MIT-WPU

Communication Systems (ECE2007A)

 Module 4: Digitization of Analog Signal

Sampling Theorem, sampling techniques, analog

pulse modulation methods: PAM, PWM and PPM,
Digital Communication system block diagram, Digital
pulse modulation methods: PCM, Non-uniform
Quantization and Companding, Differential Pulse
Code Modulation, Delta Modulation, Adaptive Delta
Modulation. [7 hrs]
 Classification of Modulation Techniques
 A. Continuous Wave (CW): Analog-to-Analog Converter (AAC)
 A.1 Linear/Amplitude Modulation
Variants: DSB-SC, SSB-SC, ISB, VSB
 A.2 Angle/Exponential
A.2.1 Frequency Modulation (FM)
A.2.2 Phase Modulation (PM)

3
 B. Pulse Modulation : Analog-to-Digital Converter (ADC)
 B.1 Pulse Analog
B.1.1 PAM
B.1.2 PWM
B.1.3 PPM
 B.2 Pulse Digital:
B.2.1 PCM
B.2.2 DPCM
B.2.3 DM
B.2.4 ADM

4
Time Limited Signals

A time-limited signal is the one that is non-zero

only for a finite length of time interval.
Band Limited Signals
Time-Limited and Band-Limited Signals

A time - limited signal cannot be also band limited.

Band-Limited and Time-Limited Signals

A band - limited signal cannot be also time-

limited.
Analog-to-Digital Converter (ADC)-
Pulse Code Modulation (PCM)
Sampling
 The process of transmitting signals in the form of pulses (discontinuous
signals) by using special techniques.
 The signal is sampled at regular intervals such that each sample is
proportional to the amplitude of signal at that instant. This technique is
called “sampling”.
 Sampling is common in all pulse modulation techniques.
Sampling
 Analog signal is sampled every TS secs.
 Ts is referred to as the sampling interval.
 fs = 1/Ts is called the sampling rate or sampling frequency.
Sampling
Sampling Theorem

1) A band limited signal of finite energy, which has no frequency

components higher than W Hz, is completely described by specifying the
1
values of the signal at instants of time separated by sec.
2𝑊

2) A band limited signal of finite energy, which has no frequency

components higher than W Hz, may be completely recovered from the
knowledge of its samples taken at the rate of 2W per sec.

• Link between analog waveform and its samples version

• Can be implemented using sample and hold operation (a switch and
storage mechanism)
• O/P is called PAM(Pulse Amplitude Modulation)
• The analog waveform can be approximately retrieved from a PAM
waveform by simple low-pass filtering.
 Nyquist Criterion or Uniform sampling
Theorem
Fs = 2fm is also called as Nyquist rate.
A continuous time signal can be completely represented in its
samples & recovered back if sampling frequency 𝑓𝑠 is greater
than or equal to the twice of highest frequency component of
message signal.
Nyquist criterion ensures an analog signal to be
reconstructed completely from a set of uniformly
spaced discrete time samples.
Types of Sampling
Impulse or Ideal Sampling: impulse at each sampling instant

Natural Sampling :a pulse of short width with varying amplitude

Flat Top Sampling: sample and hold, like natural but with single
amplitude value
Impulse Sampling
 This is case of ideal sampling with a sequence of unit impulse functions.

 Validity of the sampling theorem using frequency convolution property of

Fourier transform.

 Sampling of x(t) can be viewed as the product of x(t) with a periodic train
of unit impulse function
xδ (t).
Imp I Impulse
Impulse Sampling…
Sampling
Impulse Sampling

 Shifting property of the impulse function states that

 Using shifting property, Sampled version of signal x(t) is given below:

 Note that Multiplication in time domain will result into convolution in

frequency domain
Impulse Sampling

 We can transform the time domain product x(t)xδ(t) to the frequency domain convolution
X(f)*Xδ(f), where

 Convolution with an impulse function simply shifts the original function as follows:

 Now, Solve for the transform Xs(f) the sample waveforms:

Impulse Sampling

Here, Sampling rate is fs = 2fm, separation between spectral replicate is fs Hz.

Analog waveform can theoretically be completely recovered from the

samples, By the use of filtering
Impulse Sampling

Here, Sampling rate is fs = 2fm, separation between spectral replicate is fs Hz.

Analog waveform can theoretically be completely recovered from the

samples, By the use of filtering
 Impulse Sampling….

