dsp mitra chap1
TRANSCRIPT
-
8/10/2019 DSP Mitra Chap1
1/125
Copyright 2002 S. K. Mitra1
Signals and Signal ProcessingSignals and Signal Processing
Signals play an important role in our dailylife
A signal is a function of independent
variables such astime,distance,position,
temperature,andpressure
Some examples of typical signals are shownnext
-
8/10/2019 DSP Mitra Chap1
2/125
Copyright 2002 S. K. Mitra2
Examples of Typical SignalsExamples of Typical Signals Speech and music signals- Represent air
pressureas a function of timeat a point in
space
Waveform of the speech signal I like
digital signal processing is shown below
0 1 2 3-1
-0.5
0
0.5
1
Time, sec.
Amplitude
-
8/10/2019 DSP Mitra Chap1
3/125
Copyright 2002 S. K. Mitra3
Examples of Typical SignalsExamples of Typical Signals
Electrocardiography (ECG) Signal-Represents the electrical activity of the
heart
A typical ECG signal is shown below
-
8/10/2019 DSP Mitra Chap1
4/125Copyright 2002 S. K. Mitra4
Examples of Typical SignalsExamples of Typical Signals The ECG trace is a periodic waveform
One period of the waveform shown belowrepresents one cycle of the blood transfer
process from the heart to the arteries
P
R
Q
S
T
Millivolts
Seconds
-
8/10/2019 DSP Mitra Chap1
5/125Copyright 2002 S. K. Mitra5
Examples of Typical SignalsExamples of Typical Signals Electroencephalogram (EEG) Signals-
Represent the electrical activity caused bythe random firings of billions of neurons in
the brain
-
8/10/2019 DSP Mitra Chap1
6/125Copyright 2002 S. K. Mitra6
Examples of Typical SignalsExamples of Typical Signals
Seismic Signals- Caused by the movementof rocks resulting from an earthquake, a
volcanic eruption, or an underground
explosion
The ground movement generates 3types of
elastic waves that propagate through thebody of the earth in all directions from the
source of movement
-
8/10/2019 DSP Mitra Chap1
7/125Copyright 2002 S. K. Mitra7
Examples of Typical SignalsExamples of Typical Signals
Typical seismograph record
-
8/10/2019 DSP Mitra Chap1
8/125
Copyright 2002 S. K. Mitra8
Examples of Typical SignalsExamples of Typical Signals Black-and-white picture- Represents lightintensityas a function of two spatial
coordinates
I(x,y)
-
8/10/2019 DSP Mitra Chap1
9/125
Copyright 2002 S. K. Mitra9
Examples of Typical SignalsExamples of Typical Signals
Video signals - Consists of a sequence of
images, called frames, and is a function of 3variables:2spatial coordinatesand time
1Frame 3Frame 5Frame
Video
videotheonClick
-
8/10/2019 DSP Mitra Chap1
10/125
Copyright 2002 S. K. Mitra10
Signals and Signal ProcessingSignals and Signal Processing
Most signals we encounter are generatednaturally
However, a signal can also be generated
synthetically or by a computer
-
8/10/2019 DSP Mitra Chap1
11/125
Copyright 2002 S. K. Mitra11
Signals and Signal ProcessingSignals and Signal Processing A signal carries information
Objective of signal processing: Extract theuseful informationcarried by the signal
Method information extraction: Depends onthe type of signal and the nature of the
information being carried by the signal
This course is concerned with the discrete-
time representation of signals and their
discrete-time processing
-
8/10/2019 DSP Mitra Chap1
12/125
Copyright 2002 S. K. Mitra12
Characterization andCharacterization and
Classification of SignalsClassification of Signals
Types of signal: Depends on the nature ofthe independent variables and the value of
the function defining the signal
For example, the independent variables can
be continuous or discrete
Likewise, the signal can be a continuous ordiscrete function of the independent
variables
-
8/10/2019 DSP Mitra Chap1
13/125
Copyright 2002 S. K. Mitra13
Characterization andCharacterization and
Classification of SignalsClassification of Signals
Moreover, the signal can be either a real-valuedfunction or a complex-valued
function
A signal generated by a single sourceis
called a scalar signal
A signal generated by multiple sourcesiscalled a vector signalor a multichannel
signal
-
8/10/2019 DSP Mitra Chap1
14/125
Copyright 2002 S. K. Mitra14
Characterization andCharacterization and
Classification of SignalsClassification of Signals
A one-dimensional(1-D) signalis afunction of a single independent variable
A multidimensional(M-D) signalis a
function of more than one independent
variables
The speech signalis an example of a 1-Dsignal where the independent variable is
time
-
8/10/2019 DSP Mitra Chap1
15/125
Copyright 2002 S. K. Mitra15
Characterization andCharacterization and
Classification of SignalsClassification of Signals
An image signal, such as aphotograph, isan example of a 2-D signalwhere the 2
independent variables are the 2spatial
variables
A color image signal is composed of three
2-Dsignals representing the threeprimarycolors:red, greenandblue(RGB)
-
8/10/2019 DSP Mitra Chap1
16/125
Copyright 2002 S. K. Mitra16
Characterization andCharacterization and
Classification of SignalsClassification of Signals
The 3color components of a color imageare shown below
-
8/10/2019 DSP Mitra Chap1
17/125
Copyright 2002 S. K. Mitra17
