1-Digital Filters (FIR)

AGC
DSP Digital Filter Specifications

 Only the magnitude approximation problem
 Four basic types of ideal filters with magnitude
responses as shown below (Piecewise flat)
H LP (e j ) HHP (e j )
1 1
 
 – c 0 c   – c 0 c 
HBS (e j )
HBP (e j )
–1 1
 
  – c2 – c1  c1  c2    – c2 – c1  c1  c2 
Professor A G Constantinides
1
AGC
 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.
2
AGC
 Consider the ideal LP response squared
(same as actual LP response)
j
H LP (e )

 – c 0 c 
– c 
  0  c 
3
AGC
 The realisable squared amplitude response
transfer function (and its differential) is
continuous 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
4
AGC
 A realisable response would effectively
need to have an approximation of the
delta functions in the differential
 This is a necessary condition
5
AGC
 For example the magnitude response
of a digital lowpass filter may be given as
indicated below
6
AGC
 In the passband 0     p we require
that G ( e j )  1 with a deviation   p
j
1   p  G (e )  1   p ,    p
 In the stopband  s     we require
j
that G (e )  0 with a deviation  s
j
G (e )   s , s    
7
AGC
Filter specification parameters
 
p - passband edge frequency
  - stopband edge frequency
s
 p - peak ripple value in the
passband
  s - peak ripple value in the
stopband
8
AGC
 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 log10dB
( s )
9
Digital Filter Specifications
AGC
DSP
 In practice, passband edge frequency F

p
and stopband edge frequency Fs are
specified in Hz
 For digital filter design, normalized bandedge
frequencies need to be computed from
specifications in Hz using
 p 2 Fp
p    2 FpT
FT FT
 s 2 Fs
s    2 Fs T
FT FT Professor A G Constantinides
10
AGC
 Example - Let Fp  7 kHz, Fs  3
kHz, and FT  25 kHz
 Then
2 (7  10 )
3
p  3
 0.56
25  10
2 (3  103 )
s  3
 0.24
25  10
11
Selection of Filter Type
AGC
DSP
 The transfer function H(z) meeting the

specifications must be a causal transfer
function
 For IIR real digital filter the transfer function
is a real rational function of z 1
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
AGC
DSP Selection of Filter Type
 FIR real digital filter transfer function is a
polynomial in z 1 (order N) with real
coefficients
N
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:

n]   h[ N  n]
(More on thish[later)
13
AGC
DSP Selection of Filter Type
 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
AGC
DSP 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
AGC
DSP Filters
 The transfer function is given by
N 1
n
H ( z )   h(n).z
n 0
 The length of Impulse Response is N

 All poles are at z  0 .
 Zeros can be placed anywhere on the z-
plane
16
AGC
DSP FIR: Linear phase
For phase linearity the FIR transfer

function must have zeros outside
the unit circle
17
AGC
 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
1 1
H ( z )  K  (1  i z ). (1  i z )
i 1 i 1
 Where  i  1,  i  1 K , is
real & zeros occur in conjugates
18
AGC
 Let H ( z )  KN1 ( z ) N 2 ( z )
where
n1 n2
1
N1 ( z )   (1  i z ) N 2 ( z )   (1  i z 1 )
i 1 i 1
 Thus n1 n2
1 1
ln( H ( z ))  ln( K )   ln(1   i z )   ln(1   i z )
i 1 i 1
19
AGC
 Expand in a Laurent Series
convergent within the unit circle
 To do so modify the second sum as
n2
1
n2
1
n2 1
 ln(1  i z )   ln( i z )   ln(1  z )
i 1 i 1 i 1 i
20
AGC
 So that
n1
1 1 n2
ln( H ( z ))  ln( K )  n2 ln( z )   ln(1   i z )   ln(1  z )
i 1 i 1 
 Thus i
N1 N2
 sm m s m m
ln(H ( z ))  ln(K )  n2 ln( z )   z  z
m 1 m m
 where n1 n1
smN1   i  m
sNm2    im
i 1 i 1
21
AGC
 smN1 are the root moments of the
minimum phase component
N2
 s m are the inverse root moments of
the maximum phase component
 Now on the unit circle we have z  e j
and H (e j )  A( )e j ( )
22
AGC
DSP Fundamental Relationships
N1 N2
 s s
ln( H (e j ))  ln( K )  jn2   m e  jm  m e jm
m 1 m m
ln(H (e j ))  ln( A( )e j ( ) )  ln( A( ))  j ( )
 hence (note Fourier form)

smN1 sNm2
ln( A( ))  ln( K )   (  ) cos m
m 1 m m
 smN1 sNm2
 ( )   n2   (  ) sin m
m 1 m m
23
AGC
 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 Professor A G Constantinides
24
AGC
 It follows that zeros of linear phase FIR
trasfer functions not on the
circumference of the unit circle occur in
the form
 i e 
 ji 1
25
AGC
 For Linear Phase t.f. (order N-1)

h( n)   h( N  1  n)
 so that
N
for N even:
1 N 1
2 n
H ( z )   h(n).z   h(n).z  n
n 0 n N
2
N 1 N 1
2 2
  h(n).z  n   h( N  1  n).z ( N 1 n )
n 0 n 0
N 1
2

  h( n) z  n  z  m
n 0
 m  NProfessor
 1AGnConstantinides
26
AGC
 for N odd:
N 1 N 1 
1