 If fs >2fm , replications will move further apart in frequency.

 Easier to perform filtering operations.

 Typical low pass filter might be used to separate the baseband spectrum from
those at higher frequency
 Impulse Sampling…

If fs < 2fm , the replicates will overlap.

As a result, information lost will be there.
This phenomenon is Aliasing occur because of undersampling.

 Note that to avoid aliasing, the Nyquist

criteria fs >2fm must be satisfied.
Natural Sampling

Why Natural???
Top of each pulse in the xs(t) sequence retains the shape of its corresponding
analog segment during the pulse interval

Multiplication of x(t) and xp(t)

xp(t) can be viewed as opening and closing switch.
Natural Sampling
Flat Top Sampling
 Flat Top Sampling

 In frequency domain, flat top sampled signal has same output as in

natural sampled case.

 Most obvious effect of hold operation is significant attenuation of

the higher frequency spectral replicates.
 Flat Top Sampling

 The sampled signal in the flat top sampling has the attenuated high
frequency components. This effect is called the Aperture Effect.
 The aperture effect can be compensated by:
1. selecting the pulse width τ as very small.
2. by using an equalizer circuit.

 Equalizer decreases the effect of the in-band loss of the interpolation filter (LPF).
 Aliasing
Aliasing
 To avoid Aliasing..
To avoid aliasing
1) One way is Higher Sampling rate
 To avoid Aliasing..
To avoid aliasing
2) Antialiasing filter.
 Analog signal is prefiltered so that new maximum
frequency is reduced
Cont…
 Another way is use of postfiltering after sampling.

35
Why Oversampling????
 Most economy solution for the task of transforming an analog
signal to a digital signal.

 Because signal processing performed with high performance

analog equipment is typically much more costly than using
digital signal processing equipment

36
 Applications
Applications of of Sampling theorem
Sampling Theorem

 Signal Analysis
 Signal Processing
 Signal Transmission
 PulseMethods
Pulse Modulation Modulation

• Pulse Analog Modulation

 Pulse Amplitude Modulation
 Pulse Width Modulation
 Pulse Position Modulation
 Pulse Digital Modulation
 Pulse Code Modulation
 Delta Modulation etc.
Pulse Modulation
Pulse Modulation
 Pulse Amplitude Modulation(PAM)

The carrier is in the form of narrow pulses having frequency fs. Samples are placed
Ts sec away from each other.
Uniform sampling takes place in multiplier to generate PAM signal.
In PAM, amplitude of pulses is varied in accordance with instantaneous value of
modulating signal.

Modulatin Low Pass Multiplier

Modulating PAM
PAM
g Signal
Signal Filter Signal
Signal

Pulse train
generator
Pulse Amplitude Modulation(PAM)

 Types :
 Natural PAM
 Flat top PAM

 Flat top PAM is preferred

for transmission
 Pulse Amplitude Modulation
Advantages:
• Simple process for both modulation and demodulation technique.
• Simple Circuit
• PAM can generate other pulse modulation signals and can carry the
message or information at same time.

Disadvantages :
• Bandwidth required is large
• Noise immunity is less when compared to other modulation
techniques. (like amplitude modulation).
• Pulse amplitude signal amplitude varies, so transmission power
required will be more, peak power is also more, even at receiving
more power is required to receive the pulse amplitude signal.
Applications of Pulse Amplitude
Modulation
• Base for all digital modulation techniques
• Used for Ethernet communications which is type of
computer network communication.
• Used as electronic driver for LED lighting.
• Used in many micro controllers for generating the
control signals etc.
 Pulse Width Modulation(PWM)

 The amplitude is maintained constant but the width

of each pulse is varied in accordance with
instantaneous value of the analog signal.(PDM)
Pulse Width Modulation
 Pulse Width Modulation (PWM)

 In this case a sawtooth signal of frequency fs is a sampling

signal.
 It is applied to inverting terminal of a comparator with
modulating signal at non inverting terminal.
 O/P remains high as long as modulating signal is higher than
that of ramp signal.
 PWM

Advantages :
 As amplitude is constant,effect of noise is less.
 Synchronization between transmitter and receiver
is not required unlike pulse position modulation.

Disadvantages :
• Power will be variable because of varying in width of pulse.
• Bandwidth should be large to use in communication,
should be huge even when compared to the pulse
amplitude modulation.

