Unit - 1 : Chapter - 1

Waveform Coding Techniques

School of Electronics and Communication Engineering

1
 Message
Signal

Analog signal Digital Signal

An analog signal is one in which both In a digital signal, the information

amplitudes and time vary continuously bearing signal is processed so that
over their respective intervals. A it can be represented in both
speech signal, a television signal, and a amplitude and time as discrete
signal representing atmospheric values. Computer data and
temperature or pressure at some telegraph signals are examples of
location are examples of analog signals digital signals.

2
The purpose of a communication system is to transport an information-
bearing signal from a source to a user destination via a communication
channel.
Communication
System

Analog CS Digital CS

In an ACS, the information-bearing In a DCS, the information bearing

signal is continuously varying in both signal is processed so that it can be
amplitude and time, and it is used represented by a sequence of
directly to modify some characteristic discrete messages.
of a sinusoidal carrier wave, such as
amplitude, phase, or frequency. 3
The discipline of digital communications has experienced a phenomenal
growth in both scope and application. The growth of digital communications is
largely due to the following factors:

1. The impact of the computer, not only as a source of data but also as a tool
for communications, and the demands other digital services such as telex and
facsimile.

2. The use of digital communications offers flexibility and compatibility in that

the adoption of a common digital format makes it possible for a transmission
system to sustain many different sources of information in a flexible manner.

3. The improved reliability made possible by use of digital communications.

4. The availability of wide-band channels provided by geostationary satellites,
optical fibers, and coaxial cables.

5. The ever-increasing availability of integrated solid-state electronic

technology, which has made it possible to increase system complexity by
orders of magnitude in a cost-effective manner.

Indeed, the trend toward digital communications (and away from analog
communications) will continue, so much so that the second half of the
twentieth century will be recorded in history as the era of digital
communications.
A source of information generates a message
Ex: human voice
television picture
teletype data
atmospheric temperature and pressure.

In these examples, the message is not electrical in nature, and so a

transducer is used to convert it into an electrical waveform called the
message signal. The waveform is also referred to as a baseband signal; the
term "baseband" is used to designate the band of frequencies
representing the message signal generated at the source.
Block diagram of a digital communication system
In source coding, the encoder maps the digital signal generated at the
source output into another signal in digital form.

The mapping is one-to-one, and the objective is to eliminate or reduce

redundancy so as to provide an efficient representation of the source
output.

Since the source encoder mapping is one- to-one, the source decoder
simply performs the inverse mapping and thereby delivers to the user
destination a reproduction of the original digital source output.

The primary benefit thus gained from the application of source coding is a
reduced bandwidth requement.
In channel coding, the objective is for the encoder to map the incoming
digital signal into a channel input and for the decoder to map the channel
output into an output digital signal in such a way that the effect of channel
noise is minimized.

That is, the combined role of the channel encoder and decoder is to provide
reliable communication over a noisy channel.

This provision is satisfied by introducing redundancy in a prescribed fashion

in the channel encoder and exploiting it in the decoder to reconstruct the
original encoder input as accurately as possible.

Thus, in source coding, we remove redundancy, whereas in channel coding,

we introduce controlled redundancy.
Clearly, we may perform source coding alone, channel coding alone, or the
two together.

In the latter case, naturally, the source encoding is performed first, followed
by channel encoding in the transmitter as illustrated in Fig.

In the receiver we proceed in the reverse order: channel decoding is

performed first, followed by source decoding.
Modulation, it is performed with the purpose of providing for the efficient
transmission of the signal over the channel.

In particular, the modulator operates by key shifts in the amplitude, frequency,

or phase of a sinusoidal carrier wave to the channel encoder output.

The digital modulation techniques referred to as amplitude-shift keying,

frequency-shift keying, or phase-shift keying, respectively.

The detector performs demodulation (the inverse of modulation), thereby

producing a signal that follows the time variations in the channel encoder
output (except for the effects of noise).
The combination of modulator, channel, and detector, enclosed inside the
dashed rectangle shown in Fig., is called a discrete channel. It is so called since
both its input and output signals are in discrete form.

Traditionally, coding and modulation are performed as separate operations, and

the introduction of redundant symbols by the channel encoder appears to
imply increased transmission bandwidth.

In some applications, however, these, two operations are performed as one

function in such a way that the transmission bandwidth need not be increased.
Pulse Modulation:

When any one characteristics (such as amplitude, width or position) of a

relatively higher frequency carrier signal comprising of discrete pulses is
varied in accordance with the amplitude of the analog modulating signal, it is
called pulse modulation.

• To transmit an analog information data using digital signals & digital-

transmission techniques, pulse modulation is necessary.
• Pulse modulation essentially consists of sampling analog information data
at regular intervals and then converting those samples into discrete
pulses.
• These pulses can then be transmitted from a source to a destination over
a physical communication channel.
 PAM - Pulse amplitude modulation
PTM - Pulse time modulation
PWM - Pulse width modulation
PPM - Pulse position modulation
Fig: Classification of pulse PCM - Pulse code modulation
modulation techniques DPCM - Differential pulse code modulation
DM - Delta modulation
ADM - Adaptive delta modulation
CVSDM - Continuous variable slope delta modulation
Pulse Code Modulation (PCM):

A sampled analog message signal needs to be represented in digital form in

the process of digital transmission. PCM provides a method to convert the
analog signal in to digital singal.

