Signal Spectra,

Signal Processing
Engr. Ernesto Vergara
[email protected]
 Signals
• A signal is a physical quantity, or quality, which
conveys information.

• A (scalar-valued) signal is a physical quantity of

interest that is modeled as a function of (or
sequence on) one or more independent
variables (or indices).

• A signal is a single valued function of time that

plays the role of the dependent variable; at
every instant time, the function has a unique
value.
 Signals
• A signal is any physical quantity that varies with
time, space or any other independent variables.

• A signal can represent by mathematical

equations, it is a function of one or more
independent variables.

• A signal is represented by a function of one or

more independent variables
 Signals
• Examples of signals are: temperature,
pressure, voltage, brightness, speech,
audio, video, electrocardiogram (ECG).

• Example: voice of my friend is a signal

which causes me to perform certain
actions or react in a particular way
My friend's voice is called an excitation
My action or reaction is called a response
 Systems
• System is an assemblage of entities/objects, real or abstract,
comprising a whole with each every component/element interacting
or related to another one.

Systems process input signals to produce output signals

• Examples
- Motion, sound, picture, video, traffic light…
- Natural system (ecosystem), human-made system (machines,
computer storage system), abstract system (traffic, computer
programs), descriptive system (plans)
 Systems
• A physical device that performs an operation on a signal. It may be
classified as linear or non-linear system.

• These operations are what we call SIGNAL PROCESSING.

Software Hardware
Algorithms
- set of rules for implementing the
system by a program that performs
the operations.
 What is signal processing?
• The conversion from excitation to response is called
signal processing

• A typical reason for signal processing is to eliminate or

reduce an undesirable signal

• We convert the original signal into a form that is suitable

for further processing

• One fundamental representation of a signal is as a

function of at least one independent variable
 Types of signal
processing

I. Analog Signal Processing

II. Digital Signal Processing
 BASIC ELEMENTS OF SIGNAL
PROCESSING
Analog Signal Processing

• Most of the signals are analog in nature.

• Analog signals, the signals are functions of a continuous variable,
such as time or space.
• Analog signal processors can be filters or frequency analyzers, or
frequency multipliers in order to change their characteristics or
extracting some desired information.
 BLOCK DIAGRAM OF A DIGITAL
SIGNAL PROCESSING SYSTEM

• DSP – provides an alternative method for processing the analog signal.

Processing digital signal by performing some type of algorithm or
numerical computation.
• A/D Converter – interface between the analog signal and the digital
processor. This is very important in order to perform the processing
digitally. Its output is a digital signal that is appropriate as an input to the
digital processor.
• Digital signal Processor – this may be a large programmable digital
computer or a small microprocessor programmed to perform the desired
operations on the input signal. It may also be a hardwired digital processor
configured to perform a specified set of operations on the input signal.
 – Hardwired machine – difficult to configure
– Programmable machines – flexible, you can change the signal processing
operations through a change in the software.
• Advantage of Hardwired over Programmable Machines (this is when signal
processing operations are well defined)
– Can be optimized
– Cheaper signal processor
– Runs faster than programmable counter part.

• D/A Converter – interface from the digital domain to the analog domain.
Sometimes, this is not essential anymore.

Example: Digital processing of Radar signals, the information extracted

from the radar signal, such as the position of the aircraft and its speed,
may simply be printed on paper.
 Advantage of Digital over Analog
Signal Processing
1. Digital programmable system allows flexibility in reconfiguring the
digital signal processing operations simply by changing the
program.
• Reconfiguration of an analog system means:
• redesign of the hardware
• testing
• verification to see that it operates properly

2. Digital system provides much better control of accuracy

requirements. Accuracy, very hard for the designer to control in
analog system because of the tolerance in analog circuit
components.
 Advantage of Digital over Analog
Signal Processing
3. Digital signals are easily stored on magnetic media (tape or disk)
without deterioration or loss of signal fidelity. Here, signals
become transportable and can be processed off-line in a remote
laboratory.

• DSP allows the implementation of more sophisticated signal

processing algorithms.

5. Digital implementation of the signal processing system is cheaper

than its analog counterpart, because digital hardware is cheaper.

6. Very easy to perform precise mathematical operations on signals.

Limitation: speed of operation of A/D converters and DSP processors.

