Scribd
Upload
38 views

FIR Filter Design: - Linear Phase Filter Design - Magnitude Filter Design - Equalizer Design

This document summarizes lecture material on FIR filter design. It discusses linear phase filter design by imposing symmetry constraints on the impulse response. It also describes magnitude filter design using autocorrelation coefficients and spectral factorization to cast it as a linear program. Finally, it covers equalizer design problems to modify an existing system's frequency response, which can be formulated as LPs through variable changes. The examples shown include designing lowpass filters and equalizing multiple systems.

Uploaded by

Manas Bhatt
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
38 views

FIR Filter Design: - Linear Phase Filter Design - Magnitude Filter Design - Equalizer Design

This document summarizes lecture material on FIR filter design. It discusses linear phase filter design by imposing symmetry constraints on the impulse response. It also describes magnitude filter design using autocorrelation coefficients and spectral factorization to cast it as a linear program. Finally, it covers equalizer design problems to modify an existing system's frequency response, which can be formulated as LPs through variable changes. The examples shown include designing lowpass filters and equalizing multiple systems.

Uploaded by

Manas Bhatt
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 20

L.

Vandenberghe

EE236A (Fall 2013-14)

Lecture 10
FIR filter design

linear phase filter design


magnitude filter design
equalizer design

101

Finite impulse response (FIR) filter


y(t) =

n1
X
=0

h u(t )

u : Z R is input signal; y : Z R is output signal

hi R are filter coefficients; n is filter order or length


frequency response: a function H : R C defined as
H() = h0 + h1e
=

n1
X
t=0

+ + hn1e

ht cos t j

n1
X

j(n1)

(with j =

1)

ht sin t

t=0

periodic and conjugate symmetric; we only need to consider [0, ]


design problem: choose hi so that H satisfies/optimizes specifications
FIR filter design

102

Example: lowpass FIR filter


impulse response (order n = 21)

h(t)

0.2
0.1
0
0.1
0.2
0

10

12

14

16

18

20

frequency response: magnitude |H()| and phase 6 H()


1

3
2

H()

10

10

1
0
1

10

|H()|

10

10

0.5

1.5

FIR filter design

2.5

3
0

0.5

1.5

2.5

103

Linear-phase filters
suppose n = 2N + 1 is odd and impulse response is symmetric about hN :
ht = hn1t,

t = 0, . . . , n 1

frequency response
H() = h0 + h1ej + + hn1ej(n1)

= ejN (2h0 cos N + 2h1 cos(N 1) + + hN )


= ejN G()

term ejN represents N -sample delay


G() is real-valued and |H()| = |G()|
linear phase:
FIR filter design

H() is linear except for jumps of


104

Lowpass filter specifications


1
1/1

2
p s

maximum passband ripple (20 log10 1 in dB):


1/1 |H()| 1

for [0, p]

minimum stopband attenuation (20 log10 2 in dB):


|H()| 2
FIR filter design

for [s, ]
105

Linear-phase lowpass filter design


sample the frequency axis: k = k/K, k = 0, . . . , K 1
assume without loss of generality that G(0) > 0, so ripple spec. is
1/1 G(k ) 1
maximum stopband attenuation (for given passband ripple 1)
minimize 2
subject to 1/1 G(k ) 1 for k [0, p]
2 G(k ) 2 for k [s, ]
a linear program in variables hi, 2
known and used since 1960s
can add other constraints, e.g., |hi|
FIR filter design

106

Example
linear-phase filter of order n = 31
passband [0, 0.12]; stopband [0.24, ]
maximum ripple 1 = 1.059 (0.5dB)
0.2
0.15

h(t)

0.1
0.05
0
0.05
0.1

10

15

20

25

30

t
1

10

|H()|

10

10

10

10

10

0.5

1.5

2.5

FIR filter design

107

Variations
minimize passband ripple (variables 1, h)
minimize 1
subject to 1/1 G(k ) 1 for k [0, p]
2 G(k ) 2 for k [s, ]
minimize transition bandwidth (variables s, h)
minimize s
subject to 1/1 G(k ) 1 for k [0, p]
2 G(k ) 2 for k [s, ]
minimize filter order (variables N , h)
minimize N
subject to 1/1 G(k ) 1 for k [0, p]
2 G(k ) 2 for k [s, ]
not LPs, but can be solved by bisection/LP feasibility problems
FIR filter design

108

Outline

linear phase filter design


magnitude filter design
equalizer design

Filter magnitude specifications


magnitude specification: a constraint
L() |H()| U ()

