Digital Filter Specifications

AG
C
DS
P
Digital Filter Specifications

Only the magnitude approximation problem
Four basic types of ideal filters with magnitude
responses as shown below (Piecewise flat)
HLP (e j )
HHP (e j )
c 0
HBP (e j )
c
0
HBS (e j )
c2 c1
c1 c2
c2 c1
c1 c2
Professor A G Constantinides
AG
C
DS
P

These filters are unealisable because
(one of the following is sufficient)
their impulse responses infinitely long
non-causal
Their amplitude responses cannot be
equal to a constant over a band of
frequencies
Another perspective that provides some
understanding can be obtained by
looking at the ideal amplitude squared.
AG
C
DS
P

Consider the ideal LP response squared
(same as actual LP response)
j
HLP (e
c 0
AG
C
DS
P

The realisable squared amplitude
response transfer function (and its
differential) iscontinuous in
Such functions
if IIR can be infinite at point but
around that point cannot be zero.
if FIR cannot be infinite anywhere.
Hence previous defferential of ideal
response is unrealisable
AG
C
DS
P

A realisable response would
effectively need to have an
approximation of the delta
functions in the differential
This is a necessary condition
AG
C
DS
P

For example the magnitude response
of a digital lowpass filter may be
given as indicated below
AG
C
DS
P

In the passband0 p
we
require G
that
with a
( e j ) 1
deviation
j
1 p G (e ) 1 p , p
j
G
(
e
) 0
In the stopband
require that
j
G (e ) s ,
we s
with a deviation
AG
C
DS
P

Filter specification parameters
p - passband edge frequency
s - stopband edge frequency

- peak ripple value in the
p
passband
s - peak ripple value in the

stopband
AG
C
DS
P
Practical specifications are often given

in terms of loss function (in dB)
G ( ) 20 log10 G (e j )
Peak passband ripple
p 20 log10 (1 p )
dB
Minimum stopband attenuation
s 20 log10 ( s )
dB
AG
C
DS
P
In practice, passband edge frequency

Fp
and stopband edge frequency
are
Fs
specified in Hz
For digital filter design, normalized
bandedge frequencies need to be
computed from specifications in Hz
p 2 Fp
using
2 FpT
FT
FT
s 2 Fs
s
2 Fs T
FT
FT
10
AG
C
DS
P
Example - LetFp 7
kHz,Fs 3
kHz, and FT 25
kHz
Then
2 (7 103 )
p
0.56
3
25 10
2 (3 103 )
s
0.24
3
25 10
11
Selection of Filter Type
AG
C
DS
P
The transfer function H(z) meeting the

specifications must be a causal transfer
function
For IIR real digital filter the transfer
1
function is a real rational function
of
z
p0 p1 z 1 p2 z 2 pM z M
H ( z)
d 0 d1 z 1 d 2 z 2 d N z N
H(z) must be stable and of lowest order N

or M for reduced computational
complexity
12
AG
C
DS
P

FIR real digital filter transfer function is a
polynomial in
(order N) with real
1
z
coefficients
N
H ( z ) h[n] z
n 0
For reduced computational complexity,

degree N of H(z) must be as small as possible
If a linear phase is desired then we must
have:
(More on thish[later)
n] h[ N n]
13
AG
C
DS
P

Advantages in using an FIR filter (1) Can be designed with exact linear
phase
(2) Filter structure always stable with
quantised coefficients
Disadvantages in using an FIR filter - Order
of an FIR filter is considerably higher than
that of an equivalent IIR filter meeting the
same specifications; this leads to higher
computational complexity for FIR
14
AG
C
DS
P
FIR Design
FIR Digital Filter Design
Three commonly used approaches
to FIR filter design (1) Windowed Fourier series
approach
(2) Frequency sampling approach
(3) Computer-based optimization
methods
15
Finite Impulse Response

Filters
AG
C
DS
P
The transfer function is given by

N 1
H ( z ) h(n).z
n 0
The length of Impulse Response is

N
z0
All poles are at
.
Zeros can be placed anywhere on
the z-plane
16
AG
C
DS
P
FIR: Linear phase