Characterization andCharacterization and
Classification of SignalsClassification of Signals
The full color image obtained by displayingthe previous 3color components is shown
below
-
8/10/2019 DSP Mitra Chap1
18/125
Copyright 2002 S. K. Mitra18
Characterization andCharacterization and
Classification of SignalsClassification of Signals Each frameof a black-and-white digital
video signalis a 2-D image signalthat is afunction of 2discrete spatial variables, with
each frameoccurring at discrete instants oftime
Hence, black-and-white digital video signal
can be considered as an example of a 3-D
signalwhere the 3independent variables are
the 2spatial variables andtime
-
8/10/2019 DSP Mitra Chap1
19/125
Copyright 2002 S. K. Mitra19
Characterization andCharacterization and
Classification of SignalsClassification of Signals A color video signal is a 3-channel signal
composed of three3-D signalsrepresentingthe three primary colors:red,greenandblue
(RGB) For transmission purposes, the RGB
television signalis transformed into another
type of 3-channel signalcomposed of a
luminancecomponent and 2chrominance
components
-
8/10/2019 DSP Mitra Chap1
20/125
Copyright 2002 S. K. Mitra20
Characterization andCharacterization and
Classification of SignalsClassification of Signals
For a 1-D signal, the independent variable isusually labeled as time
If the independent variable is continuous,
the signal is called a continuous-time signal
If the independent variable is discrete, the
signal is called a discrete-time signal
-
8/10/2019 DSP Mitra Chap1
21/125
Copyright 2002 S. K. Mitra21
Characterization andCharacterization and
Classification of SignalsClassification of Signals A continuous-time signalis defined at every
instant of time
A discrete-time signalis defined at discrete
instants of time, and hence, it is a sequenceof numbers
A continuous-time signal with a continuousamplitude is usually called an analog signal
A speech signal is an example of an analog
signal
-
8/10/2019 DSP Mitra Chap1
22/125
Copyright 2002 S. K. Mitra22
Characterization andCharacterization and
Classification of SignalsClassification of Signals A discrete-time signal with discrete-valued
amplitudes represented by a finite numberof digits is referred to as the digital signal
An example of a digital signal is thedigitized music signal stored in a CD-ROMdisk
A discrete-time signal with continuous-valued amplitudes is called a sampled-data
signal
-
8/10/2019 DSP Mitra Chap1
23/125
Copyright 2002 S. K. Mitra23
Characterization andCharacterization and
Classification of SignalsClassification of Signals
A digital signal is thus a quantized sampled-data signal
A continuous-time signal with discrete-
value amplitudes is usually called a
quantized boxcar signal
The figure in the next slide illustrates the 4types of signals
-
8/10/2019 DSP Mitra Chap1
24/125
Copyright 2002 S. K. Mitra24
Characterization andCharacterization and
Classification of SignalsClassification of Signals
signaltime-continuousA
signaldata-sampledAsignalboxcarquantizedA
tTime,
Amplitude
tTime,
Amplitude
tTime,
Amplitude
tTime,
Amplitude
signaldigitalA
-
8/10/2019 DSP Mitra Chap1
25/125
-
8/10/2019 DSP Mitra Chap1
26/125
Copyright 2002 S. K. Mitra26
Characterization andCharacterization and
Classification of SignalsClassification of Signals
For a discrete-time 1-Dsignal, the discreteindependent variable is usually denoted by
n
For example, {v[n]}represents a discrete-
time 1-D signal
Each member, v[n], of a discrete-time signalis called a sample
-
8/10/2019 DSP Mitra Chap1
27/125
Copyright 2002 S. K. Mitra27
Characterization andCharacterization and
Classification of SignalsClassification of Signals
In many applications, a discrete-time signalis generated by samplinga parent
continuous-time signal at uniform intervals
of time
If the discrete instants of time at which a
discrete-time signal is defined are uniformlyspaced, the independent discrete variable n
can be normalized to assume integer values
-
8/10/2019 DSP Mitra Chap1
28/125
Copyright 2002 S. K. Mitra28
Characterization andCharacterization and
Classification of SignalsClassification of Signals
In the case of a continuous-time 2-Dsignal,the 2independent variables are the spatial
coordinates, usually denoted byxandy
For example, the intensityof a black-and-
white image at location (x,y) can be
expressed asu(x,y)
-
8/10/2019 DSP Mitra Chap1
29/125
Copyright 2002 S. K. Mitra29
Characterization andCharacterization and
Classification of SignalsClassification of Signals
On the other hand, a digitized image is a 2-Ddiscrete-time signal, and its 2independent
variables are discretized spatial variables,
often denoted by mandn
Thus, a digitized image can be represented as
v[m,n] A black-and-white video signal is a 3-D
signal and can be represented asu(x,y,t)
-
8/10/2019 DSP Mitra Chap1
30/125
Copyright 2002 S. K. Mitra30
Characterization andCharacterization and
Classification of SignalsClassification of Signals
A color video signalis a vector signalcomposed of 3signals representing the 3
primary colors:red,green,andblue
!!
!
