Signal processing

Chapter 2 Discrete-Time Signals and

Systems

1
Chapter 2 Discrete-Time Signals and Systems

2.0 Introduction
2.1 Discrete-Time Signals: Sequences
2.2 Discrete-Time Systems
2.3 Linear Time-Invariant (LTI) Systems
2.4 Properties of LTI Systems
2.5 Linear Constant-Coefficient
Difference Equations

2 08/13/20
Chapter 2 Discrete-Time Signals and Systems

2.6 Frequency-Domain Representation

of Discrete-Time Signals and systems

3 08/13/20
 2.0 Introduction
Signal: something conveys information
Signals are represented mathematically as
functions of one or more independent variables.
Continuous-time (analog) signals, discrete-
time signals, digital signals
Signal-processing systems are classified along the
same lines as signals: Continuous-time (analog)
systems, discrete-time systems, digital systems
Discrete-time signal
Sampling a continuous-time signal
Generated directly by some discrete-time process

4 08/13/20
 2.1 Discrete-Time Signals: Sequences

Discrete-Time signals are represented as

x   x n ,    n  , n : integer
Cumbersome, so just use x  n 
In sampling,
x n  xa  nT  , T : sampling period

1/T (reciprocal of T) : sampling frequency

5 08/13/20
 Figure 2.1 Graphical representation
of a discrete-time signal

Abscissa: continuous line

x  n  : is defined only at discrete instants
6 08/13/20
 x [ n ]  x a ( t ) |t  n T  x a ( n T )
EXAMPLE Sampling the analog waveform

7 Figure 2.2
 Basic Sequence Operations
Sum of two sequences
x[n]  y[n]
Product of two sequences
x[n]  y[n]
Multiplication of a sequence by a numberα
  x[n]
Delay (shift) of a sequence
y[n]  x[n  n0 ] n0 : integer

8 08/13/20
 Basic sequences

Unit sample sequence 0 n  0

(discrete-time impulse,   n  
1 n  0
impulse)

9 08/13/20
 Basic sequences

A sum of scaled, delayed impulses

p n  a3  n  3  a1  n  1  a2  n  2  a7  n  7

arbitrary
sequence
x[n]   x[k ] [n  k ]
k  

10 08/13/20
 Basic sequences
1 n  0
Unit step sequence u[n]  
0 n  0

 k 
 n

0, when n  0
    k   1, when n  0 ,




u[n]   k  
k    since   k   0 k  0 
 1 k 0 

u[n]   [n]   [n  1]   [n  2]      [n  k ]
k 0
 [n]  u[n]  u[n  1] First backward difference
11 08/13/20
 Basic Sequences
Exponential sequences x[n]  A n

A and α are real: x[n] is real

A is positive and 0<α<1, x[n] is positive and
decrease with increasing n
-1<α<0, x[n] alternate in sign, but decrease
in magnitude with increasing n
   1: x[n] grows in magnitude as n increases

12 08/13/20
 EX. 2.1 Combining Basic sequences
If we want an exponential sequences that is zero for n <0, then

 A n
n0
x[n]   Cumbersome
0 n0

x[n]  A u[n] n
simpler

13 08/13/20
 Basic sequences
Sinusoidal sequence

x[n]  A cos w0 n    for all n

14 08/13/20
 Exponential Sequences
j
A  Ae  e jw0

j n n j  w0 n  
x[n]  A  A e  e
n jw0 n
 A e
 A  cos w0 n     j A  sin  w0 n   
n n

Exponentially weighted sinusoids

 1 Exponentially growing envelope

 1 Exponentially decreasing envelope

 1 x[n]  Ae jw0 n
is refered to
Complex Exponential Sequences
15 08/13/20
 Frequency difference between
continuous-time and discrete-time
complex exponentials or sinusoids

j  w0  2  n j 2n
x[n]  Ae  Ae jw0 n
e  Ae jw0 n

x[n]  A cos  w0  2 r  n     A cos  w0n   

 w0 : frequency of the complex sinusoid

or complex exponential
  : phase

16 08/13/20
 Periodic Sequences

A periodic sequence with integer period N

x[n]  x[n  N ] for all n

A cos w0 n     A cos w0 n  w0 N   

w0 N  2 k , where k is integer

N  2 k / w0 , where k is integer

17 08/13/20
 EX. 2.2 Examples of Periodic Sequences
x1[n]  cos  n / 4
Suppose it is periodic sequence with period N
x1[n]  x1[n  N ]
cos  n / 4  cos  n  N  / 4
 n / 4  2 k   n / 4  N / 4, k : integer
N  2 k / ( / 4)  8 k
k  1,  N  8  2 / w0
18 08/13/20
 EX. 2.2 Examples of Periodic Sequences
2
8