 Applications of DSP
a. Speech processing
b. Signal transmission
c. Image processing and transmission
d. Seismology and Geophysics
e. Oil exploration
f. Detection of Nuclear explosions
g. Processing of signals received from
outer space.
TYPES OF
SIGNALS
 Types of Signal
1. Analog/Continous – signals that have
values at any point of time.

2. Discrete Time Signals – signals that have

values at intervals of time.

3. Discrete Signals – if you force the

amplitudes of DT signals to be zeros and
1’s only.
 Analog vs Digital
• Analog signal is a signal that has an
infinite number of levels

• A digital signal is a signal with a fixed

number of possible levels
 Signal Representation in
MATLAB
3
t = 0:30; analog signal
x = 0.2+2*sin(0.245*t+0.15); digital (quantized)
2
d = 0.5;
xq = d*round(x/d);
plot(t,x)

x(t), xq(kT)
1
hold on
stem(t,xq,'r')
hold off 0
ylabel('x(t), x_q(kT)')
xlabel('t') -1
legend('analog signal',...
'digital (quantized)')
-2
0 10 20 30
t
CLASSIFICATIONS
OF SIGNALS
 Multichannel and/or
Multidimensional Signals
• Multichannel Signals - Signals generated by
multiple sources or multiple sensors which can be
represented in vector form (i.e. electrocardiogram,
ECG, which may be 3-lead or 12-lead device.).

• Multidimensional Signals - A signal which

represents several parameters or dimensions as a
result of several independent variables. An M-
dimensional signal is a result of a function of M
independent variables (i.e. the color TV picture is a
three-channel, three-dimensional signal).
ECG
ECG
Picture Tube
Picture Tube
Component Nature of Color

G B
Video Color Palette
 Color Television
R = Red
G = Green
B = Blue
B + G = Cyan
G + R = Yellow
B + R = Magenta
NTSC Color Bars
Continuous-Time and/or Discrete-
Time Signals
• A signal is continuous if the independent
variable is continuous also known as
continuous time signal. Example of
continuous-time signals are: telephone or radio
signal as a function of time, atmospheric
pressure as a function of altitude

• Signals which are defined for every value of time

and may take on values in the continuous
interval.
x1(t) = cos  t
x2(t) = e- t, -∞ < t < ∞
Continuous-Time and/or Discrete-
Time Signals
• A signal is said to be discrete-time signal if the
independent variable takes on only discrete values ,
where TS is a fixed positive real numbers and k ranges
over the set of integers, . Example of discrete-time
signals are amount of loan payment in the nth month, the
weekly Philippine Stock Exchange index, and the output
of an information source that produces one of the digits
1,2,…,M every TS seconds.

• Signals which are defined only at a certain specific value

of time (either equidistant or not) and is denoted by x(n)
instead of x(t). Discrete time signals are generated in
two ways: (1) by selecting values of an analog signal at
discrete-time instants (sampling), and (2) by
accumulating a variable over a period of time.
 CONTINUOUS VS DISCRETE-TIME SIGNALS
Continuous-time signal Discrete-time signal
 Continuous-Valued and/or
Discrete-Valued Signals
• Continuous-Valued Signals - These are
signals which take on all possible values on a
finite or an infinite range.
• Discrete-Valued Signals - These are signals
which take on values from a finite set of possible
x(n)
values.
Digital Signal
... with three
different
amplitude levels
-1 0 1 2 3 4 5 6 n

• A discrete-time signal having a set of discrete

values is called a DIGITAL SIGNAL
 Deterministic and/or Random
Signals
• Deterministic signal is a signal that can be
expressed by a mathematical expression

• Deterministic signals these are any signals

that can be uniquely described by an explicit
mathematical expression, a table data, or a well-
defined rule. It is also any signal which is about
which there is no uncertainty with respect to its
value at any time, past, present or future.
 Deterministic and/or Random
Signals
• Random Signals these are signals that
had evolved in time in an unpredictable
manner which is usually generated by
statistical techniques.

• Random signal is a signal that can not be

expressed by a formula and it evolves in
time in an unpredictable manner
 FREQUENCY CONCEPTS IN
CONTINUOUS-TIME AND
DISCRETE-TIME SIGNALS
Continuous-Time Sinusoidal Signal
xa(t) = A cos (Ωt + θ), -∞ < t < ∞

where: A = amplitude
Ω = 2  F (rad/sec)
θ = phase angle (rad)
TP = 1 / F
A

A cos θ

t
 Properties Of Continuous-Time
Sinusoidal Signal
1. For every fixed value of the frequency, F, xa(t) is
periodic.
xa(t + TP) = xa(t)where: TP = 1 / F - fundamental period

2. Continuous-time sinusoidal signals with distinct

(different) frequencies are themselves distinct.