"
#
$$
$
%
&
'
),,(),,(
),,(
),,(tyxbtyxg
txr
tyxu
-
8/10/2019 DSP Mitra Chap1
31/125
Copyright 2002 S. K. Mitra31
Characterization andCharacterization and
Classification of SignalsClassification of Signals
A signal that can be uniquely determined bya well-defined process, such as a
mathematical expression or rule, or table
look-up, is called adeterministic signal
A signal that is generated in a random
fashion and cannot be predicted ahead oftime is called arandom signal
-
8/10/2019 DSP Mitra Chap1
32/125
Copyright 2002 S. K. Mitra32
Typical Signal ProcessingTypical Signal Processing
ApplicationsApplications
Most signal processing operations in thecase of analog signals are carried out in the
time-domain
In the case of discrete-time signals, both
time-domainor frequency-domain
operations are usually employed
-
8/10/2019 DSP Mitra Chap1
33/125
Copyright 2002 S. K. Mitra33
Elementary Time-DomainElementary Time-Domain
OperationsOperations Three most basic time-domain signal
operations arescaling,delay, andaddition
Scalingis simply the multiplicationof a
signal either by a positive or negativeconstant
In the case of analog signals, the operationis usually called amplification if themagnitude of the multiplying constant,
called gain, is greater than1
-
8/10/2019 DSP Mitra Chap1
34/125
Copyright 2002 S. K. Mitra34
Elementary Time-DomainElementary Time-Domain
OperationsOperations
If the magnitude of the multiplying constant
is less than 1, the operation is called
attenuation
Ifx(t)is an analog signal that is scaled by a
constant (, then the scaling operation
generates a signaly(t) = (x(t) Two other elementary operations are
integrationanddifferentiation
-
8/10/2019 DSP Mitra Chap1
35/125
Copyright 2002 S. K. Mitra35
Elementary Time-DomainElementary Time-Domain
OperationsOperations
The integrationof an analog signalx(t)
generates a signal
The differentiationof an analog signalx(t)
generates a signal
) **' +,
t
dxty )()(
dt
tdxtw
)()( '
-
8/10/2019 DSP Mitra Chap1
36/125
Copyright 2002 S. K. Mitra36
Elementary Time-DomainElementary Time-Domain
OperationsOperations The delayoperation generates a signal that is
a delayed replicaof the original signal
For an analog signalx(t),
is the signal obtained by delayingx(t)by the
amount of time which is assumed to be apositivenumber
If is negative, then it is an advance
operation
)()( ottxty ,'
ot
ot
-
8/10/2019 DSP Mitra Chap1
37/125
Copyright 2002 S. K. Mitra37
Elementary Time-DomainElementary Time-Domain
OperationsOperations
Many applications require operations
involving two or more signals to generate a
new signal
For example,
is the signal generated by the additionof thethree analog signals, , , and
)()()()( 321 txtxtxty --'
)(3tx)(1tx )(2tx
-
8/10/2019 DSP Mitra Chap1
38/125
Copyright 2002 S. K. Mitra38
Elementary Time-DomainElementary Time-Domain
OperationsOperations
Theproductof 2signals, and ,
generates a signal
The elementary operations discussed so far
are also carried out on discrete-time signals
More complex operations operations areimplemented by combining two or more
elementary operations
)(1tx )(
2tx
)()()( 21 txtxty .'
-
8/10/2019 DSP Mitra Chap1
39/125
-
8/10/2019 DSP Mitra Chap1
40/125
Copyright 2002 S. K. Mitra40
FilteringFiltering
The range of frequencies that is allowed to
passthrough the filter is called the
passband, and the range of frequencies that
isblockedby the filter is called thestopband
In most cases, the filtering operation foranalog signals is linear
-
8/10/2019 DSP Mitra Chap1
41/125
Copyright 2002 S. K. Mitra41
FilteringFiltering
The filtering operation of a linear analog
filter is described by the convolution
integral
wherex(t)is the input signal,y(t)is theoutputof the filter, and h(t)is the impulse
responseof the filter
) ***,' ++,
dxthty )()()(
Filt i
-
8/10/2019 DSP Mitra Chap1
42/125
Copyright 2002 S. K. Mitra42
FilteringFiltering
A lowpass filterpasses all low-frequency
components below a certain specified
frequency , called the cutoff frequency,and blocks all high-frequency components
above A highpass filterpasses all high-frequency
components a certain cutoff frequency
and blocks all low-frequency components
below
cf
cf
cf
-
8/10/2019 DSP Mitra Chap1
43/125
Copyright 2002 S. K. Mitra43
FilteringFiltering Abandpass filterpasses all frequency
components between 2cutoff frequencies,and , where , and blocks allfrequency components below the frequency
and above the frequency Abandstop filterblocks all frequency
components between 2cutoff frequencies,and , where , and passes all
frequency components below the frequency
and above the frequency
1cf 2cf 21 cc ff /
1cf 2cf 21 cc ff /
1cf 2cf
1cf 2cf
-
8/10/2019 DSP Mitra Chap1
44/125
Copyright 2002 S. K. Mitra44
FilteringFiltering Figures below illustrate the lowpassfiltering of an input signal composed of 3
sinusoidal components of frequencies 50Hz, 110Hz, and 210 Hz
0 20 40 60 80 100
-4
-2
0
2
4
Time, msec
Amplitude
Input signal
0 20 40 60 80 100
-1
-0.5
0
0.5
1
Time, msec
mp
u
e
Lowpass filter output
-
8/10/2019 DSP Mitra Chap1
45/125
Copyright 2002 S. K. Mitra45
FilteringFiltering
Figures below illustrate highpassand