L, U : R R+ are given and


H() =

n1
X
t=0

ht cos t j

n1
X

ht sin t

t=0

arises in many applications, e.g., audio, spectrum shaping


not equivalent to linear inequalities in hi (linear inequalities can not
express the lower bound on absolute value)
can change variables and convert to set of linear inequalities
FIR filter design

109

Autocorrelation coefficients
definition: autocorrelation coefficients of h = (h0, . . . , hn1) Rn
rt =

n1t
X

h h +t

=0

(with hk = 0 for k < 0 or k n)

rt = rt and rt = 0 for |t| n; hence suffices to specify r = (r0, . . . , rn1)


Fourier transform of autocorrelation coefficients:
R() =

ej r = r0 +

n1
X
t=1

2rt cos t = |H()|2

magnitude specifications are linear inequalities in coefficients rt:


L()2 R() U ()2
FIR filter design

for [0, ]
1010

Spectral factorization
when is r Rn the vector of autocorrelation coefficients of some h Rn?
spectral factorization theorem: if and only if R() 0 for all
condition is an infinite set of linear inequalities in r
many algorithms for spectral factorization (find h s.t. R() = |H()|2)
consequence: to cast magnitude design problem as an LP,
use r = (r0, . . . , rn1) as variable instead of h = (h0, . . . , hn1)
add spectral factorization condition as constraint: R() 0 for all
discretize the frequency axis
optimize over r and use spectral factorization to recover h
FIR filter design

1011

Magnitude lowpass filter design


maximum stopband attenuation design (with variables r)
minimize 2
subject to 1/1 R() 1 for [0, p]
R() 2 for [s, ]
R() 0 for [0, ]
(i corresponds to i2 in original problem)
discretization: impose constraints at finite set of frequencies k
minimize 2
subject to 1/1 R(k ) 1 for k [0, p]
R(k ) 2 for k [s, ]
R(k ) 0 for k [0, ]
this is a linear program in r, 2
FIR filter design

1012

Outline

linear phase filter design


magnitude filter design
equalizer design

Equalizer design
g(t)

h(t)

(time-domain) equalization
given g (unequalized impulse response), gdes (desired impulse response)
design FIR equalizer h so that convolution g = h g approximates gdes
example
gdes is pure delay D: gdes(t) =

1 t=D
0 t=
6 D

find equalizer h by solving


minimize

max |
g (t)|
t6=D

subject to g(D) = 1
this can be cast as an LP in the coefficients hi
FIR filter design

1013

Example
unequalized system (10th order FIR)
impulse response
1
0.8

g(t)

0.6
0.4
0.2
0
0.2
0.4
0

frequency response magnitude


1

10

G()

10

1
0
1

|G()|

2
1

10

0.5

1.5

FIR filter design

2.5

3
0

0.5

1.5

2.5

1014

time-domain equalization (30th order FIR, D = 10)


minimize max |
g (t)|
t6=10

equalized system impulse response


1

g
(t)

0.8
0.6
0.4
0.2
0
0.2
0

10

15

20

25

30

35

equalized frequency response


1

10

e
G()

10

1
0
1

e
|G()|

2
1

10

0.5

1.5

FIR filter design

2.5

3
0

0.5

1.5

2.5

1015

Magnitude equalizer design


H()

G()

problem
given system frequency response G : [0, ] C
design FIR equalizer H so that |G()H()| 1:
minimize



2
max |G()H()| 1

[0,]

LP formulation: use autocorrelation coefficients as variables


minimize


2

subject to
|G()| R() 1 for [0, ]
R() 0 for [0, ]
after discretizing the frequency axis, we obtain an LP in r and
FIR filter design

1016

Multi-system magnitude equalization


problem
we are given M frequency responses Gk : [0, ] C
design FIR equalizer H so that |Gk ()H()| constant:
minimize

max



2

max |Gk ()H()| k

k=1,...,M [0,]

subject to k 1,

k = 1, . . . , M

LP formulation: use autocorrelation coefficients as variables


minimize


2

subject to |Gk ()| R() k
R() 0 for [0, ]
k 1, k = 1, . . . , M

for [0, ],

k = 1, . . . , M

after discretizing the frequency axis, we obtain an LP in k , r,


FIR filter design

1017

Example
M = 2 systems, equalizer of order n = 25

2.5

2.5

|Gk ()H()|2

|Gk ()|2

unequalized and equalized frequency responses

1.5

0.5

0
0

0.5

1.5

FIR filter design

2.5

1.5

0.5

0
0

0.5

1.5

2.5

1018

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy