Digital Signal

Processing
Naqib Sad Pathan
Assistant Professor
Department of Electrical and Electronic Engineering
Chittagong University of Engineering and Technology
 Reference Books

Authors: Alan V.
Oppenheim and others
Chapter 2 : Discrete Time Signals and
Systems

Three most important

topics in this chapter:

1. Causality of a system
2.Convolution operation
3. Convolutional
properties of unit sample
function(δ[n]).
 What is Signal?
Signal is a physical quantities that varies with time or other independent
variables. 
A signal is a function that conveys information about a
phenomenon.
Example: Examples of signals include as temperature over time or
space, sound (speech, music, etc) over time, images over space, etc.
A signal carries information and contains energy.
Analog vs Digital Signal

Analog Signal Digital Signal

Continuous Time (CT) vs Discrete Time
(DT) Signal

CT Signal DT Signal
Example of Signals
CT- Analog CT- Digital
Signal Signal

DT- Analog DT-Digital

Signal Signal
A/D convertion
Sampling
Quantization
Encoding
 Sampling Rate and
Nyquist Sampling
Theorem
The Nyquist Sampling Theorem states that:
A bandlimited continuous-time signal can be sampled and perfectly
reconstructed from its samples if the waveform is sampled over twice as
fast as it's highest frequency component.
Nyquist limit:
the highest frequency component that can be accurately represented:
 Sampling Rate and
Nyquist Theorem
Nyquist frequency: sampling rate required to accurately represent up
to  :

No information is lost if sampling above 

No information is gained by sampling much faster than 

 Sampling Rate and Nyquist
Theorem: Sampling at Nyquist rate
Sampling Rate and
Nyquist Theorem:
Oversampled
Sampling Rate and Nyquist
Theorem: Under-Sampled
 Sampling Rate and Nyquist Theorem:
Special case when Nyquist theorem fails
 Digital to
Analog
Conversion
by
Reconstruction
Filter
 These steps will be
revisited after studying
frequency domain
analysis of signals
Representation of Digital/
Discrete time signal
 Functional and Graphical
Representation
For Signal 1,

For Signal 2,

Functional representation
Graphical representation
Tabular and Sequence Representation
Tabular and Sequence Representation
Basic Sequences: Unit sample
Sequence
 Basic Sequences: Unit sample Sequence
[n]
2
seq

x[n]= [n]
0

-1
-3 -2 -1 0 1 2 3
Time, t (ms)
[n-1] [n-2]
2 2
seq seq

x[n]= [n-2]
1
x[n]= [n-1]

0 0

-1 -1
-2 -1 0 1 2 3 4 -1 0 1 2 3 4 5
Time, t (ms) Time, t (ms)
 Basic Sequences: Unit sample Sequence
[n]
2
seq

x[n]= [n]
0

-1
-3 -2 -1 0 1 2 3
Time, t (ms)
[n+1] [n+2]
2 2
seq seq
x[n]= [n+1]

x[n]= [n+2]
1 1

0 0

-1 -1
-4 -3 -2 -1 0 1 2 -4 -2 0 2
Time, t (ms)
Basic Sequences: Unit Step
Sequence
 Basic Sequences: Other
representation of u[n] and δ[n]
Basic Sequences: Some other common sequences
Power signal vs Energy signal
Power signal or Energy signal ?
Periodic signal vs Aperiodic signal
Discrete-Time System
Example Discrete-Time System
Example Discrete-Time System
Example Discrete-Time System
Example Discrete-Time System
Example Discrete-Time System

Linear or Non-linear?
Example Discrete-Time System
Example Discrete-Time System

Time variant
Systems:
Upsampling/downsampling
System.
See later
Causality
Stability
 FIR and IIR System : Definition
FIR System: If Output of a system depends on current and past input only then that system is called FIR system.
Example y[n]=x[n]-x[n-1]
y[n]={x[n]}^2

IIR System: If Output of a system depends on past output then that system is called IIR system.
Example y[n]=y[n-1]+x[n]
 Example of an FIR system

We can consider every system as a filtering

operation.