• In PCM method of signal coding, the message signal is sampled and the
amplitude of each sample is rounded off (approximated) to the nearest
one of a finite set of discrete levels, so that both time and amplitude are
represented in discrete form.
• This allows the message to be transmitted by means of a digital (coded)
waveform, thereby distinguishing pulse-code modulation from all analog
modulation techniques.
• With PCM, the continuously-varying analog signals are converted into
pulses of fixed amplitude and fixed duration.
• PCM is a binary system where a pulse or no pulse within a prescribed time
slot represents either a logic 1 or logic 0 condition.
• Importantly, PCM is a digital pulse modulation technique also known as
time domain waveform coding technique.
• It is the only digitally encoded modulation technique used for baseband
digital transmission.
Where PCM is used ?

• PCM is used in digital telephone systems (trunk lines).

• It is also the standard form for digital audio in computers and various
compact disc formats, digital videos, etc.
• PCM is the preferred method of communications within Public Switched
Telephone Network (PSTN) because with PCM it is easy to combine
digitized voice and digital data into a single, high-speed digital signal and
transmit it over either coaxial or optical fiber cables.
The basic operations performed in the transmitter of a PCM system in the
process of converting analog signal into digital form are:
1. Sampling
2. Quantizing
3. Encoding, as shown in the block diagram

Analog Digital
Signal Signal
Sampling Quantizing Encoding

Fig: Analog to digital Conversion

Fig: The basic elements of a PCM system.
PCM Transmitter

LPF  pre-alias filter, exclude frequencies > fm (W)

Sampling of x(t) with narrow rectangular pulse train
fs ≥ 2fm for perfect reconstruction
Signal is sampled, each sample amplitude is rounded off (Quantization)
Discretized in both time and amplitude
Quantized samples are binary coded
continuous time message signal  as a sequence of binary coded pulses
 Regenerative Repeater

Important feature of PCM systems

Control the effects of distortion and noise introduced by the channel
Performs 3 basic functions: Equalization, Timing and Decision making
Equalizer – Pulse shaping
Timing circuit – provides periodic pulse train
Sampling of equalized pulses
Decision device is enabled at the sampling times
Makes decision based on whether the amplitude of the quantized pulse
plus noise exceeds a predetermined threshold
PCM Receiver

--------------------------------------------- OR
--------------------------------------------------
The LPF is included to prevent aliasing of the message signal.

The quantizing and encoding operations are usually performed in the same
circuit, which is called an analog-to-digital converter.

The basic operations in the receiver are regeneration of impaired signals,

decoding, and reconstruction of the train of quantized samples.

Regeneration also occurs at intermediate points along the transmission

path as necessary.

When time-division multiplexing is used, it becomes necessary to

synchronize the receiver to the transmitter for the overall system to operate
satisfactorily.
Analog
Signal Sampled/Discrete
Signal

Quantized
Signal

Digital
Signal
In the sampling operation, only sample values of the analog signal at uniformly
spaced discrete instants of time are retained.

In the quantizing operation, each sample value is approximated by the nearest

level in a finite set of discrete levels.

In the encoding operation, the selected level is represented by a code word

that consists of a prescribed number of code elements.
Part (a) of the figure
shows a segment of
an analog
waveform.

Part (b) shows the

corresponding
digital waveform,
based on the use of
a binary code.
In this example, symbols 0 and 1 of the binary code are represented by zero and
1 volt, respectively.

The code word consists of four binary digits (bits), with the last bit assigned the
role of a sign bit that signifies whether the sample value is positive or negative.

The remaining three bits are chosen to provide a numerical representation for
the absolute value of a sample in accordance with Table.
Binary representation of quantized levels
As a result of sampling and quantizing operations, errors are introduced into
the digital signal.

These errors are nonreversible in that it is not possible to produce an exact

replica of the original analog signal from its digital representation.

However, the errors are under a designer's control. Indeed, by proper

selections of the sampling rate and code-word length (i.e., number of
quantizing levels)

The errors due to sampling and quantizing can be made so small that the
difference between the analog signal and its digital reconstruction is not
discernible by a human observer.
Sampling:
The sampling operation is performed in accordance with the sampling theorem. Specifically,
we may state the sampling theorem for band-limited signals of finite energy in two equivalent
parts.

1. A band-limited signal of finite energy, which has no frequency components higher than W
hertz, is completely described by specifying the values of the signal at instants of time
separated by seconds.

2. A band-limited signal of finite energy, which has no frequency components higher than W
hertz, may be completely recovered from a knowledge of its samples taken at the rate of 2W
per second.

