222 - EC8501, EC6501 Digital Communication - Notes 1

Click Here for Digital Communication full study material.
EC6501
DIGITAL COMMUNICATION
www.BrainKart.com
OBJECTIVES:
Toknow the principles of sampling &
quantization
Tostudy the various waveform coding
schemes
Tolearn the various baseband transmission
schemes
Tounderstand the various Band pass signaling
schemes
Toknow the fundamentals of channel coding
www.BrainKart.com
SYLLABUS www.BrainKart.com
UNIT I SAMPLING & QUANTIZATION 9

Low pass sampling – Aliasing- Signal Reconstruction-Quantization - Uniform & non-uniform
quantization - quantization noise - Logarithmic Companding of speech signal- PCM - TDM 56
UNIT II WAVEFORM CODING 9

Prediction filtering and DPCM - Delta Modulation - ADPCM & ADM principles-Linear Predictive
Coding
UNIT III BASEBAND TRANSMISSION 9

Properties of Line codes- Power Spectral Density of Unipolar / Polar RZ & NRZ – Bipolar NRZ -
Manchester- ISI – Nyquist criterion for distortionless transmission – Pulse shaping – Correlative
coding - Mary schemes – Eye pattern - Equalization
UNIT IV DIGITAL MODULATION SCHEME 9

Geometric Representation of signals - Generation, detection, PSD & BER of Coherent BPSK,
BFSK & QPSK - QAM - Carrier Synchronization - structure of Non-coherent Receivers - Principle
of DPSK.
UNIT V ERROR CONTROL CODING 9

Channel coding theorem - Linear Block codes - Hamming codes - Cyclic codes - Convolutional
codes - Vitterbi Decoder
TOTAL: 45 PERIODS
www.BrainKart.com
www.BrainKart.com
OUTCOMES
Upon completion of the course, students will be
able to
Design PCM systems
Design and implement base band transmission
schemes
Design and implement band pass signaling
schemes
Analyze the spectral characteristics of band
pass signaling schemes and their noise
performance
Design error control coding schemes
www.BrainKart.com
www.BrainKart.com
EC6501
DIGITAL COMMUNICATION
UNIT - 1
www.BrainKart.com
www.BrainKart.com
INTRODUCTION
www.BrainKart.com
www.BrainKart.com
www.BrainKart.com
BrainKart.com
Get useful study materials from BrainKart.com

www.BrainKart.com

www.BrainKart.com
UNIT I
SAMPLING & QUANTIZATION (9)
 Low pass sampling
 Aliasing
 Signal Reconstruction
 Quantization
 Uniform & non-uniform quantization
 Quantization Noise
 Logarithmic Companding of speech signal
 PCM
 TDM
BrainKart.com

BrainKart.com
Digital communication system
Input Low
Signal Source Channel
Pass Sampler Quantizer Multiplexer
Analog/ Encoder Encoder
Filter
Digital
Carrier
Pulse
Line
To Channel Modulator Shaping
Encoder
Filters
De- Receiver
From Channel Detector
Modulator Filter
Carrier Ref.
Signal
Digital-to-Analog Channel De-
at the
Converter Decoder Multiplexer
user end

BrainKart.com
Key Questions
 How can a continuous wave form be

converted into discrete samples?
 How can discrete samples be converted back

into a continuous form?

BrainKart.com
Low Pass Sampling
Sampling (in time) is
 Measure amplitude at regular intervals
 How many times should we sample?

BrainKart.com
Nyquist Theorem
For lossless digitization, the sampling rate
should be at least twice the maximum
frequency of the signal to be sampled.
 In mathematical terms:
fs > 2*fm
 where fs is sampling frequency and fm is the

maximum frequency in the signal
BrainKart.com
Limited Sampling
 But what if one cannot sample fast
enough?

BrainKart.com
Limited Sampling
 Reduce signal frequency to half of
maximum sampling frequency
 low-pass filter removes higher-frequencies
 (e.g.) If max sampling frequency is 22kHz, the it

is a must to low-pass filter a signal down to
11kHz
BrainKart.com

BrainKart.com
BrainKart.com

BrainKart.com

BrainKart.com

BrainKart.com

BrainKart.com
Aliasing effect
LP filter
Nyquist rate
aliasing

BrainKart.com
Three different sampling methods
Practical Sampling Methods are Natural Sampling
and Flat-top Sampling

BrainKart.com
Natural Sampling

BrainKart.com

BrainKart.com
Pulse-Amplitude Modulation
• Pulse-Amplitude Modulation (PAM)

– The amplitude of regularly spaced pulses are
varied in proportion to the corresponding sample
values of a continuous message signal.
– Two operations involved in the generation of the
PAM signal
• Instantaneous sampling of the message signal m(t)
every Ts seconds,
• Lengthening the duration of each sample, so that it
occupies some finite value T. Fig. 5
Click Here for Digital Communication full study material. 28

BrainKart.com
Back Next
Fig.5

BrainKart.com
Fig.6 Back Next

BrainKart.com
Back Next
Fig.7

• BrainKart.com
The advantages offered by digital pulse modulation
– Performance
• Digital pulse modulation permits the use of regenerative repeaters,
when placed along the transmission path at short enough distances,
can practically eliminate the degrading effects of channel noise and
signal distortion.
– Ruggedness
• A digital communication system can be designed to withstand the
effects of channel noise and signal distortion
– Reliability
• Can be made highly reliable by exploiting powerful error-control
coding techniques.
– Security
• Can be made highly secure by exploiting powerful encryption
algorithms
– Efficiency
• Inherently more efficient than analog communication system in the
tradeoff between transmission bandwidth and signal-to-noise ratio
– System integration
• To integrate digitized analog signals with digital computer data
Quantization ProcessBrainKart.com
• Amplitude quantization
– The process of transforming the sample amplitude m(nTs) of a
baseband signal m(t) at time t=nTs into a discrete amplitude
v(nTs) taken from a finite set of possible levels.
I k : {mk  m mk 1}, k  1,2,..., L (17) Fig. 9
– Representation level (or Reconstruction level)

• The amplitudes vk , k=1,2,3,……,L
– Quantum (or step-size)
• The spacing between two adjacent representation levels
v  g(m) (18) Fig. 10

BrainKart.com
Back Next
Fig.9

BrainKart.com
Fig.10 Back Next
Two types of quantization are

a) Mid-tread
b) Mid-rise Click Here for Digital Communication full study material. 35
BrainKart.com
Linear Quantization
• Applicable when the signal is in a
finite range (fmin, fmax)
• The entire data range is divided
into L equal intervals of length Q
(known as quantization interval or
quantization step-size)
• Q=(fmax-fmin)/L Interval i is
mapped to the middle value of this
interval
• We store/send only the index of
quantized value min

BrainKart.com
Signal Range is Symmetric

BrainKart.com
Quantization Noise

Non-Uniform Quantizawtiwown.rejinpaul.com
 Many signals such as speech have a nonuniform distribution.
– The amplitude is more likely to be close to zero than to be at higher levels.
 Nonuniform quantizers have unequally spaced levels
– The spacing can be chosen to optimize the SNR for a particular type of signal.
Output sample
XQ 6
2 Example: Nonuniform 3 bit quantizer
-8 -6 -4 -2 2 4 6 8
-2
Input sample
X
-4
-6
www.BrainKart.com
Non-Linear Quantization
• The quantizing intervals are not of equal size
• Small quantizing intervals are allocated to small
signal values (samples) and large quantization
intervals to large samples so that the signal-to-
quantization distortion ratio is nearly independent of
the signal level
• S/N ratios for weak signals are much better but are
slightly less for the stronger signals
• “Companding” is used to quantize signals

www.BrainKart.com
Function representation

BrainKart.com
Uniform and Non-uniform Quantization
BrainKart.com
Companding
• Formed from the words compressing and
expanding.
• A PCM compression technique where analogue
signal values are rounded on a non-linear scale.
• The data is compressed before sent and then
expanded at the receiving end using the same
non-linear scale.
• Companding reduces the noise and crosstalk
levels at the receiver.

BrainKart.com
u-LAW and A-LAW definitions

• A-law and u-law are companding schemes used in
telephone networks to get more dynamics to the
8 bit samples that is available with linear coding.
• Typically 12..14 bit samples (linear scale) sampled
at 8 kHz sample are companded to 8 bit
(logarithmic scale) for transmission over 64 kbit/s
data channel.
• In the receiving end the data is then converted
back to linear scale (12..14 bit) and played back.
converted back
– Compressor
BrainKart.com
• A particular form of compression law : μ-law
log(1  m )
v (5.23)
log(1  )

d m log(1  )
 (1  m ) (5.24)
dv 
• μ-law is neither strictly linear nor strictly logarithmic
• A-law :

 Am 1
1 log A , 0  m 
v   A (5.25)
1
1 log( Am ) , m1
 1 log A A
1
1 log A 0  m 
,
d m  A A
d v   1 (5.26)
(1 log A) m , Fig. 5.11
 m 1
 A
BrainKart.com
Fig.11 Back Next

BrainKart.com
Example: -law Companding
1
0.5
x[n]=speech /song/ 0
-0.5
-1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0.5
y[n]=C(x[n]) 0
-0.5
Companded Signal -1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0.5
Close View of the Signal
Segment of x[n] 0
-0.5
-1
2200 2300 2400 2500 2600 2700 2800 2900 3000
0.5
Segment of y[n] 0
Companded Signal -0.5
-1
2200 2300 2400 2500 2600 2700 2800 2900 3000

A-law and law Compandi.ng
• These two are standard companding methods.
• u-Law is used in North America and Japan
• A-Law is used elsewhere to compress digital telephone signals
Eeng 360 48
Presenter
2015-08-02 19:50:34
BrainKart.com
--------------------------------------------
Quantization - why do we need such classification ?!A-good
(3)Idea is to use a non-uniform
quantizer.A non-uniform quantizer can
provide fine
quantization levels for weak signals (
Comparison – Uniform Vs. Non-Uniform Usage 50
%)and coarse levels for strong signals
(15%) .
The goal is decrease the SQNR . And
theS
Q
Ris proportional to the number of
N
 Speech signals doesn’t require high quantization resolution forlevels,
specially at the weak signal part.
high amplitudes (50% Vs. 15%).
 wasteful to use uniform quantizer ?
The goal is decrease the SQNR, more levels for low amplitudes, less levels for
high ones.
 Maybe use a Non-uniform quantizer ?

Presenter
2015-08-02 19:50:35
BrainKart.com
--------------------------------------------
Example from the lecture of Prof.
S.N.Merchant.htp:/www.youtbe.com/w
atch?v=DBPAubw
UrQYTasks:
Put the Example for Comapding
Quantization Explain Mu-Law and A-Law
Understanding from where does the 13
kBitCome From.
More About Non-Uniform Quantizers (Companding) Explain
It should be noted that, A-Law and M-
 Uniform quantizer = use more levels when you need it. Lawareused to compress the 13 or 14 bit
signed linear PCM samples to
 The human ear follows a logarithmic process in which high amplitude sound
logarithmic doesn’t 8 bit samples
require the same resolution as low amplitude sounds.
 One way to achieve non-uniform quantization is to use what is called as “Companding”
 Companding = “Compression + Expanding”
Compressor Uniform
Expander
Function Qua ntization
Function
(-1)

Pulse-Code Modulation
• PCM (Pulse-Code Modulation)
– A 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
– The basic operation
• Transmitter : sampling, quantization, encoding
• Receiver : regeneration, decoding, reconstruction
• Operation in the Transmitter

1. Sampling
1. The incoming message signal is sampled with a train of rectangular pulses
2. The reduction of the continuously varying message signal to a limited
number of discrete values per second
2. Nonuniform Quantization
1. The step size increases as the separation from the origin of the input-
output amplitude characteristic is increased, the large end-step of the
quantizer can take care of possible excursions of the voice signal into the
large amplitude ranges that occur relatively infrequently.
BrainKart.com
Fig.11 Back Next

BrainKart.com
3. Encoding
1.To translate the discrete set of sample vales to a
more appropriate form of signal Fig. 11
2.A binary code
 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.
 The binary code is easy to generate and regenerate
Table. 2

BrainKart.com
• Regeneration Along the Transmission Path
– The ability to control the effects of distortion and noise produced by
transmitting a PCM signal over a channel
– Equalizer
• Shapes the received pulses so as to compensate for the effects of
amplitude and phase distortions produced by the transmission
– Timing circuitry
• Provides a periodic pulse train, derived from the received pulses
• Renewed sampling of the equalized pulses
– Decision-making device Fig. 13
• The sample so extracted is compared o a predetermined threshold
– ideally, except for delay, the regenerated signal is exactly the same as the
information-bearing signal
1. The unavoidable presence of channel noise and interference causes
the repeater to make wrong decisions occasionally, thereby
introducing bit errors into the regenerated signal
2. If the spacing between received pulses deviates from its assigned
value, a jitter is introduced into the regenerated pulse position,
thereby causing distortion.