Putting, x[n]=δ[n] we get,

h[n]= δ[n] - δ[n-1]
Here, h[n] has only two values. So this is an FIR
filter
 Example of an FIR system

We can consider every system as a filtering operation.

x[n]=[1 2 -2 3]
h[n]= δ[n] - δ[n-1]
find
x[-n+3]= δ[(n-3)] +2δ[(n-2)]- 2δ[(n-1)]+3 δ[n]

Calculate y[n]=x[-n+3]*h[n]
 Example of an IIR system

Putting, x[n]=δ[n] we get,

T=y[n-1]+x[n]

Putting, x[n]=δ[n] we get,

h[n]= h[n-1] +δ[n]
 Example of an IIR system

h[n]= h[n-1] +δ[n] =h[n-2]+ δ[n-1] +δ[n]=h[n-3]+ δ[n-2] +δ[n-1] +δ[n]

h[n-1]=h[n-2]+δ[n-1]
h[n-2]=h[n-3]+ δ[n-2]
X(t)=sinwt+sin2wt
I/O Relationship of an LTI System

Output of an LTI system

can be described as,
y[n]=x[n]*h[n]
Where h[n] is the
response of the system
 I/O Relationship of an LTI System

Output of an LTI system

can be described as,
y[n]=x[n]*h[n]
Where h[n] is the
response of the system
What is response?
What is convolution?
And most importantly,
why convolution?
 Response of a system h[n]
Consider a system, y[n]=x[n]-x[n-1]
Response of a system can be found by setting input, x[n]= δ[n].
Then,
y[n]= δ[n]*h[n]=h[n]

[n]
2
seq

1 y[0]=x[0]-x[-1]=δ[0]-δ[-1]=1
x[n]= [n]

0 y[1]=x[1]-x[0]= δ[1]-δ[0]=-1
-1
-3 -2 -1 0 1 2 3
Time, t (ms)
Response of a system h[n]
Consider a system, y[n]=x[n]-x[n-1]
Response of a system is defined by setting,
x[n]= δ[n]

y[-1]=x[-1]-x[-2]=δ[-1]-δ[-2]=0
y[0]=x[0]-x[-1]=δ[0]-δ[-1]=1
y[1]=x[1]-x[0]= δ[1]-δ[0]=-1
y[2]=x[2]-x[1]= δ[2]-δ[1]=0
y[n]=h[n]= […..0,1,-1,0,0,0……]
Response of a system h[n]
y[-1]=x[-1]-x[-2]=δ[-1]-δ[-2]=0
y[0]=x[0]-x[-1]=δ[0]-δ[-1]=1
y[1]=x[1]-x[0]=δ δ[1]-δ[0]=-1
y[2]=x[2]-x[1]= δ[2]-δ[1]=0
y[n]=h[n]= […..0,1,-1,0,0,0……]

h[n]= δ[n]- δ[n-1]

Response of a system h[n]
So we know response, h[n]= δ[n]- δ[n-1]
Consider, x[n]=[….2,3,-2,0,0,0….].

Then x[n]=2 δ[n+1]+3δ[n]-2δ[n-1]

Using the property of linearity,
y[n]=2 h[n+1]+3h[n]-2h[n-1]
=2(δ[n+1]- δ[n])+3(δ[n]- δ[n-1])-2(δ[n-1]-
δ[n-2])=………………………………..
Response of a system h[n]

So for, x[n]=[….0,0,2,3,-2,0,0,0….].

y[n]=[…0,2,1,-5,2,0]
Response of a Moving average system

x[n]=[….0,0,2,3,-2,0,0,0….].

x[n]=2 δ[n+1]+3δ[n]-2δ[n-1]
h[n]=δ[n]+δ[n-1]
Find y[n].
Response of a Moving average system

x[n]=[….0,0,2,3,-2,0,0,0….].

x[n]=2 δ[n+1]+3δ[n]-2δ[n-1]
h[n]= (1/3)⤫(δ[n+1] +δ[n]+δ[n-1])
Find y[n].
Comments on the property of system from
h[n]
• Causality
• Stability
• Memory/memoryless
• Time variance
Example:
h1[n]=δ[n]-δ[n-1]
h2[n]=δ[n+1]-δ[n-1]
h3[n]=nδ[n]
h4[n]= (accumulator system)
 Properties of δ[n]
δ[n] is an even function:
δ[n] = δ[-n]

