Transform analysis of LTI systems

Oppenheim and Schafer, Second edition pp. 240339.

kD1
.k|hn k|.
Alternatively, this relationship can be expressed in the z-transform domain as
Y(:) = H(:)X(:).
where H(:) is the system function, or the z-transform of the system impulse
response.
Recall that a LTI system is completely characterised by its impulse response,
or equivalently, its system function.
1 Frequency response of LTI systems
The frequency response H(e
j!
) of a system is dened as the gain that the
system applies to the complex exponential input e
j!n
. The Fourier transforms
of the system input and output are therefore related by
Y(e
j!
) = H(e
j!
)X(e
j!
).
In terms of magnitude and phase,
[Y(e
j!
)[ = [H(e
j!
)[[X(e
j!
)[
^Y(e
j!
) = ^H(e
j!
) ^X(e
j!
).
In this case [H(e
j!
)[ is referred to as the magnitude response or gain of the
system, and ^H(e
j!
) is the phase response or phase shift.
1
1.1 Ideal frequency-selective lters
Frequency components of the input are suppressed in the output if [H(e
j!
)[ is
small at those frequencies. The ideal lowpass lter is dened as the LTI
system with frequency response
H
lp
(e
j!
)
_
_
_
1 [o[ _ o
c
0 o
c
< [o[ _ .
Its magnitude and phase are
0
0
1
0
0
o
o
o
c
o
c
o
c
o
c

[
H
(
e
j
!
)
[
^
H
(
e
j
!
)
This response, as for all discrete-time signals, is periodic with period 2. Its
impulse response (for o< n < o) is
h
lp
n| =
1
2
_
!
c
!
c
e
j!n
Jo =
1
2
_
1
n
e
j!n
_
!
c
!
c
=
1
n
1
2
(e
j!
c
n
e
j!
c
n
) =
sin(o
c
n)
n
.
which for o
c
= ,4 is
2
20 15 10 5 0 5 10 15 20
0.1
0
0.1
0.2
0.3
n
h
l
p

n
|
The ideal lowpass lter is noncausal, and its impulse response extends from
o< n < o. The system is therefore not computationally realisable. Also,
the phase response of the ideal lowpass lter is specied to be zero this is a
problem in that causal ideal lters have nonzero phase responses.
The ideal highpass lter is
H
hp
(e
j!
) =
_
_
_
0 [o[ _ o
c
1 o
c
< [o[ _ .
Since H
hp
(e
j!
) = 1 H
lp
(e
j!
), its frequency response is
h
hp
n| = n| h
lp
n| = n|
sin(o
c
n)
n
.
1.2 Phase distortion and delay
Consider the ideal delay, with impulse response
h
id
n| = n n
d
|
and frequency response
H
id
(e
j!
) = e
j!n
d
.
3
The magnitude and phase of this response are
[H
id
(e
j!
)[ = 1.
^H
id
(e
j!
) = on
d
. [o[ < .
The phase distortion of the ideal delay is therefore a linear function of o. This
is considered to be a rather mild (and therefore acceptable) form of phase
distortion, since the only effect is to shift the sequence in time. In other words,
a lter with linear phase response can be viewed as a cascade of a zero-phase
lter, followed by a time shift or delay.
In designing approximations to ideal lters, we are therefore frequently willing
to accept linear phase distortion. The ideal lowpass lter with phase distortion
would be dened as
H
lp
(e
j!
) =
_
_
_
e
j!n
d
[o[ _ o
c
0 o
c
< [o[ _ .
with impulse response
h
lp
n| =
sin(o
c
(n n
d
)
(n n
d
)
.
A convenient measure of linearity of the phase is the group delay, which
relates to the effect of the phase on a narrowband signal. Consider the
narrowband input .n| = sn| cos(o
0
n), where sn| is the envelope of the
signal. Since X(e
j!
) is nonzero only around o = o
0
, the effect of the phase of
the system can be approximated around o = o
0
by
^H(e
j!
) ~
0
on
d
.
Thus the response of the system to .n| = sn| cos(o
0
n) is approximately
,n| = sn n
d
| cos(o
0
n
0
o
0
n
d
). The time delay of the envelope sn|
of the narrowband signal .n| with Fourier transform centred at o
0
is therefore
given by the negative of the slope of the phase at o
0
. The group delay of a
4
system is therefore dened as
t(o) = grdH(e
j!
)| =
J
Jo