BrainKart.com
Back Next
Fig.13

BrainKart.com
• Operations in the Receivers
1. Decoding and expanding

1. Decoding : regenerating a pulse whose amplitude is
the linear sum of all the pulses in the code word
2. Expander : a subsystem in the receiver with a
characteristic complementary to the compressor
1. The combination of a compressor and an expander is a
compander
2. Reconstruction
1.Recover the message signal : passing the expander
output through a low-pass reconstruction filter
BrainKart.com
Categories of multiplexing

BrainKart.com
Time Division Multiplexing (TDM)
TDM is a technique used for

transmitting several message signals
over a single communication channel
by dividing the time frame into slots,
one slot for each message signal

www.rejinpaul.com
Time Division Multiplexing
• Entire spectrum is allocated for a channel (user) for a limited time.
• The user must not transmit until its
next turn.
k1 k2 k3 k4 k5 k6
• Used in 2nd generation
Frequency
• Advantages: Time
– Only one carrier in the medium at any given time
– High throughput even for many users
– Common TX component design, only one power amplifier
– Flexible allocation of resources (multiple time slots).
BrainKart.com
Time Division Multiplexing
• Disadvantages
– Synchronization
– Requires terminal to support a much higher data
rate than the user information rate therefore
possible problems with intersymbol-
interference.
• Application: GSM
 GSM handsets transmit data at a rate of 270 kbit/s in
a 200 kHz channel using GMSK modulation.
 Each frequency channel is assigned 8 users, each
having a basic data rate of around 13 kbit/s
Time Division Multiplexiwnwgw.rejinpaul.com
At the Transmitter
 Simultaneous transmission of several signals on a time-sharing basis.
 Each signal occupies its own distinct time slot, using all frequencies, for
the duration of the transmission.
 Slots may be permanently assigned on demand.
At the Receiver
 Decommutator (sampler) has to be synchronized with the incoming
waveform  Frame Synchronization
 Low pass filter
 ISI – poor channel filtering
 Feedthrough of one channel's signal into another channel -- Crosstalk
Applications of TDM: Digital Telephony, Data communications, Satellite

Access, Cellular radio.
Get useful study materials from BrainKart.com 61
Time Division Multiplexiwnwgw.rejinpaul.com
Conceptual diagram of multiplexing-demultiplexing.
PAM TDM System

Get useful study materials from BrainKart.com 62
BrainKart.com
TDM-PAM: Transmitter

BrainKart.com
TDM-PAM : Receiver

BrainKart.com
Samples of Signal -1
g1(t)
time
0 Ts 2Ts

BrainKart.com
Samples of signal - 2
g2(t)
Ts Ts

BrainKart.com
Multiplexing of TWO signals
0 Ts 2Ts

BrainKart.com
TDM-PAM for 4 signals.
4
4
4
1 1 1
2 2 2
3 3 3
Time

BrainKart.com
Problem
Two low-pass signals of equal bandwidth
are sampled and time division
multiplexed using PAM. The TDM signal is
passed through a Low-pass filter & then
transmitted over a channel with a
bandwidth of 10KHz.
Continued….

BrainKart.com
Problem (continued…)
a) What is maximum Sampling rate for each

Channel?
b) What is the maximum frequency content
allowable for each signal?

BrainKart.com
Problem: Solution
Channel Bandwidth = 10 KHz.

Number of samples that can be transmitted
through the channel = 20K
Maximum Sampling rate for each channel =
10K Samples/sec.
Maximum Frequency for each Signal = 5KHz

BrainKart.com
End of Unit-1

BrainKart.com
Unit – II
Waveform Coding

BrainKart.com
Syllabus
Prediction filtering and DPCM - Delta

Modulation - ADPCM & ADM principles-Linear
Predictive Coding

BrainKart.com
Introduction to Waveform Coding

• Waveform coding is some kind of approximately
lossless coding, as it deals with speech signal as
any kind of ordinary data.
• The resulting signal is close as possible as the
original one.
• Codecs using this techniques have generally low
complexity and give high quality at rates 16 Kbps.
• The simplest form of waveform coding is Pulse
Code Modulation (PCM).
BrainKart.com
Pulse Code Modulation (PCM)

• It involves sampling and quantizing the input
waveform.
• PCM consists of three steps to digitize an
analog signal:
1. Sampling
2. Quantization
3. Binary encoding

BrainKart.com

BrainKart.com
Prediction Filtering
• Linear prediction is a mathematical operation
where future values of a discrete-time signal
are estimated as a linear function of previous
samples.
• In digital signal processing, linear prediction is

often called linear predictive coding (LPC).
• linear prediction can be viewed as a part of

mathematical modelling or optimization.
BrainKart.com
The Prediction Model

• The most common representation is
• Where is the predicted signal value, x(n-i)

the previous observed values, and the
predictor coefficients.
• The error generated by this estimate is
• Where x(n) is the true value.

BrainKart.com
Differential pulse-code modulation
(DPCM)
• Differential pulse-code modulation (DPCM) is

a signal encoder that uses the baseline of
pulse-code modulation (PCM) but adds some
functionalities based on the prediction of the
samples of the signal.
• The input can be an analog signal or a digital

signal.
BrainKart.com
Block Diagram of DPCM

BrainKart.com
• DPCM code words represent differences between samples

unlike PCM where code words represented a sample value.
• Basic concept of DPCM - coding a difference, is based on

the fact that most source signals show significant
correlation between successive samples so encoding uses
redundancy in sample values which implies lower bit rate.
• Realization of basic concept (described above) is based on a

technique in which we have to predict current sample value
based upon previous samples (or sample) and we have to
encode the difference between actual value of sample and
predicted value.

BrainKart.com
Principle of DPCM

BrainKart.com
Delta Modulation
• A Delta modulation (DM or Δ-modulation) is an analog-to-digital
and digital-to-analog signal conversion technique used for
transmission of voice information where quality is not of primary
importance.
• To achieve high signal-to-noise ratio, delta modulation must use

oversampling techniques, that is, the analog signal is sampled at a
rate several times higher than the Nyquist rate.
• Derived forms of delta modulation are continuously variable slope

delta modulation, delta-sigma modulation, and differential
modulation.
• Differential pulse-code modulation is the super-set of DM.

BrainKart.com
Features
• the analog signal is approximated with a series of
segments
• each segment of the approximated signal is compared to
the original analog wave to determine the increase or
decrease in relative amplitude
• the decision process for establishing the state of
successive bits is determined by this comparison
• only the change of information is sent, that is, only an
increase or decrease of the signal amplitude from the
previous sample is sent whereas a no-change condition
causes the modulated signal to remain at the same 0 or 1
state of the previous sample.

BrainKart.com
Principle of delta modulation

BrainKart.com
Differential Pulse Code Modulation

(DPCM)
• What if we look at sample differences, not the
samples themselves?
– dt = xt-xt-1
– Differences tend to be smaller
• Use 4 bits instead of 12, maybe?

BrainKart.com
Differential Pulse Code Modulation

(DPCM)
• Changes between adjacent samples small

• Send value, then relative changes
– value uses full bits, changes use fewer bits
– E.g., 220, 218, 221, 219, 220, 221, 222, 218,.. (all values between 218
and 222)
– Difference sequence sent: 220, +2, -3, 2, -1, -1, -1, +4....
– Result: originally for encoding sequence 0..255 numbers need 8 bits;
– Difference coding: need only 3 bits

BrainKart.com
Adaptive Differential Pulse Code

Modulation (ADPCM)
• Adaptive similar to DPCM, but adjusts the width of the
quantization steps
• Encode difference in 4 bits, but vary the mapping of bits to
difference dynamically
– If rapid change, use large differences
– If slow change, use small differences

BrainKart.com
Adaptive Delta Modulation (ADM)

BrainKart.com
A large step size is required when sampling those parts

of the input waveform of steep slope. But a large
step size worsens the granularity of the sampled
signal when the waveform being sampled is changing
slowly.
• A small step size is preferred in regions where the
message has a small slope. This suggests the
need for a controllable step size – the control
being sensitive to the slope of the sampled signal
• Hence ADM is prefered.

BrainKart.com
Adaptive Delta Modulation

BrainKart.com
VCA

BrainKart.com
Linear Predictive Coding (LPC)

BrainKart.com
Basic Concepts of LPC
• It is a parametric de-convolution algorithm

• x(n) is generated by an unknown sequence e(n)
exciting a unknown system V(Z) which is supposed to
be a linear non time-variant system.
• V(Z) = G(Z)/A(Z), E(Z)V(Z) = X(Z)
• G(Z) = Σj=0Q gjZ-j, A(Z) = Σi=0P aiZ-i
• Where ai and gj are parameters, real and a0 = 1
• If an algorithm could estimate all these parameters,
then V(Z) could be found, and E(Z) could be found
also. This finishes de-convolution.
BrainKart.com
• There are some limitations for the model

• (1) G(Z) = 1 then V(Z) = 1/A(Z) this is so called “Full
Poles εodels” and the parametric de-convolution
became coefficients(ai) estimation problem.
• (2) e(n) sequence is of form Ge(n), where e(n) is a
periodic pulse or a Gaussian white noise sequence.
For the first case e(n) = Σ6(n-rNp) and for the second
case R(k) = E[e(n)e(n+k)] = 6(k) and the value of
e(n) satisfied with Normal distribution. G is a non-
negative real number controlling the amplitude.
• The way is x(n)->V(Z)(P,ai)->e(n),G->type of e(n)

BrainKart.com
• Suppose x(n) and type of e(n) are known, what is

the optimized estimation of P and ai, e(n) and G? It is
the LMS algorithm.
• Suppose x(n) is the predicted value of x(n), it is the
linear sum of previous P’ known values of x:
• x(n) = Σi=1P’ ai x(n-i)
• The predicted error
• s(n) = x(n)-x(n) = x(n) - Σi=1P’ ai x(n-i)
• It is a stochastic sequence. The variance of it could
be used to evaluate the quality of prediction.

BrainKart.com
• σ2 = Σnc2(n) (time average replaced means)

• It could be proved that if x(n) is generated by “full
poles” model : x(n) = -Σi=1P ai x(n-i) + Ge(n) and
optimized P’ = P, optimized ai = ai, σ2 is minimal.
• σ2 = Σn [x(n) -Σi=1P ai x(n-i)]2
• ={Σn x2(n)}-2Σi=1P ak{Σn x(n-k)x(n)}+
• Σk=1PΣi=1P akai{Σn x(n-k)x(n-i)}
• By setting ð(σ2 )/ ðak = 0 we can get
• -2 {Σn x(n-k)x(n)}+2Σi=1P ai{Σn x(n-k)x(n-i)}=0
• Or Σi=1P aiφ(k,i) = φ(k,0)
• if φ(k,i) =Σn x(n-k)x(n-i) 1<=i<=P and 1<=k<=P
BrainKart.com
• Σ P a φ(k,i) = φ(k,0), k=1~P is called δPC

i=1 i
canonical equations. There are some different
algorithms to deal with the solution.
• [σ2]min = Σk=0P ak φ(k,0) with a0 = 1
• So if we have x(n), φ(k,i) could be calculated,
and equations could be solved to get ai and
[σ2]min also could be obtained. For short-time
speech signal according to different lower and
upper limitation of the summary we could
have different types of equations. We will
discuss these different algorithms later.
BrainKart.com
Auto-Correlated Solution of LPC
• Suppose windowed signal is xw(n)

• φ(k,i) = Σn xw(n-k)xw(n-i)
• If window length is N-1 then the summation
range will be 0~N+P-1
• φ(k,i) = Σm xw(m+(i-k))xw(m) = R(i-k) if n-i
=m
• φ(k,i) = R(i-k) = R(k-i) = R(|i-k|) <= R(0)
• The equations became Σ P a R(|i-k|)= - R(k)
i=1 i
1<=k<=P
• These are Toplitz equations and have high
efficient soluGtieot n
us.eful study materials from BrainKart.com
BrainKart.com
• |R(0) R(1) …… R(P-1)| |a1| | R(1) |

• |R(1) R(0) …… R(P-2)| |a2| | R(2) |
• |………………………………….| |...| = …...
• |R(P-1) R(P-2) … R(0) | |ap| | R(P) |
• 6.2.1 Durbin Algorithm

• 1. E(0) = R(0)
• 2. Ki = [ R(i) - Σ aj(i-1)R(i-j)]/E(i-1) 1<=i<=p
• 3. ai(i) = Ki
• 4. aj(i) = aj(i-1) – Kiai-j(i-j) 1<=j<=i-1
• 5. E(i) = (1-Ki2)E(i-1)
• Final solution is aj = aj(p) 1<=j<=p
BrainKart.com
• For iteration i we got a set of

coefficients for the predictor of i-th
order and the minimal predicted error
energy E(i). We also can get it by {R(k)}
:
• E(i) = R(0) –Σk=1i akR(k), 1<=i<=p
• Ki is the reflect coefficient : -1<=Ki<=1
It is a sufficient and necessary condition
for stable HG(ezt u)sedfuu
l sr
tuidn
ygmaitetre
iar
lsafrtoim
ownw.w.rejinpaul.com
BrainKart.com
• 6.2.2 Schur algorithm