3
8

x1[n]  cos 3 n / 8 
Suppose it is periodic sequence with period N
x1[n]  x1[n  N ]
cos 3 n / 8  cos 3  n  N  / 8
3 n / 8  2 k  3 n / 8  3 N / 8, k : integer
N  2 k / w0  2 k / (3 / 8) k  3,  N  16
N  2 3 / w0  2 / w0 ( for continuous signal)
19 08/13/20
 EX. 2.2 Non-Periodic Sequences
x2 [n]  cos n 
Suppose it is periodic sequence with period N

x2 [ n ]  x 2 [ n  N ]

cos n   cos(n  N )
for n  2 k  n  N , k : integer,
there is no integer N

20 08/13/20
High and Low Frequencies in Discrete-time signal
x[n]  A cos( w0 n)
(a) w0 = 0 or 2

(b) w0 = /8 or 15/8

(c) w0 = /4 or 7/4

(d) w0 = 
21 08/13/20
 2.2 Discrete-Time System

Discrete-Time System is a trasformation

or operator that maps input sequence
x[n] into a unique y[n]
y[n]=T{x[n]}, x[n], y[n]: discrete-time
signal

x[n] y[n]
T{‧}

Discrete-Time System
22 08/13/20
 EX. 2.3 The Ideal Delay System

y[n]  x[n  nd ],    n  
If nd is a positive integer: the delay of the
system. Shift the input sequence to the
right by nd samples to form the output .

If nd is a negative integer: the system will

shift the input to the left by nd samples,
corresponding to a time advance.
23 08/13/20
 EX. 2.4 Moving Average
M2
1
y  n   x n  k
M 1  M 2  1 k  M1
1

M1  M 2  1
 x  n  M 1   x  n  M 1  1  ...  x  n   x  n  1  ...  x  n  M 2  

for n=7, M1=0, M2=5 dummy index

m x[m
]

n-5
n m

24 08/13/20
 Effect of a moving average filter. (Sample values are
connected by straight lines to enable easier viewing of stock
exchange trends)
25
 Properties of Discrete-time systems
2.2.1 Memoryless (memory) system

Memoryless systems:
the output y[n] at every value of n depends
only on the input x[n] at the same value of n

y n   x[n]
2

26 08/13/20
 Properties of Discrete-time systems
2.2.2 Linear Systems
 If x1  n T{‧} y1 n 
x2  n  T{‧} y2  n
 and only If:

x1  n  x2  n T{‧} y1  n  y2  n additivity property

ax n T{‧} ay n homogeneity or scaling

property

 principle of superposition

x3  n  ax1  n  bx2  n T{‧} y3  n  ay1  n  by2  n

27 08/13/20
 Example of Linear System
n
Ex. 2.6 Accumulator system y n    x k 
for arbitrary x1  n and x2  n k  
n n
y1  n   x k 1 y2  n   x k 2
k   k  

when x3  n  ax1  n  bx2  n

n n
y3  n    x  k     ax  k   bx  k  
3 1 2
k   k  
n n
 a  x1  k   b  x2  k   ay1  n  by2  n
k   k  
28 08/13/20
 Example 2.7 Nonlinear Systems
Method: find one counterexample
y n   x[n]
2
 For

1  1  1  1
2 2 2
 counterexample

 For y n  log10  x[n] 

 counterexample

10  log10  1   log10  10 1 
29 08/13/20
 Properties of Discrete-time systems
2.2.3 Time-Invariant Systems
Shift-Invariant Systems

x1  n T{‧} y1  n

x2  n  x1  n  n0  T{‧} y2  n  y1  n  n0 

30 08/13/20
 Example of Time-Invariant System
Ex. 2.8 Accumulator system
n
y n    x k 
k  

x1[n]  x n  n0 

n n n  n0
y1  n   x  k    x  k  n    x k   y  n  n 
1 0 1 0
k   k   k1  

31 08/13/20
 Properties of Discrete-time systems

2.2.4 Causality

A system is causal if, for every choice

of n0 , the output sequence value at
the index n  n0 depends only on the
input sequence value for n  n0

32 08/13/20
 Ex. 2.10 Example for Causal System

Forward difference system is not Causal

y n  x n  1  x n

Backward difference system is Causal

y n  x n  x n  1

33 08/13/20
 Properties of Discrete-time systems
2.2.5 Stability
Bounded-Input Bounded-Output (BIBO)
Stability: every bounded input sequence
produces a bounded output sequence.

