DIGITAL SIGNAL

PROCESSING
Introduction
Lecture 1
Solomon T.

Mekelle Institute of Technology-Mekelle Uni 10/29/2024 1

DEFINITION
 Signal
 A function of independent variables such as time, distance, position, temperature
and pressure
 Flow of information: A signal carries information,
 Digital
 Uses discrete signals to represent data in the form of numbers
 Processing
 Operating in some fashion on signal to extract some useful information
 Digital Signal Processing
 Concerned with the representation of signals by sequence of numbers or
symbols and the processing of these sequence
 the algorithmic operation carried out on sequence of numbers to extract the
information
 Changing or analysing information which is measured as discrete sequence of
numbers
Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 2
 TIME AND FREQUENCY DOMAIN
SIGNALS
 A signal can be defined simply as a mathematical
function: y=f(x), x is independent variable which
specifies the domain
 y=sin(wt) is a time domain
 X(ω)=1/(-mω2+icω+k) is a frequency domain
signal;

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 3

SIGNAL PROCESSING
 Signals may have to be transformed in order to
 amplify or filter out embedded information
 detect patterns
 prepare the signal to survive a transmission channel
 prevent interference with other signals sharing a medium
 undo distortions contributed by a transmission channel
 find information encoded in a different domain
 To do so, we also need
 methods to measure, characterize, model and simulate
transmission channels
 mathematical tools that split common channels and
transformations into easily manipulated building blocks
Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 4
 ELEMENTS OF DIGITAL SIGNAL PROCESSING

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 5

 CONT‘D
 Input signal
 Generated by communication systems or tranducer
 Represented by x(t)--- Analog signal
 Anti-aliasing filter
 It is a low pass filter used to remove high frequency component and/or band
limit the signal
 Sample and hold circuit
 Take samples from the input signals and keeps the voltage level relatively
constant which is the requirment of the ADC
 ADC
 Convert analog signal into digital

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 6

 CONT‘D
 Digital Signal Processor
 Process the input signal as required
 DAC
 Convert the signal into analog
 Reconstruction Filter
 Remove high frequency component if available
 Output Signal
 Represented by y(t)---analog signal

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 7

ADVANTAGE OF DIGITAL OVER ANALOG
SIGNAL PROCESSING
 Versatility: reprogramable and easly portable to different HW
 Repeatability: easily duplicated.
 Simplicity: easy to build
 Accuracy: use resistors, capacitors and inductors and is difficult to have
accurate design. In DSP ....better accurace
 Remote processing : Digital signals are easily stored on magnetic tapes
and disks without deterioration or loss of signal fidelity
 So the signals become transportable and can be processed offline
 Impelentation of Algorithms: Mathematical lgorithms are easy to
impelement
 Easy upgradations: Software ...DSP systems are easy to upgrade
 Cheaper: In many applications DSP compenents are cheaper than their
digital counterpart
 Multiplexing:The same digital signal processor can be used to perform
two operations by time multiplexing, since digital signals are defined only
at finite number of time instants.
Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 8
LIMITATIONS OF DIGITAL SIGNAL
PROCESSING SYSTEMS
 System Complexity: increases coplexity as it uses ADC & DAC
and it requires more time
 Bandwidth Limitation: If the inpute signal has wide bandwidth
it requires fast ADC.
 To avoid aliasing effect---Sampling theorem . Such signals requires fast DS
processors.
 In practice, difficult to have fast ADC and DS processors
 Power Consumption: power dissipation is more as a typical
DSP chip contains more than 4 transistors.
 In analog signal processor, passive elements (resistors, inductors,
capacitor) are used, which dissipate very less power.
 In digital signal processor, active elements like transistors are used, which
dissipate more power.
 Expensive for small applications

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 9

APPLICATION OF DSP
 Image Processing
 Pattern Recognition
 Animation
 Robotic Vision
 Image enhancement
 Image restoration
 Instrumentation and Control
 Spectral analysis
 Data Compression
 Noise reduction
 Speech/Audio
 Speech Recognition
 Automotive Applications
 Vibration analysis
 Voice commands
Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 10
APPLICATION OF DSP
 Biomedical
 Scanner
 Electrocardiogram (ECG) analysis –monitoring electrical activity of a heart
 Patient monitoring
 Telecommunications
 Eco canclation
 Data Communication
 Military
 Sonar processing
 Radar processing
 Secure communications
 Industrial Application
 Robotics
 Power line monitor

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 11

 BASIC SIGNALS
 To analyize coommunication systems standard signals are used:
test performance of a system
 Apply an input and check the output
 Input-output characterstics help us to identify the properties of the
system/operation of the system
 Common basic signals
 Delta ou unit impulse denoted by
 Unit Step Signal
 Ramp signal
 Exponential Signal
 Sinosodial Signal

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 12

 DISCRETE CONVOLUTION

 a method of finding the response of linear time-invariant (LTI) systems.