• At first an auxilary sequence is defined. Its properties
are :
• (1) qi(j) = R(j) when i = 0
• (2) qi(j) = 0 when i > 0, j=1~p
• (3) qp(0) = E(p) is the predicted error energy.
• (4) |qi(j)| <= R(0), it is equal only if i=j=0
• The algorithm is as following:
• 1. r(j) = R(j)/R(0), r(-j) = r(j), j=0~p
• 2. a0 = 1, E(0) = 1
• 3. q0(j) = r(j) Get -upse<fujl<stpudy materials from BrainKart.com
BrainKart.com
• 4. i = 1, k1 = r(1)
• 5. For i-p<=j<=p
qi(j) = qi-1(j) + ki *qi-1 (i-j)
ki = qi-1(j)/qi(0)
aj(i) = qi-1(i-j)
E(i) = E(i-1)(1-ki2)
• 6. If i<p, back to step 5
• 7. Stop
• If we only calculate ki, then only first two expressions
in step 5 are enough. It is suitable for fix-point
calculation (r<=1) or hardware implementation.
BrainKart.com
Covariance Solution of LPC
• If not using windowing, but limiting the range of

summation, we could get :
• σ2 = Σn=0N-1c2(n) n=0~N-1
• φ(k,i) = Σn=0N-1 x(n-k)x(n-i) k=1~p, i=0~p
• = Σm=-iN-i-1 x(m+(i-k))x(m) let n-i=m, m=-i~N-i-1
• The equations will be like following :
• |φ(1,1) φ(1,2) …… φ(1,p)| |a1| |φ(1,0)|
• |φ(2,1) φ(2,2) …… φ(2,p)| |a2| |φ(2,0)|
• .………………………………………………=…………
• |φ(p,1) φ(p,2) …… φ(p,p)| |ap| |φ(p,0)|

BrainKart.com
Covariance Solution of LPC
• The matrix is a covariance matrix and it is positive

determined, but not Toplitz. There is no high efficient
algorithm to solve. Only common used LU algorithm
could be applied. Its advantage is not having big
predicted error on the two ends of window. So when
N~P the estimated parameters have more accuracy
than auto-correlated method. But in speech
processing very often N>>P, so the advantage is not
obvious.

BrainKart.com
LPC parameters and their relationships
• (1) Reflect Coefficients

• Also known as PARCOR coefficients
• If {aj} are known, ki could be found as following :
• aj(p) = aj 1<=j<=p
• ki = ai(i)
• aj(i-1) = (aj(i) + aj(i) ai-j(i))/(1-ki2) 1<=j<=i-1
• The inverse process :
• aj(i) = ki
• aj(i) = aj(I-1) - kj ai-j(i-1) at last aj= aj(p) 1<=j<=p
• -1<=ki<=1 is the sufficient and necessary condition
for stable systeGm et ufusenfucltsiotundy materials from BrainKart.com
BrainKart.com
• (2) Coefficients of Logarithm Area Ratio

• gi = log(Ai+1/Ai) = log[(1-ki)/1+ki]) i=1~p
• Where A is the intersection area of i-th
segment of the lossless tube.
• ki = (1-exp(gi))/(1+exp(gi)) i=1~p
• (3) Cepstrum Coefficients

• cn = an + Σk=1n kckan-k/n, 1<=n<=p+1
• = an + Σ k=n-pn-1 kck an-k /n, n>p+1
BrainKart.com
• (4) The Roots of Predictor

• A(z) = 1 – Σk=1p akz-k = Πk=1p (1-zk z-1) = 0
Σ
• Transfer to S-plane: zi = exp(siT)
• Suppose si = σi + jΩi , zi = zir + jzii , then
• Ωi = tan(zii/zir)/T andσi= log(zii2 + zir2 )/(2T)
• (5)The impulse response of full poles system

• h(n) = Σk=1p akh(n-k)+σ(n) n>=0
• =0 n<0
BrainKart.com
• (6) Auto-correlated Coefficients of impulse

response of the full poles system
• H(z) = S(z)/U(z) = G/(1- Σ p a z-k)
k=1 k
• The auto-correlated coefficients of h(n) is :
• R(i) = Σn=0 h(n)h(n-i) = R(-i)
• It could be proved that :
• R(i) = Σ p a R(|i-k|) 1<=i<=p
k=1 k p
• And R(0) = Σ a R(k) + G2
k=0 k
• {ak} -> {R(i)} and {R(i)} -> {ak} are
equivalent
BrainKart.com
• (7) Auto-correlated coefficients of impulse

response of the predicted error filter (inverse
filter)
• A(z) = 1 - Σ akz-k
• The impulse response is :
• a(n) = 6(n) - Σk=1p ak6(n-k)
• = 1, n = 0; an , 0<n<=p; 0, otherwise
• Its auto-correlated function is :
• R(i) = Σk=1p a(k)a(k+i) 0<=i<=p
BrainKart.com
• (8)Line Spectrum Pair (LSP) or Line

Spectrum Frequency (LSF)
• A(p)(z)=1-Σk=1p akz-k ( p is even)
• Define P(z) = A(p)(z)+z-(p+1)A(p)(z-1)
Q(z) = A(p)(z)- z-(p+1)A(p)(z-1)
• It could be proved that : All roots of P(z) and
Q(z) are on the unit circle and alternatively
arranged on it provided the roots of A(z) are
inside the unGitetcuisrecfulles.tudy materials from BrainKart.com
BrainKart.com
• Replace z with expjω:

• P(expjω)=|A(P)(expjω)|expjφ(ω)[1+exp[-j((p+1)ω+
2φ(ω))]
• Q(expjω)=|A(P)(expjω)|expjφ(ω)[1+exp[-j((p+1)ω+
2φ(ω)+π)]
• If the roots of A(P)(z) are inside the unit circle, whenωis
0~π, φ(ω) changes from 0 and returns to 0, the amount
[(p+1)ω+2φ(ω)] will be 0~(p+1)π
• P(expjω)=0 : [(p+1)ω+2φ(ω)]=kπ, k=1,3,…P+1
• Q(expjω)=0 : [(p+1)ω+2φ(ω)]=kπ, k=0,2,…P
• The roots of P and Q : Zk = expjωk [(p+1)ω+2φ(ω)]=kπ,
k=0,1,2,…P+1
• And ω0 < ω1 < ω2 < … < ωP < ωP+1
BrainKart.com
• If a1(p)~a p(p) are known, LSP could be found by

A(z) -> P(z) -> p(w) -> f1, f2, … fp
• If f1~fp are known, ai(p) could be found by P(z)
and Q(z) -> A(z) = P(z) + Q(z) -> a (p) ~a (p)
1 p

BrainKart.com
Thank you

BrainKart.com
Unit 3
Baseband Transmission

BrainKart.com
Syllabus
Properties of Line codes- Power Spectral Density of Unipolar /
Polar RZ & NRZ – Bipolar NRZ - Manchester- ISI – Nyquist
criterion for distortionless transmission – Pulse shaping –
Correlative coding - Mary schemes – Eye pattern - Equalization

BrainKart.com
Baseband Transmission
• The digital signal used in baseband
transmission occupies the entire bandwidth
of the network media to transmit a single data
signal.
• Baseband communication is bidirectional,
allowing computers to both send and receive
data using a single cable.

BrainKart.com
Baseband Modulation
• An information bearing-signal must conform to the limits of its channel
• • Generally modulation is a two-step process
• – baseband: shaping the spectrum of input bits to fit in a limited spectrum
• – passband: modulating the baseband signal to the system rf carrier
• • Most common baseband modulation is Pulse Amplitude Modulation
(PAM)
• – data amplitude modulates a sequence of time translates of basic pulse
• – PAM is a linear form of modulation: easy to equalize, BW is pulse BW
• – Typically baseband data will modulate in-phase [cos] and quadrature
[sine] data
• streams to the carrier passband
• • Special cases of modulated PAM include
• – phase shift keying (PSK)
• – quadrature amplitude modulation (QAM)

BrainKart.com
Need for Baseband Modulation

• An analog signal has a finite bandwidth.
• · A digital stream or signal, with sharp transitions, has an infinite
• bandwidth.
• · Due to the limited available system bandwidth, only the major
portion of
• a digital signal spectrum can be transmitted and restored. Even if
there is
• no loss or noise in the communication system, the received signal
will
• have distortion due to the limited channel bandwidth.
• To avoid or to reduce this signal distortion, we
• use baseband modulation techniques

BrainKart.com
Line Codes
• In telecommunication, a line code (also called digital baseband
modulation or digital baseband transmission method) is a code
chosen for use within a communications system for baseband
transmission purposes.
• Line coding is often used for digital data transport.
• Line coding consists of representing the digital signal to be

transported by an amplitude- and time-discrete signal that is
optimally tuned for the specific properties of the physical channel
(and of the receiving equipment).
• The waveform pattern of voltage or current used to represent the

1s and 0s of a digital data on a transmission link is called line
encoding.

BrainKart.com
Common types of Line Codes

• The common types of line encoding are
• unipolar
• polar
• bipolar
• Manchester encoding

BrainKart.com
Need for Line Codes

BrainKart.com
Properties of Line Codes

BrainKart.com
Unipolar coding
• Unipolar encoding is a line code. A positive voltage represents a
binary 1, and zero volts indicates a binary 0. It is the simplest line
code, directly encoding the bitstream, and is analogous to on-off
keying in modulation.
• Its drawbacks are that it is not self-clocking and it has a significant

DC component, which can be halved by using return-to-zero, where
the signal returns to zero in the middle of the bit period.
• With a 50% duty cycle each rectangular pulse is only at a positive

voltage for half of the bit period.
• This is ideal if one symbol is sent much more often than the other
and power considerations are necessary, and also makes the signal
self-clocking.

BrainKart.com
• NRZ(Non-Return-to-Zero) - Traditionally, a unipolar scheme was

designed as a non-return-to-zero (NRZ) scheme, in which the positive
voltage defines bit 1 and the negative voltage defines bit 0.
• It is called NRZ because the signal does not return to zero at the
middle of the bit.
• Compared with its polar counterpart, Uni Polar NRZ, this scheme is
very expensive.
• The normalized power (power required to send 1 bit per unit line
resistance) is double that for polar NRZ.
• For this reason, this scheme is not normally used in data

communications today.

BrainKart.com

BrainKart.com

BrainKart.com
Return-to-zero
• Return-to-zero (RZ) describes a line code used in
telecommunications signals in which the signal
drops (returns) to zero between each pulse.
• This takes place even if a number of consecutive
0s or 1s occur in the signal.
• The signal is self-clocking. This means that a
separate clock does not need to be sent
alongside the signal, but suffers from using twice
the bandwidth to achieve the same data-rate as
compared to non-return-to-zero format.

BrainKart.com

BrainKart.com
Polar RZ

BrainKart.com
BiPolar Signalling

BrainKart.com
Manchester Encoding

BrainKart.com
Inter Symbol Interference (ISI)

• In telecommunication, inter symbol
interference (ISI) is a form of distortion of
a signal in which one symbol interferes with
subsequent symbols.
• This is an unwanted phenomenon as the previous
symbols have similar effect as noise, thus making
the communication less reliable.
• ISI is usually caused by multipath propagation or
the inherent non-linear frequency response of
a channel causing successive symbols to "blur"
together.

BrainKart.com
Nyquist criterion for distorionless
transmission
• To design under the following two conditions:
(a). There is no ISI at the sampling instants (Nyquist criterion).

(b). A controlled amount of ISI is allowed (correlative coding)

BrainKart.com
Design of Bandlimited Signals for Zero
ISI - Nyquist criterion
• Recall the output of the receiving filter,
sampled at t = kT, is given by
• Thus, in time domain, a sufficient condition

for µp(t) such that it is ISI free is

BrainKart.com
• Theorem: (Nyquist) A necessary and and sufficient condition

for p(t) to satisfy (1) is that the Fourier transform P(f) satisfies
• This is known as the Nyquist pulse-shaping criterion or

Nyquist condition for zero ISI.

BrainKart.com

BrainKart.com
Investigate possible pulses which satisfy the
Nyquist criterion

BrainKart.com

BrainKart.com

BrainKart.com

BrainKart.com
Disadvantages:
(a) an ideal LPF is not physically realizable.
(b) Note that
Thus, the rate of convergence to zero is low since the

tails of p(t) decay as 1/|t|.
• Hence, a small mistiming error in sampling the
output of the matched filter at the demodulator
results in an infinite series of ISI components.

