1

Digital Communication Systems

Lecture-2, Prof. Dr. Habibullah
Jamal
Under Graduate, Spring 2008

2
Formatting
3
Example 1:
In ASCII alphabets, numbers, and symbols are encoded using a 7-
bit code
A total of 2
7
= 128 different characters can be represented using
a 7-bit unique ASCII code (see ASCII Table, Fig. 2.3)
4
Formatting

Transmit and Receive Formatting
Transition from information source digital symbols
information sink
5
Character Coding (Textual Information)
A textual information is a sequence of alphanumeric characters
Alphanumeric and symbolic information are encoded into digital bits
using one of several standard formats, e.g, ASCII, EBCDIC
6
Transmission of Analog Signals

Structure of Digital Communication Transmitter
Analog to Digital Conversion
7
Sampling

Sampling is the processes of converting continuous-time analog
signal, x
a
(t), into a discrete-time signal by taking the samples at
discrete-time intervals
Sampling analog signals makes them discrete in time but still
continuous valued
If done properly (Nyquist theorem is satisfied), sampling does not
introduce distortion
Sampled values:
The value of the function at the sampling points
Sampling interval:
The time that separates sampling points (interval b/w samples), T
s
If the signal is slowly varying, then fewer samples per second will
be required than if the waveform is rapidly varying
So, the optimum sampling rate depends on the maximum
frequency component present in the signal
8
Analog-to-digital conversion is (basically) a 2 step process:
Sampling
Convert from continuous-time analog signal x
a
(t) to discrete-
time continuous value signal x(n)
Is obtained by taking the samples of x
a
(t) at discrete-time
intervals, T
s

Quantization
Convert from discrete-time continuous valued signal to discrete
time discrete valued signal



9
Sampling

Sampling Rate (or sampling frequency f
s
):
The rate at which the signal is sampled, expressed as the
number of samples per second (reciprocal of the sampling
interval), 1/T
s
= f
s

Nyquist Sampling Theorem (or Nyquist Criterion):
If the sampling is performed at a proper rate, no info is lost about
the original signal and it can be properly reconstructed later on
Statement:
If a signal is sampled at a rate at least, but not exactly equal to
twice the max frequency component of the waveform, then the
waveform can be exactly reconstructed from the samples
without any distortion


max
2
s
f f >
10
Ideal Sampling ( or Impulse Sampling)

Therefore, we have:

Take Fourier Transform (frequency convolution)

1
( ) ( ) e
s
jn t
s
n
s
x t x t
T
e

=
| |
=
|
\ .

{ }
1 1
( ) ( )* ( )*
s s
jn t jn t
s
n n
s s
X f X f e X f e
T T
e e

= =

= =
`
)

1
( ) ( ) * ( ),
2
s
s s s
n
s
X f X f f nf f
T
e
o
t

=
= =

1 1
( ) ( ) ( )
s s
n n
s s s
n
X f X f nf X f
T T T

= =
= =

11
Sampling

If R
s
< 2B, aliasing (overlapping of the spectra) results
If signal is not strictly bandlimited, then it must be passed through
Low Pass Filter (LPF) before sampling
Fundamental Rule of Sampling (Nyquist Criterion)
The value of the sampling frequency f
s
must be greater than twice
the highest signal frequency f
max
of the signal
Types of sampling
Ideal Sampling
Natural Sampling
Flat-Top Sampling


12
Ideal Sampling ( or Impulse Sampling)

Is accomplished by the multiplication of the signal x(t) by the uniform
train of impulses (comb function)
Consider the instantaneous sampling of the analog signal x(t)
Train of impulse functions select sample values at regular intervals



Fourier Series representation:
( ) ( ) ( )
s s
n
x t x t t nT o

=
=

1 2
( ) ,
s
jn t
s s
n n
s s
t nT e
T T
e
t
o e

= =
= =

13
Ideal Sampling ( or Impulse Sampling)

This shows that the Fourier Transform of the sampled signal is the
Fourier Transform of the original signal at rate of 1/T
s

14
Ideal Sampling ( or Impulse Sampling)

As long as f
s
> 2f
m
,no overlap of repeated replicas X(f - n/T
s
) will
occur in X
s
(f)
Minimum Sampling Condition:

