You are on page 1of 17

DIGITAL SIGNAL PROCESSING

LECTURE #1, 2, 3
13TH – 15TH FEBRUARY 2023
 Instructor’s introduction:
 Amena Ejaz Aziz (PhD, Hong Kong
Polytechnic University.
 Area of specialization is visible light
communication).

PREFACE
 Contact information:
 Email: amena.ejaz@ist.edu.pk
 Mobile: 0332-5169909

2
COURSE MATERIAL

Textbook:

 J. G. Proakis and D. G. Manolakis, Digital Signal Processing –


Principles, algorithms and Applications, 4th Edition, Pearson, Prentice
Hall, 2007.

Class notes/ lectures


3
COURSE OUTLINE
Chapter Chapter names Weeks Weeks
numbers (tentative) (Cumulative)
1 Introduction 1 1
2 Discrete time signals and systems 2 3
3 Z – transforms 2 5
4 Fourier analysis of signals 1 6
5 Fourier analysis of systems 3 9
6 Sampling and reconstruction 1 10
7 Discrete Fourier transform 1 11
8 Fast Fourier transform 1 12
9 Structures 1 13
10 FIR filter designing 1 14 4

10 IIR filter designing 1 15


ASSESSMENT CRITERIA

Criteria Expected Number Weightage

Assignment
(both hand-written and software 3~4 12%
based)

Quizzes Best of 6 18%

OHTs 2 30%

5
Final 1 40%
GENERAL INSTRUCTIONS

 Please don’t expect teacher to change the “assessment criteria” if for your any reason YOU couldn’t perform well in a certain
domain. Assessment weightages have been conveyed to you on the VERY first day.
 Please don’t request me AT ALL if you suffer from short attendance. It is your own responsibility to ensure more than 80%
attendance in this subject.
 Please don’t request your teacher to delay the submission of any assignment or postpose a quiz if YOU have a commitment.
 About choices in quizzes, it is given to FACILITATE you so that you might not suffer if you miss a quiz, it is not teacher’s
and whole class’s responsibility to ensure that one appears in every quiz.
 If the whole class has any genuine issue, and wants to delay/postpone an assessment, please contact me through CR.
 Your one-hour tests and final exam will be OPEN NOTES, CLOSED BOOK exams.
 There will be NO RETAKES of any quiz/one-hour tests.
 In your FIRST (WhatsApp/SMS/email) message, please write your complete name and subject, before discussing the issue.
 Be participative, it is for your own good!!! 6
Digital Signal Processing

(1) Not defined on every Performing a series of actions in order to achieve a particular
instant. result.
(2) Defined as numbers For example, signal security, multiplexing, data compression,
speaker/system identification, reconstruction, filtering etc.

A physical quantity that conveys


information. The “quantity” usually
varies with an independent variable. 7
APPLICATIONS OF DSP

 Radar and sonar (target detection, position and velocity estimation)

 Biomedical stuff (analysis of biomedical signals, patient monitoring, preventive

healthcare, ECG signals’ analysis etc.)

 Speech applications (noise reduction, speech recognition, synthesis of artificial speech)

8
 Multimedia system (data compression, efficient transmission, storage etc.)
CHAPTER 1 – INTRODUCTION

 (1.1) Signals, systems and signal processing

 (1.2) Classification of signals

 (1.3) The concept of frequency in continuous-time and discrete time signals

 (1.4) Analog-to-digital and digital-to-analog conversion 9


1.1 SIGNALS, SYSTEMS AND SIGNAL PROCESSING

1.1.2 Advantages of digital over analog


1.1.1 Basic elements of a DSP processing
 Precise signal level does not remain important.
 Digital signals are fairly immune to the
imperfections of electronics/hardware.
 Can be coded (to combat channel noise) and
encrypted (to be sent to intended user).
 Enables transmission of signal over longer
distances (concept of regenerative repeater with
elimination of noise).
 Storage of digital signals is easy and reliable. 10
1.2 CLASSIFICATION OF SIGNALS
1.2.2 Continuous-time versus discrete-time signals
1.2.3 Continuous-valued versus discrete-valued signals