BrainKart.com

BrainKart.com

BrainKart.com

BrainKart.com

BrainKart.com
Correlative Coding and Equalization

• Correlative Coding
 For zero ISI, the symbol rate R = 1/T < 2W, the
Nyquist rate.
 We may relax the condition of zero ISI in order to
achieve R = 2W.
• The schemes which allow a controlled amount
of ISI to achieve the symbol rate 2W are called
correlative coding or partial response signaling
schemes.
BrainKart.com

BrainKart.com

BrainKart.com
2nd Nyquist Criterion

• Values at the pulse edge are distortionless
• p(t) =0.5, when t= -T/2 or T/2; p(t)=0, when t=(2k-1)T/2, k≠0,1
where -1/T ≤ f ≤ 1/T

Pr ( f )  Re[  (1)
n
n
P( f  n / T )]  T cos( fT / 2)

PI ( f )  Im[  (1) n P( f  n / T )]  0
n

BrainKart.com

BrainKart.com
3rd Nyquist Criterion

• Within each symbol period, the integration of signal (area) is proportional to
the integration of the transmit signal (area)
(wt) / 2 
sin(wT ,w
P(w)  / 2) T
 0, 
w 
 T
1  /T
p(t)  
(wt / 2) e jwt dw
2  / T sin (wT / 2)

1,
A   n 0
2 n1
p(t)dt  
T
2
2 n1
T 0
2
0, n
BrainKart.com
Eye Diagram
• Eye diagram is a means of evaluating the quality of a received
“digital waveform”
– By quality is meant the ability to correctly recover symbols and timing
– The received signal could be examined at the input to a digital receiver
or at some stage within the receiver before the decision stage
• Eye diagrams reveal the impact of ISI and noise
• Two major issues are 1) sample value variation, and 2) jitter
and sensitivity of sampling instant
• Eye diagram reveals issues of both
• Eye diagram can also give an estimate of achievable BER
• Check eye diagrams at the end of class for participation

BrainKart.com
Interpretation of Eye Diagram

BrainKart.com
Raised Cosine Eye Diagram

• The larger , the wider the
opening.
• The larger , the larger
bandwidth (1+ )/Tb
• But smaller  will lead to
larger errors if not sampled
at the best sampling time
which occurs at the center
of the eye.

BrainKart.com
Cosine rolloff filter: Eye pattern
2nd Nyquist 
1st Nyquist:
  1st Nyquist:

2nd Nyquist:

2nd Nyquist:

1st Nyquist







1st Nyquist:
  1st Nyquist:

2nd Nyquist:

2nd Nyquist:


BrainKart.com
Eye Diagram Setup
• Eye diagram is a retrace display of
data waveform
– Data waveform is applied to
input channel
– Scope is triggered by data
clock
– Horizontal span is set to cover
2-3 symbol intervals
• Measurement of eye opening is
performed to estimate BER
– BER is reduced because of
additive interference and noise
– Sampling also impacted by
jitter

BrainKart.com
Partial Response Signals
• Previous classes: Sy(w)=|P(w)|^2 Sx(w)
– Control signal generation methods to reduce Sx(w)
– Raise Cosine function for better |P(w)|^2
• This class: improve the bandwidth efficiency
– Widen the pulse, the smaller the bandwidth.
– But there is ISI. For binary case with two symbols, there is
only few possible interference patterns.
– By adding ISI in a controlled manner, it is possible to achieve
a signaling rate equal to the Nyquist rate (2W symbols/sec)
in a channel of bandwidth W Hertz.

BrainKart.com
Thank you

BrainKart.com
UNIT IV
DIGITAL MODULATION SCHEME

BrainKart.com
Geometric Representation of
Signals
• Objective: To represent any set of M energy
signals {si(t)} as linear combinations of N
orthogonal basis functions, where N ≤ M
• Real value energy signals s1(t), s2(t),..sM(t),
each of duration TOrtshoegocnal basis
function
N
0  t  T 
si (t)   sij j (t),   4.1
(5.5)
j 1  i==1,2, ... ,M

coefficient
Energy signal
BrainKart.com







• Coefficients:
T i=1,2,....,M 
sij   si (t) j (t)dt,   (5.6)

0
j=1,2, ... ,M
• Real-valued basis functions:
1 if i  j 
  (t) (t)dt
T
     (5.7)
 0 if i j
 
i j ij
0

BrainKart.com







• The set of coefficients can be viewed as a

N-dimensional vector, denoted by si
• Bears a one-to-one relationship with the
transmitted signal si(t)

BrainKart.com
(a) Synthesizer for generating the signal si(t). (b) Analyzer for
generating the set of signal vectors si.

BrainKart.com
So,
• Each signal in the set si(t) is completely
determined by the vector of its coefficients
si1 
 s 
 i 2 
. 
si   , i  1,2,. .. ,M (5.8)
. 
. 
 
 siN 

BrainKart.com



Finally,
• The signal vector si concept can be extended to 2D, 3D
etc. N-dimensional Euclidian space
• Provides mathematical basis for the geometric
representation of energy signals that is used in noise
analysis
• Allows definition of
– Length of vectors (absolute value)
– Angles between vectors
– Squared value (inner product of si with itself)
si 2  sT s Matrix
i i Transposition
N
= s 2ij , i  1,2,. .. ,M (5.9)

j 1

BrainKart.com
Figure 5.4
Illustrating the geometric
representation of signals
for the case when N  2
and M  3.
(two dimensional space,
three signals)

BrainKart.com
Also,
What is the relation between the vector
representation of a signal and its energy value?
• …start with the T
Ei  s2i (t)dt
definition of average
energy in a signal…  0
(5.10)
• Where si(t) is N
si (t)   sij j (t), (5.5)

j 1

BrainKart.com
 N 
 T N
• After substitution: Ei    sij j (t)    sikk (t)  dt
0  j 1   k 1 



N N T
• After regrouping: Ei    s s   (t) (t)dt

ij ik j k (5.11)
j 1 k 1 0
• Φj(t) is orthogonal, so N
E i  s
2
2
finally we have: ij = si (5.12)
j 1
The energy of a
signal is equal to the
squared length of its
vector
BrainKart.com
Formulas for two signals

• Assume we have a pair of signals: si(t) and sj(t),
each represented by its vector,
• Then:
s  T
T
s (t)s (t)dt  s s
ij  0 i k i k
(5.13)
Inner product is invariant

to the selection of basis
Inner product of the functions
signals is equal to the
inner product of their
vector representations
[0,T]
BrainKart.com
Euclidian Distance
• The Euclidean distance between two points
represented by vectors (signal vectors) is equal
to
||si-sk|| and the squared
N
value is given by:
= (s -s )2
2
s s (5.14)
i k ij kj
j 1
T (t)  s (t))2 dt
= (s
0 i k

BrainKart.com
BPSK - BINARY PHASE
SHIFT KEYING
Generation of BPSK:
• Consider a sinusoidal carrier. If it is modulated
by a bi-polar bit stream, its polarity will be
reversed every time the bit stream changes
polarity.
• This, for a sinewave, is equivalent to a phase
reversal (shift). The multiplier output is a BPSK 1
signal.

BrainKart.com

BrainKart.com
BPSK signal in time domain

BrainKart.com
sychronous demodulation of BPSK

BrainKart.com
Frequency Shift Keying (FSK)

• Binary FSK
• Frequency of the constant amplitude carrier is changed

according to the message state  high (1) or low (0)
s1 (t)  Acos(2fc  2f )t 0  t  Tb (bit  1)

s2 (t)  Acos(2fc  2f )t 0  t  Tb (bit  0)
• Discontinuous / Continuous Phase

BrainKart.com
Discontinuous Phase FSK

Switching between 2 independent oscillators for binary 1
&0
input data phase jumps
cos w1t
switch
cos w2t
binary 1 sBFSK(t)= vH(t)

2Eb
= cos(2fH t 1) 0  t  Tb
Tb
binary 0 sBFSK(t)= vL(t)

2Eb
= cos(2f Lt   2 ) 0  t  Tb
Tb
• results in phase discontinuities
• discontinuities causes spectral spreading & spurious transmission
• not suited for tigG
htely
t udseesfuigl n
steuddysymsatetem
ris
als from BrainKart.com
BrainKart.com
Continuous Phase FSK

single carrier that is frequency modulated using m(t)
2Eb
sBFSK(t) = cos(2fct   (t))
Tb
 t

2Eb FSK 
= cos2fct  2k m()d 
Tb   
t
where (t) = 2kFSK m()d
 


• m(t) = discontinuous bit stream

• (t) = continuous phase function proportional to integral of m(t)

BrainKart.com
FSK Example
Data
1 1 0 1
FSK
Signal
0 1 1
x
a0 0 VCO modulated composite
a1 1 signal
cos wct
BrainKart.com
Spectrum & Bandwidth of BFSK Signals
• complex envelope of BFSK is nonlinear function of m(t)

• spectrum evaluation - difficult - performed using actual time
averaged measurements
PSD of BFSK consists of discrete frequency components at

• fc
• fc  nf , n is an integer
PSD decay rate (inversely proportional to spectrum)

1
• PSD decay rate for CP-BFSK  f 4
1
• PSD decay rate for non CP-BFSK  f 2
f = frequency offset from fc

BrainKart.com
Spectrum & Bandwidth of BFSK Signals
Transmission Bandwidth of BFSK Signals (from Carson’s Rule)

• B = bandwidth of digital baseband signal
• BT = transmission bandwidth of BFSK signal
BT = 2f +2B
• assume 1st null bandwidth used for digital signal, B

- bandwidth for rectangular pulses is given by B = Rb
- bandwidth of BFSK using rectangular pulse becomes
BT = 2(f + Rb)
if RC pulse shaping used, bandwidth reduced to:

BT = 2f +(1+) Rb
BrainKart.com
General FSK signal and orthogonality
• Two FSK signals, VH(t) and VL(t) are orthogonal if

T
VH (t) VL (t)dt  0

0
?
• interference between VH(t) and VL(t) will average to 0 during
demodulation and integration of received symbol
• received signal will contain VH(t) and VL(t)
• demodulation of VH(t) results in (VH(t) + VL(t))VH(t)

T
T
0
V H (t)VL (t)dt  0 V
0
H (t )VH (t )dt  0 ?
BrainKart.com
An FSK signal for 0 ≤ t ≤ Tb
vH(t) = 2Eb 2Eb
cos(2 ( fc  f )t) and vL(t) = cos(2 ( fc  f )t)
Tb Tb
2Eb
then vH(t) vL(t) = cos(2 ( f c  f )t)cos(2 ( f c  f )t)
Tb
=
Eb
cos(2 (2 fc )t)  cos(2 (2f )t)
Tb

 T Tb
and
0 VH (t)VL (t)dt  
Eb
cos(4fct)  cos(4ft )dt
0
Tb
sin(4ft ) 
Tb
E sin(4f t) E sin(4f T ) sin(4fT ) 
b  c  b  c
=    = 
b
b 

Tb  4fc 4f 0 Tb  4fc 4f 




vH(t) vL(t) are orthogonal if Δf sin(4πfcTb) = -fc(sin(4πΔf Tb)

BrainKart.com
QPSK
• Quadrature Phase Shift Keying (QPSK)
can be interpreted as two independent
BPSK systems (one on the I-channel
and one on Q-channel), and thus the
same performance but twice the
bandwidth (spectrum) efficiency.

BrainKart.com
QPSK Constellation Diagram

Q
Q
I I
Carrier phases
Carrier phases
{0, /2, , 3/2}
{/4, 3/4, 5/4, 7/4}
• Quadrature Phase Shift Keying has twice the

bandwidth efficiency of BPSK since 2 bits are
transmitted in a single modulation symbol
BrainKart.com
Types of QPSK
Q Q Q
I I I
Conventional QPSK Offset QPSK /4 QPSK
• Conventional QPSK has transitions through zero (i.e. 1800 phase

transition). Highly linear amplifiers required.
• In Offset QPSK, the phase transitions are limited to 900, the transitions on
the I and Q channels are staggered.
• In /4 QPSK the set of constellation points are toggled each symbol, so
transitions through zero cannot occur. This scheme produces the lowest
envelope variations.
• All QPSK schemes require linear power amplifiers
BrainKart.com
Quadrature Phase Shift Keying (QPSK):

• Also a type of linear modulation scheme
• Quadrature Phase Shift Keying (QPSK) has twice the bandwidth efficiency of
BPSK, since 2 bits are transmitted in a single modulation symbol.
• The phase of the carrier takes on 1 of 4 equally spaced values, such as
where each value of phase corresponds to a unique pair of message bits.
• The QPSK signal for this set of symbol states may be defined as:

BrainKart.com
QPSK
• The striking result is that the bit error probability of QPSK is identical to
BPSK, but twice as much data can be sent in the same bandwidth. Thus,
when compared to BPSK, QPSK provides twice the spectral efficiency with
exactly the same energy efficiency.
• Similar to BPSK, QPSK can also be differentially encoded to allow non-
coherent detection.