Part 1 of the sampling theorem is exploited in the transmitter; part 2 of the theorem is
exploited in the receiver. The sampling rate 2W is called the Nyquist rate, and its reciprocal is
called the Nyquist interval.
The sampling process is usually described in the time domain. It is an
operation that is basic to digital signal processing and digital
communications.

Using the sampling process, an analog signal is converted into a

corresponding sequence of samples that are usually spaced uniformly in
time.

Clearly, for such a procedure to have practical utility, it is necessary that we

choose the sampling rate properly, so that the sequence of samples uniquely
defines the original analog signal. This is the essence of the sampling
theorem
Consider an arbitrary signal g(t) of finite energy, which is specified for all
time. A segment of the signal g(t) is shown in Fig. (a).

Suppose that we sample the signal g(t) instantaneously and at a uniform

rate, once every Ts seconds.

Consequently, we obtain an infinite sequence of samples spaced Ts seconds

apart and denoted by {g(nTs)}, where n takes on all possible integer values.

We refer to Ts as the sampling period, and to its reciprocal fs = as the

sampling rate. This ideal form of sampling is called instantaneous
sampling.
 Fig. (a)

FIGURE The sampling process

(a) Analog signal
(b)Instantaneously sampled version of
the signal.

Fig. (b)
Let gδ(t) denote the signal obtained by individually weighting the elements of
a periodic sequence of Dirac delta functions spaced Ts seconds apart by the
sequence of numbers { g(nTs) }, as shown by (see Fig. b)

We refer to gδ(t) as the ideal sampled signal. The term δ(t - nTs) represents a
delta function positioned at time t = nTs.
From the definition of the delta function, we recall that such an idealized
function has unit area.

We may therefore view the multiplying factor g(nTs) in the above Eq. as a
“mass” assigned to the delta function δ(t - nTs).

A delta function weighted in this manner is closely approximated by a

rectangular pulse of duration Δt and amplitude ; the smaller we make Δt, the
better will be the approximation.
The ideal sampled signal gδ(t) has a mathematical form similar to that of the
Fourier transform of a periodic signal.

This correspondence suggests that we may determine the Fourier transform

of the ideal sampled signal gδ(t) by applying the duality property of the
Fourier transform to the transform pair.

By applying, and using the fact that a delta function is an even function of
time, we get the result:

where G(f) is the Fourier transform of the original signal g(t), and f s is the
sampling rate. Above Eq. states that the process of uniformly sampling a
continuous-time signal of finite energy results in a periodic spectrum with a
period equal to the sampling rate.
Problem 1
Find the nyquist rate and nyquist interval for following signals.

Solution:
Solution Contd. :
Types of Sampling/PAM
Types of Sampling/PAM
 Fig. (a)

Figure: (a) Analog Signal

(b) Sampling Function
(c) Sampled Signal

Fig. (b)

Fig. (c)
Figure: Flat-top Samples
Quantizing/Quantization:

The conversion of an analog (continuous) sample of the signal into a digital

(discrete) form is called the quantizing process.

Graphically, the quantizing process means that a straight line representing

the relation between the input and the output of a linear analog system is
replaced by a transfer characteristic that is staircase-like in appearance.

Following figure depicts one such characteristic.

The quantizing process has a two-fold effect:

(1) The peak-to-peak range of input sample values is subdivided into a finite
set of decision levels or decision thresholds that are aligned with the "risers"
of the staircase

(2) The output is assigned a discrete value selected from a finite set of
representation levels or reconstruction values that are aligned with the
"treads" of the staircase.

In the case of a uniform quantizer, characterized as in previous Fig., the

separation between the decision thresholds and the separation between the
representation levels of the quantizer have a common value called the step
size.
Quantization Process: process of transforming Sampled amplitude values of a message signal into
a discrete amplitude value
Types of Quantizers:
1. Uniform Quantizer  the quantization levels are
uniformly spaced,

2. Non- Uniform Quantizer

the spacing between the levels will be unequal and

mostly the relation is logarithmic
 Types of Quantizer
 
Uniform Quantizer  Mid Tread and Mid Rise
Mid-tread type - the origin lies in the middle of a tread of the stair-case
like graph
 The quantization levels in this type are odd in number

Mid-Rise type - the origin lies in the middle of a raising part

The quantization levels in this type are even in number

Both - symmetric about the origin.

Midtread
Quantizer
Midriser
Quantizer
Nonuniform Quantization:

In non-uniform quantization, the spacing between the quantization levels

is not uniform and step size varies in accordance with the relative
amplitude level of the sampled value.

• In the use of PCM for the transmission of speech signals, the quantizer
has to accommodate input signals with widely varying power levels.
• For example, the range of voltage amplitude levels covered by normal
speech signals, from the peaks of loud speech levels to the lows of weak
speech levels, is on the order of 1000 to 1.
• So it is highly desirable that signal-to-quantization noise ratio should
remain essentially constant for a wide range of input signal levels.
• This requirement is met by a non-uniform quantizer, as explained with
the help of following example.
Example: Consider a 4 bit PCM coded system. The normalized peak-to-
peak input voltage range is ±16 V for a uniform quantizer. Justify that
non-uniform quantization would have yielded better results.