bandpassfiltering of the same input signal
0 20 40 60 80 100-1
-0.5
0
0.5
1
Time, msec
mp
ue
Highpass filter output
0 20 40 60 80 100-1
-0.5
0
0.5
1
Time, msec
mp
ue
Bandpass filter output
-
8/10/2019 DSP Mitra Chap1
46/125
Copyright 2002 S. K. Mitra46
FilteringFiltering
There are various other types of filters
A filter blocking a single frequency
component is called a notch filter
A multiband filterhas more than one
passband and more than one stopband
A comb filterblocks frequencies that areintegral multiples of a low frequency
-
8/10/2019 DSP Mitra Chap1
47/125
Copyright 2002 S. K. Mitra47
FilteringFiltering
In many applications the desired signal
occupies a low-frequency band from dc tosome frequency Hz, and gets corrupted
by a high-frequency noisewith frequency
components above Hz with
In such cases, the desired signal can be
recovered from the noise-corrupted signalby passing the latter through a lowpass filter
with a cutoff frequency where
Lf
f Lff 0
cf cL fff //
-
8/10/2019 DSP Mitra Chap1
48/125
Generation of ComplexGeneration of Complex
-
8/10/2019 DSP Mitra Chap1
49/125
Copyright 2002 S. K. Mitra49
Generation of ComplexGeneration of Complex
SignalsSignals
A signal can be real-valuedor complex-
valued
For convenience, the former is usually
called a real signalwhile the latter is calleda complex signal
A complex signalcan be generated from areal signalby employing a Hilbert
transformer
Generation of ComplexGeneration of Complex
-
8/10/2019 DSP Mitra Chap1
50/125
Copyright 2002 S. K. Mitra50
Generation of ComplexGeneration of Complex
SignalsSignals Theimpulse responseof a Hilbert
transformer is given by
Consider a real signalx(t)with a
continuous-time Fourier transform(CTFT)
X(j1)given by
tthHT 2'
1)(
)'1+
+,
1, dtetxjX tj)()(
-
8/10/2019 DSP Mitra Chap1
51/125
Generation of ComplexGeneration of Complex
-
8/10/2019 DSP Mitra Chap1
52/125
Copyright 2002 S. K. Mitra52
Generation of ComplexGeneration of Complex
SignalsSignals Thus we can write
where is the portion ofX(j1)occupying thepositive frequency rangeand
is the portion ofX(j1) occupying
the negative frequency range
)()()( 1-1'1 jXjjXjX np)( 1jX
p
)( 1jn
Generation of ComplexGeneration of Complex
-
8/10/2019 DSP Mitra Chap1
53/125
Copyright 2002 S. K. Mitra53
Generation of ComplexGeneration of Complex
SignalsSignals Ifx(t)is passed through a Hilbert
transformer, its outputy(t)is given by:
The spectrum Y(j1)ofy(t)is given by the
product of the CTFTs of andx(t)
) ***,'
+
+,dxthty
HT)()()(
)(th T
Generation of ComplexGeneration of Complex
-
8/10/2019 DSP Mitra Chap1
54/125
Copyright 2002 S. K. Mitra54
Generation of ComplexGeneration of Complex
SignalsSignals The CTFT of is given by
Therefore
)( 1jHT
)(thHT
3
45
/1
01,'1
0,
0,)(
j
jjHHT
)()()( 11'1 jXjHjY HT
)()( 1-1,' jXjjXj np
Generation of ComplexGeneration of Complex
-
8/10/2019 DSP Mitra Chap1
55/125
Copyright 2002 S. K. Mitra55
Generation of ComplexGeneration of Complex
SignalsSignals As the magnitude and phase of Y(j1)are an
even and odd function, respectively, it
follows from
thaty(t)is also a real function
Consider the complex signalg(t):
g(t) =x(t) +j y(t)
)()()( 1-1,'1 jXjjXjjY np
Generation of ComplexGeneration of Complex
-
8/10/2019 DSP Mitra Chap1
56/125
Copyright 2002 S. K. Mitra56
Generation of ComplexGeneration of Complex
SignalsSignals The CTFT ofg(t)is thus given by
In other words, the complex signalg(t),
called an analytic signal, has onlypositive
frequency components
)(2)()()( 1'1-1'1 jXjYjjXjG p
6)(tx
)(tx
)(tyHilbertrTransforme
partreal
partimaginary
)()()( tyjtxtg -'
-
8/10/2019 DSP Mitra Chap1
57/125
Copyright 2002 S. K. Mitra57
Modulation and DemodulationModulation and Demodulation For efficient transmission of a low-
frequency signal over a channel, it is
necessary to transform the signal to a high-
frequencysignal by means of amodulationoperation
At the receiving end, the modulated high-frequency signal is demodulatedto extract
the desired low-frequency signal
-
8/10/2019 DSP Mitra Chap1
58/125
Copyright 2002 S. K. Mitra58
Modulation and DemodulationModulation and Demodulation There are 4major types of modulation of
analog signals:
(1) Amplitude modulation
(2) Frequency modulation
(3) Phase modulation
(4) Pulse amplitude modulation
Amplitude ModulationAmplitude Modulation
-
8/10/2019 DSP Mitra Chap1
59/125
Copyright 2002 S. K. Mitra59
Amplitude ModulationAmplitude Modulation
Amplitude modulation is conceptuallysimple
Here, the amplitude of a high-frequencysinusoidal signal , called thecarrier signal, is varied by thelow-frequencysignalx(t), called the modulatingsignal
Process generates a high-frequency signal,called modulated signal,y(t)given by:
)cos( otA 1
)cos()()( ottAxty 1'
-
8/10/2019 DSP Mitra Chap1
60/125
Copyright 2002 S. K. Mitra60
Amplitude ModulationAmplitude Modulation Thus, amplitude modulation can be
implemented by forming theproductof the
modulating signal with the carrier signal
To demonstrate the frequency translatingproperty, let
x(t) =where
)cos( 1t1
o1 1//1
-
8/10/2019 DSP Mitra Chap1
61/125
Copyright 2002 S. K. Mitra61
Amplitude ModulationAmplitude Modulation Then
The CTFT of Y(j1)ofy(t)is given by
whereX(j1