if x n  Bx  , for all n

then y  n   B y  , for all n

34 08/13/20
 Ex. 2.11 Test for Stability or Instability

y n   x[n]
2
is stable

if x n  Bx  , for all n

then y  n   B y  B  ,
2
x for all n

35 08/13/20
 Ex. 2.11 Test for Stability or Instability
n
Accumulator system y n    x k 
k  

0 n  0
x n   u  n    : bounded
1 n  0

n n
0 n0
y n   x k    x k    : not bounded
k   k   n  1 n  0

Accumulator system is not stable

36 08/13/20
 Properties of Discrete-time systems (Repeat)
2.2.2 Linear Systems
 If  
x1 n T{‧} y1 n  
x2  n  T{‧} y2  n
 and only If:

x1  n  x2  n T{‧} y1  n  y2  n additivity property

ax n T{‧} ay n homogeneity or scaling

property

 principle of superposition

x3  n  ax1  n  bx2  n T{‧} y3  n  ay1  n  by2  n

37 08/13/20
 2.3 Linear Time-Invariant (LTI)
Systems
Impulse response

  n T{‧} h  n

  n  n0  h  n  n0 
T{‧}

38 08/13/20
 LTI Systems: Convolution
Representation of general sequence as a linear
combination of delayed impulse

x n   x k   n  k 
k  
principle of superposition
  
y n  T   x k   n  k     x k T   n  k 
k    k  

  x k  h n  k   x n  h n
k  

An Illustration Example （ interpretation 1 ）

39 08/13/20
40 08/13/20
 Computation of the Convolution
（ interpretation 2 ）

y n    x k h n  k 
k  

h n  k   h   k  n  
h k  h  k 
reflecting h[k] about the origion to obtain h[-k]
Shifting the origin of the reflected sequence to
k=n

41 08/13/20
 Ex. 2.12

42 08/13/20
Convolution can be realized by
–Reflecting h[k] about the origin to obtain h[-k].
–Shifting the origin of the reflected sequences to k=n.
–Computing the weighted moving average of x[k] by
using the weights given by h[n-k].

43
 Ex. 2.13 Analytical Evaluation
of the Convolution
For system with impulse response
1 0  n  N  1
h n  u n  u n  N   
0 otherwise

input x n   a u  n 
n h(k)

Find the output at index n

44 08/13/20
 1 0  n  N  1
h n   x n  a u n
n

0 otherwise

h(-k) h(k)

0 0

h(n-k) x(k)

0
y  n  0 n  0
45
 h(-k) h(k)

0 0

n  0, n   N  1  0  0  n  N  1
n 1
n n
1  a
y n   x k  h   k  n     a k 
k 0 k 0 1 a
46 08/13/20
 h(-k) h(k)

0 0

n   N  1  0  n  N  1
n n
y n    x k  h   k  n     a k

k  n  N 1 k  n  N 1

a n  N 1
a n 1

n  N 1 1  a
N

 a  
1 a  1 a 
47 08/13/20
 
 0, n0
 n 1
1  a
y n    , 0  n  N 1
 1 a

 a n  N 1  1  a
N

 1 a , N 1  n

  

48 08/13/20
 2.4 Properties of LTI Systems
Convolution is commutative

x n  h n  h n  x n

x[n] h[n] y[n]

h[n] x[n] y[n]

Convolution is distributed over addition

x n   h1  n  h2  n   x n  h1  n  x n  h2  n
49 08/13/20
 Cascade connection of systems
h n  h1  n  h2  n

x [n] h1[n] h2[n] y [n]

x [n] h2[n] h1[n] y [n]

x [n] h1[n] ]h2[n] y [n]

50 08/13/20
 Parallel connection of systems
h n  h1  n  h2  n

51 08/13/20
 Stability of LTI Systems
LTI system is stable if the impulse response
is absolutely summable . 
S  h k   
k  
 
y n    h k  x n  k    h k  x n  k 
k   k  

x n   B x y  n   Bx  h k  
k 

Causality of LTI systems h n  0, n  0

HW: proof, Problem 2.62
52
 Impulse response of LTI systems

 Impulse response of Ideal Delay systems

h  n     n  nd  , nd a positive fixed integer

Impulse response of Accumulator

n
1, n  0
h n     k     u n
k   0, n  0
53 08/13/20
 Impulse response of Moving
Average systems

M2
1
h n    n  k
M 1  M 2  1 k  M1
 1
 ,  M1  n  M 2
  M1  M 2  1
 0 , otherwise

54 08/13/20
 Impulse response of Forward Difference

h n    n  1    n

Impulse response of Backward Difference

h n    n    n  1

55
 Finite-duration impulse
response (FIR) systems
The impulse response of the system has
only a finite number of nonzero samples.
M2
1
such as: h n    n  k
M 1  M 2  1 k  M1
 1
 ,  M1  n  M 2
  M1  M 2  1
 0 , otherwise
The FIR systems always are stable.