Convolution Properties:
x[n]*δ[n] = x[n]
The output of a system with response δ[n] is
same as input.
 Properties of δ[n]
Convolution Properties:
x[n]*δ[n- nd] = x[n- nd]
x[n-nd1]* δ[n- nd2]= x[n- (nd1+nd2)]
y[n]= x[n] * h[n]
Multiplication with δ[n]:
x[n] δ[n]=x[0] δ[n]
x[n] δ[n- 1] = x[1] δ[n- nd]
 Properties of δ[n]
Convolution Properties:
δ[n]*δ[n- nd] = δ[n- nd]
δ[n-nd1]* δ[n- nd2]= δ[n- (nd1+nd2)]
Examples:
δ[n]* δ[n]= δ[n]
δ[n]* δ[-n]= δ[n]* δ[n]=δ[n]
δ[n]* δ[n-3]= δ[n-3]
δ[n-1]* δ[n-3]= δ[n-4]
Properties of δ[n]
Examples:
δ[n+1]* δ[n]= δ[n+1]
δ[n+1]* δ[n-3]= δ[n-2]
δ[n+3]* δ[n-2]= δ[n+1]
δ[n+3]* δ[n+2]= δ[n+5]
 Properties of δ[n]
Examples:
δ[n+1]* δ[-n]= δ[n+1]
δ[n+1]* δ[-n-3]= δ[n+1]* δ[-(n+3)]=
δ[n+1]* δ[n+3]= δ[n+4]
 Properties of δ[n]
Exercise:
Given that, x1[n]=δ[n+1]-2δ[n]+3δ[n-1]
And
x2[n]=-2δ[n+1]+δ[n]+2δ[n-1]
Plot x1[n] and x2[n] and then
Find y[n]= x1[n]*x2[n]
Exercise of Convolution
x1[n]
3

-1

-2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

x2[n]
2

-1

-2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
 Exercise of Convolution
y[n]= x1[n]*x2[n] =
( δ[n+1] -2δ[n] + 3δ[n-1] ) *
( -2δ[n+1] + δ[n] + 2δ[n-1] )=
δ[n+1]* -2δ[n+1] + δ[n+1]* δ[n]+ δ[n+1]*2δ[n-1]
-2δ[n]* -2δ[n+1] -2δ[n]* δ[n]-2δ[n]* 2δ[n-1] +
3δ[n-1]* -2δ[n+1] + 3δ[n-1]* δ[n] + 3δ[n-1]*
2δ[n-1]
Exercise of Convolution
y[n]= x1[n]*x2[n] =
-2 δ[n+2] +5 δ[n+1] -6 δ[n] - δ[n-1] +6 δ[n-1]

y[n]=[-2 5 -6 -1 6]
 Auto-correlation

Auto-correlation of a sequence x[n] is defined as

r[n]= x[n]*x[-n]
where x[-n] is time reversal of x[n].
Exercise of Time Reversal
Exercise:
Given that, x[n]=δ[n+1]-2δ[n]+3δ[n-1]
Find x[-n].
Exercise of Time Reversal
x[n]
4

-2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Time reversal of x[n]: x[-n]

4

-2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Exercise of Time Reversal
x[n]= δ[n+1]-2δ[n]+3δ[n-1]
Putting n=-n we get,
x[-n]=δ[-n+1]-2δ[-n]+3δ[-n-1]
=δ[-(n-1)]-2δ[-(n)]+3δ[-(n+1)]
=δ[n-1]-2δ[n]+3δ[n+1]
=3δ[n+1]- 2δ[n]+ δ[n-1]
Exercise of Autocorrelation
Exercise:
Given that, x[n]=δ[n+1]-2δ[n]+3δ[n-1]
Find r[n]=x[n]*x[-n].
Exercise of Autocorrelation
r[n]=x[n]*x[-n]=
( δ[n+1]-2δ[n]+3δ[n-1] )* (3δ[n+1]- 2δ[n]+ δ[n-1])
Please calculate it.
Exercise of Autocorrelation
r[n]=x[n]*x[-n]=
( δ[n+1]-2δ[n]+3δ[n-1] )* (3δ[n+1]- 2δ[n]+ δ[n-1])
Please calculate it.
R[n]=[3 -8 14 -8 3]
 Topics Covered yet
• Representation of DT sequence
• Systems
• Different representation of systems
• Response of a system
• Types of system from response
• Convolution
• Convolution of two sequence with unit sample sequence
• Autocorrelation
Extend of Convolution

