Note

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 37

Transform Analysis of LTI Systems

Dr. Sujit Kumar Sahoo


Assistant Professor,SES
IIT Goa

School of Electrical Sciences


IIT Goa

August 18, 2023

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 1 / 37
Transform analysis of LTI system

LTI System:

X
y [n] = x[k]h[n − k]
n=−∞

The z-transform of the output of an LTI system is related to the z-transform of


the input and the z-transform of the system impulse response by

Y (z) = H(z)X (z)

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 2 / 37
Systems Characterized by Linear Constant-Coefficient
Difference Eqations

Discrete-time filters are most typically realized through the implementation of a


linear constant-coefficient difference equation of the form
N
X M
X
ak y [n − k] = bk x[n − k]
k=0 k=0

Applying z-transforms on both sides


PM
Y (z) bk z −k
H(z) = = Pk=0
N
X (z) k=0 ak z
−k

It can be rearranged in the form of poles & zeros


 QM
− ck z −1 )

b0 k=1 (1
H(z) = QN
a0 k=1 (1 − dk z −1 )

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 3 / 37
Examples

Example 1: 2nd order system

(1 + z −1 )2
H(z) =
(1 − 12 z −1 )(1 + 43 z −1 )

Find the difference equation that is satisfied by the input-output of this system.

Sol:
Y (z) (1 + z −1 )2 1 + 2z −1 + z −2
H(z) = = 1 3 =
X (z) −1
(1 − 2 z )(1 + 4 z ) −1 1 + 14 z −1 − 38 z −2
 
1 3
Y (z) 1 + z −1 − z −2 = X (z) 1 + 2z −1 + z −2

4 8
1 3
y [n] + y [n − 1] − y [n − 2] = x[n] + 2x[n − 1] + x[n − 2]
4 8

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 4 / 37
Examples
Example 2: Determine all the possible ROC
5
y [n] − y [n − 1] + y [n − 2] = x[n]
2
Sol: Apply z-Transforms on both sides, then
Y (z) 1 1 − 31 4
3
H(z) = = = = +
1 − 52 z −1 + z −2 1 − 12 z −1 (1 − 2z −1 ) 1 − 12 z −1 1 − 2z −1

X (z)
1) Causal and Unstable: |z| > 2 |z| > 2&|z| > 12

 n
1 1 4
h[n] = − u[n] + 2n u[n]
3 2 3
2) Anti-causal and Unstable:|z| < 2 |z| < 2&|z| < 12
1

 n
1 1 4
h[n] = u[−n − 1] − 2n u[−n − 1]
3 2 3
3) Non-causal and stable: |z| < 2&|z| > 21
 n
1 1 4
h[n] = − u[n] − 2n u[−n − 1]
3 2 3
Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 5 / 37
Inverse systems

Inverse systems:
G (z) = H(z)Hi (z) = 1 (In time domain h[n] ∗ hi [n] = δ[n])
1
Hi (z) =
H(z)

If the system
 QM
− ck z −1 )

b0 k=1 (1
H(z) = QN ,
a0 k=1 (1 − dk z −1 )
then  QN
− dk z −1 )

a0 k=1 (1
Hi (z) = QM
b0 k=1 (1 − ck z −1 )
Note: In order to have an inverse system, the inverse system must have an
overlapping ROC.

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 6 / 37
Inverse systems: Example

Example: Inverse system for a 1st order system


Let
1 − 0.5z −1
H(z) = ,
1 − 0.9z −1
with ROC |z| > 0.9. Then find Hi (z) and impulse response of the inverse
system.
Sol:
We know that Hi (z) = 1/H(z)

1 − 0.9z −1 1 0.9z −1
Hi (z) = −1
= −1

1 − 0.5z 1 − 0.5z 1 − 0.5z −1
Case1: ROC: |z| > 0.5

hi [n] = (0.5)n u[n] − 0.9(0.5)n−1 u[n − 1]

Case2: ROC: |z| < 0.5 ; hi [n] is not possible

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 7 / 37
Inverse systems: Example

Example: Inverse for system with a Zero in the ROC


Suppose that H(z) is

z −1 − 0.5
H(z) = , |z| > 0.9
1 − 0.9z −1
What is the impulse response of the inverse system?
Sol:

1 − 0.9z −1 1.8z −1 2
Hi (z) = −1
= −1

z − 0.5 1 − 2z 1 − 2z −1
Case1: ROC: |z| < 2

hi [n] = −(1.8)2n−1 u[−n] + 2n+1 u[−n − 1]

Case2: ROC: |z| > 2 ; hi [n] is unstable.