 Applications of PWM

 In telecommunication systems
 To control the amount of power delivered to a load
 Speed control of motors
 Robotics, Embedded applications
 Pulse Position Modulation (PPM)
 The sampled waveform has fixed
amplitude and width whereas the
position of each pulse is varied as per
instantaneous value of the analog signal.
 PPM signal is further modification of a
PWM signal.
 Thus as the trailing edges of the PWM
signal keeps shifting in proportion with
the modulating signal, the PPM pulses
also keep shifting.
PPM
 Advantages
 Pulse position modulation has low noise interference
when compared to PAM because amplitude and width of
the pulses are made constant during modulation.
 Power usage is also very low when compared to other
modulations due to constant pulse amplitude and width.
 Disadvantages
• Synchronization between transmitter and receiver is
required
• Large bandwidth is required for transmission (as in pulse
amplitude modulation)
• Special circuits required
Applications of PPM

 Used in radio frequency (RF) communication.

 Used in contactless smart card, high frequency, RFID (radio frequency
ID) tags etc.
 Comparison of Pulse Modulation techniques
Comparison of PAM, PWM, and PPM
Sr.No. Pulse Amplitude Modulation Pulse Width Modulation Pulse Position Modulation
(PAM) (PWM) (PPM)
1

2 Amplitude of the pulse is Width of pulse is proportional to The relative position of the
proportional to amplitude of amplitude of modulating signal. pulse is proportional to the
modulating signal. amplitude of the modulating
signal.
3 The BW of the transmitting BW of the transmitting channel BW of the transmitting
channel depends on the width of depends on the rise time of the channel depends on the rising
the pulse. pulse. time of the pulse.
4 The instantaneous power of the The instantaneous power of the The instantaneous power of
transmitter varies. transmitter varies. the transmitter remains
constant.
5 Noise interference is high. System Noise interference is minimum. Noise interference is
is complex. minimum.
6 Similar to amplitude modulation. Simple to implement similar to Simple to implement similar
frequency modulation. to phase modulation.
Modern Digital Communication System

5/18/2022 54
Communication System Block Diagram (Basic)

 Source encoder converts message into message signal (bits)

 Transmitter converts message signal into format appropriate for channel
transmission (analog/digital signal)
 Channel conveys signal but may introduce attenuation, distortion, noise,
interference
 Receiver decodes received signal back to message signal
 Source decoder decodes message signal back into original message

5/18/2022
55
 Brief Description
 Source: analog or digital
 Transmitter: transducer, amplifier, modulator, oscillator, power
amplifier antenna
 Channel: e.g. cable, optical fiber, free space
 Receiver: antenna, amplifier, demodulator, oscillator, power
amplifier, transducer
 Recipient: e.g. person, (loud) speaker, computer

 Types of information-
Voice, data, video, music, email etc.

5/18/2022
56
 Morse Code
 Used for Telegraphy:

 A-Z, 0-9, “.”, “,”, “?” represented by sequence

of “.” (Dots) and “-”(Dashes) and Space.
 “.” - Short duration electrical pulse of current
 “-” - Long duration electrical pulse of
current
 “ ” - Space: Gap between Pulse.
 Frequent characters are transmitted using
short sequences of bits
 Ex: E is transmitted by “.”
 Rare characters are transmitted by long
sequences of bits
 Ex: Z is transmitted by “- - . . ”
1 is transmitted by “.----”
? is transmitted by “..--..”

5/18/2022
57
Communication systems

Example 1 Radio
Source: Microphone
Destination: Speakers in Radio Rx.
Channel/ Media: Space

Example 2 Television
Source: Video Camera and microphone
Destination: Picture Tube and Speakers in TV Rx.
Channel/ Media: Space

Example 3 Landline Telephone

Source: Microphone in Phone set
Destination: Speaker in phone set
Channel/ Media: Wire

All are Analog Communication Systems

5/18/2022
58
Communication Systems
Example 4 Cellular Mobile System
Source: Microphone in Phone set
Destination: Speakers in phone set
Channel/ Media: Space

Other examples of Digital Communication are LAN,

Internet

5/18/2022
59
 Information Representation

 Communication system converts information into electrical

electromagnetic/optical signals appropriate for the transmission
medium.
 Analog systems convert analog message into signals that can
propagate through the channel.