S  h  n
n 

56
 Infinite-duration impulse
response (IIR)
The impulse response of the system is
infinite in duration.
n
1, n  0
h n     k     u n
k   0, n  0

Stable IIR System: h n  a u n

n


a 1 S  h  n
n 

57
 Equivalent systems

h  n      n  1    n      n  1
   n  1     n  1    n      n     n  1
58
 Inverse system

h n  hi  n  hi  n  h n    n

h n  u n     n    n  1 
 u n  u n  1    n
59
 2.5 Linear Constant-Coefficient
Difference Equations
An important subclass of linear time-
invariant systems consist of those
system for which the input x[n] and
output y[n] satisfy an Nth-order linear
constant-coefficient difference equation.
N M

 a y n  k    b x n  m
k 0
k
m 0
m

60
 Ex. 2.14 Difference Equation
Representation of the Accumulator
n n 1
y  n   x k , y  n  1   x k
k  k 

n 1
y n  x n   x k   x n  y n  1
k  

y n  y n  1  x n
61
 Block diagram of a recursive
difference equation representing an
accumulator

y  n   y  n  1  x  n 
62
 Difference Equation
Representation of the System

An unlimited number of distinct

difference equations can be used to
represent a given linear time-invariant
input-output relation.

63
 Solving the difference equation

Without additional constraints or

information, a linear constant-
coefficient difference equation for
discrete-time systems does not provide
a unique specification of the output for
a given input.
N M

 a y n  k    b x n  m
k 0
k
m 0
m

64
 Solving the difference equation
N M
Output:
 a y n  k    b x n  m
y n   y p  n   y h  n 
k m
k 0 m 0

 y p  n Particular solution: one output sequence

for the given input x p  n
 yh  n Homogenous solution: solution for
the homogenous equation( x  n   0 ):
N
yh  n   A z
N

 ak y h  n  k   0
n
m m
m 1
k 0
N
 where zm is the roots of  k 0
a z
k 0
k

65
 Example 2.16 Recursive Computation of
Difference Equation
y  n   ay  n  1  x  n  , x  n  K  n , y  1  c

y 0  ac  K
y1  ay 0  0  a ac  K   a c  aK 2

y 2  ay1  0  a  a c  aK   a c  a K
2 3 2


y 3  ay 2  0  a a c  a K  a c  a K
3 2
 4 3

y  n  a c  a K
n 1 n
for n  0
66
 Example 2.16 Recursive Computation of
Difference Equation
y n  ay n  1  x n x n  K  n y  1  c
for n  1 y n  1  a 1  y n  x n 

y   2  a  y  1  x  1   a c
1 1

y  3  a  y  2  x  2   a a c  a c
1 1 1 2

y  4  a  y  3  x  3   a a c  a c
1 1 2 3

y  n  a c n 1
for n  1
y  n   a c  Ka u  n 
n 1 n
for all n
67
 Periodic Frequency Response
The frequency response of discrete-time
LTI systems is always a periodic function of
the frequency variable w with period 2

    h  n e

j  w  2   j  w  2  n
H e
n 
j  w  2   jwn  j 2 n  jwn
e e e e
H e  j  w  2 
  He  jw

68
H e  j  w  2 r
08/13/20

  He ,
jw
for r an integer
Signal Processing - Dr. Arif Wahla
 Periodic Frequency Response
We need only specify He  
over
jw
0  w  2
or   w  
The “low frequencies” are frequencies
close to zero
The “high frequencies” are frequencies
close to  

More generally, modify the frequency with

, r is integer.
2 r
69 08/13/20 Signal Processing - Dr. Arif Wahla
 Example 2.19 Ideal
Frequency-Selective Filters

Frequency Response of Ideal Low-pass Filter

70 08/13/20 Signal Processing - Dr. Arif Wahla

 Frequency Response of Ideal
High-pass Filter

71
 Frequency Response of Ideal
Band-stop Filter

72
 Frequency Response of Ideal
Band-pass Filter

73
Example 2.20 Frequency Response of
the Moving-Average System
 1
 ,  M1  n  M 2
h n   M 1  M 2  1
 0, otherwise

M2
1  jwn
He jw
   e
M 1  M 2  1 n M
1
jwM1  jw M 2 1

1 e e
M1  M 2  1  jw
1 e
74
 Impulse response and
Frequency response
The frequency response of a LTI system
is the Fourier transform of the impulse
response.

He jw
   h  n e  jwn

n 

h n 
1
2



He  e
jw jwn
dw
75
 Chapter 2 HW
2.1, 2.4, 2.5, 2.7, 2.11,
2.12,2.15, 2.20

76