3. Increasing the frequency, F, results in an increase in the

rate of oscillation of the signal, in the sense that more
periods are included in a given interval.

If F = 0, TP = ∞, we can increase the frequency, F,

without limit with a corresponding increase in the rate of
oscillation.
 Complex Exponential Signals
• Sinusoidal signals can also be expressed as complex exponential
signals.
xa(t) = A cos (Ωt + θ), -∞ < t < ∞
= A e j(Ω t + θ)

• From Euler’s identity, we could express e ±jΦ = cos Φ ± j sin Φ. The

resulting sinusoidal function can then be expressed as a sum of
two-equal amplitude-complex-conjugate exponential functions:
xa(t) =A cos(Ωt + θ), = A e j(Ω t + θ) + A e - j(Ω t + θ)
2 2
Im

A/2 Phasor representation of cosine

Ωt + θ
Re
function by a pair of complex-
Ωt + θ
A/2
conjugate exponentials
Ω
 FREQUENCY CONCEPTS IN
CONTINUOUS-TIME AND
DISCRETE-TIME SIGNALS

Discrete-Time Sinusoidal Signal

x(n) = A cos (ωn + θ), -∞ < n < ∞

where: n = integer variable, sample number
A = amplitude of the sinusoid
ω = 2  f (rad/sample)
θ = phase angle (rad)

Thus, x(n) = A cos (2πfn + θ), -∞ < n < ∞

 FREQUENCY CONCEPTS IN
CONTINUOUS-TIME AND
DISCRETE-TIME SIGNALS

Discrete-Time Sinusoidal Signal

For example, a sinusoid with frequency ω = π/6 rad/sample
(f = 1/12 cycles/sample) and θ = π/3.
x(n) = A cos (ωn + θ)

... ...
n

-A
 Properties of Discrete-Time
Sinusoidal Signals
1. A discrete-time sinusoid is periodic only if its frequency, f, is a
rational number.
x(n + N) = x(n) for all n
where: N – fundamental period

Proof: For a sinusoid with frequency to be periodic,

cos [2 π fo (N + n) + θ] = cos (2 π n fo + θ)

This relation is true if there exists an integer k such that,

2 π fo N = 2 k π
fo = k / N
 Properties of Discrete-Time
Sinusoidal Signals
2. Discrete time-sinusoids whose frequencies are separated by an
integer multiple of 2π are identical.

Consider: cos (ωon + θ)

cos [(ωo + 2π)n + θ] = cos (ωon + θ)

As a result, all sinusoidal sequences having the form:

xk(n) = cos (ωkn + θ), k = 0, 1, 2, . . .
where ωk = ωo + 2kπ, -π ≤ ωo ≤ π
are identical.

Thus, all sinusoids having frequencies in the range -π ≤ ω ≤ π or -

½ ≤ f ≤ ½ are distinct (unique); those with frequencies | ω | > π or
| f | > ½ are regarded as aliases.
 Properties of Discrete-Time
Sinusoidal Signals
3. The highest rate of oscillation in a discrete-time sinusoid is attained
when ω = π (or ω = - π), or equivalently, f = ½ (or f = - ½).

Proof: x(n) = cos ωon, where ω varies from 0 to π.

Assume values of ωo = 0, π/8, π/4, π/2, π

f = ωo / 2π
N=1/f
 Properties of Discrete-Time
Sinusoidal Signals
For    0  2
Consider: 1  0, 2  2  0
1   to 2 ,  2   to 0
x1(n) = Acosω1n = Acosω0n
x2(n) = Acosω2n = Acos(2π - ω0)n = Acos(-ω0n) = x1(n)
Thus, ω2 is an ALIAS of ω1
- Increasing the relative frequency of a discrete-time sinusoid from
π to 2π, its rate of oscillation.
- For ω = 2π, same with ω = 0, constant.
- In introducing negative frequency,
A j (n ) A  j (n )
x (n) = A cos(ωn + θ) = e  e
2 2
 Properties of Discrete-Time
Sinusoidal Signals
• Since from property 2, the frequencies in any interval

1    1  2 
constitutes all the existing discrete-time sinusoids or complex
exponentials.