• Sequence 1: n1:n2
• Sequence 2: n3:n4
• Convolution???
Extend of Convolution

• Sequence 1: n1:n2
• Sequence 2: n3:n4
• Convolution (n1+n3): (n2+n4)
 Cross-correlation
Cross-correlation of two sequences x1[n] and
x2[n] is defined as
c[n]= x1[n]*x2[-n]
where x2[-n] is time reversal of x2[n].
Exercise of Cross-correlation
Exercise:
Given that, x1[n]=δ[n+1]-2δ[n]+3δ[n-1]
And
x2[n]=-2δ[n+1]+δ[n]+2δ[n-1]
Find c[n]= x1[n]*x2[-n]
Exercise of Cross-correlation
Exercise:
Given that, x1[n]=δ[n+1]-2δ[n]+3δ[n-1]
And
x2[n]=-2δ[n+1]+δ[n]+2δ[n-1]
Find c[n]= x1[n]*x2[-n]
C[n]=[2 -3 2 7 -6]
Extend of Correlation

• Sequence 1: n1:n2
• Sequence 2: n3:n4
• Correlation???
Extend of Covolution

• Sequence 1: n1:n2
• Sequence 2: n3:n4
• Correlation??? (n1-n4): (n2-n3)
Noise-free transmitted signal
Transmitted signal
Transmitted signal 1
1 sig
0.5

Voltage, V (V)
0.5
Voltage, V (V)

0 0

-0.5 -0.5

-1
0 1 2 3 4 -1
Time, t (ms)
0 1 2 3 4
Time, t (ms)

Continuous- time signal Discrete- time signal

Noise-free Received signal
Recieved signal with delay
1
sig
0.5
Voltage, V (V)

-0.5

-1
0 100 200 300 400 500
Time, t (ms)
Received Power in antenna and Definition of dB

-2= log(Psig/Pnoise)
Psig/Pnoise=10^-2
100*Psig=Pnoise
 Noisy and Attenuated Received signal
Attenuated and Noise corrupted Recieved signal with delay Attenuated and Noise corrupted Recieved signal with delay
0.1
sig sig
0.04
Voltage, V (V)

0.05

Voltage, V (V)
0.02
0
0
-0.02
-0.04 SNR=20
-0.05 SNR=10
-0.06
0 100 200 300 400 500
-0.1
Time, t (ms) 0 100 200 300 400 500
Time, t (ms)
Attenuated and Noise corrupted Recieved signal with delay
0.1 Attenuated and Noise corrupted Recieved signal with delay
sig 0.2
sig
0.05
Voltage, V (V)

0.1

Voltage, V (V)
0
0
-0.05

-0.1 SNR=5 -0.1

SNR=0
0 100 200 300 400 500
-0.2
Time, t (ms) 0 100 200 300 400 500
 Theory of Auto correlation
Suppose, transmitted signal is x[n].
Then Received Signal, y[n]=x[n]+w[n]
Where w[n] is the additive white Gaussian noise or completely random noise with all possible
frequency.
Now Auto correlation of y[n] can be defined as,
Ryy[n]= y[n] * y[-n] =
(x[n]+w[n])*(x[-n]+w[-n]) = x[n]* x[-n] + x[n]* w[-n] +w[n]* x[-n] + w[n] * w[-n]
=Rxx[n] +Rxw[n] +Rwx[n] +Rww[n]
=Rxx[n] + Rww[0]
Because, from the definition of random signal, it has no correlation with other signals and
In case of autocorrelation of random signal it has value at only zeroth lag.
So Ryy[n]= Rxx[n] + W δ[0]
where W is a constant.
Notes : Autocorrelation always preserves the frequency components of a signal
 Auto-correlated Received signal
Attenuated and Noise corrupted Recieved signal with delay
SNR=20
sig
0.04
Voltage, V (V)