argH(e
j!
)|
_
.
The deviation of the group delay away from a constant indicates the degree of
nonlinearity of the phase. Note that the phase here must be considered as a
continuous function of o.
2 System response for LCCD systems
Ideal lters cannot be implemented with nite computation. Therefore we
need approximations to ideal lters. Systems described by LCCD equations
N

kD0
a
k
,n k| =
M

kD0
b
k
.n k|
are useful for providing one class of approximation.
The properties of this class of system are best developed in the z-transform
domain. The z-transform of the equation is
N

kD0
a
k
:
k
Y(:) =
M

kD0
b
k
:
k
X(:).
or equivalently
_
N

kD0
a
k
:
k
_
Y(:) =
_
M

kD0
b
k
:
k
_
X(:).
The system function for a system that satises a difference equation of the
required form is therefore
H(:) =
Y(:)
X(:)
=

M
kD0
b
k
:
k

N
kD0
a
k
:
k
=
_
b
0
a
0
_
M
kD1
(1 c
k
:
1
)

N
kD1
(1 J
k
:
1
)
.
5
Each factor (1 c
k
:
1
) in the numerator contributes a zero at : = c
k
and a
pole at : = 0. Each factor (1 J
k
:
1
) contributes a zero at : = 0 and a pole
at : = J
k
.
The difference equation and the algebraic expression for the system function
are equivalent, as demonstrated by the next example.
Example: second-order system
Given the system function
H(:) =
(1 :
1
)
2
(1
1
2
:
1
)(1
3
4
:
1
)
.
we can nd the corresponding difference equation by noting that
H(:) =
1 2:
1
:
2
1
1
4
:
1

3
8
:
2
=
Y(:)
X(:)
.
Therefore
(1
1
4
:
1

3
8
:
2
)Y(:) = (1 2:
1
:
2
)X(:).
and the difference equation is
,n|
1
4
,n 1|
3
8
,n 2| = .n| 2.n 1| .n 2|.
2.1 Stability and causality
A difference equation does not uniquely specify the impulse response of a LTI
system. For a given system function (expressed as a ratio of polynomials),
each possible choice of ROC will lead to a different impulse response.
However, they will all correspond to the same difference equation.
If a system is causal, it follows that the impulse response is a right-sided
sequence, and the region of convergence of H(:) must be outside of the
outermost pole.
6
Alternatively, if we require that the system be stable, then we must have
1

nD1
[hn|[ < o.
For [:[ = 1 this is identical to the condition
1

nD1
[hn|:
n
[ < o.
so the condition for stability is equivalent to the condition that the ROC of
H(:) include the unit circle.
Example: determining the ROC
The frequency response of the LTI system with difference equation
,n|
5
2
,n 1| ,n 2| = .n|
is
H(:) =
1
1
5
2
:
1
:
2
=
1
(1
1
2
:
1
)(1 2:
1
)
.
There are three choices for the ROC:
v Causal: ROC outside of outermost pole [:[ > 2 (but then not stable).
v Stable: ROC such that
1
2
< [:[ < 2 (but then not causal).
v If [:[ <
1
2
then the system is neither causal nor stable.
For a causal and stable system the ROC must be outside the outermost pole
and include the unit circle. This is only possible if all the poles are inside the
unit circle.
2.2 Inverse systems
The system H
i
(:) is the inverse system to H(:) if
G(:) = H(:)H
i
(:) = 1.
7
which implies that
H(:) =
1
H
i
(:)
.
The time-domain equivalent is
gn| = hn| + h
i
n| = n|.
The question of which ROC to associate with H
i
(:) is answered by the
convolution theorem for the previous equation to hold the regions of
convergence of H(:) and H
i
(:) must overlap.
Example: inverse system for rst-order system
Let H(:) be
H(:) =
1 0.5:
1
1 0.9:
1
with ROC [:[ > 0.9. Then H
i
(:) is
H
i
(:) =
1 0.9:
1
1 0.5:
1
.
Since there is only one pole, there are only two possible ROCs. The choice of
ROC for H
i
(:) that overlaps with [:[ > 0.9 is [:[ > 0.5. Therefore, the
impulse response of the inverse system is
h
i
n| = (0.5)
n
un| 0.9(0.5)
n1
un 1|.
In this case the inverse is both causal and stable.
A LTI system is stable and causal with a stable and causal inverse if and only if
both the poles and zeros of H(:) are inside the unit circle such systems are
called minimum phase systems.
The frequency response of the inverse system, if it exists, is
H(e
j!
) =
1
H
i
(e
j!
)
.
Not all systems have an inverse. For example, there is no way to recover the
8
frequency components above the cutoff frequency that were set to zero by the
action of the lowpass lter.
2.3 Impulse response for rational system functions
If a system has a rational transfer function, with at least one pole that is not
cancelled by a zero, then there will always be a term corresponding to an
innite length sequence in the impulse response. Such systems are called
innite impulse response (IIR) systems.
On the other hand, if a system has no poles except at : = 0 (that is, N = 0 in
the canonical LCCDE expression), then
H(:) =
M