 Digital systems convert bits (digits, symbols) into signals

 Computers naturally generate information as characters/bits

 Most information can be converted into bits
 Analog signals converted to bits by sampling and quantizing
(A/D conversion)

5/18/2022
60
What is Digital
 Analog signal - Continuous in time
 Digital signal - discrete in time and amplitude

 Performance evaluation criteria are different-

Analog Communication - Fidelity, SNR, Percent distortion
Digital Communication – Bit Error Rate or Probability of error

 Analog Modulation (AM, FM, PM) and

Digital Modulation (Binary shift keying and M-ary shift keying
techniques)

5/18/2022 TY ECE, MIT WPU, Pune 61

 Why digital?
 Digital techniques need to distinguish between discrete symbols allowing
regeneration versus amplification

 Digital circuits are more reliable, digital hardware provides more

flexibility in implementation

 Good processing techniques are available for digital signals, such as

medium.
 Data compression (or source coding)
 Error Correction (or channel coding)(A/D conversion)
 Equalization
 Security

5/18/2022 62
 Why Digital Communications?
 Easy to regenerate the distorted signal
 Regenerative repeaters along the transmission path can detect a
digital signal and retransmit a new, clean (noise free) signal
 Two-state signal representation
This is not possible with analog communication systems

5/18/2022
63
 Why Digital Communications?
 Easy to mix/combine/interleave signals and data using digital
techniques

 More immune to channel noise and distortion

 Hardware is more flexible

Digital hardware implementation is flexible and permits the use
of microprocessors, mini-processors, digital switching and
VLSI.
 Encryption and privacy techniques are easier to implement

 Low cost - The use of LSI and VLSI in the design of

components and systems have resulted in lower cost

5/18/2022 64
 Why Digital Communications?
 Data communication in computers is digital in nature whereas
voice communication between people is analog in nature

 The two types of communication are difficult to combine

over the same medium in the analog domain.

 Usingdigital techniques, it is possible to combine both

formats for transmission through a common medium

5/18/2022
65
 Why Digital Communications?

Disadvantages
Requires reliable “synchronization”
Requires A/D conversions at high rate
Requires larger bandwidth
Probability of error or Bit Error Rate decides
the quality of reception

5/18/2022
66
 Resources and Constraints for
a communication link

 Spectrum Availability:
Wireless, Wire line/ Fiber spectrum limited due to
1. Channel frequency response or
2. Sharing by different users and/or applications

 Transmitted Power
1. Battery Power
2. Interference

 Channel Parameters
Length, Additive White Gaussian Noise Channel (AWGN)

5/18/2022
67
 Performance criteria
Analog Communication
• Infinite set of Waveforms
• Fidelity criteria such as Signal to Noise ratio, Percent
Distortion, Expected mean square error between
Transmitted and Received Waveform

Digital Communication
• Finite set of binary words
• Probability of error or BER-Bit Error Rate

5/18/2022
68
 PCM: Pulse Code Modulation
 Message signal is represented by a sequence of coded pulses, which is
accomplished by representing the signal in discrete form in both time and
amplitude.
 Basic operations of PCM transmitter
 Sampling

 Quantizing

 Encoding

 To prevent aliasing of message signal LPF is added to prior to the

sampling block

5/18/2022
69
 Pulse code modulation(contd…)

Basic operations of PCM Receiver are

Regeneration of impaired signal
Decoding
Reconstruction of the train of quantized
samples

5/18/2022
70
Basic operations in Pulse code modulation

 SAMPLING
 Sampling with a train of narrow pulses so as to closely approximate the
instantaneous sampling process.

 Follows Nyquist criterion: fs>=2W (Sampling Theorem)

 Low pass filter (anti aliasing filter) is used before sampling to limit the
bandwidth of signal to W.

5/18/2022
71
Basic operations in Pulse code modulation
 QUANTIZATION:
 Sampled value is then quantized.
 Gives discrete signal both in time
and amplitude.

 Amplitude quantization:
Rounding off the amplitude to
the nearest pre-decided voltage
level

 Process of transforming m(nTs)

(sampled version of m(t)at a time t=
nTs) into a discrete amplitude
υ(nTs).

5/18/2022
72
Encoding in PCM
 Each quantitation level is assigned with a N bit
long binary code
 These binary codes are represented in the form
of pulses for transmission

 Example: N=3 bits, L=23=8 levels

 Binary codes : 000, 001, 010, 011, 100, ………. 110, 111

5/18/2022 73
Basic operation in Pulse code modulation
 Memory less Quantizer
 No effect of earlier or later samples of message signal.