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 8 / 37
Frequency response of LTI Systems

Y (e jω ) = H(e jω )X (e jω )

Amplitude
|Y (e jω )| = |H(e jω )||X (e jω )|
Phase
∠Y (e jω ) = ∠H(e jω ) + ∠X (e jω )

Note: H(e jω ) can be expressed in polar form as



H(e jω ) = |H(e jω )|e j∠H(e )

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 9 / 37
Frequency response of LTI Systems

Wrapped phase: −π < ARG [H(e jω )] ≤ π

Unwrapping phase: arg [H(e jω )] = ∠H(e jω ) = ARG [H(e jω )] + 2πr (ω)

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 10 / 37
Frequency response of LTI Systems

Group delay
d
τ (ω) = grd[H(e jω )] = − arg [H(e jω )]

Example: Ideal delay system

hid [n] = δ[n − nd ]

Sol:

H(e jω ) = e −jωnd

|H(e jω )| = 1; ∠H(e jω ) = −ωnd


grd[H(e jω )] = nd

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 11 / 37
Frequency response of LTI Systems
Example:

Figure: (a) Magnitude of Frequency Response.

Figure: (b) Principle value of Phase Response

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 12 / 37
Frequency response of LTI Systems

Figure: (a) Unwrapped Phase Response.

Figure: (b) Group delay of H(z)

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 13 / 37
Frequency response of LTI Systems
x1 [n] = w [n]cos(0.2πn)
x2 [n] = w [n]cos(0.4πn − π/2)
x3 [n] = w [n]cos(0.8πn + π/5)
(
0.56 − 0.46 cos( 2π
M n), for 0 ≤ n ≤ M , for M = 60.
where w [n] =
0, otherwise
Superposition of the shifted signals
x[n] = x3 [n] + x1 [n − M − 1] + x2 [n − 2M − 2]

Figure: (b) Magnitude of DTFT of x[n]

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 14 / 37
Frequency response of LTI Systems

Figure: Input x[n]

Figure: Output y [n]

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 15 / 37
Frequency Response for the Rational System Functions

PM
bk e −jωk
H(e jω ) = PN
k=0

k=o ak e −jωk
  QM −jω
b0 k=1 (1 − ck e )
H(e jω ) = QN
a0 k=1 (1 − d k e −jω )

Amplitude:
QM
jω b0 |1 − ck e −jω |
|H(e )| = | | QNk=1
a0 k=1 |1 − dk e
−jω |

Gain in dB = 20log10 |H(e jω )|


M N
b0 X X
= 20log10 | |+ 20log10 |1 − ck e −jω | − 20log10 |1 − dk e −jω |
a0
k=1 k=1

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 16 / 37
Frequency Response for the Rational System Functions

Phase:
  XM N
b0 X
arg [H(e jω )] = arg + arg [1 − ck e −jω ] − arg [1 − dk e −jω ]
a0
k=1 k=1

d
grad[H(e jω )] = arg [H(e jω )]

dw
M N
X d X d
= − (arg [1 − ck e −jω ]) + (arg [1 − dk e −jω ])
dw dw
k=1 k=1

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 17 / 37
Example: Frequency Response of 1st Order System
A single factor of the form (1 − re jθ e −jω )

Sol:
Gain in dB

(+/−)20log10 |1 − re jθ e −jω | = (+/−)10log10 1 + r 2 − 2rcos(ω − θ)


 

Wrapped phase in radian


 
rsin(ω − θ)
(+/−)ARG 1 − re jθ e −jθ = (+/−)arctan
 
1 − rcos(ω − θ)

Group delay in number of samples

r 2 − rcos(ω − θ)
(+/−)grd 1 − re jθ e −jθ =
 
(+/−)
1 + r 2 − 2rcos(ω − θ)
r 2 − rcos(ω − θ)
= (+/−)
|1 − re jθ e −jω |2

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 18 / 37
Example: Frequency Response of 1st Order System
Gain in dB

Wrapped phase in radian

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 19 / 37
Example: Frequency Response of 1st Order System
Gain in dB

Wrapped phase in radian

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 20 / 37
Example 2

2nd-Order IIR System


1
H(z) =
(1 − re jθ z −1 )(1 − re −jθ z −1 )

Sol:
1 A B
H(z) = = +
(1 − re jθ z −1 )(1 − re −jθ z −1 ) (1 − re jθ z −1 ) (1 − re −jθ z −1 )

