Digital Signal Processing L.

Demi 2019/20

Lecture 1 - Introduction

Professor: Dr. Libertario Demi

[email protected]
General information
L. Demi 2019/20
Material
• Course Slides (available online)
• Textbook:
 "The Scientist and Engineer's Guide to Digital Signal Processing", California
Technical Publishing.
 D. Manolakis and V. Ingle: Applied Digital Signal Processing, Cambridge
University Press, 2011.
 John G. Proakis and Dimitris G. Manolakis, Digital Signal Processing: Principles,
Algorithms, and Applications, 4th edition, 2007.
 Sanjit K. Mitra: Digital Signal Processing: A Computer-Based Approach, 4/e,
University of California - Santa Barbara, McGraw-Hill Education, 2011.
 Oppenheim and Schafer: Discrete-Time Signal Processing, Prentice Hall Signal
Processing.
General information
L. Demi 2019/20
Grading modality
• First - Written exam (80%)
• Next - Lab exam (20%)

Others
• Office hours: Wednesdays from 14:30 to 16:30 (please take an appointment by
email)
General information
L. Demi 2019/20
Grading modality
• First - Written exam (80%)
• Next - Lab exam (20%)

Others
• Office hours: Wednesdays from 14:30 to 16:30 (please take an appointment by
email)
• Mandatory to access the Labs
• https://www.dii.unitn.it/en/828/course-health-and-safety-in-the-workplace
• The course lasts 4 hours and it is available with the university credentials in
the website of Didattica online.
General information
L. Demi 2019/20
Grading modality
• First - Written exam (80%)
• Next - Lab exam (20%)

Others
• Office hours: Wednesdays from 14:30 to 16:30 (please take an appointment by
email)
• Mandatory to access the Labs
• https://www.dii.unitn.it/en/828/course-health-and-safety-in-the-workplace
• The course lasts 4 hours and it is available with the university credentials in
the website of Didattica online.
Digital Signal Processing
L. Demi 2019/20
Digital Signal Processing
L. Demi 2019/20
What is a signal?
Digital Signal Processing
L. Demi 2019/20
What is a signal?
An act, gesture, event, password or something that may have been agreed on and which
carries information
Digital Signal Processing
L. Demi 2019/20
What is a signal?
A function which describes the variations of a detectable quantity by which information can
be transmitted from a source to a receiver
Digital Signal Processing
L. Demi 2019/20
Acoustic signal:
teaching
Digital Signal Processing
L. Demi 2019/20
Acoustic signal:
teaching

ΔP

t
Digital Signal Processing
L. Demi 2019/20
Acoustic signal:
teaching

ΔP

t
Digital Signal Processing
L. Demi 2019/20
Acoustic signal:
teaching

ΔP

t
ΔP : detectable physical quantity
t: variable
Me: source of the signal
You: receivers
Frequencies between 20 Hz and 20 KHz
Digital Signal Processing
L. Demi 2019/20
Acoustic Signals:
Echo localization

ΔP

t
Digital Signal Processing
L. Demi 2019/20
Acoustic Signals:
Echo localization

ΔP

t
 Digital Signal Processing
L. Demi 2019/20
Acoustic Signals:
Echo localization
ΔP : detectable physical quantity
t: variable
Bat: source of the signal
The bat itself (echo-naviagation): receiver
Frequencies between 20 kHz and 80 KHz ΔP

t
Digital Signal Processing
L. Demi 2019/20
Chemical Signals:
colony behavior is mediated by pheromones
Digital Signal Processing
L. Demi 2019/20
Chemical Signals:
colony behavior is mediated by pheromones

Ph

t
 Digital Signal Processing
L. Demi 2019/20
Chemical Signals:
colony behavior is mediated by pheromones
Ph level : detectable quantity
t: variable
Ant: source of the signal
The other ants: receivers

Ph

t
Digital Signal Processing
L. Demi 2019/20
Artificial Signals:
Traffic control
Digital Signal Processing
L. Demi 2019/20
Artificial Signals:
Traffic control

green

orange
Color

red

t
 Digital Signal Processing
L. Demi 2019/20
Artificial Signals:
Traffic control
Colors: detectable “quantity”
t: variable
Traffic light: source of the signal
Drivers: receivers

green

orange
Color

red

t
 Digital Signal Processing
L. Demi 2019/20
Artificial Signals:
Traffic control
Colors: detectable “quantity”
t: variable
Traffic light: source of the signal
Drivers: receivers

green
Colors are not numbers
orange
-> need to be represented
Color

red Finite number of values

Source and Receiver are not the same system

t
Digital Signal Processing
L. Demi 2019/20
Artificial Signals:
Computer animation
Digital Signal Processing
L. Demi 2019/20
Artificial Signals:
Computer animation