BrainKart.com

Quadrature amplitude mo dulation
www.rejinpaul.com
• Quadrature amplitude modulation (QAM) is both an

analog and a digital modulation scheme.
• It conveys two analog message signals, or two

digital bit streams, by changing (modulating)
the amplitudes of two carrier waves, using
the amplitude-shift keying(ASK) digital modulation
scheme or amplitude modulation (AM) analog
modulation scheme.
• The two carrier waves, usually sinusoids, are out of

phase with each other by 90° and are thus
called quadrature carriers or quadrature components
— hence the name of the scheme.

BrainKart.com
QAM Transmitter

BrainKart.com
• First the flow of bits to be transmitted is split

into two equal parts: this process generates
two independent signals to be transmitted.
• They are encoded separately just like they
were in an amplitude-shift keying (ASK)
modulator.
• Then one channel (the one "in phase") is
multiplied by a cosine, while the other
channel (in "quadrature") is multiplied by a
sine.
• This way there is a phase of 90° between
them. They are simply added one to the other
and sent throGuegt uhsetfuhl e
stur
de aaltecrih
ym alsafn newwl.w.rejinpaul.com
rom
BrainKart.com
QAM Receiver

BrainKart.com
• The receiver simply performs the inverse

operation of the transmitter.
• Multiplying by a cosine (or a sine) and by a
low-pass filter it is possible to extract the
component in phase (or in quadrature).
• Then there is only an ASK demodulator
and the two flows of data are merged
back.

BrainKart.com
Carrier Synchronization
• Synchronization is one of the most critical
functions of a communication system with
coherent receiver. To some extent, it is the
basis of a synchronous communication
system.
• Carrier synchronization
• Symbol/Bit synchronization
• Frame synchronization
BrainKart.com
• Receiver needs estimate and compensate for frequency and phase

differences between a received signal’s carrier wave and the
receiver’s local oscillator for the purpose of coherent demodulation,
no matter it is analog or digital communication systems.
• To extract the carrier：
• 1. Pilot-tone insertion method

• Sending a carrier component at specific spectral-line along with
the signal component. Since the inserted carrier component has
high frequency stability, it is called pilot.
• 2. Direct extraction method

• Directly extract the synchronization information from the
received signal component.

BrainKart.com
1. Pilot-tone insertion method
－insert pilot to the modulated signal
x(t) Modulator Bandpass Add s(t)

filter
-asin(ct)
cos(ct) /2phase
shift
• The pilot signal is generated by shift the

carrier by 900 and decrease by several dB,
then add to the modulated signal. Assume
the modulated signal has 0 DC component,
then the pilots ist  f  t  cos ct  a sin c t
BrainKart.com
2. Direct extraction method

• If the spectrum of the received signal
already contains carrier component, then
the carrier component can be extracted
simply by a narrowband filter or a PLL.
• If the modulated signal supresses the
carrier component, then the carrier
component may be extracted by
performing nonlinear transformation or
using a PLL with specific design

BrainKart.com
DPSK
• DPSK is a kind of phase shift keying which
avoids the need for a coherent reference
signal at the receiver.
• Differential BPSK
– 0 = same phase as last signal element
– 1 = 180º shift from last signal element

BrainKart.com
DPSK modulation and
demodulation

BrainKart.com
Thank you

BrainKart.com
Unit - 5
Error Controlling Codes

BrainKart.com
ENTROPY
• The entropy of a source is defined as “the source which produces average
information per individual message or symbol in a particular interval”.
• Then the number of messages is given as
• The amount of information in message m1 is given as
• The total amount of information

BrainKart.com
Cont…
• Thus the total amount of information due to L

BrainKart.com
Cont…

BrainKart.com
PROPERTIES OF ENTROPY
• The entropy of a discrete memoryless channel
source
• Property 1
1. Entropy is zero, if the event is sure or its impossible

BrainKart.com
Cont…
Property 2
2. Entropy H= log2 K when the symbols are equally likely for K
symbols PK= 1/K

BrainKart.com
Cont…
Property 3
3. Maximum upper bound on entropy is

BrainKart.com
Cont…

BrainKart.com
Cont…

BrainKart.com
Cont..

BrainKart.com
RATE OF INFORMATION
• RATE OF INFORMATION

BrainKart.com
SOURCE CODING
• An important problem in communication system is the
efficient representation of data generated by a source, which
can be achieved by source encoding (or) source coding
process
• The device which performs source encoding is called source
encoder

BrainKart.com
EFFICIENCY OF SOURCE ENCODER
• What are the requirements that a source

encoder should satisfy for it to be efficient-----
----- Binary codewords & Discrete memory-less
source
• What are the advantages of source coding?

BrainKart.com
NOISELESS CODING THEOREM

Shanons information theory for a discrete memoryless
sources and channels involves three theorems.
(i) Shanons first theorem (or) Source coding theorem

(ii) Shanons second theorem (or) channel coding theorem (or)
shanons theorem on channel capacity
(iii) Shanons third theorem (or) information capacity theorem
(or) shanon hartley theorem.

BrainKart.com
SHANONS FIRST THEOREM
• This theorem is also called as source coding theorem. Consider a discrete
memoryless source
Encoding process
The average code word length L is given as
The coding efficiency of the source encoder is

BrainKart.com
STATEMENT
• Shanon first theorem is stated as “ Given a discrete memoryless source of
entropy H, the average codeword length L for any distortionless source
encoding is bounded as
L>= H
• According to source coding theorem the entropy H represents as the
fundamental limit on the average number of bits per source symbol
necessary to represent a discrete memoryless source
• It can be made as small as, but not smaller than the entropy H thus, with
Lmin=H, then η is represented as
η= H/L

BrainKart.com
SHANONS SECOND THEOREM

• Why channel coding?
• Goal of channel coding
• Process

BrainKart.com
Cont…
• Approach
The code rate is given as
Suppose a discrete memoryless source alphabet S, Entropy H(S) bits per

source symbol and source emits and delivers symbols for every Ts seconds

BrainKart.com
STATEMENT
Two parts

BrainKart.com
SHANONS THIRD THEOREM

The channel capacity of a white band limited
gaussian channel is
Signal power
Noise power

BrainKart.com
MUTUAL INFORMATION

BrainKart.com
MUTUAL INFORMATION
• The difference between these two values is
called mutual information

BrainKart.com
PROPERTIES OF MUTUAL INFORMATION
1. The mutual info of channel is symmetric
2. The mutual information is always non negative
3. The mutual information of a channel is related to the joint

entropy of the channel input and the channel output by

BrainKart.com
CHANNEL CAPACITY

BrainKart.com
CHANNEL CAPACITY
• What is channel capacity?
• It must satisfy two constraints

BrainKart.com
Cont…
• Transmission efficiency
• Redundancy

BrainKart.com
Error Control
Coding

BrainKart.com
Introduction
• Error Control Coding (ECC)
– Extra bits are added to the data at the transmitter
(redundancy) to permit error detection or
correction at the receiver
– Done to prevent the output of erroneous bits
despite noise and other imperfections in the
channel
– The positions of the error control coding and
decoding are shown in the transmission model

BrainKart.com
Transmission Model
Error Modulator
Digital Source Line X()
Control (Transmit
Source Encoder Coding
Coding Filter, etc)
Hc() Channel
Transmitter
N() Noise
+
Error Demod
Digital Source Line
Control (Receive
Sink Decoder Decoding Y()
Decoding Filter, etc)
Receiver

BrainKart.com
Error Models
• Binary Symmetric Memoryless Channel
– Assumes transmitted symbols are binary
– Errors affect ‘O’s and ‘1’s with equal probability
(i.e., symmetric)
– Errors occur randomly and are independent
from bit to bit (memoryless)
1-p
0 0 p is the probability of bit
p
error or the Bit Error Rate
IN OUT (BER) of the channel
p
1 1
1-p
BrainKart.com
Error Models
• Many other types

• Burst errors, i.e., contiguous bursts of bit
errors
– output from DFE (error propagation)
– common in radio channels
– Insertion, deletion and transposition errors
• We will consider mainly random errors

BrainKart.com
Error Control Techniques

• Error detection in a block of data
– Can then request a retransmission, known as
automatic repeat request (ARQ) for sensitive
data
– Appropriate for
• Low delay channels
• Channels with a return path
– Not appropriate for delay sensitive data, e.g.,
real time speech and data

BrainKart.com
Error Control Techniques

• Forward Error Correction (FEC)
– Coding designed so that errors can be corrected at
the receiver
– Appropriate for delay sensitive and one-way
transmission (e.g., broadcast TV) of data
– Two main types, namely block codes and
convolutional codes. We will only look at block
codes

BrainKart.com
Block Codes
• We will consider only binary data
• Data is grouped into blocks of length k bits
(dataword)
• Each dataword is coded into blocks of length n
bits (codeword), where in general n>k
• This is known as an (n,k) block code

BrainKart.com
Block Codes
• A vector notation is used for the datawords

and codewords,
– Dataword d = (d1 d2….dk)
– Codeword c = (c1 c2 .......... cn)
• The redundancy introduced by the code is
quantified by the code rate,
– Code rate = k/n
– i.e., the higher the redundancy, the lower the
code rate
BrainKart.com
Block Code - Example
• Dataword length k = 4
• Codeword length n = 7
• This is a (7,4) block code with code rate = 4/7
• For example, d = (1101), c = (1101001)

BrainKart.com
Error Control Process

Source code Codeword
101101 1000 data chopped (n bits)
into blocks Channel
1000 coder
Dataword
(k bits)
Codeword +
Dataword possible errors
(k bits) Channel
(n bits)
Channel
decoder
Error flags

BrainKart.com
Error Control Process

• Decoder gives corrected data
• May also give error flags to
– Indicate reliability of decoded data
– Helps with schemes employing multiple layers of
error correction

BrainKart.com
Hamming Distance
• Error control capability is determined by

the Hamming distance
• The Hamming distance between two
codewords is equal to the number of
differences between them, e.g.,
10011011
11010010 have a Hamming distance = 3
• Alternatively, can compute by adding
codewords (mod 2)
=01001001 (now count up the ones)
BrainKart.com
Hamming Distance
• The Hamming distance of a code is equal to
the minimum Hamming distance between
two codewords
• If Hamming distance is:
1 – no error control capability; i.e., a single error
in a received codeword yields another valid
codeword
XXXXXXX X is a valid codeword
Note that this representation is diagrammatic
only.
In reality each codeword is surrounded by n
codewords. That is, one for every bit that
could be cGheat unsgeefudl study materials from BrainKart.com
BrainKart.com
Hamming Distance

2 – can detect single errors (SED); i.e., a single error
will yield an invalid codeword
XOXOXO X is a valid codeword
O in not a valid codeword
See that 2 errors will yield a valid (but
incorrect) codeword

BrainKart.com
Hamming Distance

3 – can correct single errors (SEC) or can detect
double errors (DED)
XOOXOOX X is a valid codeword
O in not a valid codeword
See that 3 errors will yield a valid but incorrect
codeword

BrainKart.com
Hamming Distance - Example
• Hamming distance 3 code, i.e., SEC/DED

– Or can perform single error correction (SEC)
10011011 X
This code corrected this way
11011011 O
11010011 O This code corrected this way
11010010 X
X is a valid codeword
O is an invalid codeword

BrainKart.com
Hamming Distance
• The maximum number of detectable errors is
dmin 1
• That is the maximum number of correctable
errors is given by,
 dmin 1

t

 2 

where dmin is the minimum Hamming distance

between 2 codewords and. means the smallest
integer

BrainKart.com
Linear Block Codes
• As seen from the second Parity Code

example, it is possible to use a table to hold
all the codewords for a code and to look-up
the appropriate codeword based on the
supplied dataword
• Alternatively, it is possible to create
codewords by addition of other codewords.
This has the advantage that there is now no
longer the need to held every possible
codeword in the table.
BrainKart.com
Linear Block Codes

• If there are k data bits, all that is required is to hold k
linearly independent codewords, i.e., a set of k
codewords none of which can be produced by linear
combinations of 2 or more codewords in the set.
• The easiest way to find k linearly independent
codewords is to choose those which have ‘1’ in just
one of the first k positions and ‘O’ in the other k-1 of
the first k positions.