A = (1 − re jθ z −1 )H(z) and B = (1 − re −jθ z −1 )H(z)


z=re jθ z=re −jθ
   
1 1
1−e −j2θ 1−e j2θ
H(z) = +
1 − re jθ z −1 1 − re −jθ z −1
 
Gain in dB = −10log10 (1 + r 2 − 2rcos(θ − ω))(1 + r 2 − 2rcos(θ + ω))

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 21 / 37
Example: Frequency Response of 1st Order System
Gain in dB; r = 0.9, θ = π/4

Wrapped phase in radian

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 22 / 37
Example: 2nd- Order FIR System

2nd-Order FIR System

h[n] = δ[n] − 2rcosθδ[n − 1] + r 2 δ[n − 2]


Sol:
y [n] = x[n] − 2rcosθx[n − 1] + r 2 x[n − 2]
Y (e jω )
= H(e jω ) = 1 − 2rcosθe −jω + r 2 e −j2ω
 

X (e )

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 23 / 37
All-Pass Systems

Definition
It is a system whose magnitude response is constant for all frequencies.To satisfy
the above definition, each Hap (z) pole should be paired with a conjugate
reciprocal zero.

The causal and stable LTI system

z −1 − a∗
Hap (z) =
1 − az −1

let, a = re jθ

e −jω − re −jθ 1 − re −jθ e jω


 
Hap (e ) = jω
= e −jω
1 − re jθ e −jω 1 − re jθ e −jω
−jθ jω
Magnitude response, |Hap (e jω )| = | 1−re e
1−re jθ e −jω
|=1
 
rsin(ω−θ)
Phase response, ∠Hap (e jω ) = −ω − 2tan−1 1−rcos(ω−θ)

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 24 / 37
Example: 1st Order All-Pass System

z −1 −0.9 z −1 +0.9
Figure: H1 (z) = 1−0.9z −1
solid line and H2 (z) = 1+0.9z −1
dashed line
Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 25 / 37
Minimum-Phase Systems
Definition
1
It is a system where both the system H(z) and its inverse H(z) are stable. To have
a Minimum-Phase system, all poles and zeros of H(z) must be inside the unit
circle (for causal system).

Minimum phase and All - pass decomposition


H(z) = Hmin (z)Hap (z)
Let’s say H(z) has one zero outside the unit circle
H(z) = H1 (z)(z −1 − c ∗ )
So
(1 − cz −1 )
H(z) = H1 (z)(z −1 − c ∗ )
(1 − cz −1 )
(z −1 − c ∗ )
= H1 (z)(1 − cz −1 ) = H̃1 (z)Hap (z)
(1 − cz −1 )
It continues till H(z) = Hmin (z)Hap (z)
Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 26 / 37
Example

Example:
1 + 3z −1
H(z) =
1 + 12 z −1
Sol:
1 + 3z −1
H(z) =
1 + 12 z −1
1 3 + z −1
= 1 −1 (1 + 3z −1 )
1+ 2z
3 + z −1
!
3 + z −1 1 + 3z −1

= = Hmin (z)Hap (z)
1 + 12 z −1 3 + z −1

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 27 / 37
LTI Systems with Generalized Linear phase
Definition
A system has linear phase if its phase response ∠H(e jω ) = −ωnd for all ω and any
constant nd .

In general, a linear phase system has frequency response


H(e jω ) = |H(e jω )|e −jωnd
and delays all the frequencies by the same amount of time.
(
jω e −jωnd |ω| < ωc
Let, H(e ) =
0 otherwise
then
Z π Z ωc ωc
1 e jω(n−nd )

1 jω jωn 1 jω(n−nd )
h[n] = H(e )e dω = e dω =
2π −π 2π −ωc 2π j(n − nd ) −ωc
sin(ωc (n − nd ))
= ;
π[n − nd ]
Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 28 / 37
LTI Systems with Generalized Linear phase
The causal and stable LTI system

Hlin (z) = (z −1 − a∗ )(1 − az −1 )

let, a = re jθ

Hlin (e jω ) = (e −jω − re −jθ )(1 − re jθ e −jω )


= e −jω (1 − re −jθ e jω )(1 − re jθ e −jω )
= e −jω (1 − re −j(θ−ω) )(1 − re j(θ−ω) ) = e −jω [2 − 2r cos(θ − ω)]

Magnitude response, |Hlin (e jω )| = 2 |1 − r cos(θ − ω)|


Phase response, ∠Hlin (e jω ) = −ω
The Non-causal and stable LTI system
1
Hlin (z) =
(z −1 − a∗ )(1 − az −1 )
1
Magnitude response, |Hlin (e jω )| = 2|1−r cos(θ−ω)|

