = 2 ln tanh( ) 2 z s T = 1 1 2 1 ( ) 1 z T z − − − = + Re(Z) Im(Z) 1 S domain Z domain 1 2 1 2 T s z T s + = − jω σ
  • 26.
    Feb.2008 DISP Lab26 IIR Filter Design by Bilinear transformation method  The discrete-time filter design is obtained from the continuous-time design by means of the bilinear transformation  Unlike the impulse invariant transformation, the bilinear transformation is one-to-one, and invertible. 1 1 (2/ )(1 )/(1 ) ( ) ( ) |c s T z z H z H s − − = − + =
  • 27.
    Feb.2008 DISP Lab27 FIR Filter Design by Window function technique  Simplest FIR the filter design is window function technique  A supposition ideal frequency response may express where ( ) [ ]j j n d d n H e h n eω ω ∞ − =−∞ = ∑ 1 [ ] ( ) 2 j j n d dh n H e e d π ω ω π ω π − = ∫
  • 28.
    Feb.2008 DISP Lab28 FIR Filter Design by Window function technique  To get this kind of systematic causal FIR to be approximate, the most direct method intercepts its ideal impulse response! [ ] [ ] [ ]dh n w n h n= g ( ) ( ) ( )dH W Hω ω ω= ∗
  • 29.
    Feb.2008 DISP Lab29 FIR Filter Design by Window function technique  Truncation of the Fourier series produces the familiar Gibbs phenomenon  It will be manifested in , especially if is discontinuous. ( )H ω ( )dH ω
  • 30.
    Feb.2008 DISP Lab30 FIR Filter Design by Window function technique  1.Rectangular window  2.Triangular window (Bartett window) 1, 0 [ ] 0, n M w n otherwise ≤ ≤ =   2 , 0 2 2[ ] 2 , 2 0, n Mn M n Mw n n M M otherwise  ≤ ≤   = − < ≤  
  • 31.
    Feb.2008 DISP Lab31 FIR Filter Design by Window function technique  1.Rectangular window  2.Triangular window (Bartett window) 0 10 20 30 40 50 60 0 0.5 1 sequence (n) T(n) Rectangular window 0 10 20 30 40 50 60 0 0.5 1 sequence (n) T(n) Bartlett window 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -100 -50 0 50 100 pi units FrequencyresponseT(jw)(dB) Rectangular window 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -100 -50 0 50 100 pi units FrequencyresponseT(jw)(dB) Bartlett window
  • 32.
    Feb.2008 DISP Lab32 FIR Filter Design by Window function technique  3.HANN window  4.Hamming window 1 2 1 cos , 0 [ ] 2 0, n n M w n M otherwise π   − ≤ ≤  =     2 0.54 0.46cos , 0 [ ] 0, n n M w n M otherwise π − ≤ ≤ =  
  • 33.
    Feb.2008 DISP Lab33 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -100 -50 0 50 100 pi units FrequencyresponseT(jw)(dB) Hanning window 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -100 -50 0 50 100 pi units FrequencyresponseT(jw)(dB) Hamming window 0 10 20 30 40 50 60 0 0.5 1 sequence (n) T(n) Hanning window 0 10 20 30 40 50 60 0 0.5 1 sequence (n) T(n) Hamming window FIR Filter Design by Window function technique  3.HANN window  4.Hamming window
  • 34.
    Feb.2008 DISP Lab34 FIR Filter Design by Window function technique  5.Kaiser’s window  6.Blackman window 2 0 0 2 [ 1 (1 ) ] [ ] , 0,1,..., [ ] n I Mw n n M I β β − − = = 2 4 0.42 0.5cos 0.08cos , 0 [ ] 0, n n n M w n M M otherwise π π − + ≤ ≤ =  
  • 35.
    Feb.2008 DISP Lab35 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -100 -50 0 50 100 pi units FrequencyresponseT(jw)(dB) Blackman window 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -150 -100 -50 0 50 100 pi units FrequencyresponseT(jw)(dB) Kaiser window  5.Kaiser’s window  6.Blackman window 0 10 20 30 40 50 60 0 0.5 1 sequence (n) T(n) Blackman window 0 10 20 30 40 50 60 0 0.5 1 sequence (n) T(n) Kaiser window FIR Filter Design by Window function technique
  • 36.
    Feb.2008 DISP Lab36 FIR Filter Design by Window function technique ( / )s Mω Window Peak sidelobe level (dB) Transition bandwidth Max. stopband ripple(dB) Rectangular -13 0.9 -21 Hann -31 3.1 -44 Hamming -41 3.3 -53 Blackman -57 5.5 -74
  • 37.