X
Digital Signal Processing
L. Demi 2019/20
Artificial Signals:
Computer animation

Pixels

X
 Digital Signal Processing
L. Demi 2019/20
Artificial Signals:
Computer animation
Colors: detectable “quantity”
t: variable
2D space coordinates system: additional variables
Computer: source
Spectators: receivers

Pixels

X
Digital Signal Processing
L. Demi 2019/20
 Digital Signal Processing
L. Demi 2019/20
Signal classification

• Natural and Artificial

• Continuous or Discrete Signals

• Analog and Digital Signals

• Periodic and Non-Periodic Signals

• Energy and Power Signals

• Random and Deterministic Signals

 Digital Signal Processing
L. Demi 2019/20
Signal classification

• Natural and Artificial

• Continuous or Discrete Signals

• Analog and Digital Signals

• Periodic and Non-Periodic Signals

• Energy and Power Signals

• Random and Deterministic Signals

 Digital Signal Processing
L. Demi 2019/20
Signal classification

• Natural and Artificial

• Continuous or Discrete Signals

• Analog and Digital Signals

• Periodic and Non-Periodic Signals

• Energy and Power Signals

• Random and Deterministic Signals

 Digital Signal Processing
L. Demi 2019/20
Signal classification

• Natural and Artificial

• Continuous or Discrete Signals

• Analog and Digital Signals

• Periodic and Non-Periodic Signals

• Energy and Power Signals

• Random and Deterministic Signals

 Digital Signal Processing
L. Demi 2019/20
Signal classification

• Natural and Artificial

• Continuous or Discrete Signals

• Analog and Digital Signals

• Periodic and Non-Periodic Signals

• Energy and Power Signals

• Random and Deterministic Signals

 Digital Signal Processing
L. Demi 2019/20
Signal classification

• Natural and Artificial

• Continuous or Discrete Signals

• Analog and Digital Signals

• Periodic and Non-Periodic Signals

• Energy and Power Signals

• Random and Deterministic Signals

Continuous and discrete signals
L. Demi 2019/20

A function which describes the variations of a detectable quantity by which information

can be transmitted from a source to a receiver
Continuous and discrete signals
L. Demi 2019/20

A function which describes the variations of a detectable quantity by which information

can be transmitted from a source to a receiver

It is continuous if the signal exists for every values of the given variable(s). For example
for every value of t, where t is a continuous real valued variable representing time

s(t)

t
Continuous and discrete signals
L. Demi 2019/20

It is discrete if the signal exists only for a specific set of values of the given variable(s)

It is expressed as s[n], with n being an integer which varies discretely

dt : interval
s[n]

n
Analog and digital signals
L. Demi 2019/20
Analog and digital signals
L. Demi 2019/20

It is digital signal if it is a discrete signal which values belong only to a finite set of real
numbers. All signals stored on a computer or displayed on a computer screen are digital
signals

The opposite of a digital signal is an analog signal where the signal amplitudes varies
with respect to a continuous variable and can assume any value

s(t) : analog

s[n]

n
 From analog to discrete
L. Demi 2019/20
s(t)

Sampling s[n]
t
Here the signal is discretized
n
 From analog to digital
L. Demi 2019/20
s(t)

Sampling s[n]
t
Here the signal is discretized
n

Quantization
Here the signal amplitudes are
represented with a finite set of values

s[n]

n
Periodic and non-periodic signals
L. Demi 2019/20
Periodic and non-periodic signals
L. Demi 2019/20
s(t) = s(t+nT)
with n being an integer and T the period

s[n]=s[n+kM]
with k being an integer and M being an integer and the period

n
Energy and power signals
L. Demi 2019/20
Energy and power signals
L. Demi 2019/20

If the signal energy is finite, it is an energy signal

E = å s(n) < ¥
¥ 2

-¥

If the signal infinite average power is finite (but not zero), it is a power signal

1
å
N 2

P¥ = lim
N®¥ 2N +1
-N
s(n) < ¥

A signal can be an energy or power signal, or neither type.