 It is based on the concepts of linearity and time invariance and assumes

that the system information is known in terms of its impulse response
h[n]
 The convolution operation is defined as :
 y[n] =x[n] *h[n]
 is also called linear convolution or the convolution sum

 The order in which we perform the operation does not matter, and we can
interchange the arguments of x and h without affecting the result.
 y[n] =x[n] *h[n]= h[n] *x[n]

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 13

 CONT‘D

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 14

EXAMPLES
 Find the response y[n], given x[n] = h[n] = u[n]
 Solution y[n] =r[n+1]=(n+1)u[n]
 Example 2:
 Find the response given x[n] = h[n] = a^nu[n].
 Solution: y[n] = (n+1) a^nu[n]

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 15

SOME PROPERTIES OF CONVOLUTION

 δ[n]* x[n] = x[n]

 δ[n]* δ[n]= δ[n]
 y[n] = u[n] *x[n] is the convolution of a signal x[n] with a unit step. The
response is the running sum of the signal x[n]:
 Commutative
 h[n]* x[n]= x[n]* h[n]
 Associative
 (x[n]* h1[n])*h2[n] = x[n]* (h1[n]*h2[n])
 Distributive
 x[n]* (h1[n]+h2[n]) =x[n]* h1[n]+ x[n]* h2[n]

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 16

CONVOLUTION OF FINITE SEQUENCES

 In practice, we deal with sequences of finite length

 The convolution y[n] of two finite-length sequences x[n] and h[n]
is also of finite length and is subject to the following rules
 The starting index of y[n] equals the sum of the starting indices
of x[n] and h[n].
 The ending index of y[n] equals the sum of the ending indices
of x[n] and h[n].
 The length Ly of y[n] is related to the lengths Lx and Lh of x[n]
and h[n] by Ly = Lx + Lh - 1

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 17

THE SUM-BY-COLUMN METHOD

 This method is based on the idea that the convolution y[n] equals the sum
of the (shifted) impulse responses due to each of the impulses that make
up the input x[n].
 Line up the sequence x[n] below the sequence h[n].
 Line up with each sample of x[n], the product of the entire array h[n]
with that sample of x[n].
 Sum the columns of the (successively shifted) arrays to generate the
convolution sequence.
 Example 1: A filter has an impulse response given by h[n] = {1,2, 2, 3}.
Find its response y[n] to the input x[n] = {2,-1, 3}. Assume that both x[n]
and h[n] start at n = 0.

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 18

 CONT‘D

 y[n] = {2,3, 5, 10, 3, 9} where y[n] starts at n=0

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 19

 Example 2 :Let h[n] = {2, 5, 0, 4} and x[n] = {4,1,3}

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 20

THE FOLD, SHIFT, MULTIPLY, AND SUM CONCEPT

 Find the convolution of h[n] = {2, 5, 0, 4} and x[n] = {4, 1, 3} where

both h[n] and x [n] starts n=0.
 Answer

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 21

CONVOLUTION OF FINITE-LENGTH SIGNALS
CORRESPONDS TO POLYNOMIAL MULTIPLICATION

 The discrete convolution of two finite-length

sequences x[n] and h[n] is entirely equivalent to
multiplication of two polynomials whose
coefficients are described by the arrays x[n] and
h[n] (in ascending or descending order).
 The convolution sequence corresponds to the
coefficients of the product polynomial.
 if we insert N zeros between each pair of adjacent
samples of each sequence to be convolved, their
convolution corresponds to the original
convolution sequence with N zeros inserted
between each pair of its adjacent samples
Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 22
 CONT‘D
 Example : { 1, 1,3 }*{1, 0, 2}={1, 1, 5, 2, 6}
 y[n] = (x^2 + x + 3)(x^2 + 2) = x^4 + x^3 + 5x^2 + 2x + 6

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 23

 CONVOLUTION USING TABULATION
METHOD
 Convolve x[n] and h[n] where x[n] ={3,2,1,2} stars at n=0 and
h[n]={ ={1,2,1,2} starts at n =-1
 Steps

 Write down the sequence x[n] and h[n] as shown in the figure below
 Multiply each and every sample in h[n]with the samples of x[n] and bulate the
values
 Divide the elements in the table by drawing diagonal lines
 Starting from the left sum all elements in each strip and write down in the same order

x[n] /h[n] 3 2 1 2
1 a b c d
2 b c d e
1 c d e f
2 d e f g
Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 24
 CONT‘D
 Answer

3 2 1 2
1 3 2 1 2
2 6 4 2 4
1 3 2 1 2
2 6 4 2 4

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 25

 CORRELATION
 Correlation is nothing but the degree of similarity a signal is with respect
to another signal.
 If the signals are identical, then the correlation coefficient is 1
 if it is totally different it is then 0.
 When two independent signals are compared, the procedure is known as
cross correlation
It is a measure of similarity of two waveforms as a function of a time-lag applied to
one of them.
 when the signal is compared to itself, it is then known as autocorrelation
 Autocorrelation is the cross-correlation of a signal with itself. It is a time
domain analysis useful for determining the periodicity or repeating patterns of a
signal.

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 26

 CONT‘D

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 27

 CONT‘D
 As shown above, a pulse x(t) is transmitted from the transmitter of the
radar and there reflected signal from the target is received back by the
radar receiver.
 The received signal r(t) which is now delayed, attenuated, and added with
noise is a form of the original transmitted signal, x(t),
 It is then analyzed to estimate the time delay between the transmitted and
received signal.

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 28

CROSS-CORRELATION IS DENOTED
 rxh[n] = x[n] **h[n]
 Correlation is the convolution of one Signal with a folded version of the
other
 rxh[n] = x[n] **h[n] = x[n] * h[-n]
 Correlation length: Nx + Nh - 1

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 29

EXAMPLE DISCRETE CROSS-CORRELATION
 Let x[n] = {2,5, 0, 4} starts at n=-1and h[n] = {3, 1, 4} starts at n=0
 Solution

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 30

 CONT‘D

Mekelle Institute of Technology-Mekelle University (MIT 10/29/2024 31