Solution:
For given 4 bit PCM coded system, number of levels, M = 24 = 16 levels

The normalized peak-to-peak input voltage range = ±16 V or 32 V (given)

For uniform quantization process, step size, Δ = 32/16 = 2 V

The maximum quantization error, Qemax = Δ/2 = 1 V

For input signal amplitude level of 2 V (say), Qemax = 1 V, i.e. 50% of input
voltage 59
For input signal amplitude level of 3 V (say), Qemax = 1 V, i.e. 33% of input
voltage

For input signal amplitude level of 15 V (say), Qemax = 1 V, i.e. 6.7% of input
voltage

For input signal amplitude level of 16 V (say), Qemax = 1 V, i.e. 6.3% of input
voltage

That means, for low-signal amplitude levels, the maximum quantization error
is relatively much higher than that for high-signal amplitude levels. This is
because of uniform quantization.
Advantages of Nonuniform Quantization:

• High Average SNR value Non-uniform quantization has higher average

signal to quantization noise power ratio value than that of in the uniform
quantizer.
• Reduced Quantization Noise RMS value of the quantizer noise power of a
non-uniform quantizer is substantially proportional to the sampled value
and hence the quantization noise is also reduced.
Robust Quantization:

• A quantizer whose SNR remains essentially constant for a wide range of

input power levels is said to be robust.
• This necessitates that the stepsize must be small for low amplitude
signals and large for high amplitude signals.
• The provision for such robust performance necessitates the use of a non-
uniform quantizer.
• The non-uniform quantization technique employs an additional
logarithmic amplifier before processing the sampled speech signals by a
uniform quantizer.
• The operation of a non-uniform quantizer is equivalent to passing the analog signal
through a compressor and then applying the compressed signal to a uniform quantizer
at transmitter end.
• At the receiver, a device with a characteristic complementary to the compressor,
called expander is used to restore the signal samples to their correct relative level.
• The combination of a compressor and an expander is called a compander.

Transmitter Receiver
Input Output
Uniform
Compressor Expander
Quantizer

Fig: Nonuniform Quantizer

Nonuniform Quantizer
μ-law companding:

In the μ-law companding, the compressor characteristics are continuous,

approximating a linear dependence for low input levels and a logarithmic
one for high input levels. The compression characteristics for μ-law is given
as

65
• The μ = 0 corresponds to uniform quantization.
• For a relatively constant signal-to-quantization ratio and a 40 dB
dynamic range, the value of μ ≥ 100 is required.
• The practical value of μ is approximately 255.
Example: An analog information signal at the input to a µ-law compressor (μ
= 255) is positive, with its voltage level one-half the maximum value. What
proportion of the maximum output voltage level would be produced at the
output of compressor?
• Thus, it is observed that 87.6% of the maximum output voltage level would be
produced at the output of µ-law compressor having μ = 255.
• Figure shows the comparison of S/N versus signal level for both cases – without
companding and with µ-law companding.
• It is observed that the signal-to-noise (S/N) ratio of PCM system remains constant
with companding
A-law companding:

The compression characteristics for A-law is given as

where Vout is the compressed output amplitude

level in volts
Vmax is the maximum uncompressed analog
input amplitude level in volts
A is the parameter used to define the amount
of compression (unit less)
Vin is the amplitude of input signal at a
particular instant of time in volts
Figure shows the relationship between Vout and Vin for different values of A.
• The value A = 1 corresponds to uniform quantization.
• The practical value of A is approximately 100.
Example: An analog information signal at the input to A-law compressor (A
= 100) is positive, with its voltage level one-half the maximum value. What
proportion of the maximum output voltage level would be produced at the
output of compressor?
Firstly, we have to determine which condition is applicable. For given value
of A = 100, and Vin = 0.5 Vmax, we observe that second condition is true
(0.01 ≤ 0.5 ≤ 1). Therefore,

Thus, it is observed that 87.6% of the maximum output voltage level would
be produced at the output of A-law compressor having A = 100, which is
exactly same as that in µ-law compressor having μ = 255.
Encoding:

To exploit the advantages of sampling and quantizing for the purpose of

making the transmitted signal more robust to noise, interference, and other
channel degradations, we require the encoding process to translate the
discrete set of sample values to a more appropriate form of signal.

Any plan that represents each of this discrete set of values as a particular
arrangement of discrete events is called a code. One of the discrete events in
a code is called a code element or symbol.

The presence or absence of a pulse is a symbol. A particular arrangement of

symbols used in a code to represent a single value of the discrete set is
called a code word or character.
The maximum advantage over the effects of noise in a transmission medium
is obtained by using a binary code, because a binary symbol withstands a
relatively high level of noise and is easy to regenerate.

A line code electrically represents a binary stream of data. Any one of

several line codes can be used for the electrical representation of a binary
data stream.

The lines codes often used the terminology nonreturn-to-zero (NRZ) or

return-to-zero (RZ).