    Feb.2008 DISP Lab37 FIR Filter Design by Frequency sampling technique  For arbitrary, non-classical specifications of , the calculation of ,n=0,1,…,M, via an appropriate approximation can be a substantial computation task.  It may be preferable to employ a design technique that utilizes specified values of directly, without the necessity of determining ' ( )dH ω ( )dh n ' ( )dH ω ( )dh n
  • 38.
    Feb.2008 DISP Lab38 FIR Filter Design by Frequency sampling technique  We wish to derive a linear phase IIR filter with real nonzero . The impulse response must be symmetric where are real and denotes the integer part ( )h n [ /2] 0 1 2 ( 1/ 2) ( ) 2 cos( ) 1 M k k k n h n A A M π = + = + + ∑ kA [ / 2]M 0,1,...,n M=
  • 39.
    Feb.2008 DISP Lab39 FIR Filter Design by Frequency sampling technique  It can be rewritten as where and  Therefore, it may write where 1 / 2 / 0 /2 ( ) N j k N j kn N k k k N h n A e eπ π − = ≠ = ∑ 0,1,..., 1n N= − 1N M= + k N kA A −= / 2 / ( ) j k N j kn N k kh n A e eπ π = 1 0 /2 ( ) ( ) N k k k N h n h n − = ≠ = ∑ 0,1,..., 1n N= −
  • 40.
    Feb.2008 DISP Lab40 FIR Filter Design by Frequency sampling technique  with corresponding transform where  Hence which has a linear phase 1 0 /2 ( ) ( ) N k k k N H z H z − = ≠ = ∑ / 2 / 1 (1 ) ( ) 1 j k N N k k j k N A e z H z e z π π − − − = − ' ( 1)/2 sin / 2 ( ) sin[( / / 2)] j T N k k TN H A e k N T ω ω ω π ω − − = −
  • 41.
    Feb.2008 DISP Lab41 FIR Filter Design by Frequency sampling technique  The magnitude response which has a maximum value at where ' sin / 2 ( ) sin[( / / 2)] k k TN H A k N T ω ω π ω = − kN A /k sk Nω ω= 2 /s Tω π=
  • 42.
    Feb.2008 DISP Lab42 FIR Filter Design by Frequency sampling technique  The only nonzero contribution to at is from , and hence that  Therefore, by specifying the DFT samples of the desired magnitude response at the frequencies , and setting '( )H ω kω ω= ' ( )kH ω '( )k kH N Aω = ' ( )dH ω kω ' ( ) /k d kA H Nω= ±
  • 43.
    Feb.2008 DISP Lab43 FIR Filter Design by Frequency sampling technique  We produce a filter design from equation (5.1) for which  The desired and actual magnitude responses are equal at the N frequencies ' '( ) ( )k d kH Hω ω= kω
  • 44.
    Feb.2008 DISP Lab44 FIR Filter Design by Frequency sampling technique  In between these frequencies, is interpolated as the sum of the responses , and its magnitude does not, equal that of '( )H ω ' ( )kH ω ' ( )dH ω
  • 45.
    Feb.2008 DISP Lab45 FIR Filter Design by Frequency sampling technique  Example: For an ideal lowpass filter from , we would choose  The frequency samples are indeed equal to the desired ' 1, 0,1,..., ( ) 0, 1,...,[ / 2] d k k P H k P M ω = =  = + ' ( ) /k d kA H Nω= ± ( 1) / ( 1), 0,1,..., 0, 1,...,[ / 2] k k M k P A k P M  − + = =  = + ' ( )kH ω ' ( )d kH ω
  • 46.
    Feb.2008 DISP Lab46 FIR Filter Design by Frequency sampling technique  The response is very similar to the result form using the rectangular window, and the stopband is similarly disappointing.  We can try to search for the optimum value of the transition sample would quickly lead us to a value of approximately , k p≠0.38( 1) /( 1)p pA M= − +
  • 47.
    Feb.2008 DISP Lab47 FIR Filter Design by MSE  : The spectrum of the filter we obtain  : The spectrum of the desired filter  MSE= ( )H f ( )dH f ( ) ( )∫− − − 2/ 2/ 21 s s f f ds dffHfHf 0 0.1 0.2 0.3 0.4 0.5 -0.5 0 0.5 1 1.5
  • 48.
    Feb.2008 DISP Lab48 FIR Filter Design by MSE  Larger MSE, but smaller maximal error  Smaller MSE, but larger maximal error 0 0.1 0.2 0.3 0.4 -0.5 0 0.5 1 1.5 0 0.1 0.2 0.3 0.4 -0.5 0 0.5 H(F) H(F) - H (F)d 0 0.1 0.2 0.3 0.4 -0.5 0 0.5 1 1.5 0 0.1 0.2 0.3 0.4 -0.5 0 0.5 H(F) H(F) - H (F)d
  • 49.