• we use fundamental range,

0    2 or      
(0 ≤ f ≤ 1 or -1/2 ≤ f ≤ 1/2)
 HARMONICALLY RELATED
COMPLEX EXPONENTIALS(or
Sinusoids)
- these are sets of periodic complex exponentials with fundamental
frequencies that are multiples of a single positive frequency.

CONTINUOUS-TIME EXPONENTIALS
• Basic Signals: s k (t )  e jk0t  e jk 2F0t , k = 0, ±1, ±2,….
– periodic for each value of k
– 1/kF0 = Tp/k fundamental period
– kF0 fundamental frequency

• if k1 = k2, then Sk1(t) ≠ Sk2(t)

 HARMONICALLY RELATED
COMPLEX EXPONENTIALS(or
Sinusoids)
• From the basic signals,
 
x a (t )   k k
C S
k  
(t )   k
C e jk 0t

k  
Fourier-series Expansion for xa(t)

where Ck, k = 0, ±1, ±2,…. are arbitrary complex constants

xa(t) is periodic with fundamental period Tp = 1/F0
Ck – complex-valued constants – Fourier series constant
Sk(t) – kth harmonic of xa(t)
 HARMONICALLY RELATED
COMPLEX EXPONENTIALS(or
Sinusoids)
DISCRETE-TIME EXPONENTIALS
• Recall property 1, we choose f0 = 1/N
jk 2nfo
S k (n)  e k = 0, ±1, ±2,….

S k  N (n)  e jk 2 ( k  N ) nfo
S k  N (n)  e jk 2 ( k  N ) n / N  e jk 2n S k (n)  S k (n)
Assume the set n0 = 0,
S k (n)  e jk 2n / N k = 0, ±1, ±2,….
N 1 N 1
x(n)   C k S k (n)  C k e j 2kn / N
k 0 k 0
• A periodic signal with fundamental period N.
• The sequence Sk(n) is called the kth harmonic of x(n).
 Sample Problem
Classify the following signals according to whether they are (1) one
or multi-dimensional (2) single or multichannel, (3) continous time or
discrete time, and (4) analog or digital(in amplitude). Give a brief
explanation.
(a) closing prices of utility stocks on the New York Stock Exchange.
(b) A color movie
(c) Position of the steering wheel of a car in motion relative to car’s
reference frame.
(d) Position of the steering wheel of a car in motion relative to
ground reference frame.
(e) Weight and height measurements of a child taken every month.
 Answer
(a) closing prices of utility stocks on the New York Stock Exchange.

(a) Stocks = One dimensional

Price = Currency = multichannel
Close Price = Sample = discrete time
Set of Discrete Price Values = digital
 Answer
(b) A color movie

Multi Color = Multi dimensional

Each Color = single channel
Moving Image Frame = continous time = analog
 Answer
(c) Position of the steering wheel of a car in motion relative to car’s
reference frame.
steering wheel = One dimensional, single channel, continous time,
analog
 Answer
(d) Position of the steering wheel of a car in motion relative to
ground reference frame.
steering wheel = One dimensional, single channel, continous time,
analog
 Answer
(e) Weight and height measurements of a child taken every month.

Measurement of weight = One dimensional

Measurement of height = One dimensional
Weight ( kilograms, lbs, etc.) = multichannel
Height ( inches, meter, etc.) = multichannel
Every month = discrete time
Set of Discrete Weight Values = digital
Set of Discrete Height Values = digital
 Sample Problem
Determine which of the following sinusoids are periodic and
compute their fundamental period.
(a) cos 0.01πn
(b) cos[π(30n/105)]
(c) cos 3πn
(d) sin 3n
(e) sin[π(62n/10)]
 Answer

(a) 2πf = 0.01π; f = 0.01π/2π = 1/200; periodic and N = 200

(b) 2πf = π(30/105); f = [π(30/105)]/2π = 1/7; periodic and N =
7
(c) 2πf = 3π; f = 3π/2π = 3/2; periodic and N = 2
(d) 2πf = 3; f = 3/2π; not periodic
(e) 2πf = π(62/10); f = [π(62/10)]/2π; f = 31/10; periodic and N
= 10
 SAMPLING THEOREM
• The sampling frequency has to be at least TWICE
the maximum frequency in the continuous signal,
that is the ABSOLUTE maximum frequency, not just
the twice the maximum frequency that you are
interested in. More formally this is:
Fs ≥ 2 Fmax
• The sampling frequency is also called the NYQUIST
FREQUENCY
• If the sampling frequency was high enough (greater
than twice Fmax) then these sidebands keep away
from each other. BUT if you didn’t have a high
enough sampling frequency then the sidebands will
overlap.
 SAMPLING THEOREM
• Example when the sampling frequency is sufficient
 SAMPLING THEOREM
• Example when the sampling frequency was too low, and the
resulted in loss of information.