Return-to-zero implies that the pulse shape used to represent the bit
always returns to the 0 volts or the neutral level before the end of the bit.
Nonreturn-to-zero indicates that the pulse does not necessarily return to
the neutral level before the end of the bit.
1. Unipolar Nonreturn-to-Zero (NRZ) Signaling:

In this line code, symbol 1 is represented by transmitting a pulse of

amplitude A for the duration of the symbol, and symbol 0 is represented by
switching off the pulse, as in Figure. This line code is also referred to as on-
off signaling. A disadvantage of on-off signaling is the waste of power due to
the transmitted DC level.
2. Polar Nonreturn-to-Zero (NRZ) Signaling:

In this second line code, symbols 1 and 0 are represented by transmitting

pulses of amplitudes +A and -A, respectively, as illustrated in Figure. This
line code is relatively easy to generate and is more power-efficient than its
unipolar counterpart.
3. Unipolar Return-to-Zero (RZ) Signaling:

In this other line code, symbol 1 is represented by a rectangular pulse of

amplitude A and half-symbol width, and symbol 0 is represented by
transmitting no pulse, as illustrated in Figure.
An attractive feature of this line code is it can be used for bit-timing recovery
at the receiver. However, its disadvantage is that it requires 3 dB more power
than polar return-to-zero signaling for the same probability of symbol error.
4. Bipolar Return-to-Zero (BRZ) Signaling:

This line code uses three amplitude levels as indicated in Figure. Specifically,
positive and negative pulses of equal amplitude (i.e., +A and -A) are used
alternately for symbol 1, with each pulse having a half-symbol width. No
pulse is always used for symbol 0.

This line code is also called alternate mark inversion (AMI) signaling.
5. Split-Phase (Manchester Code):

In this method of signaling, illustrated in Figure, symbol 1 is represented by

a positive pulse of amplitude A followed by a negative pulse of amplitude -A,
with both pulses being a half-symbol wide. For symbol 0, the polarities of
these two pulses are reversed.
5. Differential Encoding:

This method is used to encode information in terms of signal transitions. In

particular, a transition is used to designate symbol 0 in the incoming binary
data stream, while no transition is used to designate symbol 1, as illustrated
in figure assuming the use of unipolar nonretum-to-zero signaling.
FIG: Illustration of the quantization process (Adapted from Bennett, 1948, with
permission of AT&T.) of a uniform quantizer of the midtread type.
 SQNR (Linear Quantization)
In PCM, there is quantization error due to the inherent errors introduced into the
signal
The step-size of the quantizer is given by

Where L = 2R indicates the total number of levels for R bits.

For a uniform quantizer, the quantization error Q e or ε will have its sample values
bounded by

If the step-size Δ is sufficiently small, it is reasonable to assume that the quantization

error Qe is a uniformly distributed random variable, and the interfering effect of the
quantization noise on the quantizer input is similar to that of thermal noise.
We may thus express the probability density function of the quantization error
Qe as follows:

With the mean of the quantization error being zero, its variance is the same as
the mean-square value
 SQNR (Linear Quantization)

Substituting fQ(q) in the above equation, we get as

Let R denote the number of bits per sample used in the construction of the
binary code. We may then write

Substituting L=2R we rewrite the step size as follows

Substituting this Δ in then rewriting as follows

We express the output signal-to-noise ratio (SNRO) of a uniform quantizer as

𝑆 𝑁𝑜𝑟𝑚𝑎𝑙𝑖𝑧𝑒𝑑 𝑠𝑖𝑔𝑛𝑎𝑙𝑝𝑜𝑤𝑒𝑟
=𝑆𝑁𝑅 𝑂 =
𝑁 𝑁𝑜𝑟𝑚𝑎𝑙𝑖𝑧𝑒𝑑 𝑛𝑜𝑖𝑠𝑒 𝑝𝑜𝑤𝑒𝑟

Let P denote the average/normalized power of the message signal m(t), and
denote the normalized noise power, then we write the SNRO as follows
If input signal is normalized to minimum and maximum value
1, then

𝑚𝑚𝑎𝑥 =1
We assume that x(t) is normalized

𝑆 2𝑅
Then, =𝑆𝑁𝑅 𝑂 =3 × 𝑃 × 2
𝑁

If signal power is also normalized, i.e., P ≤ 1

Then, 𝑆 2𝑅
=𝑆𝑁𝑅 𝑂 =3 ×2
𝑁

(𝑁) (𝑁)
For
𝑆 normalized values of power, SNR𝑆
O in decibels is, 2𝑅
=( 𝑆𝑁𝑅𝑂 )𝑑𝐵 =10 𝑙𝑜𝑔10 𝑑𝐵=10 𝑙𝑜𝑔10 (3 × 2 )=( 4 . 8+6 𝑅) 𝑑
𝑑𝐵
SQNR for Sinusoidal Wave:

Then

( )
𝑆
𝑁 𝑑𝐵
=10 𝑙𝑜𝑔10 ( )
𝑆
𝑁
3
𝑑𝐵=10 𝑙𝑜𝑔10 ( ×22 𝑅 )=(1. 8+ 6 𝑅) 𝑑𝐵
2
89
Transmission Bandwidth in PCM:

Let the quantizer use R bits to represent each level, then the number of
levels are L=2R for a L-level quantizer.