 Where L= total no of amplitude levels used in quantizer

5/18/2022
74
Uniform Distribution of L quantization
levels

5/18/2022
75
 Signaling Rate and
Transmission bandwidth in PCM

Signaling rate = No of bits per second

L = 2N or N= log2L
where,
L= total no. of quantization levels
N= no of bits per sample
No. of bits per second = (No. of bits per sample) x
(No. of samples per second)
= N x fs
Transmission Bandwidth= Signaling rate/2

5/18/2022
76
 Example on PCM Bit rate calculation

A television signal with a bandwidth of 4.2 MHz is transmitted

using binary PCM. The number of quantization levels used is
512.
Calculate i) Code word length
ii) Transmission Bandwidth
iii) Final bit rate
Answers: i) 9
ii) ≥ 37.8 MHz
iii) 75.6 x 106 Bits/second

5/18/2022
77
 Quantizer

Uniform Non Uniform

Representation
Representation
levels are levels are not
uniformly spaced uniformly spaced
but tapered

5/18/2022
78
 Uniform Quantizer

 Representation level are uniformly spaced over

the complete amplitude range.

 Two ways of implementation- It can be midtread

or midrise type

 Uniform quantizers are symmetric about origin.

5/18/2022
TY ECE, MIT WPU, Pune 79
Mid tread and Midrise quantizer characteristics
 Origin lies in the middle of a tread of staircase
 Origin lies in the middle of a rising part of staircase

5/18/2022
80
 Quantization Error
M(t)

V(t)

5/18/2022
81
 Quantization Error

 Difference between the input signal m and the output signal υ

 Let the quantization error be denoted by

or correspondingly

ASSUMPTION
 Input M(t) having zero mean and Quantizer is symmetric.

5/18/2022
82
Need of Non-uniform Quantization

 Quantization levels are not uniformly spaced

 The SQNR of usually low amplitude signal will be poor and unacceptable.
 Fine quantization of weak signals and coarse quantization of strong
signals.
 SQNR is maintained to acceptable level irrespective of input signal
level.
 Application: Telephone communication
 Range of voltages covered by voice signals, from the peaks of loud talk
(Shouting) to the weak passage of talk(whispering), is on the order of
1000 to 1.
 How to use nonuniform quantizer?
 Passing baseband signal through a compressor and then applying
compressed signal to a uniform quantizer.

5/18/2022
83
 Companding Curves
Compression at Tx end & Expansion at
Rx end

Combination of a Compressor and an expander

is called a Compandor.

5/18/2022
84
Uniform and Non uniform quantization

5/18/2022
85
Comparison of characteristics and
quantization error

5/18/2022
86
Non uniform quantization or Companding
 The two companding laws used are:
μ
law
 A law
 μ-law equation:

where m and are the normalized input and output voltages

respectively.
μ is a positive constant representing compression factor.
 Application: North America and Japan’s digital
communication system

5/18/2022
87
µ-law Compression characteristics
Slope of the
curve

Linear at low input levels

Logarithmic at high input levels

Mid Tread Quantizer

5/18/2022
88
Compression Coefficients- µ and A

5/18/2022 89
A-law Compression characteristics

Application:
European Digital
Communication System
Also used in India

Mid Rise Quantizer

5/18/2022
90
on uniform quantization using companding
A law of companding:
 v is the output and m is the input

5/18/2022
91
 Companding

 Expander: Used at receiver end

 Restore the signal samples to their corrective levels
 Complementary/inverse characteristics as compared to
compressor
 Expander output is equal to input of compressor.

 Combination of a Compressor and an expander is

called a Compandor.

5/18/2022
92
 Companding

dynamic range capability of the compander improves with

increasing μ and A respectively.

 To provide a reasonable SNR for both low and high level signals,
a compromise is usually made in choosing the value of parameter
μ and A .

 Typical values of compression coefficients are:

 μ = 255
A= 87.6

5/18/2022
93
 Encoding

 Make transmitted signal more robust to noise, interference and other

channel impairments.

 Code:
 Any plan for representing each of this discrete set of values as a particular
arrangement of discrete events.
 For ex: in a binary code, two distinct values are 0 and 1.
 Codeword consists of R bits.
 R denotes the no of bits per sample

 Using such a code, we may represent a total of 2R distinct levels.