BrainKart.com
Linear Block Codes

• For example for a (7,4) code, only four
codewords are required, e.g.,
1 0 0 0 1 1 0
0 1 0 0 1 0 1
0 0 1 0 0 1 1
0 0 0 1 1 1 1
• So, to obtain the codeword for dataword 1011,

the first, third and fourth codewords in the list
are added together, giving 1011010
• This process will now be described in more detail
BrainKart.com
Linear Block Codes

• An (n,k) block code has code vectors
d=(d1 d2….dk) and
c=(c1 c2 ...........cn)
• The block coding process can be written as
c=dG
where G is the Generator Matrix
a11 a12 ... a1n   a1 
a21 a22 ... a2n   
 a2
G      
 . . ... .   . 
   
 ak1Get uaskef2ul stu.d.y. maatekrinalsfromwawkw.rejinpaul.com
BrainKart.com
Linear Block Codes



• Thus,
k
c   di a i
i1
• ai must be linearly independent, i.e.,

Since codewords are given by summations
of the ai vectors, then to avoid 2 datawords
having the same codeword the ai vectors
must be linearly independent

BrainKart.com
Linear Block Codes

• Sum (mod 2) of any 2 codewords is
also a codeword, i.e.,
Since for datawords d1 and d2 we have;
d3  d1  d2
So,
k k k k
c3  d 3i a i   (d1i  d2i )ai  d1i a i   d 2i a i

i1 i1 i1 i1
c3  c1  c2

BrainKart.com
Linear Block Codes
• 0 is always a codeword, i.e.,

Since all zeros is a dataword then,
k
c   0 ai  0
i1

BrainKart.com
Error Correcting Power of LBC

• The Hamming distance of a linear block code (LBC) is
simply the minimum Hamming weight (number of
1’s or equivalently the distance from the all O
codeword) of the non-zero codewords
• Note d(c1,c2) = w(c1+ c2) as shown previously
• For an LBC, c1+ c2=c3
• So min (d(c1,c2)) = min (w(c1+ c2)) = min (w(c3))
• Therefore to find min Hamming distance just need
to search among the 2k codewords to find the min
Hamming weight – far simpler than doing a pair
wise check for all possible codewords.
BrainKart.com
Linear Block Codes – example 1
• For example a (4,2) code, suppose;

1 0 1 1 a1 = [1011]
G    a2 = [0101]
0 1 0 1
• For d = [1 1], then;
1 0 1 1
 0 1 0 1
c 
 _ _ _ _
 1 1 1 0

BrainKart.com

Linear Block Codes – example 2




• A (6,5) code with

1 0 0 0 0 1
 
0 1 0 0 0 1 

G  0 0 1 0 0 1
 
0 0 0 1 0 1
 1
0 0 0 0 1 
• Is an even single parity code

www.rejinpaul.com
Systematic Codes
• For a systematic block code the dataword
appears unaltered in the codeword –
usually at the start
• The generator matrix has the structure,
k R R=n-k
1 0 .. 0 p11 p12 .. p1R 

 
G  0 1 .. 0 p21 p22 .. p2R  I | P
.. .. .. .. .. .. .. .. 
 
0 0 .. 1 pk1 pk 2 .. pkR 


• P is often referred to as parity bits

BrainKart.com
Systematic Codes
• I is k*k identity matrix. Ensures dataword
appears as beginning of codeword
• P is k*R matrix.

BrainKart.com
Decoding Linear Codes
• One possibility is a ROM look-up table
• In this case received codeword is used as an address
• Example – Even single parity check code;
Address Data
000000 0
000001 1
000010 1
000011 0
……… .
• Data output is the error flag, i.e., 0 – codeword ok,
• If no error, dataword is first k bits of codeword
• For an error correcting code the ROM can also store
datawords Get useful study materials from BrainKart.com
BrainKart.com
Decoding Linear Codes
• Another possibility is algebraic decoding, i.e.,

the error flag is computed from the received
codeword (as in the case of simple parity
codes)
• How can this method be extended to more
complex error detection and correction
codes?

BrainKart.com
CYCLIC CODES

BrainKart.com
Motivation & Properties of cyclic code
• Cyclic code are a class of linear block codes.

Thus, we can find generator matrix (G) and
parity check matrix (H).
• The reason is that they can be easily
implemented with externally cost effective
electronic circuit.

BrainKart.com
Definition
• An (n,k) linear code C is cyclic if every cyclic
shift of a codeword in C is also a codeword in
C.
If c0 c1 c2 …. cn-2 cn-1 is a codeword, then
cn-1 c0 c1 …. cn-3 cn-2
cn-2 cn-1 c0 …. cn-4 cn-3
: : : : :
c1 c2 c3 …. cn-1 c0 are all codewords.

BrainKart.com
Example: (6,2) repetition code
C  000000,010101,101010,111111


is a cyclic code.

BrainKart.com
Example2: (5,2) linear block code

1 0 1 1 1
G   
0 1 1 0 1


is a single error correcting code, the set of codeword are:
0 0 0 0 0 Thus, it is not a cyclic code

 0 1 1 1 since, for example, the cyclic
C  1  shift of [10111] is [11011]
0 1 1 0 1 
  C
1 1 0 1 0 
BrainKart.com
Example 3
• The (7,4) Hamming code discussed before is
cyclic:
1010001 1110010 0000000 1111111

1101000 0111001
0110100 1011100
0011010 0101110
0001101 0010111
1000110 1001011
0100011 1100101

BrainKart.com
Generator matrix of a non-systematic (n,k) cyclic

codes
• The generator matrix will be in this form:
g0 g1  gnk1 gnk 0 0  0 

 
0 g0 g1  gnk1 gnk 0  0 
G 0 0 g0 g1  gnk1 gnk 0 0 
 
  
 0 0 0  0 g0  gnk1 
gnk 

notice that the row are merely cyclic shifts of the r n

basis vector 
g  g0g1 gnk1gnk 000 
BrainKart.com
• The code vector are:
C  mG; where m  [m0m1 mk 1 ]
 c0  m0 g 0
c1  m0 g 1  m1 g 0
c2  m0g2  m1g1  m 2g0

cn1  m 0g nk  m1gnk 1  mnk g 0
Notice thatk,1
C   m j g 1 ; where m j  0, if j  0or j  k 1
j 0
This summation is a convolution between and . m g

• It would be much easier if we deal with multiplication, this transform is done
using Polynomial Representation.

BrainKart.com
Code Polynomial
• Let c = c0 c1 c2 …. cn-1. The code polynomial
of c: c(X) = c0 + c1X+ c2 X2 + …. + cn-1 Xn-1
where the power of X corresponds to the bit position,
and
the coefficients are 0’s and 1’s.
• Example:
1010001 1+X2+X6
0101110 X+X3+X4+X5
Each codeword is represented by a polynomial of degree
deg[c(X)]  n  1
less than or equal n-1.

BrainKart.com
The addition and multiplication are as follow:
ax j  bx j  (a  b)x j Where (a+b) and (a.b) are under GF(2). But
(ax j ).(bxk )  (a.b) x j k j+k is integral addition
Example:
m(x)  m0  m1x  m2 x 2
g(x)  g0  g1x
   m(x)  g(x)  (m 0  g0 )  (m1  g1 )x  (m2  0)x 2
additi on
   m(x)g(x)  m g  (m g  m g )x  (m g  m g )x 2  m g x 3
Mul tipliaiton
0 0 0 1 1 0 1 1 2 0 2 1
Notice that in multiplication the coefficient are the

same as convolution sum

BrainKart.com
Implementing the Shift

Let c = c0 c1 c2 …. cn-1
and c(i) = cn-i cn-i+1 c0 …. cn-i-1 (i shifts to the right)
c(X) = c0 + c1X+ c2 X2 + …. + cn-1 Xn-1
c (i)(X) = cn-i + cn-i+1 X + …. + cn-1 Xi-1 + …. +
c0Xi +…. +cn-i-1 Xn-1
What is the relation between c(X) and c (i)(X)?
Apparently, shifting a bit one place to the right is
equivalent
to multiplying the term by X.
Xic(X)= c0Xi +c1X i+1 + ….+ cn-i-1 Xn-1 + cn-i Xn ….+ cn-1 Xn+i-1

BrainKart.com
Implementing the Shift (cont’d)

Xic(X) = cn-i Xn +…+cn-1 Xn+i-1 +c0Xi +c1X i+1 + …+ cn-i-1 Xn-1
The first i terms have powers n, and are not suitable
for 
representing bit locations.
Add to the polynomial the zero-valued sequence:
(cn-i + cn-i ) + (cn-i+1 + cn-i+1 )X + …. + (cn-1 + cn-1 )Xi-1
Xic(X) = cn-i (Xn +1) + cn-i+1 X (Xn +1)+…. +cn-1 Xi-1 (Xn +1)+
cn-i + cn-i+1 X +…. +cn-1 Xi-1+
c0Xi +c1X i+1 + …. + cn-i-1 Xn-1
That is:
Xic(X) = q(X)(Xn +1) + c(i)(X)

BrainKart.com
Implementing the Shift (cont’d)

c(i)(X) is the remainder from dividing Xic(X) by (Xn +1).
c(i)(X) = Rem[Xic(X)/ (Xn +1)] = Xic(X) mod (Xn +1).
Example:
c = 0101110. c(X) = X + X3 + X4 + X5.
X3c(X) = X4 + X6 + X7 + X8
Rem[X3c(X)/ (X7 +1)] = 1 + X + X4 + X6 [Show]
c(3) = 1100101
Short cut of long division:
n n
Xic(X)|Xn=1 = q(X)(Xn +1) | X =1 + c(i)(X) | X =1
That is c(i)(X) = Xic(X)|Xn=1

BrainKart.com
More on Code Polynomials

• The nonzero code polynomial of minimum degree in a
cyclic code C is unique.
(If not, the sum of the two polynomials will be a code polynomial of degree less than
the minimum. Contradiction)
• Let g(X) = g0 + g1X +….+ gr-1Xr-1 +Xr be the nonzero code

polynomial of minimum degree in an (n,k) cyclic code.
Then the constant term g0 must be equal to 1.
(If not, then one cyclic shift to the left will produce a code polynomial of degree less
than the minimum. Contradiction)
• For the (7,4) code given in the Table, the nonzero code
polynomial of minimum degree is g(X) = 1 + X + X3

BrainKart.com
Generator Polynomial
• Since the code is cyclic: Xg(X), X2g(X),…., Xn-r-1g(X)
are code polynomials in C. (Note that deg[Xn-r-
1g(X)] = n-1).
• Since the code is linear:

(a0 + a1X + …. + an-r-1 Xn-r-1)g(X) is also a code
polynomial, where ai = 0 or 1.
• A binary polynomial of degree n-1 or less is a
code polynomial if and only if it is a multiple of
g(X).
(First part shown. Second part: if a code polynomial c(X) is not a
multiple of g(X), then Rem[c(X)/g(X)] must be a code polynomial of
degree less than the minimum. Contradiction)

BrainKart.com
Generator Polynomial (cont’d)

• All code polynomials are generated from the
multiplication c(X) = a(X)g(X).
deg[c(x)] n-1, deg[g(X)] = r, ==> deg[a(x)] n-r-1
# codewords, (2k) = # different ways of forming a(x), 2n-r
Therefore, r = deg[g(X)] = n-k
• Since deg[a(X)] k-1, the polynomial a(X) may be taken
to be the information polynomial u(X) (a polynomial
whose coefficients are the information bits). Encoding
is performed by the multiplication c(X) = u(X)g(X).
• g(X), generator polynomial, completely defines the
code.

BrainKart.com
(7,4) Code Generated by 1+X+X3

Infor. Code Code polynomials
0000 0000000 0 = 0 . g(X)
1000 1101000 1 + X + X3 = 1 . g(X)
0100 0110100 X + X2 + X4 = X . g(X)
1100 1011100 1 + X2 + X3 + X4 = (1 + X) . g(X)
0010 0011010 X2 + X3 + X5 = X2 . g(X)
1010 1110010 1 + X+ X2 + X5 = (1 + X2) . g(X)
0110 0101110 X+ X3 + X4 + X5 = (X+ X2) . g(X)
1110 1000110 1 + X4 + X5 = (1 + X + X2) . g(X)
0001 0001101 X3 + X4 + X6 = X3 . g(X)

www.rejinpaul.com
3
(7,4) Code Generated by 1+X+X
(Cont’d)
Infor. Code Code polynomials
1001 1100101 1 + X + X4 + X6 = (1 + X3) . g(X)
0101 0111001 X+ X2 + X3 + X6 = (X+ X3) . g(X)
1101 1010001 1 + X2 + X6 = (1 + X + X3) . g(X)
0011 0010111 X2 + X4 + X5 + X6 = (X2 + X3). g(X)
1011 1111111 1 + X + X2 + X3 + X4 + X5 + X6
= (1 + X2 + X3) . g(X)
0111 0100011 X + X5 + X6 = (X + X2 + X3). g(X)
1111 1001011 1 + X3 + X5 + X6
= (1 + X + X2 + X3) . g(X)
BrainKart.com
Constructing g(X)
• The generator polynomial g(X) of an (n,k) cyclic
code is a factor of Xn+1.
Xkg(X) is a polynomial of degree n.
Xkg(X)/ (Xn+1)=1 and remainder r(X). Xkg(X) = (Xn+1)+ r(X).
But r(X)=Rem[Xkg(X)/(Xn+1)]=g(k)(X) =code polynomial= a(X)g(X).
Therefore, Xn+1= Xkg(X) + a(X)g(X)= {Xk + a(X)}g(X). Q.E.D.
(1) To construct a cyclic code of length n, find the
factors of the polynomial Xn+1.
(2) The factor (or product of factors) of degree n-k
serves as the generator polynomial of an (n,k)
cyclic code. Clearly, a cyclic code of length n does
not exist for every k.