Each sample is converted to R bits, i.e., No. of bits per sample = R

We know that, the No. of sample per second = fs

Therefore, the No. of bits per second

= No. of bits per sample * No. of samples per second.
= R * fs

Here the No. of bits per second is also called signaling rate of PCM, r
The signaling rate of PCM, r = R fs ; here fs ≥ 2W

The transmission bandwidth in PCM system should be greater than or

equal to half of the signaling rate.

Since, fs ≥ 2W

𝐵𝑊 𝑃𝐶𝑀 ≥ 𝑅𝑊
Problem:
Solution
Cont.….
Problem: An audio signal is required to be digitally transmitted with a
sampling rate of 40 kHz and 14 bits per sample using linear PCM system.
Calculate the minimum transmission data rate needed in the
communications channel.

Solution: We know that minimum transmission data rate in PCM system is

given by r = R fs

For given R = 14 bits, and fs = 40 kHz, we get

r = 14 × 40 kHz = 560 kbps

95
Problem: Discrete samples of an analog signal are uniformly quantized to
PCM. If maximum value of analog sample is to be represented within 0.1%
accuracy, find minimum number of binary digits required per sample.

Solution: For given accuracy of 0.1%, Δ/2 = (0.1/100) × V ….. where Δ is

the uniform step-size, and V is the maximum value of the analog sample.
or, Δ = 0.002 V

We know that Δ = 2 V/L ….. where L is the number of quantization levels,

and the discrete samples are quantized from – V to +V.
Therefore, 0.002 V = 2 V/L or, L = 1000
We know that L = 2R ….. where R is the minimum number of binary digits per
sample.
Therefore, 1000 = 2R or R = 10
Hence, minimum number of binary digits required per sample = 10 bits96
Problem: An audio signal, s(t) = 3 cos (2π 500 t) is quantized using 10 bit
PCM. Determine the signal-to-quantization noise ratio.

Solution: We know that

Average signal power,

Here Am = 3V, therefore P =

Total swing of the signal = [Am – (–Am) ] = 2 Am = 2 × 3 = 6 volts
For 10 bit PCM, the step size, Δ = = 5.86 × 10– 3

The Quantization noise, =

97
The signal-to-quantization noise ratio,
Problem: Consider an analog input signal to PCM whose bandwidth is
limited to 4 kHz and varies in amplitude from – 3.8 V to + 3.8 V, with an
average power of 30 mW. The required signal-to-quantization error ratio is
given to be 20 dB. Assuming uniform quantization, determine the number of
bits required per sample.

Solution: Given - the average signal power, P = 30 mW

The signal to quantization noise ratio, = 20 dB

The signal to quantization noise ratio,

The Quantization noise, = = 0.3 mW

99
Therefore,

We know that

Then the number of bits required per sample, R = 7 bits

Channel noise and Error Probability:

• In addition to the quantization noise, the channel noise also affects the
performance of the PCM system.
• Channel noise may get introduced anywhere along the transmission path
between transmitter and receiver.
• Due to channel noise, the receiver may not be able to reconstruct the
signal exactly all the time.
• As a result, a binary signal 0 may be decoded as binary signal 1 or vice
versa.
• Such type of errors must be minimized so as to improve the performance
of PCM system which is measured in terms of error probability.
Error Probability:

The error probability or probability of error or error rate is defined as

the probability that the symbol at the receiver output (0 or 1) differs from
the transmitted.

The error probability, Pe of PCM system is given in terms of the

complementary error function erfc as

where Pmax is the maximum or peak signal power, Tb is the bit duration, and
N0 is noise spectral density.
Rewriting the above equation

The term N0/Tb is referred as the average noise power contained in a

transmission bandwidth equal to the bit rate, fb = 1/Tb.

where Emax = Pmax Tb is the peak signal energy.

• The complementary error function erfc is a monotonically decreasing
function which means that

will decrease with increase in the ratio

• Therefore, a very small increase in transmitted signal energy or power

will enable the reception of binary signal almost free of any error. For a bit
error rate of 10-5 (1 bit in every bits), there is an error threshold at about

• Above this value, the error probability is very low whereas below this, the
error probability is high and the effect of channel noise is significant.
• The effect of channel noise can be minimized using the regenerative
repeaters.
Error Threshold in PCM:

In PCM, the main effect of channel noise is to introduce bit errors into the
received signals which may be measured in terms of the average
probability of symbol error, Pe or Bit Error Rate (BER).

Bit Error Rate (BER) is defined as the probability that the reconstructed
symbol at the receiver output is different from the transmitted symbol, on an
average.

• BER should be minimized to optimize the system performance in the

presence of channel noise (assumed to be AWGN).
• The average probability of symbol error in a binary encoded PCM receiver
depends solely on the ratio of the transmitted signal energy per bit, Eb to the
noise spectral density, N0 that is
Typical value of needed for specified value of Pe are depicted in the
following table..