Overlapping of frequency
(Frequency Aliasing)
 SAMPLING THEOREM
• If Fmax = B, Fs = 2B, we use the interpolation function,
sin 2Bt
g (t ) 
2Bt Time shifting

• Thus x a (t )   x a (n / Fs ) g (t  n / Fs )
where:
x a (n / Fs )  x a (nT )
• Consider sampling at xa(t) at Fs = 2B

 sin 2B(t  n / 2 B) 
x a (t )   x a (n / 2 B) 
n    2B(t  n / 2 B) 
• then, the Nyquist frequency is Fn = 2B = 2Fmax.
 NYQUIST SAMPLING RATE
• The sampling frequency,FS, must be at least TWICE
the higher frequency,FH, present in the signal to
prevent aliasing
FS ≥ 2 FH
 ALIAS FREQUENCY, FA
• Recovered frequency FR which is not the same as
the original frequency

FA = FS – KFH or FA = FH - kFS

where k = 1, 2, 3, 4…..; for uniformity use k = 1

 PROOF OF NYQUIST
THEOREM
• ORIGINAL DATA
 MINIMUM SAMPLING
FREQUENCY
• FS =2FH = 6 samples/ 6 seconds; 1 sec/sample

original shape was not recovered.

 Sample Problem
Consider the analog signal
xa(t) = 3cos 100πt
(a) Determine the minimum sampling rate required to
avoid aliasing.
(b) Suppose that the signal is sampled at the rate Fs =
200 Hz. What is the discrete-time signal obtained after
sampling?
(c) Suppose that the signal is sampled at the rate Fs = 75
Hz. What is the discrete-time signal obtained after
sampling?
(d) What is the frequency 0 < F < Fs/2 of a sinusoid that
yields samples identical to those obtained in part (c)?
 Answer
Consider the analog signal
xa(t) = 3cos 100πt
(a) 2πF = 100π; F = (100π)/(2π) = 50 Hz; Sampling
Frequency, Fs = 2(50Hz) = 100 Hz.
(b) Fs = 200 Hz: x(n) = 3cos(100/200)πn = 3cos(π/2)n
(c) Fs = 75 Hz: x(n) = 3cos(100/75)πn = 3cos(4π/3)n =
3cos(2π - 2π/3)n = 3cos(-2π/3)n = 3cos(2π/3)n
(d) What is the discrete-time signal obtained after
sampling? 2πf = 2π/3; f = 1/3 = F/Fs; F = (1/3)(75 Hz)
= 25 Hz; ya(t) = 3cos(50πt)
F = 50 Hz is an alias of F = 25 Hz for the sampling rate
Fs = 75 Hz.
 Input and Output Relationship

SYSTEM
INPUT OUTPUT

xa(t) Black box ya(t)

3 3

2 2

1 1

0 0

-1
FN = Fs ≥ 2Fmax = 2(50 Hz) = 100 Hz -1

-2 -2

-3 -3
0 5 10 15 20 25 30 0 5 10 15 20 25 30

xa(t) = 3cos100 πt

FN = Fs = 75 Hz
 Sample Problem
Consider the analog signal

xa(t) = 3cos50πt + 10sin300πt – cos100πt

What is the Nyquist rate of the signal?

