Transcript Document
Relationship between Magnitude and Phase
Quote of the Day
Experience is the name everyone gives to their
mistakes.
Oscar Wilde
Content and Figures are from Discrete-Time Signal Processing, 2e by Oppenheim, Shafer, and Buck, ©1999-2000 Prentice Hall
Inc.
Relation between Magnitude and Phase
• For general LTI system
– Knowledge about magnitude doesn’t provide any information
about phase
– Knowledge about phase doesn’t provide any information about
magnitude
• For linear constant-coefficient difference equations however
– There is some constraint between magnitude and phase
– If magnitude and number of pole-zeros are known
• Only a finite number of choices for phase
– If phase and number of pole-zeros are known
• Only a finite number of choices for magnitude (ignoring scale)
• A class of systems called minimum-phase
– Magnitude specifies phase uniquely
– Phase specifies magnitude uniquely
Copyright (C) 2005 Güner Arslan
351M Digital Signal Processing
2
Square Magnitude System Function
• Explore possible choices of system function of the form
He
j
H e j H e j H* 1 / z* Hz
2
z e j
• Restricting the system to be rational
1 c z
M
b0
Hz
a0
1 c z
M
1
k
k 1
N
*
1 d z
k
H 1/z
1
*
b0
a0
k 1
Cz Hz H 1 / z
*
1 d z
2
k
k 1
N
1
*
k
1 d z 1 d z
k
1
k 1
j
*
k
1 c z 1 c z
M
*
k 1
N
k 1
• The square system function
b0
a0
*
k
*
k
2
• Given H e
we can get C(z)
• What information on H(z) can we get from C(z)?
Copyright (C) 2005 Güner Arslan
351M Digital Signal Processing
3
Poles and Zeros of Magnitude Square System Function
1 c z 1 c z
M
Cz Hz H 1 / z
*
*
b0
a0
2
k
k 1
N
1
*
k
1 d z 1 d z
k
1
k 1
*
k
• For every pole dk in H(z) there is a pole of C(z) at dk and
(1/dk)*
• For every zero ck in H(z) there is a zero of C(z) at ck and
(1/ck)*
• Poles and zeros of C(z) occur in conjugate reciprocal pairs
• If one of the pole/zero is inside the unit circle the reciprocal
will be outside
– Unless there are both on the unit circle
• If H(z) is stable all poles have to be inside the unit circle
– We can infer which poles of C(z) belong to H(z)
• However, zeros cannot be uniquely determined
– Example to follow
Copyright (C) 2005 Güner Arslan
351M Digital Signal Processing
4
Example
• Two systems with
1 z 1 2z
H z
1 0.8e z 1 0.8e
2 1 z1 1 0.5z1
H1 z
1 0.8e j / 4z1 1 0.8e j / 4z1
1
2
1
j / 4 1
j / 4 1
z
• Both share the same magnitude square system function
H1 z
Copyright (C) 2005 Güner Arslan
H2 z
351M Digital Signal Processing
H z H 1 / z
Cz H1 z H1* 1 / z*
2
*
2
*
5
All-Pass System
• A system with frequency response magnitude constant
• Important uses such as compensating for phase distortion
• Simple all-pass system
z 1 a*
Hap z
1 az1
• Magnitude response constant
Hap e
j
* j
e j a*
j 1 a e
e
j
1 ae
1 ae j
• Most general form with real impulse response
z1 dk Mc z1 ck* z1 ck
Hap z A
1
1
1 ck*z1
k 1 1 dk z
k 1 1 ck z
Mr
• A: positive constant, dk: real poles, ck: complex poles
Copyright (C) 2005 Güner Arslan
351M Digital Signal Processing
6
Phase of All-Pass Systems
* j
e j a*
j 1 a e
Hap e
e
j
1 ae
1 ae j
• Let’s write the phase with a represented in polar form
j
e j re j
r sin
2
arctan
1 r cos
j j
1
re
e
• The group delay of this system can be written as
e j re j
1 r2
1 r2
grd
j j
2
2
j j
1
re
e
1
2
r
cos
r
1 re e
• For stable and causal system |r|<1
– Group delay of all-pass systems is always positive
• Phase between 0 and is always negative
argHap ej 0 for 0
Copyright (C) 2005 Güner Arslan
351M Digital Signal Processing
7