• Generally, the acceptable value of Pe in PCM system is about 10-6.

• Therefore, there is an error threshold value of exists at = 10.6 for which
Pe = 10-6.
• For < 10.6 dB, the receiver performance involves significant number of
errors.
• For > 10.6 dB, the effect of channel noise is considered to be negligible.
• The error threshold of 10.6 dB in a PCM system is much lower than that of
60 – 70 dB required for high-quality transmission of speech signals using
amplitude modulation systems.

• The average noise power in the PCM system is increased by the number of
bits used in each sample multiplied by increase in bandwidth.

• The combined presence of channel noise and interference (due to crosstalk

and impulse noise) causes the error threshold to increase for satisfactory
performance of PCM system.

• A PCM system can be made robust to channel noise and interference by

keeping an adequate margin in value above error threshold.
Let us consider the PCM waveform as shown in figure below represented by
NRZ - Unipolar scheme.

Let the amplitude of the pulses be

Let us now define the PCM wave as shown in the above figure as
The channel noise w(t) is assumed to be AWGN with zero mean and PSD of
N0/2. The received signal is given as

Following figure depicts the receiver for the detection of the binary PCM
wave.
A matched filter is a linear filter designed to provide the maximum SNR at its
output for a given transmitted symbol waveform.

To find the Probability of error, we use signal-space approach.

Let , then s(t) becomes

If is laid along the horizontal direction, we find that the signal-space diagram
is one-dimentional with two message points as shown in following figure.
The coordinates of the two message points are

Let us assume that the symbols 1 and 0 occur with equal probability. Then the
decision boundary or the threshold is set at middle, i.e, between two message
points and is .

The decision boundary divides the signal-space diagram into two regions Z1
and Z2.
If the received signal falls in the region Z1, we take decision in favor of symbol
1. If the received signal falls in the region Z2, we take decision in favor of
symbol 0.

There are two kinds of erroneous decisions possible.

1. Symbol 0 is sent, but the received signal fall inside the region Z1, and
receiver takes a decision in favour of symbol 1.
2. Symbol 1 is sent, but the received signal fall inside the region Z2, and
receiver takes a decision in favour of symbol 0.

The input to the decision device x1 will have a coordinate along axis is given by
The received signal point x1 is the sample value of a random variable X1.

The mean of X1 is that of si1 where i=1,2 and the variance of X1 is that of AWGN,
a random variable W1 having a sample w1.

If the variance of AWGN is N0/2, then

We recall the PDF of a Gaussian distributed random variable X,

Where μ is the mean and σ2 is the variance.

Hence, the RV X1 with sample value x1, having a Gaussian distribution with
zero mean and variance N0/2 has a conditional PDF or likelihood function as
follows (assuming that a 0 is sent)
The likelihood function is sketched as follows.

The likelihood function has a peak at its mean, which is zero.

The probability of error if a decision is taken in favour of symbol 1 when

symbol 0 is sent is the area shown shaded in the figure.
Then, Pe(0) becomes
We define the complementary error function, erfc as follows

Then
Similarly, the RV X1 with sample value x1, having a Gaussian distribution with
mean and variance N0/2 has a conditional PDF or likelihood function as
follows (assuming that a 1 is sent)

The above likelihood function is Gaussian distributed and peak at its mean
value . The sketch of this likelihood function is given below. The shaded area
represents the probability of error when symbol 1 is sent and decision is
taken in favour of symbol 0
Differential PCM (DPCM):

• For the digitization of analog voice or video information signal using PCM,
the sampling is usually carried out at a rate slightly higher than the Nyquist
rate.
• It is observed that the resulting sampled signal does not change much from
one sample value to the adjacent one.
• In other words, adjacent samples carry same information with a little
difference, resulting in redundant information.
• DPCM is specifically designed to take advantage of the sample-to-sample
redundancies in typical speech waveforms.
• Since the range of sample differences is typically less than the range of
individual samples, fewer bits are required for DPCM as compared to that
needed for conventional PCM.
• The number of bits required to transmit one sample will also be reduced.
• Overall bit rate will be decreased.
• Less bandwidth will be needed.
• More efficient encoded signal will be obtained.
• If a part of the redundant information is known, it is possible to predict
the future values due to correlation between the samples.

The block diagram of DPCM Tx is given below.

In the block diagram of DPCM x(nTs) is the sampled version of the analog
signal x(t) with Ts being the sampling period.

e(nTs) =
Where is the prediction of x(nTs).

The prediction error, e(nTs) is then quantized to produce eq(nTs).

The quantizer output eq(nTs) is then encoded to produce a DPCM wave.

In DPCM, the difference in the amplitude of a sample and its prediction is

transmitted rather than the actual sample.
Since the range of samples and their prediction differences is typically less
than the range of individual samples, fewer bits are required for DPCM than
conventional PCM
Where qe(nTs) is the quantization error.