BrainKart.com
Constructing g(X) (cont’d)

(3) The code generated this way is guaranteed to be cyclic.
But we know nothing yet about its minimum distance.
The generated code may be good or bad.
Example: What cyclic codes of length 7 can be
constructed?
X7+1 = (1 + X)(1 + X + X3)(1 + X2 + X3)
g(X) Code g(X) Code
(1 + X) (7,6) (1 + X)(1 + X + X3) (7,3)
(1 + X + X3) (7,4) (1 + X) (1 + X2 + X3) (7,3)
(1 + X2 + X3) (7,4) (1 + X + X3)(1 + X2 + X3) (7,6)

BrainKart.com
Circuit for Multiplying Polynomials (1)

• u(X) = uk-1Xk-1 + …. + u1X + u0
• g(X) = grXr + gr-1Xr-1 + …. + g1X + g0
• u(X)g(X) = uk-1grXk+r-1
+ (uk-2gr+ uk-1gr-1)Xk+r-2 + ….
+ (u0g2+ u1g1 +u2g0)X2 +(u0g1+ u1g0)X +u0g0
+ + + + Output
gr gr-1 gr-2 g1 g0
Input
BrainKart.com
Circuit for Multiplying Polynomials (2)
• u(X)g(X) = uk-1Xk-1(grXr + gr-1Xr-1 + …. + g1X + g0)

+ ….
+ u1X(grXr + gr-1Xr-1 + …. + g1X + g0)
+ u0(grXr + gr-1Xr-1 + …. + g1X + g0)
+ + + +
Output
g0 g1 g2 gr-1 gr
Input
BrainKart.com
Systematic Cyclic Codes

Systematic: b0 b1 b2 …. bn-k-1 u0 u1 u2 …. uk-1
b(X) = b0 + b1X+….+bn-k-1Xn-k-1, u(X) = u0+u1X+ ….+uk-1Xk-1
then c(X) = b(X) + Xn-k u(X)
a(X)g(X) = b(X) + Xn-k u(X)
Xn-k u(X)/g(X) = a(X) + b(X)/g(X)
Or b(X) = Rem[Xn-k u(X)/g(X)]
Encoding Procedure:
1. Multiply u(X) by Xn-k
2. Divide Xn-k u(X) by g(X), obtaining the remainder b(X).
3. Add b(X) to Xn-k u(X), obtaining c(X) in systematic form.

BrainKart.com
Systematic Cyclic Codes (cont’d)

Example
Consider the (7,4) cyclic code generated by
g(X) = 1 + X + X3. Find the systematic codeword for
the message 1001.
u(X) = 1 + X3
X3u(X) = X3 + X6
b(X) = Rem[X3u(x)/g(X)] = X3u(x) |g(X) = 0 = X3u(x) |X3
= X+1
= X3 (X3 +1) = (1 + X)X = X + X2
Therefore, c = 0111001
BrainKart.com
Circuit for Dividing Polynomials
Output
g0 g1 g2 gr-1 gr
+ + + +
Input

BrainKart.com
Encoder Circuit
Gate
g1 g2 gr-1
+ + +
+
• Gate ON. k message bits are shifted into the channel.

The parity bits are formed in the register.
• Gate OFF. Contents of register are shifted into the
channel.

BrainKart.com
(7,4) Encoder Based on 1 + X + X3
Gate
+
+
Input 1 1 0 1
Register : 000 110 101 100 100
initial 1st shift 2nd shift 3rd shift 4th shift
Codeword: 1 0 0 1 0 1 1

BrainKart.com
Parity-Check Polynomial
• Xn +1 = g(X)h(X)
• deg[g(x)] = n-k, deg[h(x)] = k
• g(x)h(X) mod (Xn +1) = 0.
• h(X) is called the parity-check polynomial. It
plays the rule of the H matrix for linear codes.
• h(X) is the generator polynomial of an (n,n-k)
cyclic code, which is the dual of the (n,k) code
generated by g(X).
BrainKart.com
Decoding of Cyclic Codes
• STEPS:
(1) Syndrome computation
(2) Associating the syndrome to the error pattern
(3) Error correction

BrainKart.com
Syndrome Computation
• Received word: r(X) = r0 + r1X +….+ rn-1Xn-1

• If r(X) is a correct codeword, it is divisible by g(X).
Otherwise: r(X) = q(X)g(X) + s(X).
• deg[s(X)] n-k-1.

• s(X) is called the syndrome polynomial.
• s(X) = Rem[r(X)/g(X)] = Rem[ (a(X)g(X) + e(X))/g(x)]
= Rem[e(X)/g(X)]
• The syndrome polynomial depends on the error pattern
only.
• s(X) is obtained by shifting r(X) into a divider-by-g(X)
circuit. The register contents are the syndrome bits.

BrainKart.com
Example: Circuit for Syndrome Computation
Gate
r = 0010110
+
+
Shift Input Register contents
0 0 0 (initial state) • What is g(x)?
1 0 000 • Find the syndrome using
2 1 100 long division.
3 1 110 • Find the syndrome using
4 0 011 the shortcut for the
remainder.
5 1 011
6 0 111
7 0 1 0 1 (syndrome s)
BrainKart.com
Association of Syndrome to Error
Pattern
• Look-up table implemented via a combinational logic circuit
(CLC). The complexity of the CLC tends to grow exponentially
with the code length and the number of errors to correct.
• Cyclic property helps in simplifying the decoding circuit.
• The circuit is designed to correct the error in a certain location
only, say the last location. The received word is shifted cyclically
to trap the error, if it exists, in the last location and then correct
it. The CLC is simplified since it is only required to yield a single
output e telling whether the syndrome, calculated after every
cyclic shift of r(X), corresponds to an error at the highest-order
position.
• The received digits are thus decoded one at a time.

BrainKart.com
Meggit Decoder
Shift r(X) into the buffer B and the syndrome
register R simultaneously. Once r(X) is completely
shifted in B, R will contain s(X), the syndrome of
r(X).
1. Based on the contents of R, the detection circuit
yields the output e (0 or 1).
2. During the next clock cycle:
(a) Add e to the rightmost bit of B while shifting
the contents of B. (The rightmost bit of B may be
read out). Call the modified content of B r1(1)(X).

BrainKart.com
Meggit Decoder (cont’d)

(b) Add e to the left of R while shifting the
contents of R. The modified content of R is
s1(1)(X), the syndrome of r1(1)(X) [will be shown
soon].
Repeat steps 1-2 n times.

BrainKart.com
General Decoding Circuit

.

BrainKart.com
More on Syndrome Computation

• Let s(X) be the syndrome of a received polynomial r(X)
= r0 + r1X +….+ rn-1Xn-1 . Then the remainder resulting
from dividing Xs(X) by g(X) is the syndrome of r(1)(X),
which is a cyclic shift of r(X).
• Proof: r(X) = r0 + r1X +….+ rn-1Xn-1
r(1)(X) = rn-1 + r0X +….+ rn-2Xn-1 = rn-1 + Xr(X) + rn-1Xn
= rn-1(Xn+1) + Xr(X)
c(X)g(X) + y(X) = rn-1 g(X)h(X)+ X{a(X)g(x) + s(X)}
where y(X) is the syndrome of r(1)(X) .
Xs(X) = {c(X) + a(X) + rn-1 h(X)}g(X) + y(X)
Therefore, Syndrome of r(1)(X)= Rem[Xs(X)/g(X)]. Q.E.D.

BrainKart.com
(cont’d)
Note: for simplicity of notation, let Rem[Xs(X)/g(X)]
be denoted by s(1)(X). s(1)(X) is NOT a cyclic shift of
s(X), but the syndrome of r(1)(X) which is a cyclic
shift of r(X).
Example:
r(X) = X2 + X4 + X5; g(X) = 1 + X + X3
s(X) = Rem[r(X)/g(X)] = 1 + X2
r(1)(X) = X3 + X5 + X6
s(1)(X) = Rem[r(1)(X)/g(X)] = 1 (polynomial)
Also, s(1)(X) = Rem[Xs(X)/g(X)] = 1.

More on Syndrome Compwwuwt.raejitnpiaouln.com
(cont’d)
Gate
+
r = 0010110 Gate +
Shift Input Register contents

0 0 0 (initial state)
1 0 000
2 1 100
3 1 110
4 0 011
5 1 011
6 0 111
7 0 1 0 1 (syndrome s)
8 (input gate off) - 1 0 0 (syndrome s(1) )
9 - 0 1 0 (syndrome s(2) )
BrainKart.com
(cont’d)
Let r(X) = r0 + r1X +….+ rn-1Xn-1 has the syndrome s(X).Then
r(1)(X) = rn-1 + r0 X + ….+ rn-2Xn-1 has the syndrome:
s(1)(X) = Rem[r(1)(X)/g(X)].
Define r1 (X) = r(X) + Xn-1 = r0 + r1X +….+ (rn-1+1)Xn-1
The syndrome of r1 (X), call it s1 (X):
(X)= Rem[{r(X)+ Xn-1}/g(X)] = s(X) + Rem[Xn-1/g(X)]
s1 (1)
r (X), which is one cyclic shift of r (X), has the
1 1
syndrome
s1(1)(X) = Rem[X s1 (X)/g(X)] = Rem[Xs(X)/g(X)+ Xn/g(X)]
= s(1)(X) + 1 (since Xn +1 = g(X)h(X))

BrainKart.com
Worked Example
Consider the (7,4) Hamming code generated by 1+X+X3.
Error pattern Syndrome polynomial.

X6 1 + X2 101
X 1 + X + X2 111
X4 X + X2 011
X3 1+X 110
X2 X2 001
X1 X 010
X0 1 100
Let c = 1 0 0 1 0 1 1 and r = 1 0 1 1 0 1 1

BrainKart.com
Cyclic Decoding of the (7,4) Code

.

BrainKart.com

BrainKart.com
Error Correction Capability
• Error correction capability is inferred from the

roots of g(X).
Results from Algebra of Finite Fields:
Xn +1 has n roots (in an extension field)
These roots can be expressed as powers of one
element, .
The roots are 0, 1 , …., n-1.
 roots occur in conjugates.
The
 constitute a conjugate set.
 i 2 j
mod n 

 
BrainKart.com

Designing a Cyclic Code
• Theorem:
If g(X) has l roots (out of it n-k roots) that are consecutive
powers of , then the code it generates has a minimum
distance d = l + 1.
• To design a cyclic code with a guaranteed minimum
distance of d, form g(X) to have d-1 consecutive roots. The
parameter d is called the designed minimum distance of
the code.
• Since roots occur in conjugates, the actual number of
consecutive roots, say l, may be greater than d-1. dmin = l +
1 is called the actual minimum distance of the code.

BrainKart.com
Design Example
X15 + 1 has the roots 1= 0, 1 , …., 14.
Conjugate group Corresponding
polynomial
0) X1 + X
(, 2 , 4 , 8) X 1 + X + X4
(3 , 6 , 9 , 12) X 1 + X + X2 + X3 +
X4
(5 , 10) X 1 + X + X2
(, 14 , 13 , 11) X 1 + X3 + X4
BrainKart.com
Design Example (cont’d)

• Find g(X) that is guaranteed to be a double error
correcting code.
The code must have , 2 , 3 and 4 as roots.
g(X) = X X = 1 + X4 + X6 + X7 + X8
This generator polynomial generates a (15, 7) cyclic code
of minimum distance at least 5.
Roots of g(X) = , 2, 3 , 4 , 6, 8 , 9 , 12.
Number of consecutive roots = 4.
The actual minimum distance of the code is 5.

BrainKart.com
Cyclic Codes
Some Standard Cyclic BCH Codes
Linear
Block Codes Hamming Codes
Codes
• The Hamming Codes: single-error correcting codes

which can be expressed in cyclic form.
• BCH: the Bose-Chaudhuri-Hocquenghem are among
the most important of all cyclic block codes.
Extenstion of Hamming for t-error correcting codes.
• Some Burst-Correcting Codes: good burst-
correcting codes have been found mainly by
computer search.
• Cyclic Redundancy Check Codes: shortened cyclic
error-detecting codes used in automatic repeat
request (ARQ) systems.
BrainKart.com
BCH Codes
• Definition of the codes:
• For any positive integers m (m>2) and t0 (t0 <
n/2), there is a BCH binary code of length n =
2m - 1 which corrects all combinations of t0 or
fewer errors and has no more than mt0 parity-
check bits.
Block length 2m  1
Number of parity - check bits n  k  mt 0
minimum distance d min  2t 0  1

BrainKart.com
Table of Some BCH Codes

n k d (designed) d ( actual) g(X)*
7 4 3 3 13
15 11 3 3 23
15 7 5 5 721
15 5 7 7 2463
31 26 3 3 45
31 16 5 7 107657
31 11 7 11 5423325
* Octal representation with highest order at the left.