For phase linearity the FIR
transfer function must have
zeros outside the unit circle
17
AG
C
FIR: Linear phase
DS
P
To develop expression for phase

response set transfer function (order n)
H ( z ) h0 h1 z 1 h2 z 2 ... hn z n
In factored form
n1
n2
H ( z ) K (1 i z ). (1 i z )
i 1
i 1
Where
,
1, i 1 K
i
& zeros occur in conjugates
is real
18
AG
C
FIR: Linear phase
DS
P
Let H ( z ) KN1 ( z ) N 2 ( z )
where
n1
N1 ( z ) (1 i z )
i 1
Thus
n1
n2
N 2 ( z ) (1 i z 1 )
i 1
n2
ln( H ( z )) ln( K ) ln(1 i z ) ln(1 i z )

i 1
i 1
19
AG
C
FIR: Linear phase
DS
P
Expand in a Laurent Series

convergent within the unit
circle
To do so modify the second
n2
n2
n2
1
1
1
i z ) ln( i z ) ln(1 z )
ln(1 as
sum
i 1
i 1
i 1
20
AG
C
FIR: Linear phase
DS
P
So that
1
ln( H ( z )) ln( K ) n2 ln( z ) ln(1 i z ) ln(1 z )
i 1
i 1
i
n1
n2
Thus
N1
sm
m 1
ln( H ( z )) ln( K ) n2 ln( z )
where
smN1
n1
i
i 1
sNm2
N2
s m
n1
im
i 1
21
AG
C
DS
P
FIR: Linear phase

N1
sm
are the root moments of the

minimum phase component
N2
s m
are the inverse root moments

of the maximum phase component
z e j
Now on thej unit circle
j ( ) we have
H (e ) A( )e
and
22
AG
C
Fundamental Relationships
DS
P
N1
N2
s
s
ln( H (e j )) ln( K ) jn2 m e jm m e jm
m
m 1 m
ln( H (e j )) ln( A( )e j ( ) ) ln( A( )) j ( )
hence (note Fourier form)

smN1 sNm2
ln( A( )) ln( K ) (
) cos m
m
m 1 m
smN1 sNm2
( ) n2 (
) sin m
m
m 1 m
23
AG
C
DS
P
FIR: Linear phase
Thus for linear phase the second term in

the fundamental phase relationship must
be identically zero for all index values.
Hence
1) the maximum phase factor has zeros
which are the inverses of the those of
the minimum phase factor
2) the phase response is linear with
group delay (normalised) equal to the
number of zeros outside the unit circle
24
AG
C
FIR: Linear phase
DS
P
It follows that zeros of linear phase

FIR trasfer functions not on the
circumference of the unit circle
occur in the form
ji 1
i e
25
AG
C
FIR: Linear phase
DS
P
For Linear Phase t.f. (order N-1)
h( n) h( N 1 n)
so that
for
N
even:
N
2
H ( z ) h( n).z
n 0
N 1
h( n).z n
n N
N 1
2
N 1
2
n 0
n 0
h( n).z n h( N 1 n).z ( N 1 n )
N 1
2
h( n) z n z m
n 0
m NProfessor
1AG nConstantinides
26
AG
C
FIR: Linear phase
DS
P
for N odd:
N 1
1
2
H ( z ) h(n). z
n 0
N 1
h
z
2
N 1
I) On C : z 1 we have for N
even, and +ve sign
H (e
jT
)e
N 1 N 1
2
2
jT
N 1
. 2h(n). cos T n

2
n 0
27
AG
C
DS
P
FIR: Linear phase

II) While for ve sign
H ( e j T ) e
N 1 N 1
2
2
jT
N 1
. j 2h(n).sin T n

2
n 0
/2
[Note: antisymmetric case adds
0 at
rads to phase, with discontinuity
]
N 1
jT with
III) For
N
odd
+ve
sign
1
N
jT
2
H (e
)e
N 3
2
N 1
2h( n). cos T n

2
n 0
28
AG
C
DS
P
FIR: Linear phase
IV) While with a ve sign
H (e
j T
)e
N 1
jT
2
N 3
2
N 1
2 j.h( n).sin T n
2
n 0
[Notice that for the antisymmetric case

to have linear phase we require
N 1
0.
2
The phase discontinuity is as for N even]

29
AG
C
FIR: Linear phase
DS
P
The cases most commonly used in

filter design are (I) and (III), for
which the amplitude characteristic
can be written as a polynomial in
T
cos
30
AG
C
DS
P
Design of FIR filters: Windows

h(n)
(i) Start with ideal infinite duration
(ii) Truncate to finite length. (This
produces unwanted ripples
increasing in height near
~
discontinuity.)
h (n) h(n).w(n)
(iii) Modify to
Weight w(n) is the window
31
AG
C
Windows
DS
P
Commonly used windows