Sampling Theorem: A finite energy function x(t) can be completely
reconstructed from its sampled value x(nTs) with
provided that =>
2
s m m s m
f f f f f > >
2 ( )
sin
2
( ) ( )
( )
s
s
s s
n
s
f t nT
T
x t T x nT
t nT
t
t

=

(

(


=
`

( ) sin (2 ( ))
s s s s
n
T x nT c f t nT

=
=

1 1
2
s
s m
T
f f
= s
15
Ideal Sampling ( or Impulse Sampling)

This means that the output is simply the replication of the original
signal at discrete intervals, e.g
16
T
s
is called the Nyquist interval: It is the longest time interval that can
be used for sampling a bandlimited signal and still allow
reconstruction of the signal at the receiver without distortion
17
Practical Sampling

In practice we cannot perform ideal sampling
It is practically difficult to create a train of impulses
Thus a non-ideal approach to sampling must be used
We can approximate a train of impulses using a train of very thin
rectangular pulses:
Note:
Fourier Transform of impulse train is another impulse train
Convolution with an impulse train is a shifting operation
( )
s
p
n
t nT
x t
t

| |
= H
|
\ .

18
Natural Sampling

If we multiply x(t) by a train
of rectangular pulses x
p
(t),
we obtain a gated waveform
that approximates the ideal
sampled waveform, known
as natural sampling or
gating (see Figure 2.8)
( ) ( ) ( )
s p
x t x t x t =
2
( )
s
j nf t
n
n
x t c e
t

=
=

( ) [ ( ) ( )]
s p
X f x t x t =
2
[ ( ) ]
s
j nf t
n
n
c x t e
t

=
=

[ ]
n s
n
c X f nf

=
=

19
Each pulse in x
p
(t) has width T
s
and amplitude 1/T
s

The top of each pulse follows the variation of the signal being
sampled
X
s
(f) is the replication of X(f) periodically every f
s
Hz
X
s
(f) is weighted by C
n
Fourier Series Coeffiecient
The problem with a natural sampled waveform is that the tops of the
sample pulses are not flat
It is not compatible with a digital system since the amplitude of each
sample has infinite number of possible values
Another technique known as flat top sampling is used to alleviate
this problem

20
Flat-Top Sampling

Here, the pulse is held to a constant height for the whole
sample period
Flat top sampling is obtained by the convolution of the signal
obtained after ideal sampling with a unity amplitude
rectangular pulse, p(t)
This technique is used to realize Sample-and-Hold (S/H)
operation
In S/H, input signal is continuously sampled and then the
value is held for as long as it takes to for the A/D to acquire
its value
21
Flat top sampling (Time Domain)
'( ) ( ) ( ) x t x t t o =
( ) '( )* ( )
s
x t x t p t =
( )* ( ) ( ) ( )* ( ) ( )
s
n
p t x t t p t x t t nT o o

=
(
= =
(

22
Taking the Fourier Transform will result to

where P(f) is a sinc function
( ) [ ( )]
s s
X f x t =
( ) ( ) ( )
s
n
P f x t t nT o

=
(
=
(

1
( ) ( )* ( )
s
n
s
P f X f f nf
T
o

=
(
=
(

1
( ) ( )
s
n
s
P f X f nf
T

=
=

23
Flat top sampling (Frequency Domain)
Flat top sampling becomes identical to ideal sampling as the
width of the pulses become shorter
24
Recovering the Analog Signal

One way of recovering the original signal from sampled signal X
s
(f)
is to pass it through a Low Pass Filter (LPF) as shown below
If f
s
> 2B then we recover x(t) exactly
Else we run into some problems and signal
is not fully recovered
25
Undersampling and Aliasing
If the waveform is undersampled (i.e. fs < 2B) then there will be
spectral overlap in the sampled signal
The signal at the output of the filter will be
different from the original signal spectrum

This is the outcome of aliasing!
This implies that whenever the sampling condition is not met, an
irreversible overlap of the spectral replicas is produced
26
This could be due to:
1. x(t) containing higher frequency than were
expected
2. An error in calculating the sampling rate
Under normal conditions, undersampling of signals causing
aliasing is not recommended
27
Solution 1: Anti-Aliasing Analog Filter

