You are on page 1of 17

Generalized Linear Phase

Quote of the Day


The mathematical sciences particularly exhibit
order, symmetry, and limitation; and these are
the greatest forms of the beautiful.

Aristotle
Content and Figures are from Discrete-Time Signal Processing, 2e by Oppenheim, Shafer, and Buck, 1999-2000 Prentice Hall
Inc.
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
2
Linear Phase System
Ideal Delay System

Magnitude, phase, and group delay




Impulse response


If o=n
d
is integer

For integer o linear phase system delays the input


( ) t < e =
eo e
e e H
j j
id
( )
( )
( ) | | o =
eo = Z
=
e
e
e
j
id
j
id
j
id
e H grd
e H
1 e H
| |
( ) ( )
( ) o t
o t
=
n
n sin
n h
id
| | | |
d id
n n n h o =
| | | | | | | | | | | |
d d id
n n x n n n x n h n x n y = o - = - =
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
3
Linear Phase Systems
For non-integer o the output is an interpolation of samples
Easiest way of representing is to think of it in continuous


This representation can be used even if x[n] was not originally
derived from a continuous-time signal
The output of the system is


Samples of a time-shifted, band-limited interpolation of the
input sequence x[n]
A linear phase system can be thought as


A zero-phase system output is delayed by o
( ) ( ) ( )
T j
c c
e j H and T t t h
o O
= O o o =
| | ( ) T nT x n y o =
( ) ( )
eo e e
=
j j j
e e H e H
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
4
Symmetry of Linear Phase Impulse Responses
Linear-phase systems


If 2o is integer
Impulse response symmetric



( ) ( )
eo e e
=
j j j
e e H e H
| | | | n h n 2 h = o
o=5

o=4.5

o=4.3

Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
5
Generalized Linear Phase System
Generalized Linear Phase


Additive constant in addition to linear term
Has constant group delay


And linear phase of general form




( ) ( )
| + eo e e
=
j j j j
e e A e H
( )
constants and
of function Real : e A
j
| o
e
e
( ) ( ) | | ( ) | | ( ) o =
e
= = e t
e e j j
e H arg
d
d
e H grd
( ) | | t < e s eo | =
e
0 e H arg
j
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
6
Condition for Generalized Linear Phase
We can write a generalized linear phase system response as



The phase angle of this system is



Cross multiply to get necessary condition for generalized linear
phase
( ) | | | | ( ) | | ( ) n sin n h j n cos n h e n h e H
n n
n j
n
j
e e = =


=

=
e

=
e
( ) ( ) ( ) ( ) ( ) ( ) eo | + eo | = =
e e | + eo e e
sin e jA cos e A e e A e H
j j j j j j
( )
( )
| | ( )
| | ( ) n cos n h
n sin n h
cos
sin
n
n
e
e
=
eo |
eo |

=
| | ( ) ( ) | | ( ) ( )
| | ( ) ( ) ( ) ( ) | |
| | ( ) | | ( ) | | 0 n sin n h n sin n h
0 cos n sin sin n cos n h
0 cos n sin n h sin n cos n h
n n
n
n n
= o e + | = e + eo |
= eo | e eo | e
= eo | e eo | e

=
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
7
Symmetry of Generalized Linear Phase
Necessary condition for generalized linear phase


For |=0 or t



For |= t/2 or 3t/2

| | ( ) | | 0 n sin n h
n
= o e + |

=
| | ( ) | | | | | | n h n 2 h 0 n sin n h
n
= o = o e

=
| | ( ) | | | | | | n h n 2 h 0 n cos n h
n
= o = o e

=
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
8
Causal Generalized Linear-Phase System
If the system is causal and generalized linear-phase


Since h[n]=0 for n<0 we get


An FIR impulse response of length M+1 is generalized linear
phase if they are symmetric
Here M is an even integer





| | M n and 0 n 0 n h > < =
| | | | n h n M h =
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
9
Type I FIR Linear-Phase System
Type I system is defined with
symmetric impulse response