 Answer
F1 = 25 Hz F2 = 150 Hz F3 = 50 Hz

Sampling Theorem:
Fs > 2Fmax
Nyquist Frequency:
FN = 2Fmax = 2(150 Hz) = 300 Hz
 Sample Problem
Consider the analog signal
xa(t) = 3cos2000πt + 5sin6000πt + 10 cos12000πt
(a) What is the Nyquist rate for this signal?
(b) Assume now that we sample this signal using
a sampling rate Fs = 5000 samples/s. What is
the discrete-time signal obtained after
sampling?
(c) What is the analog signal ya(t) we can
reconstruct from samples if we use ideal
interpolation?
 Answer
(a) What is the Nyquist rate for this signal?
F1 = 1000 Hz F2 = 3000 Hz F3 = 6000 Hz
Sampling Theorem:
Fs > 2Fmax ; Fs > 2(6000 Hz) = 12000 Hz
Nyquist Frequency:
FN = 2Fmax = 2(6000 Hz) = 12000 Hz
 Answer
(b) Fs = 5000 samples/s:
x(n) = 3cos(2000/5000)πn + 5sin(6000/5000)πn +
10cos(12000/5000)πn
x(n) = 3cos[2π(1/5)]n + 5sin[2π(3/5)]n +
10cos[2π(6/5)]n
x(n) = 3cos[2π(1/5)]n + 5sin[2π(1 - 2/5)]n +
10cos[2π(1 + 1/5)]n
x(n) = 3cos[2π(1/5)]n + 5sin[2π(- 2/5)]n +
10cos[2π(1/5)]n
x(n) = 3cos[2π(1/5)]n - 5sin[2π(2/5)]n +
10cos[2π(1/5)]n = 13cos[2π(1/5)]n - 5sin[2π(2/5)]n
 Answer
(c) Fs = 5000 samples/s:
x(n) = 13cos[2π(1/5)]n - 5sin[2π(2/5)]n

ya(t) = 13cos[2π(1/5)(5000)]t - 5sin[2π(2/5)(5000)]t

ya(t) = 13cos[2π(1000)]t - 5sin[2π(2000)]t

ya(t) = 13cos(2000)πt - 5sin(4000)πt

ANALOG-TO-DIGITAL
CONVERSION
 ANALOG-TO-DIGITAL
CONVERSION
Three-Step Process:

1. SAMPLING – conversion of a continuous-time

signal into a discrete-time signal obtained by taking
“samples” of the continuous-time signal at discrete-
time instants.
2. QUANTIZATION – conversion of a discrete-time
continuous valued signal into a discrete-time,
discrete-valued (digitalsignal).
3. CODING – each discrete value xq(n) is
represented by a b-bit binary sequence.
 Sampling of Continuous-Time
Signal(Periodic/uniform
sampling)
x(n) = xa(nT) -∞ < n < ∞

x(n), the discrete-time signal is obtained by “taking samples”

of the continuous-time signal xa(t) every T seconds.

Relationship of time variables, t and n:

T = nT =n/Fs

Relationship of frequency variables, F and f:

f = F/Fs or ω = ΩT
 Sampling of Continuous-Time
Signal(Periodic/uniform
sampling)
Range of frequency of continuous-time when sampled at Fs:
-1/2T = -Fs/2 ≤ F ≤ Fs/2 = 1/2T
-πFs = -π/T ≤ Ω ≤ π/T = πFs
Highest frequency values of continuous-time signal when
sampled at Fs :
Fmax = Fs/2 = 1/2T
Ωmax = πFs = π/T
Consider the sinusoid : xa(t) = A cos(2πFkt + θ) to be sample
at a rate of Fs
Where: Fk = Fo +kFs
k = ±1, ±2,…
will result to x(n)  xa(nT) = A cos(2πfon + θ)
 Quantization of Continuous-
Amplitude Signals

Quantization – the process of converting a discrete-

time continuous-amplitude signal into a digital
signal by expressing each sample value as a finite
(instead of infinite) number of digits.

Quantization error or Quantization noise – The

error introduced in representing the continuous-
valued signal by a finite set of discrete value levels.
Quantization of Continuous-
Amplitude Signals
Quantizer operation on the samples x(n) as Q[x(n)]
and let xq(n) denote the sequence of quantized
samples at the output of the quantizer.
Hence
xq(n) = Q[x(n)]

Then the quantization error is a sequence eq(n)

defined as the difference between the quantized
value and the actual sample value.
Thus
eq(n) = xq(n) – x(n)
 Quantization of Continuous-
Amplitude Signals
• In quantization, to eliminate the excess digits, it ca be
discard by using truncation or rounding the resulting
number.

• The values allowed in the digital signal are called the

quantization levels, whereas the distance Δ between two
successive quantization levels is called the quantization
step size or resolution.