)is the CTFT ofx(t)
)cos()cos()( o1 ttAty 1.1'
7 8 7 8tt AA )(cos)(cos1o21o2
1,11-1'
7 8 7 8)()()( o2o2 1-1-1,1'1 jXjXjY AA
Amplitude ModulationAmplitude Modulation
-
8/10/2019 DSP Mitra Chap1
62/125
Copyright 2002 S. K. Mitra62
Amplitude ModulationAmplitude Modulation
Spectra of the modulating signalx(t)and the
modulated signaly(t)are shown below
)( 1jY
1
o1
o1,
m1,1
o)(
o m1,1, 0
2A
)( 1j
1m1m1, 0
1
m1-1
o)(
o m1-1,
Amplitude ModulationAmplitude Modulation
-
8/10/2019 DSP Mitra Chap1
63/125
Copyright 2002 S. K. Mitra63
Amplitude ModulationAmplitude Modulation
As can be seen from the figures on the
previous slide,y(t)is a bandlimited high-
frequency signal with a bandwidth of
centered at
The portion of the amplitude-modulatedsignal between and is called
the upper sideband, whereas, the portion ofthe amplitude-modulated signal between
and is called the lower sideband
m12
o1
m1-1o
m1,1o
o1
o1
-
8/10/2019 DSP Mitra Chap1
64/125
Copyright 2002 S. K. Mitra64
Amplitude ModulationAmplitude Modulation Because of the generation of two sidebands
and the absence of a carrier component in
the modulated signal, the process is called
double-sideband suppress carrier(DSB-SC)modulation
The demodulation ofy(t)to recoverx(t)iscarried out in two stages
-
8/10/2019 DSP Mitra Chap1
65/125
Copyright 2002 S. K. Mitra65
Amplitude ModulationAmplitude Modulation First,y(t)is multiplied with a sinusoidal
signal of the same frequency as the carrier:
The result indicates that r(t)is composed of
x(t)scaled by a factor 1/2and an amplitide-modulated signal with a carrier frequency
ttAxttytr o2
o cos)(cos)()( 1'1'
)2cos()()( o22 ttxtx AA 1-'
o21
-
8/10/2019 DSP Mitra Chap1
66/125
Copyright 2002 S. K. Mitra66
Amplitude ModulationAmplitude Modulation The spectrumR(j1)of r(t)is shown below
)( 1j
1m1m1, 0
2A
o21o21,
m1,1
o2)2(
o m1,1,
c1c1,
-
8/10/2019 DSP Mitra Chap1
67/125
Copyright 2002 S. K. Mitra67
Amplitude ModulationAmplitude Modulation Thusx(t)can be recovered from r(t)by
passing it through a lowpass filterwith a
cutoff frequency at satisfying the
relation The modulation and demodulation schemes
are as shown below:
c1
mcm 1,1/1/1 o2
9)(tx )(ty
tocos1
9 LowpassFilter)(ty)(tr
)(2
txA
tocos1
-
8/10/2019 DSP Mitra Chap1
68/125
Copyright 2002 S. K. Mitra68
Amplitude ModulationAmplitude Modulation In general, it is difficult to ensure that the
demodulating sinusoidal signal has a
frequency identical to that of the carrier
To get around the above problem, themodulation process is modified so that the
transmitted signal includes the carrier signal
-
8/10/2019 DSP Mitra Chap1
69/125
Copyright 2002 S. K. Mitra69
Amplitude ModulationAmplitude Modulation This is achieved by redefining the
amplitude modulation operation as follows:
where mis a number chosen to ensure that
[1 + m x(t)]is positive for all t
As the carrier is also present in themodulated signal, the process is called
double-sideband(DSB) modulation
)cos()](1[)( ottxmAty 1-'
Amplitude ModulationAmplitude Modulation
-
8/10/2019 DSP Mitra Chap1
70/125
Copyright 2002 S. K. Mitra70
Amplitude ModulationAmplitude Modulation
Figure below shows the waveforms of a
modulating sinusoidal signal of frequency
20 Hzand the amplitude-modulated carrier
with a carrier frequency 400 Hzobtained
using the DSB modulation scheme andm=0.5
0 20 40 60 80 100-1
-0.5
0
0.5
1
Time, msec
mp
u
e
Modulating signal
0 20 40 60 80 100-2
-1
0
1
2
Time, msec
Amplitude
Modulated carrier
Amplitude ModulationAmplitude Modulation
-
8/10/2019 DSP Mitra Chap1
71/125
Copyright 2002 S. K. Mitra71
Amplitude ModulationAmplitude Modulation
In the case of the conventional DSB
amplitude modulation scheme, themodulated signal has a bandwidth of ,
whereas the bandwidth of the modulating
signal is
To increase the capacity of the transmission
medium, either the upper sidebandor thelower sidebandof the modulated signal is
transmitted
m1
m12
Amplitude ModulationAmplitude Modulation
-
8/10/2019 DSP Mitra Chap1
72/125
Copyright 2002 S. K. Mitra72
p tude odu at op
The corresponding procedure is calledsingle-sideband(SSB) modulation, a
possible implentation of which is shownbelow
The spectra of pertinent signals in the SSB
modulation scheme are shown in the next
slide
)(tx
)(ty
tocos1
9
9+6
tosin1
HilbertrTransforme
)(tu
)(1tw
)(2tw
Amplitude ModulationAmplitude Modulation
-
8/10/2019 DSP Mitra Chap1
73/125
Copyright 2002 S. K. Mitra73
pp
)(2 1W
1o1o1, 0
2A
)(1 1jW
1o1o1, 0
2A
)( 1jY
1o1o1, 0
QuadratureQuadratureAmplitudeAmplitude
-
8/10/2019 DSP Mitra Chap1
74/125
Copyright 2002 S. K. Mitra74
ModulationModulation The DSB amplitude modualtionis half as
efficient as SSB amplitude modulation
The quadrature amplitude modualtion
(QAM) method uses DSBmodulation tomodulate two different signals so that they
both occupy the same bandwidth Hence, QAMtakes up as much bandwidth
as the SSBmethod
QuadratureQuadratureAmplitudeAmplitude
-
8/10/2019 DSP Mitra Chap1
75/125
Copyright 2002 S. K. Mitra75
ModulationModulation Consider two bandlimited signals and
with a bandwidth of as indicatedbelow
)(1tx
)(2tx m1)(1 1j
1m1m1, 0
)(2 1
1m1m1, 0
QuadratureQuadratureAmplitudeAmplitude
-
8/10/2019 DSP Mitra Chap1