kD0
b
k
:
k
.
In this case the system is determined to within a constant multiplier by its
zeros, so the impulse response has a nite length:
hn| =
M

kD0
b
k
n k| =
_
_
_
b
n
0 _ n _ M
0 otherwise
In this case the impulse response is nite in length, and the system is called a
nite impulse response (FIR) system.
Example: a rst-order IIR system
Given a causal system satisfying the difference equation
,n| a,n 1| = .n|.
the system function is
H(:) =
1
1 a:
1
=
:
: a
. [:[ > [a[.
9
The condition for stability is [a[ < 1. The inverse z-transform is
hn| = a
n
un|.
Example: a simple FIR system
Consider the truncated impulse response
hn| =
_
_
_
a
n
0 _ n _ M
0 otherwise.
The system function is
H(:) =
M

nD0
a
n
:
n
=
1 a
MC1
:
M1
1 a:
1
.
The zeros of the numerator are at
:
k
= ae
j 2k=.MC1/
. k = 0. 1. . . . . M.
With a assumed real and positive, the pole at : = a is cancelled by a zero. The
pole-zero plot for the case of M = 7 is therefore given by
Re
Im zplane
7th
order
1
The difference equation satised by the input and output of the LTI system is
the convolution
,n| =
M

kD0
a
k
.n k|.
10
The input and output also satisfy the difference equation
,n| a,n 1| = .n| a
MC1
.n M 1|.
3 Frequency response for rational system
functions
If a stable LTI system has a rational system function, then its frequency
response has the form
H(e
j!
) =

M
kD0
b
k
e
j!k

N
kD0
a
k
e
j!k
.
We want to know the magnitude and phase associated with the frequency
response. To this end, it is useful to express H(e
j!
) in terms of the poles and
zeros of H(:):
H(e
j!
) =
_
b
0
a
0
_
M
kD1
(1 c
k
e
j!
)

N
kD1
(1 J
k
e
j!
)
.
It follows that
[H(e
j!
)[ =

b
0
a
0

M
kD1
[1 c
k
e
j!
[

N
kD1
[1 J
k
e
j!
[
.
Therefore [H(e
j!
)[ is the product of the magnitudes of all the zero factors of
H(:) evaluated on the unit circle, divided by the product of the magnitudes of
all the pole factors evaluated on the unit circle.
The gain in dB of H(e
j!
), also called the log magnitude, is given by
Gain in dB = 20 log
10
[H(e
j!
)[.
11
which for a rational system function takes the form
20 log
10
[H(e
j!
)[ = 20 log
10

b
0
a
0

kD1
20 log
10
[1 c
k
e
j!
[

kD1
20 log
10
[1 J
k
e
j!
[.
Also
Attenuation in dB = Gain in dB.
Thus a 60dB attenuation at frequency o corresponds to [H(e
j!
)[ = 0.001.
Also,
20 log
10
[Y(e
j!
)[ = 20 log
10
[H(e
j!
)[ 20 log
10
[X(e
j!
)[.
The phase response for a rational system function is
^H(e
j!
) = ^
_
b
0
a
0
_