All physically realizable signals are not completely bandlimited
If there is a significant amount of energy in frequencies above
half the sampling frequency (f
s
/2), aliasing will occur
Aliasing can be prevented by first passing the analog signal
through an anti-aliasing filter (also called a prefilter) before
sampling is performed
The anti-aliasing filter is simply a LPF with cutoff frequency
equal to half the sample rate

28
Aliasing is prevented by forcing the bandwidth of the sampled
signal to satisfy the requirement of the Sampling Theorem
29
Solution 2: Over Sampling and Filtering in the Digital
Domain
The signal is passed through a low performance (less costly)
analog low-pass filter to limit the bandwidth.
Sample the resulting signal at a high sampling frequency.
The digital samples are then processed by a high
performance digital filter and down sample the resulting
signal.


30
Summary Of Sampling
Ideal Sampling
(or Impulse Sampling)


Natural Sampling
(or Gating)


Flat-Top Sampling


For all sampling techniques
If fs > 2B then we can recover x(t) exactly
If fs < 2B) spectral overlapping known as aliasing will occur
( ) ( ) ( ) ( ) ( )
( ) ( )
s s
n
s s
n
x t x t x t x t t nT
x nT t nT
o
o
o

=
= =
=

2
( ) ( ) ( ) ( )
s
j nf t
s p n
n
x t x t x t x t c e
t

=
= =

( ) '( )* ( ) ( ) ( ) * ( )
s s
n
x t x t p t x t t nT p t o

=
(
= =
(

31
Example 1:

Consider the analog signal x(t) given by


What is the Nyquist rate for this signal?

Example 2:

Consider the analog signal x
a
(t) given by


What is the Nyquist rate for this signal?
What is the discrete time signal obtained after sampling, if
f
s
=5000 samples/s.
What is the analog signal x(t) that can be reconstructed from the
sampled values?

( ) 3cos(50 ) 100sin(300 ) cos(100 ) x t t t t t t t = +
( ) 3cos 2000 5sin6000 cos12000
a
x t t t t t t t = + +
32
Practical Sampling Rates
Speech
- Telephone quality speech has a bandwidth of 4 kHz
(actually 300 to 3300Hz)
- Most digital telephone systems are sampled at 8000
samples/sec
Audio:
- The highest frequency the human ear can hear is
approximately 15kHz
- CD quality audio are sampled at rate of 44,000
samples/sec
Video
- The human eye requires samples at a rate of at
least 20 frames/sec to achieve smooth motion
33
Pulse Code Modulation (PCM)

Pulse Code Modulation refers to a digital baseband signal that is
generated directly from the quantizer output
Sometimes the term PCM is used interchangeably with quantization
34
See Figure 2.16 (Page 80)
35
36
Advantages of PCM:
Relatively inexpensive
Easily multiplexed: PCM waveforms from different
sources can be transmitted over a common digital
channel (TDM)
Easily regenerated: useful for long-distance
communication, e.g. telephone
Better noise performance than analog system
Signals may be stored and time-scaled efficiently (e.g.,
satellite communication)
Efficient codes are readily available
Disadvantage:
Requires wider bandwidth than analog signals
37
2.5 Sources of Corruption in the sampled,
quantized and transmitted pulses
Sampling and Quantization Effects
Quantization (Granularity) Noise: Results when
quantization levels are not finely spaced apart enough
to accurately approximate input signal resulting in
truncation or rounding error.
Quantizer Saturation or Overload Noise: Results when
input signal is larger in magnitude than highest
quantization level resulting in clipping of the signal.
Timing Jitter: Error caused by a shift in the sampler
position. Can be isolated with stable clock reference.
Channel Effects
Channel Noise
Intersymbol Interference (ISI)
38
The level of quantization noise is dependent on how close any
particular sample is to one of the L levels in the converter
For a speech input, this quantization error resembles a noise-
like disturbance at the output of a DAC converter

Signal to Quantization Noise Ratio
39
Uniform Quantization
A quantizer with equal quantization level is a Uniform Quantizer
Each sample is approximated within a quantile interval
Uniform quantizers are optimal when the input distribution is
uniform
i.e. when all values within the range are equally
likely





Most ADCs are implemented using uniform quantizers
Error of a uniform quantizer is bounded by

