HO CHI MINH CITY UNIVERSITY OF SCIENCE

FACULTY OF ELECTRONICS AND TELECOMMUNICATIONS

DEPARTMENT OF TELECOMMUNICATIONS AND NETWORKS

Course

DIGITAL SIGNAL PROCESSING

Chapter
INTRODUCTION TO DIGITAL
1 SIGNAL PROCESSING

Lecturer: Tran Thi Thao Nguyen, Ph.D. Email: tttnguyen@hcmus.edu.vn

 Outline
1. Course overview
2. Digital Signal Processing
3. Basic operations & block diagrams
4. Classes of sequences

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 2
 1. Course overview
Digital signal processing:
Modifying signals with computers
Book:
Mitra “Digital Signal Processing” (3rd ed., 2005)

Lecturer:
Tran Thi Thao Nguyen, Ph.D.
Contact:
Email: tttnguyen@hcmus.edu.vn
Microsoft Teams

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 3
 Grading structure
Class assignments: 20 points

Group project: 20 points

Final examination: 50 points

Class Activities: 10 points

Questions for Lecture
Attendance >= 8 days

Total = 100 points

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 4
 Course project

Course project
Work in Group (5-6 members/group)
oTheory
 Choose a theory topic from the lecturer or
yourself selected
oPractice
 Goal: hands-on experience with DSP
 Any ideas/topics are good
 Practical implementation
 Recommend MATLAB
Submit report and presentation

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 5
 Schedule
 45 periods.

 4-5 periods / day.

 Totally in 10 days:
 9 days for lecture
 1 day for Group presentation

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 6
 Course content

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 7
 2. Digital Signal Processing

Signals: Information-bearing function

E.g. sound: air pressure variation at a point as a

function of time p(t)

Dimensionality:
Sound: 1-Dimension
Greyscale image i(x,y): 2-D
Video: 3 x 3-D: {r(x,y,t) g(x,y,t) b(x,y,t)}
1

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 8
 Example signals

Noise - all domains

Spread-spectrum phone - radio
ECG - biological
Music
Image/video - compression
….

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 9
 Signal processing

Modify a signal to extract/enhance/rearrange the

information

Origin in analog electronics e.g. radar

Examples
Noise reduction
Data compression
Representation for recognition/classification

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 10
 Digital Signal Processing

DSP = signal processing on a computer

Two effects: discrete-time, discrete level

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 11
 DSP vs. analog SP

Conventional signal processing:

Digital SP system:

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 12
 Digital vs. analog

Pros
Noise performance - quantized signal
Use a general computer - flexibility, upgrade
Stability/duplicability
Novelty

Cons
Limitations of A/D & D/A
Baseline complexity / power consumption

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 13
 DSP example

Speech time-scale modification: extend duration

without altering pitch

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 14
 3. Operations on signals

Discrete time signal often obtained by sampling a

continuous-time signal

Sequence {x[n]} = xa(nT), n=…-1,0,1,2…

T= samp. period; 1/T= samp. frequency

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 15
 Sequences

Can write a sequence by listing values:

Arrow indicates where n=0

Thus,

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 16
 Left- and right-sided

x[n] may be defined only for certain n:

N1 ≤ n ≤ N2: Finite length (length = …)
N1 ≤ n: Right-sided (Causal if N1 ≥ 0)
n ≤ N2: Left-sided (Anticausal)

Can always extend with zero-padding

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 17
 Operations on sequences

Addition operation:
Adder

Multiplication operation
Multiplier

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 18
 More operations

Product (modulation) operation:

Modulator

E.g. Windowing:
Multiplying an infinite-length sequence by a finite-
length window sequence to extract a region

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 19
 Time shifting

Time-shifting operation:
where N is an integer

If N > 0, it is delaying operation

Unit delay

If N < 0, it is an advance operation

Unit advance

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 20
 Combination of basic operations

Example

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 21
 Up- and down-sampling

Certain operations change the effective sampling rate

of sequences by adding or removing samples

Up-sampling = adding more samples = interpolation

Down-sampling = discarding samples = decimation

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 22
 Down-sampling

In down-sampling by an integer factor M > 1,

every M-th sample of the input sequence is kept and
M - 1 in-between samples are removed:

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 23
 Down-sampling
An example of down-sampling

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 24
 Up-sampling

Up-sampling is the converse of downsampling: L-1

zero values are inserted between each pair of original
values.

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 25
 Up-sampling
An example of up-sampling

3
Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 26
 Complex numbers

 A mathematical convenience that lead to simple

expressions
 A second “imaginary” dimension (j≡√-1) is added to
all values.
Rectangular form: x = xre + j·xim
where magnitude |x| = √(xre2 + xim2)
and phase θ = tan-1(xim/xre)
Polar form: x = |x| ejθ = |x|cosθ + j· |x|sinθ
(ejθ= cosθ + j sinθ)

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 27
 Complex math

When adding, real and

imaginary parts add:
(a+jb) + (c+jd) =
(a+c) + j(b+d)

When multiplying,
magnitudes multiply
and phases add:
rejθ·sejφ = rsej(θ+φ)

Phases modulo 2π

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 28
 Complex conjugate

Flips imaginary part / negates phase:

Conjugate x* = xre – j·xim = |x| ej(–θ)
Useful in resolving to real quantities:
x + x* = xre + j·xim + xre – j·xim = 2xre
x·x* = |x| ej(θ) |x| ej(–θ) = |x|2

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 29
 Classes of sequences

Useful to define broad categories

Finite/infinite (extent in n)

Real/complex:
x[n] = xre[n] + j·xim[n]

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 30
 Classification by symmetry

Conjugate symmetric sequence:

if x[n] = xre[n] + j·xim[n]
then xcs[n] = xcs*[-n]
= xre[-n] – j·xim[-n]

Conjugate antisymmetric:
xca[n] = –xca*[-n] = –xre[-n] + j·xim[-n]
Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 31
 Conjugate symmetric decomposition
Any sequence can be expressed as conjugate symmetric
(CS) / antisymmetric (CA) parts:
x[n] = xcs[n] + xca[n]
where:
xcs[n] = 1/2(x[n] + x*[-n]) = xcs*[-n]
xca[n] = 1/2(x[n] – x*[-n]) = -xca*[-n]

When signals are real,

CS → Even (xre[n] = xre[-n]), CA → Odd

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 32
 Basic sequences

 Unit sample sequence:

 Shift in time:

Can express any sequence with :

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 33
 More basic sequences

Unit step sequence:

Relate to unit sample:

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 34
 Exponential sequences

Exponential sequences are eigenfunctions of LTI

systems
General form: x[n] = A·an
If A and a are real (and positive):

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 35
 Complex exponentials

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 36
 Complex exponentials

Complex exponential sequence can ‘project down’

onto real & imaginary axes to give sinusoidal
sequences

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 37
 Periodic sequences

Lecturer: Tran Thi Thao Nguyen, Ph.D. - Department of Telecommunications and Networks, FETEL, HCMUS 38
 QA

Lecturer: Tran Thi Thao Nguyen, Ph.D.

Ho Chi Minh City University of Science
Faculty of Electronics - Communications
Department of Telecommunication - Networks
Email: tttnguyen@hcmus.edu.vn