kD1
^1 c
k
e
j!
|
N

kD1
^1 J
k
e
j!
|.
The zero factors contribute with a plus sign and the pole factors with a minus.
In the above equation, the phase of each term is ambiguous, since any integer
multiple of 2 can be added at each value of o without changing the value of
the complex number. When calculating the phase with a computer, the angle
returned will generally be the principal value ARGH(e
j!
)|, which lies in the
range to . This phase will generally be a discontinuous function,
containing jumps of 2 radians whenever the phase wraps. Appropriate
multiples of 2 can be added or subtracted, if required, to yield the continuous
phase function argH(e
j!
)|.
12
0
0
ARGH(e
j!
)|
argH(e
j!
)|
o
o

2
3
4
3.1 Frequency response of a single pole or zero
Consider a single zero factor of the form
(1 re
j
e
j!
)
in the frequency response. The magnitude squared of this factor is
[1 re
j
e
j!
[
2
= (1 re
j
e
j!
)(1 re
j
e
j!
)
= 1 r
2
2r cos(o 0).
so the log magnitude in dB is
20 log
10
[1 re
j
e
j!
[ = 10 log
10
1 r
2
2r cos(o 0)|.
The principle value of the phase for the factor is
ARG1 re
j
e
j!
| = arctan
_
r sin(o 0)
1 r cos(o 0)
_
.
These functions are periodic in o with period 2.
13
The following plot shows the frequency response for r = 0.9 and three
different values of 0:
0
20
10
0
10
G
a
i
n

(
d
B
)
0
0
P
h
a
s
e

(
r
a
d
i
a
n
s
)
o
o

3
2
3
2
2
2
0 = 0
0 =

2
0 =
Note that
v The gain dips at o = 0. As 0 changes, the frequency at which the dip
occurs changes.
v The gain is maximised for o 0 = , and for r = 0.9 the magnitude of
the resulting gain is
10 log
10
(1 r
2
2r) = 20 log
10
(1 r) = 5.57dB.
v The gain is minimised for o = 0, and for r = 0.9 the resulting gain is
10 log
10
(1 r
2
2r) = 20 log
10
[1 r[ = 20dB.
v The phase is zero at o = 0.
Note that if the factor (1 re
j
e
j!
) occurs in the denominator, thereby
representing a pole factor, then the entire analysis holds with the exception that
the sign of the log magnitude and the phase changes.
14
The frequency response can be sketched from the pole-zero plot using a simple
geometric construction. Note rstly that the frequency response corresponds to
an evaluation of H(:) on the unit circle. Secondly, the complex value of each
pole and zero can be represented by a vector in the z-plane from the pole or
zero to a point on the unit circle.
Take for example the case of a single zero factor
H(:) = (1 re
j
:
1
) =
: re
j
:
r < 1.
which corresponds to a pole at : = 0 and a zero at : = re
j
.
zplane
Im
Re
v
1
v
2
v
3
0
o

3
If the vectors v
1
, v
2
, and v
3
= v
1
v
2
represent respectively the complex
numbers e
j!
, re
j
, and e
j!
re
j
, then
[H(e
j!
)[ = [1 re
j
e
j!
[ =

e
j!
re
j
e
j!

=
[v
3
[
[v
1
[
= [v
3
[.
The phase is
^H(e
j!
) = ^(1 re
j
e
j!
) = ^(e
j!
re
j
) ^(e
j!
)
= ^(v
3
) ^(v
1
) =
3

1
=
3
o.
A vector such as v
3
from a zero to the unit circle is referred to as a zero
15
vector, and a vector from a pole to the unit circle is called a pole vector.
Consider now the pole zero system depicted below:
zplane
Im
Re
v
1
v
1
v
2
v
3
v
3
o

3
The frequency response for the single zero at different values of r and 0 = is
0
20
10
0
10
G
a
i
n

(
d
B
)
r=0.5
r=0.7
r=0.9
r=1
0
0
P
h
a
s
e

(
r
a
d
i
a
n
s
)
o
o

3
2
3
2
2
2

2
Note that the log magnitude dips more sharply as r approaches 1 (and at
o = tends to oas r tends towards 1). The phase function has positive
slope around o = 0. This slope increases as r approaches 1, and becomes
16
innite for r = 1. In this case the phase function is discontinuous, with a jump
of radians at o = 0.
If r increases still further, to lie outside of the unit circle,
zplane
Im
Re
v
1
v
2
v
3