2 2
q q
e < s
40
The mean-squared value (noise variance) of the quantization error is
given by:
/ 2 / 2 / 2
2 2 2
/ 2 / 2 / 2
1
1
( )
2
q q q
q q q
e p e de e de e de
q
q
o

| |
= =
} } }
|
\ .
=
3
/ 2
/ 2
2
1
3 12
q
q
q
e
q

= =
Signal to Quantization Noise Ratio
41
The peak power of the analog signal (normalized to 1Ohms )can be
expressed as:




Therefore the Signal to Quatization Noise Ratio is given by:


2
2
2 2
2 4
1
pp
p
V
V
L q
P
| |
| |
|
|
|
|
\ .
\ .
= = =
2 2
2
/ 4
/12
2
3
q
L q
q
SNR L = =
42





where L = 2
n
is the number of quantization levels for the converter.
(n is the number of bits).

Since L = 2
n
, SNR = 2
2n
or in decibels


pp
V
L
q =
2
10log (2 ) 6
10
n
S
n dB
N
dB
| |
|
|
\ .
= =
If q is the step size, then the maximum quantization error that can
occur in the sampled output of an A/D converter is q
43
Nonuniform Quantization

Nonuniform quantizers have unequally spaced levels
The spacing can be chosen to optimize the Signal-to-Noise Ratio
for a particular type of signal
It is characterized by:
Variable step size
Quantizer size depend on signal size
44
Many signals such as speech have a nonuniform distribution
See Figure on next page (Fig. 2.17)
Basic principle is to use more levels at regions with large probability
density function (pdf)
Concentrate quantization levels in areas of largest pdf
Or use fine quantization (small step size) for weak signals and
coarse quantization (large step size) for strong signals
45
Statistics of speech Signal Amplitudes
Figure 2.17: Statistical distribution of single talker speech signal
magnitudes (Page 81)
46
Nonuniform quantization using companding
Companding is a method of reducing the number of bits required in
ADC while achieving an equivalent dynamic range or SQNR
In order to improve the resolution of weak signals within a converter,
and hence enhance the SQNR, the weak signals need to be
enlarged, or the quantization step size decreased, but only for the
weak signals
But strong signals can potentially be reduced without significantly
degrading the SQNR or alternatively increasing quantization step size
The compression process at the transmitter must be matched with an
equivalent expansion process at the receiver
47

The signal below shows the effect of compression, where the
amplitude of one of the signals is compressed
After compression, input to the quantizer will have a more uniform
distribution after sampling
At the receiver, the signal is
expanded by an inverse
operation
The process of COMpressing
and exPANDING the signal is
called companding
Companding is a technique
used to reduce the number of bits
required in ADC or DAC while
achieving comparable SQNR
48


Basically, companding introduces a nonlinearity into the signal
This maps a nonuniform distribution into something that more
closely resembles a uniform distribution
A standard ADC with uniform spacing between levels can be used
after the compandor (or compander)
The companding operation is inverted at the receiver

There are in fact two standard logarithm based companding
techniques
US standard called -law companding
European standard called A-law companding

49
Input/Output Relationship of Compander

Logarithmic expression Y = log X is the most commonly
used compander
This reduces the dynamic range of Y
50
Types of Companding
-Law Companding Standard (North & South
America, and Japan)

where
x and y represent the input and output voltages
is a constant number determined by experiment
In the U.S., telephone lines uses companding with = 255
Samples 4 kHz speech waveform at 8,000 sample/sec
Encodes each sample with 8 bits, L = 256 quantizer levels
Hence data rate R = 64 kbit/sec
= 0 corresponds to uniform quantization
| |
max
max
log 1 (| | /
sgn( )
log (1 )
e
e
x x
y y x

+
=
+
51
A-Law Companding Standard (Europe, China, Russia, Asia,
Africa)

where
x and y represent the input and output voltages
A = 87.6
A is a constant number determined by experiment

max
max
max
max
max
max
| |
| | 1
sgn( ), 0
(1 )
( )
| |
1 log
1 | |
sgn( ), 1
(1 log )
e
e
x
A
x x
y x
A x A
y x
x
A
x
x
y x
A A x

< s
+

=

(
| |

+
( |

\ .