721 is 111 010 001 representing 1+X4+X6+X7+X8

BrainKart.com
Burst Correcting Codes

• good burst-correcting codes have been found mainly
by computer search.
• The length of an error burst, b, is the total number of
bits in error from the first error to the last error,
inclusive.
• The minimum possible number of parity-check bits
required to correct a burst of length b or less is given
by the Rieger bound. r  2b
• The best understood codes for correcting burst errors
are cyclic codes.
• For correcting longer burst interleaving is used.

BrainKart.com
Table of Good Burst-Correcting Codes
n k b g(X) (octal)
7 3 2 35 (try to find dmin!)
15 10 2 65
15 9 3 171
31 25 2 161
63 56 2 355
63 55 3 711
511 499 4 10451
1023 1010 4 22365

BrainKart.com
Cyclic Redundancy Check Codes

• Shortened cyclic codes
• Error-detecting codes
• used in automatic repeat request (ARQ)
systems.
• Usually concatenated with error correcting code
Error Correction Error Correction CRC
CRC To
Encoder Decoder Syndrom
To Info Sink
Encoder e Checker
Transmitter

BrainKart.com
Performance of CRC Codes

• CRC are typically evaluated in terms of their
– error pattern coverage
– Burst error detection capability
– Probability of undetected error
• For a (n,k) CRC the coverage, λ, is the ratio of the number of invalid
blocks of length n to the total number of blocks of length n.
• This ratio is a measure of the probability that a randomly chosen block is
not a valid code block. By definition,
  1 2 r
Code Coverage
• where r is the number of check bits
• For some near-optima CRC codes, CRC-12 0.999756
see table 5.6.5 CRC-ANSI 0.999985
CRC-32A 0.99999999977
BrainKart.com
Basic Definitions
• k =1, n = 2 , (2,1) Rate-1/2 convolutional code

• Two-stage register ( M=2 )
• Each input bit influences the output for 3 intervals
(K=3)
• K = constraint length of the code = M + 1

BrainKart.com
Generator Polynomial
• A convolutional code may be defined by a set
of n generating polynomials for each input bit.
• For the circuit under consideration:
g1(D) = 1 + D + D2
g2(D) = 1 + D2
• The set {gi(D)} defines the code completely.
The length of the shift register is equal to the
highest-degree generator polynomial.

BrainKart.com
State Diagram Representation

• The output depends on the current input and
the state of the encoder ( i. e. the contents of
the shift register).

BrainKart.com
Trellis Diagram Representation

• Expansion of state diagram in time.

BrainKart.com
Decoding
• A message m is encoded into the code

sequence c.
• Each code sequence represents a path in
the trellis diagram.
• Minimum Distance Decoding
– Upon receiving the received sequence r, search
for the path that is closest ( in Hamming
distance) to r .
BrainKart.com
The Viterbi Algorithm

• Walk through the trellis and compute the Hamming
distance between that branch of r and those in the
trellis.
• At each level, consider the two paths entering the
same node and are identical from this node onwards.
From these two paths, the one that is closer to r at this
stage will still be so at any time in the future. This path
is retained, and the other path is discarded.
• Proceeding this way, at each stage one path will be
saved for each node. These paths are called the
survivors. The decoded sequence (based on MDD) is
guaranteed to be one of these survivors.

BrainKart.com
The Viterbi Algorithm (cont’d)

• Each survivor is associated with a metric of
the accumulated Hamming distance (the
Hamming distance up to this stage).
• Carry out this process until the received
sequence is considered completely. Choose
the survivor with the smallest metric.

BrainKart.com
The Viterbi Algorithm:
• The viterbi algorithm is used to decode convolutional codes

and any structure or system that can be described by a trellis.
• It is a maximum likelihood decoding algorithm that selects
the most probable path that maximizes the likelihood
function.
• The algorithm is based on add-compare-select the best path
each time at each state.

BrainKart.com
Example: For the convolutional code example in the previous lecture,
starting from state zero, Decode the following received sequence.
At the end of the

trellis, select the
path with the
minimum
cumulative
Hamming weight
This is the
survival
path in
this
example
Decoded
sequence is
m=[10 1110]

Compute the two possible paths at
Add the weight of the each state and select the one This is called the
path at each state with less cumulative Hamming survival path
Geweight
t useful study materials from www.rejinpaul.com
BrainKart.com
Distance Properties of Conv. Codes

• Def: The free distance, dfree, is the minimum Hamming
distance between any two code sequences.
• Criteria for good convolutional codes:
– Large free distance, dfree.
– Small Hamming distance (i.e. as few differences as possible
) between the input information sequences that produce
the minimally separated code sequences. dinf
• There is no known constructive way of designing a
conv. code of given distance properties. However, a
given code can be analyzed to find its distance
properties.

BrainKart.com
Distance Prop. of Conv. Codes (cont’d)

• Convolutional codes are linear. Therefore, the
Hamming distance between any pair of code
sequences corresponds to the Hamming distance
between the all-zero code sequence and some nonzero
code sequence. Thus for a study of the distance
properties it is possible to focus on the Hamming
distance between the all-zero code sequence and all
nonzero code sequences.
• The nonzero sequence of minimum Hamming weight
diverges from the all-zero path at some point and
remerges with the all-zero path at some later point.

BrainKart.com
Distance Properties: Illustration
• sequence 2: Hamming weight = 5, dinf = 1

• sequence 3: Hamming weight = 7, dinf = 3.

BrainKart.com
Modified State Diagram

• The span of interest to us of a nonzero path starts
from the 00 state and ends when the path first
returns to the 00 state. Split the 00 state (state a)
to two states: a0 and a1.
• The branches are labeled with the dummy
variables D, L and N, where:
The power of D is the Hamming weight (# of 1’s) of the
output corresponding to that branch.
The power of N is the Hamming weight (# of 1’s) of the
information bit(s) corresponding to that branch.
The power of L is the length of the branch (always = 1).

BrainKart.com
Modified State Diagram (cont’d)

.

BrainKart.com
Properties of the Path
Sequence 2:
code sequence: .. 00 11 10 11 00 ..
state sequence: a0 b c a1
Labeled: (D2LN)(DL)(D2L) = D5L3N
Prop. : w =5, dinf =1, diverges from the allzero path by 3
branches.
Sequence 3:
code sequence: .. 00 11 01 01 00 10 11 00 ..
state sequence: a0 b d c b c a1
Labeled: (D2LN)(DLN)(DL)(DL)(LN)(D2L) = D7L6N3
Prop. : w =7, dinf =3, diverges from the allzero path by 6
branches.

BrainKart.com
Transfer Function
• Input-Output relations:
a0 = 1
b = D2LN a0 + LNc
c = DLb + DLNd
d = DLNb + DLNd
a1 = D2Lc
• The transfer function T(D,L,N) = a1 /a0
D5 L3
T(D, L, N) 
1 DNL(1 L)

BrainKart.com
Transfer Function (cont’d)

• Performing long division:
T = D5L3N + D6L4N2 + D6L5N2 + D7L5N3 + ….
• If interested in the Hamming distance property of
the code only, set N = 1 and L = 1 to get the
distance transfer function:
T (D) = D5 + 2D6 + 4D7
There is one code sequence of weight 5. Therefore
dfree=5.
There are two code sequences of weight 6,
four code sequences of weight 7, ….

BrainKart.com
Decoding of Convolutional Codes
 Let Cm be the set of allowable code sequences of length m.
 Not all sequences in {0,1}m are allowable code sequences!

 Each code sequence c C can be represented by a unique path
 m
through the trellis diagram
 What is the probability that the code sequence is sent and the
c
binary sequence is received? 

y

 
   
dH y,
 c  mdH y,
c 
c  p
Pr y |  





.(1 p) 





where p is the probability of bit error of BSC from modulation

BrainKart.com
Decoding Rule for Convolutional Codes
 Maximum Likelihood Decoding Rule:
max
cC
  
c  min
Pr y, 
 cC
  
d y, 
H 
c
  m  m
Choose the code sequence through the trellis which has the
smallest Hamming distance to the received sequence!

BrainKart.com

 The Viterbi Algorithm (Viterbi, 1967) is a clever way of
implementing Maximum Likelihood Decoding.
Computer Scientists will recognize the Viterbi Algorithm as an
example of a C“ technique called “ Dynamic Programming”
 ‘eference: G. D. Forney, “ The Viterbi Algorithm”,

Proceedings of the IEEE, 1973
 Chips are available from many manufacturers which
implement the Viterbi Algorithm for K < 10
 Can be used for either hard or soft decision decoding
We consider hard decision decoding initially

BrainKart.com
Basic Idea of Viterbi Algorithm

 There are 2rm code sequences in Cm .
 This number of sequences approaches infinity as m
becomes large
 Instead of searching through all possible sequences,
find the best code sequence "one stage at a time"

BrainKart.com
(Hamming Distance Metric)
 Initialization:
Let time i = 0.
We assign each state j a metric Z j 0 at time 0.
We know that the code must start in the state 0.
Therefore we assign:
Z j 0  
Z j 0   for all other states

BrainKart.com
The Viterbi Algorithm (continued)

Consider decoding of the ith segment:
 Let y i be the segment of n bits received between times i
and i + 1
 There are several code segmecnits of n bits which lead
into state j at time i+1. We wish to find the most likely one.
 Let s c i  be the state from which the code segm
c ient
emerged
 For each state j, we assume thcai t is the path leading
into i d
Z s 
ci 
c ,y
H i 
state j if:
is the smallest of all the code segments leading into state
j.

BrainKart.com
The Viterbi Algorithm (continued)

 Iteration: 
• Let
Z
j
i 1  Z

i
sc
i d

c ,y
H i 
• Let i=i+1 



• Repeat previous step

•Incorrect paths drop out as i approaches infinity.

BrainKart.com
Viterbi Algorithm Decoding Example

 r =1/2, K = 3 code from previous example
 c = (0 0 1 1 01 00 10 10 1 1) is sent
 y
 = (0 1 1 1 01 00 10 10 1 1) is received.
What path through the trellis does the Viterbi Algorithm choose?

BrainKart.com
Viterbi Algorithm Decoding Example
(continued)

BrainKart.com
SUMMARY
• Learnt about the concepts of Entropy and the
source coding techniques
• Statement and the theorems of Shanon
• Concepts about mutual information and
channel capacity
• Understand error control coding techniques
and the concepts about linear block codes,
cyclic codes, convolution codes & viterbi
decoding algorithm

BrainKart.com
Thank you

222 - EC8501, EC6501 Digital Communication - Notes 1

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

222 - EC8501, EC6501 Digital Communication - Notes 1

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

222 - EC8501, EC6501 Digital Communication - Notes 1

Uploaded by

Copyright:

Available Formats

Click Here for Digital Communication full study material.

UNIT I SAMPLING & QUANTIZATION 9

UNIT II WAVEFORM CODING 9

UNIT III BASEBAND TRANSMISSION 9

UNIT IV DIGITAL MODULATION SCHEME 9

UNIT V ERROR CONTROL CODING 9

Get useful study materials from BrainKart.com

Get useful study materials from BrainKart.com

Get useful study materials from BrainKart.com

Digital communication system

Click Here for Digital Communication full study material.

 How can a continuous wave form be

 How can discrete samples be converted back

Click Here for Digital Communication full study material.

Low Pass Sampling

Sampling (in time) is

 Measure amplitude at regular intervals

 How many times should we sample?

Click Here for Digital Communication full study material.

 where fs is sampling frequency and fm is the

Click Here for Digital Communication full study material.

 low-pass filter removes higher-frequencies

 (e.g.) If max sampling frequency is 22kHz, the it

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

• Pulse-Amplitude Modulation (PAM)

Click Here for Digital Communication full study material. 28

Click Here for Digital Communication full study material. 29

Click Here for Digital Communication full study material. 30

Click Here for Digital Communication full study material. 31

I k : {mk  m mk 1}, k  1,2,..., L (17) Fig. 9

– Representation level (or Reconstruction level)

v  g(m) (18) Fig. 10

Click Here for Digital Communication full study material. 33

Click Here for Digital Communication full study material. 34

Two types of quantization are

Click Here for Digital Communication full study material.

Signal Range is Symmetric

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

2 Example: Nonuniform 3 bit quantizer

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

u-LAW and A-LAW definitions

Click Here for Digital Communication full study material. 46

Companded Signal -0.5

Click Here for Digital Communication full study material. 47

Click Here for Digital Communication full study material.

Click Here for Digital Communication full study material.

• Operation in the Transmitter

Click Here for Digital Communication full study material. 52