• The rounding quantizer assigns each sample of x(n) to the

nearest quantization level. In contrast, a quantizer that
performs truncation would have assign each sample of
x(n) to the quantization level below it.
 Quantization of Continuous-
Amplitude Signals
• Uniform quantizer - quantization levels are
uniformly distributed between the minimum and
maximum ranges of the input sequence.

• The peak-to-peak range of the input sequence x[k]

is divided uniformly into (L + 1) quantization levels
{d0, d1, . . . , dL} such that the separation Δ = (dm+1
– dm) is the same between any two consecutive
levels. The separation Δ between two quantization
levels is referred to as the quantile interval or
quantization step size.
 Quantization of Continuous-
Amplitude Signals
• For a given input, the output of the quantizer is
calculated from the following relationship:
1
y[k ]  rm  [d m  d m1 ]
2
for d m  x[k ]  d m 1 and 0  m  L
• The quantized value of the input lying within the
levels dm and dm+1 is given by rm, which equals
0.5(dm + dm+1). The quantization levels {d0, d1, . . . ,
dL} are referred to as the decision levels, while the
output levels {r0, r1, . . . , rL−1} are referred to as the
reconstruction levels.
 Quantization of Continuous-
Amplitude Signals

(a) Uniform quantizer (b) non-uniform quantizer

 Quantization of Continuous-
Amplitude Signals
• The quantization error eq(n) in rounding is limited
to the range of –Δ/2 to Δ/2, that is
 
  e q ( n) 
2 2
• If xmin and xmax represent the minimum and
maximum value of x(n) and L is the number of
quantization levels then xmax  xmin

L 1
where xmax – xmin is the dynamic range of the signal
 Sample Problem
A digital communication link carries binary-coded words
representing samples of an input signal
xa(t) = 3cos600 πt + 2cos1800 πt
The link is operated at 10,000 bits/s and each input
sample is quantized into 1024 different voltage levels.
(a) What is the sampling frequency and the folding
frequency?
(b) What is the Nyquist rate for the signal xa(t)?
(c) What are the frequencies in the resulting discrete-time
signal x(n)?
(d) What is the resolution Δ?
 Answer
(a) Number of bits/sample = log21024 = 10
Fs = (10000 bits/sec)/(10 bits/sample) = 1000 samples/sec
Ffold = Fs/2 = 500 Hz

(b) Fmax = 1800π/2π = 900 Hz

FN = 2Fmax = 1800 Hz
 Answer
(c) x(n) = 3cos[2π(300/1000)n] + 2cos[2π(900/1000)n]
x(n) = 3cos[2π(3/10)n] + 2cos[2π(9/10)n]
x(n) = 3cos[2π(3/10n] + 2cos[2π(1 - 1/10)n]
x(n) = 3cos[2π(3/10n] + 2cos[2π(-1/10)n]
x(n) = 3cos[2π(3/10)n] + 2cos[2π(1/10)n]

(d) resolution, Δ = (xmax -xmin)/(m - 1)

resolution, Δ = (5 - [-5])/(1024 - 1) = 10/1023
 Quantization of Continuous-
Amplitude Signals
• The quality of the output of the A/D converter is usually
measured by the signal-to-quantization noise ratio (SQNR),
which provides the ratio of the signal power to the noise
power:
Px 3 2b
SQNR   2
Pq 2
• Expressed in decibels(dB), the SQNR is
SQNR(dB)  10 log10 SQNR  1.76  6.02b
• This implies that the SQNR increases approximately 6dB
for every bit added to the word length, that is, for each
doubling of the quantization levels.
 DIGITAL-TO-ANALOG
CONVERSION
-uses interpolation
- linear interpolation
-quadratic interpolation
-most common type is the zero-order
hold
-also known as the sample-hold
-simply holds constant the value of one
sample unit the next one is taken.
 DSP Application areas
DSP is one of the fastest growing fields in modern electronics being used in any
area where information is handled in a digital form or controlled by a digital
processor. Application areas include the following:

Image processing Instrument/control Speech/audio

Pattern recognition Spectrum analysis Speech recognition
Robotic vision Position and rate control Speech synthesis
Image enhancement Noise reduction Text to speech
Facsimile Data compression Digital audio
Satellite weather map Equalization
Animation