2n
1
N 1
Rectangular
n
N
2
2n
Bartlett
1 cos
N
Hann
2n
Hamming 0.54 0.46 cos N
2n 0.08 cos 4n
0
.
42
0
.
5
cos
Blackman
N
N
Kaiser
J 0 1
2n
N 1
J 0 ( )
32
AG
C
Kaiser window
DS
P
Kaiser window
Transition Min. stop

width
attn dB
(Hz)
2.12
1.5/N
30
4.54
2.9/N
50
6.76
4.3/N
70
8.96
5.7/N
90
33
AG
C
Example
DS
P
Lowpass filter of length 51 and
LowpassFilterDesignedUsingHammingwindow
0
G ain,dB
50
100
100
0
0.2
0.4
0.6
0.8
0.2
0.4
LowpassFilterDesignedUsingBlackmanwindow
0
Gain,dB
G ain,dB
LowpassFilterDesignedUsingHannwindow
0
50
c / 2
0.6
0.8
50
100
0
0.2
0.4
0.6
0.8
34
AG
C
DS
P
Frequency Sampling Method
In this approach we are given H (k ) and

need to find H (z )
This is an interpolation problem and the
solution is given in the DFT part of the
course
N
1 N 1
1 z
H ( z ) H ( k ).
2
N k 0
j k
1 e N .z 1
It has similar problems to the windowing

approach
35
Linear-Phase FIR Filter

Design by Optimisation
AG
C
DS
P
Amplitude response for all 4 types

of linear-phase FIR filters can be
expressed as H ( ) Q( ) A( )
where
Q( )
1,
for Type 1
cos(/2), for Type 2
sin( ), for Type 3
sin( / 2), for Type 4
36
AG
C
DS
P

Design by Optimisation
Modified form of weighted error

function
E ( ) W ( )[Q( ) A( ) D ( )]
D ( )
W ( )Q( )[ A( ) Q ( ) ]
~
~
W ( )[ A( ) D( )]
where
~
W ( ) W ( )Q( )
~
D( ) D ( ) / Q ( )
37
AG
C
DS
P

Design by Optimisation ~
a [k ]
Optimisation Problem - Determine
which minimise the peak
L
~
~
absolute
value
of
~
E ( ) W ( )[ a [k ] cos( k) D ( )]
k 0
R
over the specified frequency bands
~[k ]
a
After
has been determined,
j
A(e )
construct the original
and
hence h[n]
38
AG
C
DS
P

Design
Optimisation
Solutionby
is obtained
via the
Alternation Theorem
The optimal solution has equiripple
behaviour consistent with the total
number of available parameters.
Parks and McClellan used the Remez
algorithm to develop a procedure
for designing linear FIR digital
filters.
39
AG
C
DS
P
FIR Digital Filter Order

Estimation
Kaisers Formula:
N
20 log10 ( p s )
14.6( s p ) / 2
ie N is inversely proportional to
transition band width and not on
transition band location
40

Estimation
Hermann-Rabiner-Chans Formula:
AG
C
DS
P
D ( p , s ) F ( p , s )[( s p ) / 2 ]2
N
( s p ) / 2
where
D ( p , s ) [a1 (log10 p ) 2 a2 (log10 p ) a3 ] log10 s

[a4 (log10 p ) 2 a5 (log10 p ) a6 ]
F ( p , s ) b1 b2 [log10 p log10 s ]
with a1 0.005309, a2 0.07114, a3 0.4761
a4 0.00266, a5 0.5941, a6 0.4278
b1 11.01217, b2 0.51244
41
AG
C
DS
P

Estimation
p s
Formula valid for
For p s , formula to be used is
obtained by interchanging
and
p
s
Both formulae provide only an
estimate of the required filter order N
If specifications are not met, increase
filter order until they are met
42
AG
C
DS
P

Estimation
Fred Harris guide:
A
N
20(s p ) / 2
where A is the attenuation in dB

Then add about 10% to it
43

Digital Filter Specifications

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Digital Filter Specifications

Hochgeladen von

Copyright:

Verfügbare Formate

AG