< s
+

52
Pulse Modulation
Recall that analog signals can be represented by a sequence of discrete
samples (output of sampler)
Pulse Modulation results when some characteristic of the pulse (amplitude,
width or position) is varied in correspondence with the data signal

Two Types:
Pulse Amplitude Modulation (PAM)
The amplitude of the periodic pulse train is varied in proportion to the
sample values of the analog signal
Pulse Time Modulation
Encodes the sample values into the time axis of the digital signal
Pulse Width Modulation (PWM)
Constant amplitude, width varied in proportion to the signal
Pulse Duration Modulation (PDM)
sample values of the analog waveform are used in determining the
width of the pulse signal
53
54
PCM Waveform Types
The output of the A/D converter is a set of binary bits
But binary bits are just abstract entities that have no physical definition
We use pulses to convey a bit of information, e.g.,





In order to transmit the bits over a physical channel they must be
transformed into a physical waveform
A line coder or baseband binary transmitter transforms a stream of bits
into a physical waveform suitable for transmission over a channel
Line coders use the terminology mark for 1 and space to mean 0
In baseband systems, binary data can be transmitted using many kinds of
pulses


55

There are many types of waveforms. Why? performance criteria!
Each line code type have merits and demerits
The choice of waveform depends on operating characteristics of a
system such as:
Modulation-demodulation requirements
Bandwidth requirement
Synchronization requirement
Receiver complexity, etc.,

56
Goals of Line Coding (qualities to look for)
A line code is designed to meet one or more of the following goals:
Self-synchronization
The ability to recover timing from the signal itself
That is, self-clocking (self-synchronization) - ease of clock lock
or signal recovery for symbol synchronization
Long series of ones and zeros could cause a problem
Low probability of bit error
Receiver needs to be able to distinguish the waveform associated
with a mark from the waveform associated with a space
BER performance
relative immunity to noise
Error detection capability
enhances low probability of error

57
Spectrum Suitable for the channel
Spectrum matching of the channel
e.g. presence or absence of DC level
In some cases DC components should be avoided
The transmission bandwidth should be minimized
Power Spectral Density
Particularly its value at zero
PSD of code should be negligible at the frequency near zero
Transmission Bandwidth
Should be as small as possible
Transparency
The property that any arbitrary symbol or bit pattern can be
transmitted and received, i.e., all possible data sequence should
be faithfully reproducible

58
Line Coder

The input to the line encoder is
the output of the A/D converter
or a sequence of values a
n
that
is a function of the data bit
The output of the line encoder
is a waveform:

where f(t) is the pulse shape and T
b
is the bit period (T
b
=T
s
/n for n
bit quantizer)
This means that each line code is described by a symbol mapping
function a
n
and pulse shape f(t)
Details of this operation are set by the type of line code that is
being used
( ) ( )
n b
n
s t a f t nT

=
=

59
Summary of Major Line Codes

Categories of Line Codes
Polar - Send pulse or negative of pulse
Unipolar - Send pulse or a 0
Bipolar (a.k.a. alternate mark inversion, pseudoternary)
Represent 1 by alternating signed pulses
Generalized Pulse Shapes
NRZ -Pulse lasts entire bit period
Polar NRZ
Bipolar NRZ
RZ - Return to Zero - pulse lasts just half of bit period
Polar RZ
Bipolar RZ
Manchester Line Code
Send a 2- | pulse for either 1 (high low) or 0 (low high)
Includes rising and falling edge in each pulse
No DC component
60
When the category and the generalized shapes are combined, we have
the following:
Polar NRZ:
Wireless, radio, and satellite applications primarily use Polar
NRZ because bandwidth is precious
Unipolar NRZ
Turn the pulse ON for a 1, leave the pulse OFF for a 0
Useful for noncoherent communication where receiver cant
decide the sign of a pulse
fiber optic communication often use this signaling format
Unipolar RZ
RZ signaling has both a rising and falling edge of the pulse
This can be useful for timing and synchronization purposes

61
Bipolar RZ
A unipolar line code, except now we alternate
between positive and negative pulses to send a 1
Alternating like this eliminates the DC component
This is desirable for many channels that cannot
transmit the DC components
Generalized Grouping
Non-Return-to-Zero: NRZ-L, NRZ-M NRZ-S
Return-to-Zero: Unipolar, Bipolar, AMI
Phase-Coded: bi-f-L, bi-f-M, bi-f-S, Miller, Delay
Modulation
Multilevel Binary: dicode, doubinary