Deterministic and random signals
L. Demi 2019/20
Deterministic and random signals
L. Demi 2019/20
The signal is deterministic if it can assume one and only one value for a given value
of the variable to which it is referred

s(t)

The signal is random if it can assume multiple values, each with a certain probability,
for a given value of the variable to which it is referred

s(t)

t
Digital Signal Processing
L. Demi 2019/20
Processing concerns in essence all the possible operations which can be performed
with signals

A way to describe one or a set of operations is by means of systems

The system can thus also be seen as the set of operations, or the process, which
links an unprocessed signal (input) to a processed signal (output)

Input Output
System
Systems
L. Demi 2019/20
As for signals, systems can be analog and digital depending on whether they have as
input and output an analog or digital signal.

t
x(t) t
y(t) = ò x(t ) dt
ò -¥
-¥

n
x[n] n
y[n] = å x[k]
å -¥
-¥
Systems
L. Demi 2019/20
Signals can be recorded, stored, played and visualized analogically and digitally

0101100

The bit (binary digit), is the binary unit of

information for digital signals and can have
only one of two values.
Analog to Digital
L. Demi 2019/20

Analog world Interface Digital world

0101100
ADC

0101100
DAC
 Analog to Digital Conversion
F=1/dt
30213 L. Demi 2019/20

Samplin 11 00 10 01 11
Quantization Coder
s(t) g

s(t)

t
 Analog to Digital Conversion
F=1/dt
30213 L. Demi 2019/20

Samplin 11 00 10 01 11
Quantization Coder
s(t) g s[n]=s(nT)

s(t)

dt t
 Analog to Digital Conversion
F=1/dt
30213 L. Demi 2019/20

Samplin 11 00 10 01 11
Quantization Coder
s(t) g s[n]=s(nT) sd[n]

dx dx=Signal Dynamic Range/(2k-1)

Signal Dynamic Range

s(t)

dt t
 Analog to Digital Conversion
F=1/dt
30213 L. Demi 2019/20

Samplin 11 00 10 01 11
Quantization Coder
s(t) g s[n]=s(nT) sd[n]
7
dx dx=Signal Dynamic Range/(2k-1)
5 5

Signal Dynamic Range

4 4 4
s(t) 3 3

2 2

1
0

dt t
 Analog to Digital Conversion
F=1/dt
30213 L. Demi 2019/20

Samplin 11 00 10 01 11
Quantization Coder
s(t) g s[n]=s(nT) sd[n] k=3
7 111
dx 111 dx=Signal Dynamic Range/(2k-1) 110
5 5
101
101 101 4 4 4 100
s(t) 3 100 100 100 3 011
011 2 2 011 010
010 1 010 001
0 001 000

dt 000 t
Problems?
L. Demi 2019/20
sd(n)

s(t)

t
 Problems?
L. Demi 2019/20
sd(n)

s(t)

We are going from an infinite precision (not quantized signal) to a finite precision (quantized)
Quantization Error
L. Demi 2019/20

The dynamic range (DR) equals = Max[s(t)]-Min[s(t)]

If DR = 14 and we can use only 3 bits, since 23 = 8, we have 8 possible

quantization levels

Thus our quantization step is DR/(23-1) = 2, and the error that we make when
quantizing the discrete signal is:

0 <= Error < DR/(2n-1)

Note: the larger the DR the bigger the error, and the higher the amount of bits
the smaller the error
Quantization Error
L. Demi 2019/20

DR = 14, 3 bits DR = 14, 4 bits

Digital to Analog Conversion
L. Demi 2019/20
Ideal DAC

11 00 10 01 11 Mathematical
Pulser
s’d[n] Interpolation s’(t)
Digital to Analog Conversion
L. Demi 2019/20
Ideal DAC

11 00 10 01 11 Mathematical
Pulser
s’d[n] Interpolation s’(t)

Practical DAC

11 00 10 01 11 More analog
DAC signal
s’(t) processing s’’(t)

Note: in practice the final analog signal is only a reconstruction of the original
analog signal, and it is a modified version of it.

In this sense part of the original information is lost.

Digital to Analog Conversion
L. Demi 2019/20
Ideal DAC

11 00 10 01 11 Mathematical
Pulser
s’d[n] Interpolation s’(t)

Practical DAC

11 00 10 01 11 More analog
DAC signal
s’(t) processing s’’(t)

When digitizing a signal it is thus very important to select the right parameters
in order to guarantee a sufficiently good signal reconstruction, where good is
dependent of the application
Why going digital?
L. Demi 2019/20

Digital signals are easier to store and transport, and suffer less from
deterioration or loss of signal
Why going digital?
L. Demi 2019/20

Digital signal processing operations are much more easily reconfigurable

compared to analog signal processing, which makes a digital system much
more flexible
Why going digital?
L. Demi 2019/20

Digital systems are less sensitive to the components tolerance and

environmental changes, and are thus more stable
Areas of application of DSP
L. Demi 2019/20
Areas of application of DSP
L. Demi 2019/20

Biomedical engineering
Areas of application of DSP
L. Demi 2019/20

Communication
Areas of application of DSP
L. Demi 2019/20

Photography
Areas of application of DSP
L. Demi 2019/20

Music
Next Lecture
L. Demi 2019/20

Discrete Time Signals

• characterization and classification of

discrete time signals

• elementary operations on them