M is an even integer

The frequency response can
be written as




Where


| | | | M n 0 for n M h n h s s =
( ) | |
| | ( )
(

e =
=

=
e
e
=
e
n cos n a e
e n h e H
2 / M
0 n
2 / M j
n j
M
0 n
j
| | | |
| | | | M/2 1,2,..., k for k 2 / M h 2 k a
2 / M h 0 a
= =
=
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
10
Type II FIR Linear-Phase System
Type I system is defined with
symmetric impulse response


M is an odd integer

The frequency response can
be written as




Where


| | | | M n 0 for n M h n h s s =
( ) | |
| |
( )
(

|
|
.
|

\
|
|
.
|

\
|
e =
=

+
=
e
e
=
e
2
1
n cos n b e
e n h e H
2 / 1 M
1 n
2 / M j
n j
M
0 n
j
| | ( ) | |
( )/2 1 M 1,2,..., k for
k 2 / 1 M h 2 k b
+ =
+ =
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
11
Type III FIR Linear-Phase System
Type I system is defined with
symmetric impulse response


M is an even integer

The frequency response can
be written as




Where


| | | | M n 0 for n M h n h s s =
( ) | |
| | ( )
(

e =
=

=
e
e
=
e
n sin n c je
e n h e H
2 / M
1 n
2 / M j
n j
M
0 n
j
| | | |
M/2 1,2,..., k for
k 2 / M h 2 k c
=
=
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
12
Type IV FIR Linear-Phase System
Type I system is defined with
symmetric impulse response


M is an odd integer

The frequency response can
be written as




Where


| | | | M n 0 for n M h n h s s =
( ) | |
| |
( )
(

|
|
.
|

\
|
|
.
|

\
|
e =
=

+
=
e
e
=
e
2
1
n sin n d je
e n h e H
2 / 1 M
1 n
2 / M j
n j
M
0 n
j
| | ( ) | |
( )/2 1 M 1,2,..., k for
k 2 / 1 M h 2 k d
+ =
+ =
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
13
Location of Zeros for Symmetric Cases
For type I and II we have

So if z
0
is a zero 1/z
0
is also a zero of the system
If h[n] is real and z
0
is a zero z
0
*
is also a zero
So for real and symmetric h[n] zeros come in sets of four
Special cases where zeros come in pairs
If a zero is on the unit circle reciprocal is equal to conjugate
If a zero is real conjugate is equal to itself
Special cases where a zero come by itself
If z=1 both the reciprocal and conjugate is itself
Particular importance of z=-1

If M is odd implies that

Cannot design high-pass filter with symmetric FIR filter and M odd
| | | | ( ) ( )
1 M z
z H z z H n M h n h

= =
( ) ( ) ( ) 1 H 1 1 H
M
=
( ) 0 1 H =
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
14
Location of Zeros for Antisymmetric Cases
For type III and IV we have


All properties of symmetric systems holds
Particular importance of both z=+1 and z=-1
If z=1

Independent from M: odd or even
If z=-1

If M+1 is odd implies that

| | | | ( ) ( )
1 M z
z H z z H n M h n h

= =
( ) ( ) ( ) 1 H 1 1 H
1 M
=
+
( ) 0 1 H =
( ) ( ) ( ) 0 1 H 1 H 1 H = =
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
15
Typical Zero Locations
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
16
Relation of FIR Linear Phase to Minimum-Phase
In general a linear-phase FIR system is not minimum-phase
We can always write a linear-phase FIR system as


Where


And M
i
is the number of zeros
H
min
(z) covers all zeros inside the unit circle
H
uc
(z) covers all zeros on the unit circle
H
max
(z) covers all zeros outside the unit circle
( ) ( ) ( ) ( ) z H z H z H z H
max uc min
=
( ) ( )
i
M 1
min max
z z H z H

=
Copyright (C) 2005 Gner Arslan
351M Digital Signal Processing
17
Example
Problem 5.45

You might also like