5/18/2022
94
 Encoding
 Binary number system for R=4 bits/sample

5/18/2022
95
 3 bit PCM- Example

5/18/2022
96
5/18/2022 97
Regeneration: Detection of presence of a
pulse

Sampling
Instant

Use of chain of regenerative repeaters located at sufficiently close spacing

along the transmission route.

5/18/2022 98
 Regeneration
• First amplifies the signal
• Equalizer has a transfer function which is
close approximation of the inverse of the
channel transfer function.
• Equalizes shape (despread the signal) of
received pulse
• In other words, compensate for the effects of
amplitude and phase distortion produced by
the non ideal transmission characteristics of
channel.
• Timing Circuit:
Provides a periodic pulse train, derived from
the received pulses, for sampling the equalized
pulses at the instants of time where the SNR
is a maximum.
• Decision Making Device: To decide
whether the bit is 1 or 0,
• Consists of sample and hold cct and
comparator
• Sample and hold has 2 inputs, incoming
signal and Clock, Comparator has 2 inputs ,
o/p of sample and hold and threshold
voltage 5/18/2022 99
Regeneration

5/18/2022 100
Sources of degradation in PCM signal

Sources of Corruption

Sampling and Channel effects

quantizing effects

Quantization Noise
Channel Noise
Quantizer Saturation

Inter-symbol Interference
Timing jitter
SNR considering channel
noise and quantization
noise
5/18/2022
101
 Nq, Quantization Noise

 Inherent distortion in PCM, round off or truncation error.

 Process of encoding of the PAM signals into a quantized PAM signals

involves discarding some of the original analog information.

 Amount of Noise is inversely proportional to the no. of levels

employed in the quantization process.

 Nq=Quantization noise power=s2/12

 where s= step size

5/18/2022
102
Quantizer (A to D converter) saturation

 Allocates L levels for approximating the continuous range of

inputs to a finite a set of outputs.
 Operating range of converter: Range of inputs for which the
difference between the input and output is small.
 If the input exceeds this range, converter is operating in
saturation.

 REMEDIES
 Use of Automatic gain Control (AGC), which effectively extends
the operating range of the converter.

5/18/2022
103
 Timing Jitter

 Slight jitter in the position of the sample will result into non
uniform sampling.
 Although exact reconstruction is still possible if the sample
positions are accurately known, the jitter is usually random
process and thus the sample positions are not accurately known.

 REMEDIES

 Stable clock signal used as reference.

5/18/2022
104
 Channel Noise

 Errors in detecting the pulse carrying the digitized samples due to

 Thermal Noise
 Interference from other users

 Interference from switching transients

 Rapid degradation of output signal quality with channel induced

errors is called a THRESHOLD EFFECT.

 Large channel errors affect ability to detect the signal, the

resulting detection errors causes reconstruction errors.

5/18/2022
105
 Inter-symbol interference: ISI

 Channel is always band limited.

 Band limited channel disperses or spreads a pulse waveform passing
through it.
 When the channel BW is much greater than the pulse BW, the
spreading of the pulse will be slight.
 When the channel BW is close to signal BW , the spreading will exceed
a symbol duration and cause signal pulses to overlap.
 This overlapping is Inter-symbol Interference.

 ISI causes system degradation.

5/18/2022
106
Noise considerations in PCM system
 Performance of PCM system is influenced by two major
sources of noise.
1. Channel Noise
 Introduced between Tx and Rx.
 Always present once the equipment is switched on.
 Affects Average probability of symbol error, also referred as bit error
rate (BER).
 For optimizing system performance, need to minimize BER.
 channel noise as additive, white and Gaussian.
 Using regenerative repeater at short enough spacing, effect of
channel noise can be made practically negligible.
2. Quantization Noise
 Introduced in the transmitter.
 Reduced with increase in no. of quantization levels or step size
reduction
5/18/2022
107
 Decoding
 Clean pulses after regeneration are then regrouped into code words
 And decoded into a quantized PAM signals
 Generate amplitude of which is the linear sum of all the pulses in the code
word, with each pulse being weighted by its place value in the code.

Filtering
Pass decoder output through a low pass filter whose cut off frequency is equal to the
message bandwidth.

5/18/2022
108
 PCM Transmitter

v= 3 bits/sample
Binary M=2;
levels = 2^3 = 8

4-ary system : M = 4
V =3 bits/sample
Levels = 4^3 = 64

Antialiasing filter :
bandlimits the
signal
Ts= 1/fs

5/18/2022
109
 Quantization Noise and PCM performance
 Quantization error is represented by

 is mean square Quantization noise.