From the DPCM Tx, the input to the prediction filter is

xq(nTs) =

Substituting e(nTs) and eq(nTs) in xq(nTs), we get

Where xq(nTs) is the quantized version of the input sample x(nTs).

By taking appropriate number of quantization levels, it is possible to adjust

the average power of the prediction error.

The receiver for constructing the quantized version of the input is given
below.
Delta Modulation:

Delta modulation (DM) uses a single-bit DPCM code to achieve digital

transmission of analog signals.

The present sample value is compared with the previous sample value.

Two possibilities emerge:

• If the value of the current sample is smaller than the value of the previous
sample, a logic 0 is transmitted.
• If the value of the current sample is larger than the value of the previous
sample, a logic 1 is transmitted.

Thus, the difference between the actual input signal and the approximation
signal is quantized into two representation levels only: +Δ and -Δ,
corresponding to positive and negative difference respectively.
• Hence, for each sample, only one binary bit is transmitted.
• Following figure shows an ideal delta modulation waveform.
Where Ts is sampling instant,
xq(nTs-Ts) is recent approximation of
x(nTs),
e(nTs) is error signal representing the
difference between the present sample
x(nTs) and latest approximation to it,
and
eq(nTs) is quantized version of e(nTs)
The quantizer output eq(nTs) is the desired delta modulated wave for
different values of n. The rate of information transmission is 1/Ts.

The DM transmitter contains a comparator, a quantizer and an accumulator

as shown in figure below.
The comparator gives the difference e(nTs) between x(nTs) and xq(nTs - Ts).

The output of quantizer is +Δ , if e(nTs) > 0 and -Δ, if e(nTs) < 0, i.e, eq(nTs) is
Δsgn{e(nTs)}. The quantizer output is applied to accumulator producing

At the sampling instant nTs , the accumulator increments xq(nTs - Ts) by a step
±Δ depending on the sign of the error signal e(nTs).
Following figure shows the block diagram of a DM receiver.

LPF is used to remove step variations and to get a smooth reconstructed

message signal x(t).
Delta modulation systems are subject to two types of quantization error:
(1) slope-overload distortion
(2) granular noise.
The key parameters staircase step size Δ, and sampling rate fs are chosen
properly, so that the staircase signal is a close approximation to the actual
analog waveform.

Every signal has a definable upper frequency, that decides the fastest rate at
which it can change.

To get the fastest change in the signal, both Δ and fs must be increased or one
of them must be increased.

Increasing sampling frequency requires larger bandwidth, increasing step size

increases the quantization error.

If the step size is small, it results in slope overload condition, where the staircase
cannot track rapid changes in the analog signal.
If steps are too large compared to small variations in the input signal,
Granular noise occurs.

This means that for very small variations in the input signal, the staircase
signal is changed by a large amount because of the large step size.

Granular noise is similar to quantization noise in PCM system.

To avoid slope overload condition, we require that

Granular noise can be reduced by reducing the step size.

Slope overload is very common in analog signals that have steep slopes or
whose amplitudes vary rapidly.

The delta modulation with uniform step size, is called as “Linear delta
modulation”
Quantization Noise in Delta Modulation:

We define the quantization error or quantization noise as the difference

between the original signal and the quantized (staircase) approximation.

Let the random variable Q denote the quantization error and qe its sample
value. Then,

Q = x(t) - xq(t)

Let us assume that the sampling rate fs and step size Δ are chosen in order to
avoid slope overload.

Under this assumption, magnitude of the quantization noise is always less

than or equal to Δ.
To simplify the analysis, let us make the assumption that all the signal
amplitudes are equally likely.

Then, we conclude that the error is uniformly distributed over the range
between -Δ and +Δ as shown in figure below.

Fig: Probability density function of quantization error

Mean of the quantization error is
Variance or Mean square value of the quantization error is
Let us assume that the quantization noise power Pq is uniformly distributed
over the frequency band upto fs.

The output quantizing noise power within the bandwidth fM of the LPF in the
receiver is
If P is the average signal power, then output signal-to-noise ratio is
Adaptive delta modulation (ADM)

ADM is a delta modulation where the step size δ of the staircase waveform is
varied depending upon the slope or amplitude characteristics of the analog
input signal.

The effects of slope overload distortion and granular noise can be overcome
by making the staircase step size δ(nTs), a function of the slope of the input
analog signal.

When the slope of the input analog signal is large, δ(nTs) is increased and
when the slope is fairly constant, δ(nTs) is decreased. In practice, the
staircase step size δ(nTs) is bounded by lower and upper limits δmin and δmax,
respectively.
δmin ≤ δ(nTs) ≤ δmax
References:

• Simon Haykin, Digital Communications, John Wiley & Sons, 1998.

• Simon Haykin, Michael Moher, Communication Systems, 5th Edition,
Wiley, 2009.
• T L Singal, Analog and Digital Communications, TMH, 2012.
• K.N. Hari Bhat, D. Ganesh Rao, Digital Communications Theory and
Lab Practice, Pearson, 3rd Edition, 2010.

146
THANK YOU

147