Phase response, ∠Hlin (e jω ) = ω


Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 29 / 37
FIR Linear Phase Systems
Type I Causal FIR Generalized Linear-Phase Systems:
M is even (Odd number of coefficients)
Symmetric; h[n] = h[M − n]; 0 ≤ n ≤ M
M
M −1
2 M
M
X X X
−jωn

H(e ) = h[n]e = h[n]e −jωn + h [M/2] e −jω 2 + h[n]e −jωn
n=0 n=0 n= M2 +1
M
−1
2
−jω M2 −jω M2 M
X
= e h [M/2] + e h[n]e −jω(n− 2 )
n=0
M
M M
X
+e −jω 2 h[n]e −jω(n− 2 )
n= M2 +1
| {z }
putting n=M−m
 M M

−1
2 −1
2
−jω(n− M2 −jω( M2
X X
= e −jωM/2 h [M/2] + h[n]e )
+ h[M − m]e −m) 

n=0 m=0

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 30 / 37
FIR Linear Phase Systems
 M

−1
2 h i
−jω M2 M M
X
H(e jω ) = e h [M/2] + h[m] e −jω(m− 2 ) + e jω(m− 2 ) 
m=0
 M

−1
2
−jω M2
X
= e h [M/2] + h[m]2 cos ω (m − M/2)
m=0
 M

−1
2
−jω M2
X
= e h [M/2] + 2h[m] cos ω (M/2 − m)
m=0

substituting k = M/2 − m
 M

2
M
X
= e −jω 2 h [M/2] + 2h [M/2 − k] cos ωk 
k=1
M
Frequency response, jω −jω M2
2
X
H(e ) = e a[k] cos ωk,
k=0

where a[0] = h[M/2], and a[k] = 2h[M/2 − k].


Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 31 / 37
FIR Linear Phase Systems
Type II Causal FIR Generalized Linear-Phase Systems:
M is odd (even number of coefficients)
Symmetric; h[n] = h[M − n]; 0 ≤ n ≤ M
M−1
M
X 2
X M
X
−jωn
H(e ) jω
= h[n]e = h[n]e −jωn + h[n]e −jωn
n=0 n=0 n= M+1
2
M−1
2 M
−jω M2 M M M
X X
= e h[n]e −jω(n− 2 ) + e −jω 2 h[n]e −jω(n− 2 )
n=0 n= M+1
2
| {z }
putting n=M−m
 M−1 M−1

2 2
M M M
X X
= e −jω 2  h[n]e −jω(n− 2 ) + h[M − m]e −jω( 2 −m) 
n=0 m=0
M 
2 −1 h i
M M M
X
= e −jω 2  h[m] e −jω(m− 2 ) + e jω(m− 2 ) 
m=0

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 32 / 37
FIR Linear Phase Systems
 M−1 
2  
−jω M2
X M
H(e jω ) = e  h[m]2 cos ω m − 
m=0
2
 M−1 
2  
M
X M
= e −jω 2  2h[m] cos ω −m 
m=0
2

substituting k = (M + 1)/2 − m
 M 
2    
M
X M + 1 1
= e −jω 2  2h − k cos ω k − 
2 2
k=1

Frequency response,
 
(M+1)/2   
M
X 1 
H(e jω ) = e −jω 2  b[k] cos ω k −
2
k=1

where, b[k] = 2h[(M + 1)/2 − k]


Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 33 / 37
FIR Linear Phase Systems

Type III Causal FIR Generalized Linear-Phase Systems:


M is even (odd number of coefficients)
Anti-Symmetric; h[n] = −h[M − n]; 0≤n≤M

Frequency response,
 M 
2
−jω M2
X
H(e jω ) = je  c[k] sin ωk 
k=1

where, c[k] = 2h[( M2 ) − k]

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 34 / 37
FIR Linear Phase Systems

Type IV Causal FIR Generalized Linear-Phase Systems:


M is odd (even number of coefficients)
Anti-Symmetric; h[n] = −h[M − n]; 0≤n≤M

Frequency response,
 
(M+1)/2   
M
X 1 
H(e jω ) = je −jω 2  d[k] sin ω k −
2
k=1

where, d[k] = 2h[(M + 1)/2 − k]

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 35 / 37
FIR Linear Phase Systems: Summary
Type Impulse response symmetry Length
I Symmetry Even M (Odd length)
II Symmetry Odd M (Even length)
III Anti-symmetry Even M (Odd length)
IV Anti-symmetry Odd M (Even length)

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 36 / 37
The End

Dr. Sujit Kumar Sahoo (IIT GOA) Transform Analysis of LTI Systems August 18, 2023 37 / 37

You might also like