 Furthermore

 Quantization error has zero mean and its PDF is

 Quantization Noise power is given by

5/18/2022
110
 SNR in PCM
 Performance in terms of destination signal power

Analog message be a voltage waveform normalized such that

Now signal to Noise ratio becomes

5/18/2022
111
PCM Signal to Noise Ratio (SNR) in dB
 For sinusoidal input signal, the output SNR in dB is
given by
(SNR)o= 1.8+6v
Where v is no. of bits/sample

 For Non-sinusoidal input signal, the output SNR in dB

is given by
(SNR)o=4.8+6v
Where v is no. of bits/sample

5/18/2022 112
 PCM System Limitations

 Increased system complexity

 Can be solved by available and/or custom made VLSI
chips

 Increased Bandwidth
 BW=N*fs/2
 Bit rate = N*fs
BW = bit rate/2
5/18/2022
113
 Differential pulse code modulation
(DPCM)
High degree of correlation between adjacent samples in sampled
signal.
Smaller variance between adjacent samples compares to variance of
signal itself.

 Why DPCM ?
Redundant information in PCM
More Bandwidth requirement of PCM
Basic idea behind DPCM is to remove redundancy
Fewer levels are required to quantize the successive difference
Fewer bits will be needed for encoding and saving in bandwidth

5/18/2022
114
Differential pulse code modulation (DPCM)

Baseband signal x(t) sampled at rate fs

Results into discrete signal{m[n]} with highly Correlated
Samples
Difference between adjacent samples has a smaller variance than
the variance of signal itself.
Working on “Prediction” principle.
Past behavior may be used to make an estimate of future value of
the signal.

5/18/2022
115
Differential pulse code modulation (DPCM)
 DPCM Transmitter

Input signal to quantizer is defined by

 where
e[n]= prediction error or difference signal
Where m[n]= unquantized input sample, mˆ[n]= predicted sample at n

5/18/2022
116
Differential pulse code modulation (DPCM)

5/18/2022
117
Differential pulse code modulation (DPCM)

 Quantizer output may be expressed as

Where q[n]= quantization error

5/18/2022
118
 Cont…..

Put value of in above eq.

5/18/2022
119
 Cont…..

From error,

Which represent quantized version of the input sample m[n]

5/18/2022
120
Differential pulse code modulation (DPCM)
 DPCM receiver

 Reconstruction of quantized error signal by decoder

 Same prediction filter as Tx end is used
 Give quantized version of original input signal
 In absence of channel noise, encoded at Rx=encoded at Tx
 Accordingly, Rx o/p = mq[n].

5/18/2022
121
5/18/2022 122
5/18/2022 123
 Delta Modulation (DM)

Special case of DPCM

Incoming message signal is oversampled

Purposely increase correlation between adjacent samples.

DM provides staircase approximation for oversampled input

signal

Difference between the input and the approximation is quantized

into only two levels, namely delta, ± Δ.

5/18/2022
124
 Delta Modulation

5/18/2022
125
 Delta Modulation
 DM Transmitter

5/18/2022
126
 Delta Modulation
 We may write

 Difference between the presents sample of input signal and latest

approximation of it.

 Where eq [n] is quantized version of e[n]

 Finally quantizer output is encoded

 Rate of transmission of Information is equal to Sampling rate fs

5/18/2022
127
 Delta Modulation
 DM receiver

Staircase approximation obtained from decoder o/p.

Accumulator adds up received quantized difference
LPF remove Out of band quantization noise

5/18/2022
128
Disadvantages of Delta Modulation

 Errors in DM
1) Slope overload distortion
2) Granular noise
3) Startup error

o We may write quantized output sample

o We observe from

5/18/2022
129
 Slope overload distortion

Following condition must be satisfied in a region of maximum

slope

But Δ is constant and small

5/18/2022
130
 Let x(t) = A coswt
 dx/dt = -wAsinwt
 Condition for no overload
 Max. value of Slope < 𝛥/Ts
 Or Max. value of Slope = wA < 𝛥fs
 Maximum value of signal Amax that can be tolerated
without overload is
 Amax = 𝛥fs / w ;
 w = 2π fm

131
 Granular noise

Occurs when step size is too large relative to slope characteristics

of input
As a result staircase approximation to hunt around a flat segment
of the input waveform

5/18/2022
132
 Practice Problems

1. A TV signal has BW of 4.5 MHz . This signal is sampled , quantized &

binary coded to obtain a PCM signal. What will be the sampling rate if
sampled at 20% above the Nyquist rate? If the samples are quantized into
1024 levels, find the number of pulses to encode each sample. What is the
binary pulse rate generated? Find the minimum transmission BW required to
transmit the signal.