3
then the frequency response becomes
0
20
10
0
10
G
a
i
n

(
d
B
)
r=1/0.9
r=1.25
r=2.0
0
0
P
h
a
s
e

(
r
a
d
i
a
n
s
)
o
o

3
2
3
2
2
2
17
3.2 Frequency response with multiple poles and zeros
In general, the z-transform of a LTI system can be factorised as
H(:) =
_
b
0
a
0
_
M
kD1
(1 c
k
:
1
)

N
kD1
(1 J
k
:
1
)
=
_
b
0
a
0
_
:
NM

M
kD1
(: c
k
)

N
kD1
(: J
k
)
.
Depending on whether N is greater than or less than M, the factor :
NM
represents either N M zeros at the origin, or M N poles at the origin. In
either case, the z-transform can be written in the form
H(:) = 1

M
0
iD0
(: :
k
)

N
0
iD0
(:
k
)
where :
1
. . . . . :
M
0
are the zeros, and
1
. . . . .
N
0
the poles of H(:). This
representation could also be obtained by merely factorising H(:) in terms of :
rather than :
1
.
The frequency response of this system is
H(e
j!
) = 1

M
0
iD0
(e
j!
:
k
)

N
0
iD0
(e
j!

k
)
.
The magnitude is therefore
[H(e
j!
)[ = [1[

M
0
iD0
[e
j!
:
k
[

N
0
iD0
[e
j!

k
[
.
and the phase is
^H(e
j!
) =
M
0

iD0
^(e
j!
:
k
)
N
0

iD0
^(e
j!

k
).
In the z-plane, (e
j!
:
k
) is simply the vector from the zero :
k
to the point on
the unit circle. The term [e
j!
:
k
[ is the length of this vector, and
^(e
j!
:
k
) is the angle that it makes with the positive real axis. Similarly, the
18
term (e
j!

k
) corresponds to the vector from the pole
k
to the point on the
unit circle.
It follows then that the the magnitude response is the product of the lengths of
the zero vectors, divided by the product of the lengths of the pole vectors. The
phase response is the sum of the angles of the zero vectors, minus the sum of
the angles of the pole vectors. Thus, for the two pole, two zero system
zplane
Im
Re

2
:
1
:
2

2
0
1
0
2
U
1
U
2
V
1
V
2
the frequency response is
[H(e
j!
)[ = [1[
U
1
U
2
V
1
V
2
^H(e
j!
) = 0
1
0
2
(
1

2
).
Here 1 is a constant factor which cannot be determined from the pole zero
diagram alone, but only serves to scale the magnitude.
4 Realisation structures for digital lters
The difference equation, the impulse response, and the system function are all
equivalent characterisations of the input-output relation for a LTI discrete-time
system. For implementation purposes, systems described by LCCDEs can be
19
implemented by structures consisting of an interconnection of the basic
operations of addition, multiplication by a constant, and delay. The desired
form for the interconnections depends on the technology to be used.
Discrete-time lters are often represented in the form of block or signal ow
diagrams, which are convenient for representing the difference equations or
transfer functions.
For example, the system with the difference equation
,n| = .n 1| b
1
,n 1| b
2
,n 2| b
3
,n 3|
can be represented in a block diagram form as
x[n1]
y[n]
y[n1]
y[n3]
y[n2]
x[n]
:
1
:
1
:
1
:
1
b
1
b
2
b
3
The symbol :
1
represents a delay of one unit of time, and the arrows
represent multipliers (with the constant multiplication factors next to them).
The equivalent signal ow diagram is
20
y[n1]
y[n3]
y[n2]
x[n] y[n]
:
1
:
1
:
1
:
1
b
1
b
2
b
3
The relationship between the diagrams and the difference equation is clear.
Many alternative lter structures can be developed, and they differ mainly with
respect to their numerical stability and the effects of quantisation on their
performance. A discussion of these effects can be found in many DSP texts.
21