0.02
250
0
200
-0.02
150
-0.04
100
-0.06
0 100 200 300 400 500 50
Time, t (ms) 0

-50

-100

-150

-200
-500 -400 -300 -200 -100 0 100 200 300 400 500
 Auto-correlated Received signal
Attenuated and Noise corrupted Recieved signal with delay SNR=10
0.1
sig
0.05
Voltage, V (V)

0
300
-0.05 250

200
-0.1
0 100 200 300 400 500 150

Time, t (ms) 100

50

-50

-100

-150

-200
-500 -400 -300 -200 -100 0 100 200 300 400 500
 Auto-correlated Received signal
SNR=5
Attenuated and Noise corrupted Recieved signal with delay
0.1
sig
0.05
Voltage, V (V)

-0.05 400

-0.1
300
0 100 200 300 400 500
Time, t (ms) 200

100

-100

-200
-500 -400 -300 -200 -100 0 100 200 300 400 500
 Auto-correlated Received signal
SNR=0
Attenuated and Noise corrupted Recieved signal with delay
0.2
sig
0.1
Voltage, V (V)

0
700

-0.1 600

500
-0.2
0 100 200 300 400 500 400
Time, t (ms)
300

200

100

-100

-200
-500 -400 -300 -200 -100 0 100 200 300 400 500
 Auto-correlated Received signal
SNR=-10 and -20
6000

5000

4000

3000

104
5
2000

1000 4

0
3

-1000
-500 -400 -300 -200 -100 0 100 200 300 400 500 2

-1
-500 -400 -300 -200 -100 0 100 200 300 400 500
 Theory of Cross-correlation
Suppose, transmitted signal is x[n].
Then Received Signal, y[n]=x[n]+w[n]
Now cross correlation of y[n] with x[n] can be defined as,
Ryy[n]= y[n] * x[-n] =
(x[n]+w[n])*x[-n] =
x[n]* x[-n] + w[n] * x[-n]
=Rxx[n]+Rwx[n]
=Rxx[n]
 Cross-correlated Received signal
Attenuated and Noise corrupted Recieved signal with delay
SNR=20
sig
0.04
Voltage, V (V)

0.02
0
-0.02
-0.04 200

-0.06 150
0 100 200 300 400 500
Time, t (ms) 100

50

-50

-100

-150

-200
-500 -400 -300 -200 -100 0 100 200 300 400 500
 Cross-correlated Received signal
Attenuated and Noise corrupted Recieved signal with delay SNR=10
0.1
sig
0.05
Voltage, V (V)

250
-0.05
200

-0.1 150
0 100 200 300 400 500
100
Time, t (ms)
50

-50

-100

-150

-200
-500 -400 -300 -200 -100 0 100 200 300 400 500
 Cross-correlated Received signal
Attenuated and Noise corrupted Recieved signal with delay
SNR=5
0.1
sig
0.05
Voltage, V (V)

-0.05
200

-0.1
150

0 100 200 300 400 500

100
Time, t (ms)
50

-50

-100

-150

-200
-500 -400 -300 -200 -100 0 100 200 300 400 500
 Cross-correlated Received signal
SNR=0
Attenuated and Noise corrupted Recieved signal with delay
0.2
sig
0.1
Voltage, V (V)

0
200
-0.1
150

-0.2
0 100 200 300 400 500 100

Time, t (ms) 50

-50

-100

-150

-200
-500 -400 -300 -200 -100 0 100 200 300 400 500
 Auto-correlated Received signal
250 SNR=-10 and -20
200

150

100

50

-50

-100 150

-150
100
-200

-250 50
-500 -400 -300 -200 -100 0 100 200 300 400 500

-50

-100

-150

-200
-500 -400 -300 -200 -100 0 100 200 300 400 500