76/125
Copyright 2002 S. K. Mitra76
ModulationModulation The two modulating signals are individually
modulated by the two carrier signals
and , respectively,
and are summed, resulting in
The two carrier signals have the samecarrier frequency but have a phase
difference of90o
)cos( otA 1 )sin( otA 1
)sin()()cos()()( o2o1 ttAxttAxty 1-1'
o1
QuadratureQuadratureAmplitudeAmplitude
-
8/10/2019 DSP Mitra Chap1
77/125
Copyright 2002 S. K. Mitra77
ModulationModulation The carrier is called the in-phase
componentand the carrier is
called the quadrature component
The spectrum Y(j1)of the composite signaly(t)is given by
)cos( otA 1)sin( otA 1
7 8 7 8}(({)( o1o12 1-1-1,1'1 jXjXjY A
7 8 7 8}(({ o2o22
1-1,1,1- jXjXj
A
QuadratureQuadratureAmplitudeAmplitude
-
8/10/2019 DSP Mitra Chap1
78/125
Copyright 2002 S. K. Mitra78
ModulationModulation y(t) is seen to occupy the same bandwidth
as the modulated signal obtained by a DSB
modulation
To recover and ,y(t)is multipliedby both the in-phase and the quadrature
components of the carrier separately,
resulting in
)(1tx )(2tx
)cos()()( o1 ttytr 1'
)sin()()( o2 ttytr 1'
QuadratureQuadratureAmplitudeAmplitude
-
8/10/2019 DSP Mitra Chap1
79/125
Copyright 2002 S. K. Mitra79
ModulationModulation Substituting the expression fory(t)in both
of the last two equations, we obtain aftersome algebra
Lowpass filtering of and by filters
with a cutoff at yields and
)2sin()()2cos()()()( o22o12121 ttxttxtxtr AAA
1-1-'
)2cos()()2sin()()()( o22o12222 ttxttxtxtr AAA 1,1-'
)(1tr )(2tr
m1 )(1tx )(2 tx
QuadratureQuadratureAmplitudeAmplitude
-
8/10/2019 DSP Mitra Chap1
80/125
Copyright 2002 S. K. Mitra80
ModulationModulation The QAM modulation and demodulation
schemes are shown below
)(12 txA
)(22 txA
-phase90o
shifter
69
9
)(1tx
)(2 tx
)(ty
)cos( otA 1
phase90o
shifter
69
9
Lowpass
filter
Lowpass
filter)(ty
)cos( ot1
6
Multiplexing andMultiplexing and
-
8/10/2019 DSP Mitra Chap1
81/125
Copyright 2002 S. K. Mitra81
DemultiplexingDemultiplexing For an efficient utilization of a wideband
transmission channel, many narrow-bandwidth low-frequency signals are
combined for a composite wideband signalthat is transmitted as a single signal
The process of combining the low-
frequency signals is called multiplexing
Multiplexing andMultiplexing and
-
8/10/2019 DSP Mitra Chap1
82/125
Copyright 2002 S. K. Mitra82
DemultiplexingDemultiplexing Multiplexing is implemented to ensure that
a replica of each of the original narrow-bandwidth low-frequency signal can be
recovered at the receiving end The recovery process of the low-frequency
signals is called demultiplexing
Multiplexing andMultiplexing and
-
8/10/2019 DSP Mitra Chap1
83/125
Copyright 2002 S. K. Mitra83
DemultiplexingDemultiplexing One method of combining different voice
signals in a telephone communicationsystem is the frequency-divisionmultiplexing(FDM) scheme
Here, each voice signal, typicallybandlimited to a low-frequency band of
width 1 , is frequency-translated into ahigher frequency band using the amplitudemodulation method
m
Multiplexing andMultiplexing and
-
8/10/2019 DSP Mitra Chap1
84/125
Copyright 2002 S. K. Mitra84
DemultiplexingDemultiplexing The carrier frequency of adjacent
amplitude-modulated signals is separated by, where to ensure that there is
no overlap in the spectra of the individualmodulated signals after they are added toform the baseband composite signal
The composite signal is then modulatedonto the main carrier developing the FDMsignal and transmitted
o1 m101 2o
Multiplexing andMultiplexing and
-
8/10/2019 DSP Mitra Chap1
85/125
Copyright 2002 S. K. Mitra85
DemultiplexingDemultiplexing)(1 1j
1m1m1, 0
)(2 1
1m1m1, 0
)(3 1j
1m1
111 21 31011,
)( 1jYcomp
, m1, 0
lowtheofSpectra signalsfrequency,
signalcompositemodulatedtheofSpectra
Multiplexing andMultiplexing and
-
8/10/2019 DSP Mitra Chap1
86/125
Copyright 2002 S. K. Mitra86
DemultiplexingDemultiplexing At the receiving end, the composite
baseband signal is first recovered from theFDM signal by demodulation
Then each individual frequency-translatedsignal is demultiplexedby passing the
composite signal through a bank of
bandpass filters
-
8/10/2019 DSP Mitra Chap1
87/125
Ad t f DSPAd t f DSP
-
8/10/2019 DSP Mitra Chap1
88/125
Copyright 2002 S. K. Mitra88
Advantages of DSPAdvantages of DSP Absence of drift in the filter characteristics
Processing characteristics are fixed, e.g. by
binary coefficients stored in memories
Thus, they are independent of the externalenvironment and of parameters such as
temperature
Aging has no effect
Ad t f DSPAd t f DSP
-
8/10/2019 DSP Mitra Chap1
89/125
Copyright 2002 S. K. Mitra89
Advantages of DSPAdvantages of DSP Improved quality level Quality of processing limited only by economic
considerations
Arbitrarily low degradations achieved with
desired quality by increasing the number of bitsin data/coefficient representation
An increase of 1bit in the representation results
in a 6dB improvement in the SNR
Ad t f DSPAd t f DSP
-
8/10/2019 DSP Mitra Chap1
90/125
Copyright 2002 S. K. Mitra90
Advantages of DSPAdvantages of DSP Reproducibility
Component tolerances do not affect system