    Feb.2008 DISP Lab49 FIR Filter Design by MSE  1. ( ) ( ) ( ) ( )∫∫ −− − −=−= 2/1 2/1 22/ 2/ 21 dFFHFRdffHfRfMSE d f f ds s s ( ) ( ) dFFHFnns d k n ∫ ∑− = −= 2/1 2/1 2 0 || 2cos][ π ( ) ( ) ( ) ( ) dFFHFnnsFHFnns d k n d k n ∫ ∑∑− ==       −      −= 2/1 2/1 00 2cos][2cos][ ππ ( ) ( ) 1/2 1/2 0 0 [ ]cos 2 [ ]cos 2 k k n s n n F s F dF τ π τ π τ − = = = ∑ ∑∫ ( ) ( ) ( ) 1/2 1/2 2 1/2 1/2 0 2 [ ]cos 2 k d d n s n n F H F dF H F dFπ − − = − +∑∫ ∫
  • 50.
    Feb.2008 DISP Lab50 FIR Filter Design by MSE  2. when n ≠ τ, when n = τ, n ≠ 0, when n = τ, n = 0,  3. The formula can be repressed as: ( ) ( ) 02cos2cos 2/1 2/1 =∫− dFFFn τππ ( ) ( ) 2/12cos2cos 2/1 2/1 =∫− dFFFn τππ ( ) ( ) 12cos2cos 2/1 2/1 =∫− dFFFn τππ ( ) ( ) ( )dFFHdFFHFnnsnssMSE dd k n k n ∫∫ ∑∑ −− == +−+= 2/1 2/1 22/1 2/1 01 22 2cos][22/][]0[ π
  • 51.
    Feb.2008 DISP Lab51 FIR Filter Design by MSE  4. Doing the partial differentiation:  5. Minimize MSE: for all n’s ( )∫− −= ∂ ∂ 2/1 2/1 2]0[2 ]0[ dFFHs s MSE d ( ) ( )∫− −= ∂ ∂ 2/1 2/1 2cos2][ ][ dFFHFnns ns MSE dπ 0 ][ = ∂ ∂ ns MSE ( )∫− = 2/1 2/1 ]0[ dFFHs d ( ) ( )∫− = 2/1 2/1 2cos2][ dFFHFnns dπ [ ] [0] [ ] [ ]/ 2 for n=1,2,...,k [ ] [ ]/ 2 for n=1,2,...,k [ ] 0 for n<0 and n N h k s h k n s n h k n s n h n = + = − = = ≥
  • 52.
    Feb.2008 DISP Lab52 Conclusions  FIR advantage: 1. Finite impulse response 2. It is easy to optimalize 3. Linear phase 4. Stable  FIR disadvantage: 1. It is hard to implementation than IIR
  • 53.
    Feb.2008 DISP Lab53 Conclusions  IIR advantage: 1. It is easy to design 2. It is easy to implementation  IIR disadvantage: 1. Infinite impulse response 2. It is hard to optimalize than FIR 3. Non-stable
  • 54.
    Feb.2008 DISP Lab54 References  [1]B. Jackson, Digital Filters and Signal Processing, Kluwer Academic Publishers 1986  [2]Dr. DePiero, Filter Design by Frequency Sampling, CalPoly State University  [3]W.James MacLean, FIR Filter Design Using Frequency Sampling  [4] 蒙以正 , 數位信號處理 , 旗標 2005  [5]Maurice G.Bellanger, Adaptive Digital Filters second edition, Marcel dekker 2001
  • 55.
    Feb.2008 DISP Lab55 References  [6] Lawrence R. Rabiner, Linear Program Design of Finite Impulse Response Digital Filters, IEEE 1972  [7] Terrence J mc Creary, On Frequency Sampling Digital Filters, IEEE 1972

Editor's Notes

  • #8 有三種古典continuous
  • #19 because the former are monotonic in the stopband , while the latter are not
  • #20 The impulse invariant transformation is not usually performed directly in the form of (2.1) the parameters of H(z) may be obtained directly from H(s)
  • #23 優點:To avoid aliasing of the frequency response 缺點:It is nonlinear between discrete-time frequency and continuous-time frequency.
  • #31 1.main lobe越窄 resolution越高 side lobe 越低越好 2.統計上常用 resolution降一半 main lobe 變寬(trade off) Main lobe變寬(trade off) side lobe降一半
  • #33 1.In fact, the length of window is M-1 2.main lobe和HANN差不多但side lobe降了10dB 3.Hamming 常用在語音處理
  • #35 有參數可調,能得適當的組合