Military Telecommunications Biomedical

Secure communication Echo cancellation Patient monitoring
Radar processing Adaptive equalization Scanners
Sonar processing ADPCM transcoders EEG brain mappers
Missile guidance Spread spectrum ECG analysis
Video conferencing X-ray storage/
Data communication enhancement
 Digital signal processing and its
benefits
Speech, which we encounter from example in telephony, radio and every day life.
Biomedical signals, such as the electroencephalogram(brain signals),
Sound and music, such as reproduce by the compact disc,player,video and
image,which most people watch on the television, and radar signals which are
used to determine the range and bearing distant targets.
The specific reason for processing a digital signal may be, for example, to remove
interference or noise from the signal, to obtain the spectrum of the data, or to
transform the signal into a more suitable form. DSP is now used in many areas
where analogue methods were previously used and in entirely new application
which were difficult or impossible with analogue methods. The attraction of DSP
comes from key advantage such as the following.
• Guaranteed accuracy. Accuracy is only determined by the number of bits used.
• Perfect reproducibility. Identical performed from unit to unit is obtained since
there are no variations due to component tolerances. For example using DSP
technique, a digital recording can be copied or reproduced several times over
without any degradation in the signal; quality.
• No drift in performance with temperature or age.
 Digital signal processing and its
benefits
Advantages is always taken of the tremendous advances in semiconductor
technology to achieve greater reliability, smaller size, lower cost, low power
consumption, and higher speed. For example, in the last two or three years it
has become possible to produce high speed low power ICs(integrated circuits)
using SCOMS technology. As a result newer DSP chips are predominantly
CMOS devices instead of bipolar, while NMOS devices are being upgraded to
CMOS.
• Greater flexibility. DSP system can be programmed and reprogramme to
perform a variety of functions, without modifying the hardware. This is perhaps
one of the most important features of DSP.
• Superior performance. DSP can be used to perform functions not possible with
analogue signal processing. For example, linear phase response can be
achieved, and complex adaptive filtering algorithms can be implemented using
DSP techniques.
In some cases information may already be in digital form and DSP offers the only
viable option.
 Digital signal processing and its
benefits
DSP is not without disadvantages. However, the significance of these
disadvantages is being continually diminished by new technology.

• Speed and cost. DSP designs can be expensive especially when large
bandwidth signals are involved. At the present, fast ADC’s/DAC’s(analogue-to-
digital converters/digital-to-analogue converters) either are too expensive or to
not have sufficient resolution for wide bandwidth DSP applications.

• Design time. Unless you are knowledgeable in DSP techniques and have the
necessary resources(software packages and so on), DSP designs can be time
consuming and some in cases almost impossible. The acute shortage of
suitable engineers in this area is widely recognized.

• Finite wordlength problems. In real-time situations, economic considerations

often mean that DSP algorithms are implemented using only a limited number
of bits.
 Seatwork
1. A continuous-time sinusoid xa(t) with fundamental period
Tp = 1/Fo is sampled at a rate Fs = 1/T to produce a
discrete-time sinusoid x(n) = xa(nT).
(a) Show that x(n) is periodic if T/Tp = k/N (i.e., T/Tp is a
rational number).
(b) If x(n) is periodic, what is its fundamental period Tp in
seconds?
(c) Explain the statement: x(n) is periodic if its fundamental
period Tp in seconds is equal to an integer number of
periods of xa(t).
 Seatwork
2. An analog signal contains frequencies up to 10
kHz.
(a) What range of sampling frequencies allows
exact reconstruction of this signal from its
samples?
(b) Suppose that we sample this signal with a
sampling frequency Fs = 8 kHz. Examine what
happens to the frequency F1 = 5kHz.
(c) Repeat part (b) for a frequency F2 = 9 kHz.
 Seatwork
3. An analog electrocardiogram (ECG)
signal contains useful frequencies up to
100 Hz.
(a) What is the Nyquist rate for this signal?
(b) Suppose that we sample this signal at a
rate of 250 samples/s. What is the
highest frequency that can be
represented uniquely at this sampling
rate?
 Seatwork
4. An analog signal
xa(t) = sin(480πt) + 3sin(720πt)
is sampled 600 times per second.
(a) Determine the Nyquist sampling rate for xa(t).
(b) Determine the folding frequency.
(c) What are the frequencies, in radians, in the
resulting discrete time signal x(n)?
(d) If x(n) is passed through an ideal D/A converter,
what is the reconstructed signal ya(t)?