performance with correct operation
No adjustments necessary during fabrication No realignment needed over lifetime of
equipment
Ad t f DSPAd t f DSP
-
8/10/2019 DSP Mitra Chap1
91/125
Copyright 2002 S. K. Mitra91
Advantages of DSPAdvantages of DSP Ease of new function development
Easy to develop and implement adaptive filters,
programmable filters and complementary filters
Illustrates flexibility of digital techniques
Ad t f DSPAd t f DSP
-
8/10/2019 DSP Mitra Chap1
92/125
Copyright 2002 S. K. Mitra92
Advantages of DSPAdvantages of DSP Multiplexing
Same equipment can be shared between several
signals, with obvious financial advantages for
each function
Modularity
Uses standard digital circuits for
implementation
Ad t f DSPAd t f DSP
-
8/10/2019 DSP Mitra Chap1
93/125
Copyright 2002 S. K. Mitra93
Advantages of DSPAdvantages of DSP Total single chip implementation using
VLSI technology
No loading effect
Limitations of DSPLimitations of DSP
-
8/10/2019 DSP Mitra Chap1
94/125
Copyright 2002 S. K. Mitra94
Limitations of DSPLimitations of DSP Lesser Reliability
Digital systems are active devices, and thus use
more power and are less reliable
Some compensation is obtained from thefacility for automatic supervision and
monitoring of digital systems
Limitations of DSPLimitations of DSPLi i d F R f O i
-
8/10/2019 DSP Mitra Chap1
95/125
Copyright 2002 S. K. Mitra95
Limited Frequency Range of Operation
Frequency range technologically limited to
values corresponding to maximum computingcapacities that can be developed and exploited
Additional Complexity in the Processing ofAnalog Signals
A/D and D/A converters must be introduced
adding complexity to overall system
DSP Application ExamplesDSP Application Examples
-
8/10/2019 DSP Mitra Chap1
96/125
Copyright 2002 S. K. Mitra96
Cellular Phone Discrete Multitone Transmission
Digital Camera Digital Sound Synthesis
Signal Coding & Compression
Signal Enhancement
Cellular Phone Block DiagramCellular Phone Block Diagram
-
8/10/2019 DSP Mitra Chap1
97/125
Copyright 2002 S. K. Mitra97
DSP Core
S/W
DSP Core
S/W
ARM RISCCore
S/W
ARM RISCCore
S/W
SINGLE CHIP
DIGITALBASEBAND
ASICB
ackplane
User DisplayUser Display
KeyboardKeyboard
SIM CardSIM Card
RF
Interface
RF
InterfaceAudioInterface
AudioInterface
SpeakerSpeaker MicMic
Op AmpsOp Amps
SwitchesSwitches
RegulatorsRegulators
ReceiverReceiver
ModulatorModulator DriverDriver
RFSECTION
PowerAmp
Power
AmpSynthesizerSynthesizer
Touch ScreenTouch Screen
Cellular Phone Block DiagramCellular Phone Block Diagram
sInstrumentTexas:Courtesy
Cellular Phone Baseband
System on a Chip
-
8/10/2019 DSP Mitra Chap1
98/125
Copyright 2002 S. K. Mitra98
100-200 MHz DSP +
MCU
ASIC Logic
Dense Memory
Analog
System on a Chip
sInstrumentTexas:Courtesy
DiscreteDiscreteMultitoneMultitone
Transmission (DMT)Transmission (DMT)
-
8/10/2019 DSP Mitra Chap1
99/125
Copyright 2002 S. K. Mitra99
Transmission (DMT)Transmission (DMT)
Core technology in the implementation oftheasymmetric digital subscriber line
(ADSL) andvery-high-rate digital
subscriber line(VDSL)
Closely related to: Orthogonal frequency-
division multiplexing(OFDM)
ADSLADSL
-
8/10/2019 DSP Mitra Chap1
100/125
Copyright 2002 S. K. Mitra100
A local transmission system designed tosimultaneously support three services on a
single twisted-wire pair: Data transmission downstream (toward the
subscriber) at bit rates of upto9 Mb/s
Data transmission upstream (away from thesubscriber) at bit rates of upto1 Mb/s
Plain old telephone service(POTS)
ADSL
-
8/10/2019 DSP Mitra Chap1
101/125
Copyright 2002 S. K. Mitra101
Band-allocations for an FDM-based ADSL
system
powerTransmit
bandstream-down
rate-bitHigh
bandPOTS
bandGuard
band
upstream
rate-bitLow
Frequency
ADSLADSL
-
8/10/2019 DSP Mitra Chap1
102/125
Copyright 2002 S. K. Mitra102
Asymmetry in the frequency bandallocation:
to bring movies, television, video catalogs,
remote CD-ROMs, corporate LANs, and theInternet into homes and small businesses
VDSLVDSL
i l k i f i d
-
8/10/2019 DSP Mitra Chap1
103/125
Copyright 2002 S. K. Mitra103
Optical network emanating from twisted
pair provides data rates of 13to 26 Mb/s
downstream and 2to 3 Mb/supstream over
short distances less than about 1 km
Allows the delivery of digital TV, super-fastWeb surfing and file transfer, and virtual
offices at home
DiscreteDiscreteMultitoneMultitone
TransmissionTransmission
-
8/10/2019 DSP Mitra Chap1
104/125
Copyright 2002 S. K. Mitra104
TransmissionTransmission
Advantages in using DMTfor ADSLandVDSL
The ability to maximize the transmitted bit rate Adaptivity to changing line conditions
Reduced sensitivity to line conditions
OFDMOFDM
-
8/10/2019 DSP Mitra Chap1
105/125
Copyright 2002 S. K. Mitra105
Applications:
Wireless communications- an effectivetechnique to combat multipath fading
Digital audio broadcasting
Uses a fixed number of bits per subchannel
while DMT uses loading for bit allocation
-
8/10/2019 DSP Mitra Chap1
106/125
Digital CameraDigital CameraCMOS I i S
-
8/10/2019 DSP Mitra Chap1
107/125
Copyright 2002 S. K. Mitra107
CMOS Imaging Sensor
Increasingly being used in digital cameras
Single chip integration of sensor and other