2. A DM system is designed to operate at 3 times the Nyquist rate for a signal

with a 3 KHz bandwidth. The quantization step size =250mV.
 Find the maximum amplitude of a 1 KHz input sinusoid for which the
delta modulator does not show slope overload.

5/18/2022 133
 Solution for Problem 1
PCM System

 Fs= 20% more than Nyquist rate , 2*fm

 Fs=1.2* 2*4.5*106 = 10.8MHz

 L= 1024, L=2 N
Find N, N= 10 bits
Binary Pulse rate= N* Fs =108 Mbits/s
Transmission Bandwidth= N* Fs/2= 54MHz

5/18/2022 134
Solution for Problem 2: DM System
BW= 3.3 KHz ,Fm = 1 KHz, 𝛥 = 250 mV
Fs= 3 * (2* 3.3 KHz) = 19.8 KHz
Amax = 𝛥*Fs/2*π * fm
A < = 𝛥fs / w ;for no slope overload
Amax = 𝛥fs / w ; for no slope overload
= 𝛥fs / 2π fm
=(250 * 10-3 * 19.8 * 103 )/( 2* π* 1 * 103 )
Amax. = 0.787 V
SNR in DM system

5/18/2022 136
SNR in DM system

5/18/2022 137
DM Problem
 Given : Bitrate= N*fs= 20 Kbps, Input speech bandwidth 3.4KHz, A=1V
 Find step size delta.

 with N=1, Fs= 20 KHz

 Delta >=2.Pi 3.4KHz.1/20Khz = 1.068V

5/18/2022 138
DM SNR Problem
 Given: Sampling frequency, Fs=30KHz
 Input signal frequency, Fa=2KHz
 LPF Bandwidth at Rx , B=4KHz
 Find the output SNR in dB.

 Answer:
 SNR=64.1 (unitless ratio)
 SNR= 10 log10 (64.1) = 18.07 dB

5/18/2022 139
 Comparison of DPCM & DM

Difference:
Replacement of prediction filter by a single delay element
in DM
(zero order prediction filter)

SIMILARITIES

Both involve the use of feedback

5/18/2022
140
DM Example: errors

Disadvantages of delta modulation can be overcome by two

methods:
• Delta Sigma modulation
• Adaptive delta modulation
5/18/2022 141
 ADM
Why Adaptive?
 There is a need to have a large step size to accommodate a wide dynamic
range of input signal,
 whereas a small step size is required for the accurate representation of
relatively low-level signals.
 It is therefore clear that if we are to choose an optimum step size that
minimizes the average power of the quantization error in a delta
modulator, we need to make the DM system adaptive.
 This requirement, in turn, means that the step size has to vary in
accordance with the incoming message signal.

5/18/2022 142
 Adaptive delta modulation (ADM)

 Additional hardware and logic is designed to

provide variable- adaptive- step size
 Signal to Noise ratio for ADM is typically 8-14dB
better than DM
 Start up error and Slope overload error are
reduced

5/18/2022
143
 ADM
Transmitter

5/18/2022
144
 ADM Algorithm

 Figure above portrays the action of an ADM transmitter in which the step
size in the feedback loop is adjusted by a variable gain g(k) such that

The step-size controller carries out the adjustment algorithm

where K is a constant taken to be in the range 1 < K < 2 . Thus, the

effective step size increases by successive powers of K during slope-
overload conditions
5/18/2022 145
 ADM System
Waveforms

Example for K=1.5

1 1 1 1 1 1 0 1 0 1 0 0 0 1

ADM signal

5/18/2022
146
Adaptive Delta Modulation

5/18/2022 147
5/18/2022 148
Adaptive Step Size Algorithm in ADM system
Logic for step size control used in the digital
processor

5/18/2022 149
Advantages of ADM over DM
• Reduction in slope overload and granular noise
• Improvement in SNR
• Wide dynamic range due to variable step size
• Better BW utilization than DM (similar to the effect of
companding in PCM)
• Low signalling rate (1 bit /sample )

150