2
H ( z )   h(n). z
n 0
 n
z m
 
 h
N  1
 2 
z  2 

 I) On C : z  1 we have for N even,

and +ve sign
N 1  N 1
 jT  N  1 
jT
 2  
H (e )e  2 
.  2h(n). cos T  n  
n 0   2 
27
AGC
 II) While for –ve sign
N 1  N 1
 jT  N  1 

.  j 2h(n).sin  T  n 
 2
H (e jT )  e  2 

n 0   2 
 [Note: antisymmetric case adds  / 2 rads
to phase, with discontinuity at   0 ]
 III) For N odd with +ve sign
N 1
 j T  N  1
jT 
 2   
H (e )e h 
  2 
N 3

 N  1  
  2h( n). cos T 
2
 n   
n 0   2  
 28
AGC
 IV) While with a –ve sign
 N 1  N 3 
 jT 
 2  
2   N  1 
H (e jT )  e   2 j.h(n).sin T  n   
 n 0   2  
 
 [Notice that for the antisymmetric case to

have linear phase we require
N  1
h
   0.
 2 
The phase discontinuity is as for N even]

29
AGC
 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
2
30
AGC
DSP Design of FIR filters: Windows
(i) Start with ideal infinite duration  h(n)
(ii) Truncate to finite length. (This
produces unwanted ripples increasing in
height near discontinuity.)
~
(iii) Modify to h (n)  h(n).w(n)
Weight w(n) is the window
31
AGC
DSP Windows
Commonly used windows
 Rectangular 1  2 n n 
N 1
N
 Bartlett  2n  2
1  cos 
 N 
 Hann 2n 

 Hamming 0.54  0.46 cos N 

0 . 42  0 . 5 cos 2n   0.08 cos 4n 
 Blackman    
 N   N 

 2 n
2
J 0  1  


  J 0 ( )
 Kaiser 
  N  1 
32
AGC
DSP Kaiser window
 Kaiser window
β Transition Min. stop
width (Hz) attn dB
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
AGC
DSP Example
• Lowpass filter of length 51 and c   / 2
Lowpass Filter Designed Using Hann window Lowpass Filter Designed Using Hamming window
0 0
G ain, dB
G ain, dB
-50 -50
-100 -100
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

w /p w /p
Lowpass Filter Designed Using Blackman window
0
Gain, dB
-50
-100
0 0.2 0.4 0.6 0.8 1 Professor A G Constantinides

34
w /p
AGC
DSP 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 Professor A G Constantinides
35
AGC
Linear-Phase FIR Filter
DSP Design by Optimisation
 Amplitude response for all 4 types of
linear-phase FIR filters can be

expressed as H ( )  Q( ) A( )
 1, for Type 1
where
 cos(/2), for Type 2

Q( )  
 sin( ), for Type 3
sin( / 2), for Type 4
36
AGC
 Modified form of weighted error function
E ( )  W ( )[Q( ) A( )  D( )]
 W ( )Q( )[ A( )  QD(( )) ]
~ ~
 W ( )[ A( )  D( )]
where
~
W ( )  W ( )Q( )
~
D( )  D( ) / Q( )
37
AGC
 Optimisation Problem - Determine a~[k ]
which minimise the peak absolute
value of ~ L
~ ~
E ( )  W ( )[  a [k ] cos( k)  D( )]
k 0
over the specified frequency bands   R

 ~
After [k ] has been determined,
a
j
construct the original A(e ) and
hence h[n]
38
AGC
Solution 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
AGC
FIR Digital Filter Order
DSP Estimation
Kaiser’s Formula:
 20 log10 (  p s )
N
14.6( s   p ) / 2
 ie N is inversely proportional to
transition band width and not on
transition band location
40
AGC
DSP Estimation
 Hermann-Rabiner-Chan’s Formula:
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 Professor A G Constantinides
41
AGC
DSP Estimation
 Formula valid for  p   s
 For  p   s , formula to be used is
obtained by interchanging  p and  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
AGC
DSP Estimation
 Fred Harris’ guide:
A
N
20(s   p ) / 2
where A is the attenuation in dB

 Then add about 10% to it
43

1-Digital Filters (FIR)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1-Digital Filters (FIR)

Uploaded by

Copyright:

Available Formats

AGC

DSP Digital Filter Specifications

to a constant over a band of frequencies

that point cannot be zero.

 Hence previous defferential of ideal response

 In practice, passband edge frequency F

 The transfer function H(z) meeting the

 The length of Impulse Response is N

For phase linearity the FIR transfer

 hence (note Fourier form)

 I) On C : z  1 we have for N even,

 [Notice that for the antisymmetric case to

The phase discontinuity is as for N even]

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

0 0.2 0.4 0.6 0.8 1 Professor A G Constantinides

over the specified frequency bands   R

where A is the attenuation in dB

You might also like