image processing algorithms needed to generate
final image Can be manufactured at low cost
Less expensive cameras use single sensor with
individual pixels in the sensor covered with
either ared, agreen, or ablueoptical filter
Digital CameraDigital Camera
-
8/10/2019 DSP Mitra Chap1
108/125
Copyright 2002 S. K. Mitra108
Image Processing Algorithms
Bad pixel detection and masking
Color interpolation
Color balancing Contrast enhancement
False color detection and masking
Image and video compression
Digital CameraDigital Camera
Bad Pixel Detection and Masking
-
8/10/2019 DSP Mitra Chap1
109/125
Copyright 2002 S. K. Mitra109
g
-
8/10/2019 DSP Mitra Chap1
110/125
-
8/10/2019 DSP Mitra Chap1
111/125
Digital Sound SynthesisDigital Sound Synthesis
-
8/10/2019 DSP Mitra Chap1
112/125
Copyright 2002 S. K. Mitra112
Wavetable Synthesis
- Recorded or synthesized musical events stored ininternal memory and played back on demand
- Playback tools consists of various techniques forsound variation during reproduction such aspitch
shifting, looping, envelopingandfiltering
- Example: Giga Sampler
Digital Sound SynthesisDigital Sound Synthesis
S l S h i
-
8/10/2019 DSP Mitra Chap1
113/125
Copyright 2002 S. K. Mitra113
Spectral Synthesis
- Produces sounds from frequency domain models- Signal represented as a superposition of basis
functions with time-varying amplitudes
- Practical implementation usually consist of acombination of additive synthesis, subtractive
synthesisand granular synthesis
- Example: Kawaii K500 Demo
Digital Sound SynthesisDigital Sound Synthesis
-
8/10/2019 DSP Mitra Chap1
114/125
Copyright 2002 S. K. Mitra114
Nonlinear Synthesis
- Frequency modulation method: Time-
dependent phase terms in the sinusoidal basis
functions- An inexpensive method frequently used in
synthesizers and in sound cards for PC
- Example: Variation modulation index complex
algorithm(Pulsar)
Digital Sound SynthesisDigital Sound Synthesis
Physical Modeling
-
8/10/2019 DSP Mitra Chap1
115/125
Copyright 2002 S. K. Mitra115
Physical Modeling
- Models the sound production method
- Physical description of the main vibrating
structures by partial differential equations
- Most methods based on wave equationdescribing the wave propagation in solids and in
air
- Examples: (CCRMA, Stanford) Guitar with nylon strings
Marimba
Tenor saxophone
Signal Coding & CompressionSignal Coding & Compression
-
8/10/2019 DSP Mitra Chap1
116/125
Copyright 2002 S. K. Mitra116
Concerned with efficient digital
representationof audio or visual signalforstorageandtransmissionto provide
maximum quality to the listener orviewer
Signal Compression ExampleSignal Compression Example
Original speech0 5
1
16-bit, sampled at 44.1 kHz rate
-
8/10/2019 DSP Mitra Chap1
117/125
Copyright 2002 S. K. Mitra117
Data size 330,780 bytes
Compressed speech (GSM 6.10)
- Sampled at 22.050 kHz,Data size 16,896 bytes
Compressed speech (Lernout & Hauspie
CELP 4.8kbit/s)
Sampled at 8 kHz,Data size 2,302 bytes
0 1 2 3-1
-0.5
0
0.5
Time, sec.
mp
u
e
Signal Compression ExampleSignal Compression Example
-
8/10/2019 DSP Mitra Chap1
118/125
Copyright 2002 S. K. Mitra118
Original music
Audio Format:PCM 16.000 kHz, 16 Bit(Data size 66206 bytes)
Compressed music
Audio Format:GSM 6.10, 22.05 kHz(Data size 9295 bytes)
Courtesy: Dr. A. Spanias
Signal Compression ExampleSignal Compression Example
-
8/10/2019 DSP Mitra Chap1
119/125
Copyright 2002 S. K. Mitra119
Compressed Image
Average bit rate - 0.5 bits per pixel
Original Lena
8 bits per pixel
Signal EnhancementSignal Enhancement
Purpose:To emphasizespecific signal
-
8/10/2019 DSP Mitra Chap1
120/125
Copyright 2002 S. K. Mitra120
p p p g
features to provide maximum quality to the
listener or viewer
For speech signals, algorithms include
removal of background noise or interference For image or video signals, algorithms
include contrast enhancement, sharpeningand noise removal
Signal Enhancement ExampleSignal Enhancement Example
0.5
1
-
8/10/2019 DSP Mitra Chap1
121/125
Copyright 2002 S. K. Mitra121
Noisy speech signal
(10% impulse noise)
Noise removed speech
0 1 2 3-1
-0.5
0
0.5
1
Time, sec.
mp
u
e
0.4 0.6 0.8 1 1.2-1
-0.5
0
Time, sec.
mp
u
e
Signal Enhancement ExampleSignal Enhancement Example
-
8/10/2019 DSP Mitra Chap1
122/125
Copyright 2002 S. K. Mitra122
EKG corrupted with EKG after filtering with
60 Hzinterference a notch filter
0 500 1000 1500 2000-50
0
50
100
150
time
EKG After Noise Removal
mp
tu
e
0 500 1000 1500 2000-50
0
50
100
150
time
mp
tu
e
EKG Corrupted With 60 Hz Interference
Signal Enhancement ExampleSignal Enhancement Example
-
8/10/2019 DSP Mitra Chap1
123/125
Copyright 2002 S. K. Mitra123
Original image and its contrast enhanced version
Original Enhanced
Signal Enhancement ExampleSignal Enhancement Example
Original image and its contrast enhanced version
-
8/10/2019 DSP Mitra Chap1
124/125
Copyright 2002 S. K. Mitra124Original Enhanced
Signal Enhancement ExampleSignal Enhancement Example
Noise corrupted image and its noise-removed
version
-
8/10/2019 DSP Mitra Chap1
125/125
Copyright 2002 S. K. Mitra125
version
20% pixels corrupted with
additive impulse noise
Noise-removed version