Note:There are many other variations of line codes (see Fig. 2.22,
page 80 for more)
62
Commonly Used Line Codes

Polar line codes use the antipodal mapping


Polar NRZ uses NRZ pulse shape
Polar RZ uses RZ pulse shape
, 1
, 0
n
n
n
A when X
a
A when X
+ =

=

=

63
Unipolar NRZ Line Code
Unipolar non-return-to-zero (NRZ) line code is defined by
unipolar mapping


In addition, the pulse shape for unipolar NRZ is:
where T
b
is the bit period


, 1
0, 0
n
n
n
A when X
a
when X
+ =

=

=

Where X
n
is the n
th
data bit
( ) , NRZ Pulse Shape
b
t
f t
T
| |
= H
|
\ .
64
Bipolar Line Codes
With bipolar line codes a space is mapped to zero and
a mark is alternately mapped to -A and +A

It is also called pseudoternary signaling or alternate mark inversion
(AMI)
Either RZ or NRZ pulse shape can be used
, when 1 andlast mark
, when 1 andlast mark
0, when 0
n
n n
n
A X A
a A X A
X
+ =

= = +

65
Manchester Line Codes
Manchester line codes use the antipodal mapping
and the following split-phase pulse shape:
4 4
( )
2 2
b b
b b
T T
t t
f t
T T
| | | |
+
| |
= H H
| |
| |
\ . \ .
66
Summary of Line Codes

67
68
Comparison of Line Codes

Self-synchronization
Manchester codes have built in timing information because they
always have a zero crossing in the center of the pulse
Polar RZ codes tend to be good because the signal level always
goes to zero for the second half of the pulse
NRZ signals are not good for self-synchronization
Error probability
Polar codes perform better (are more energy efficient) than
Unipolar or Bipolar codes
Channel characteristics
We need to find the power spectral density (PSD) of the line
codes to compare the line codes in terms of the channel
characteristics

69
Comparisons of Line Codes

Different pulse shapes are used
to control the spectrum of the transmitted signal (no DC value,
bandwidth, etc.)
guarantee transitions every symbol interval to assist in symbol timing
recovery
1. Power Spectral Density of Line Codes (see Fig. 2.23, Page 90)
After line coding, the pulses may be filtered or shaped to further
improve there properties such as
Spectral efficiency
Immunity to Intersymbol Interference
Distinction between Line Coding and Pulse Shaping is not easy
2. DC Component and Bandwidth
DC Components
Unipolar NRZ, polar NRZ, and unipolar RZ all have DC components
Bipolar RZ and Manchester NRZ do not have DC components
70
First Null Bandwidth
Unipolar NRZ, polar NRZ, and bipolar all have 1st null bandwidths
of Rb = 1/Tb
Unipolar RZ has 1st null BW of 2Rb
Manchester NRZ also has 1st null BW of 2Rb, although the
spectrum becomes very low at 1.6Rb
71
Generation of Line Codes

The FIR filter realizes the different pulse shapes
Baseband modulation with arbitrary pulse shapes can be
detected by
correlation detector
matched filter detector (this is the most common detector)

72
Section 2.8.4: Bits per PCM Word and Bits per Symbol
L=2
l


Section 2.8.5: M-ary Pulse Modulation Waveforms
M = 2
k

Problem 2.14: The information in an analog waveform, whose
maximum frequency f
m
=4000Hz, is to be transmitted using a 16-level
PAM system. The quantization must not exceed 1% of the peak-to-
peak analog signal.
(a) What is the minimum number of bits per sample or bits per PCM
word that should be used in this system?
(b) What is the minimum required sampling rate, and what is the
resulting bit rate?
(c) What is the 16-ary PAM symbol Transmission rate?
Bits per PCM word and M-ary Modulation
73
max
2 2
2
2
| | | |
2
1
2
2
1
log log (50) 6
2
8000 48000 16
48000
12000 / sec
log ( ) 4
pp
pp
l
pp
q
e pV e
V
V Lq q L
L p
l l
p
fs Rs M
R
R symbols
M
s =
= = = >
| |
> > =
|
\ .
= = =
= = =
Solution to Problem 2.14