11
1.2 CLASSIFICATION OF SIGNALS
sampling

quantization

12

reconstruction
1.3 THE CONCEPT OF FREQUENCY IN CONTINUOUS-TIME
AND DISCRETE-TIME SIGNALS

1.3.1 Continuous-time sinusoidal signals


A continuous time signal is mathematically defined as:
𝑥𝑎 𝑡 = 𝐴 cos Ω𝑎 𝑡 + 𝜃 (1)
(i) The signal 𝑥𝑎 𝑡 is always periodic and thus satisfy the condition:
𝑥𝑎 𝑡 + 𝑇 = 𝑥𝑎 𝑡 (2)
(ii) Two continuous-time sinusoid signals 𝑥𝑎 𝑡 and 𝑥𝑏 𝑡 with two distinct frequencies
Ω𝑎 and Ω𝑏 will always be different.
(iii) From (ii), it is true that if 𝑥𝑎 𝑡 = cos Ω𝑎 𝑡 , and 𝑥𝑏 𝑡 = cos Ω𝑏 𝑡 = cos൫(Ω𝑎 +
2𝜋)𝑡൯, then these two signals will be different with 𝑥𝑏 𝑡 a higher frequency component 13

than 𝑥𝑎 𝑡 .
1.3 THE CONCEPT OF FREQUENCY IN CONTINUOUS-TIME
AND DISCRETE-TIME SIGNALS

1.3.2 Discrete-time sinusoidal signals

A discrete time signal is mathematically defined as:


𝑥𝑎 𝑛 = 𝐴 cos 𝜔𝑎 𝑛 + 𝜃 (3)
(i) A discrete-time sinusoidal may or may not be periodic and hence may not satisfy:
𝑥𝑎 𝑛 + 𝑁 = 𝑥𝑎 𝑛 (4)
(ii) Two discrete-time sinusoid signals 𝑥𝑎 𝑛 and 𝑥𝑏 𝑛 with two distinct frequencies 𝜔𝑎
and ω𝑏 may or may not be different.
(iii) From (ii), it is true that if 𝑥𝑎 𝑛 = cos 𝜔𝑎 𝑛 , and 𝑥𝑏 𝑛 = cos 𝜔𝑏 𝑛 = cos൫(𝜔𝑎 + 14

2𝜋)𝑛൯, then these two signals will always be similar/ indistinguishable.


CONCEPT OF FUNDAMENTAL/PRINCIPAL RANGE
 The range 0 ≤ 𝜔 < 2𝜋 or −𝜋 ≤ 𝜔 < 𝜋 is the principal range for a discrete-time signal. (Signals existing in
this range are indistinguishable.)
 Of these two, we prefer to use the range −𝜋 ≤ 𝜔 < 𝜋. Why?

15
1.3 THE CONCEPT OF FREQUENCY IN CONTINUOUS-TIME
AND DISCRETE-TIME SIGNALS

1.3.3 Harmonically related complex exponentials

Consider a discrete time complex exponential signal defined as:


𝑥 𝑛 = 𝑒 𝑗𝜔𝑛 (5)

Its 𝑘-th harmonic will be defined as:


𝑥𝑘 𝑛 = 𝑒 𝑗𝜔𝑘 𝑛 = 𝑒 𝑗𝜔𝑘𝑛 (6)

While infinite number of exponentials are possible for a continuous time signal, only 𝑁
number of harmonics are possible for a discrete time signal, including the signal with
fundamental frequency. 16
THINGS TO INVESTIGATE:

 What might be some disadvantages of using DSP?

 Why do we still want to use it then?

 Try plotting following signals through MATLAB, and see if they are periodic or not:
2𝜋
cos 𝑛, cos 0.35𝜋𝑛 , cos 𝑛 , cos 2𝜋𝑛 , cos log 2𝑛
7
 Try to see (through MATLAB) if signals with different frequencies are indeed distinct in the principal range.
 Try plotting of ALL harmonics for the following signal:

2𝜋
𝑥 𝑛 = cos 𝑛
5 17

You might also like