1
Contents
Line coding
Reasons For Using Line Coding
Properties of a Line Code
Line coding schemes
DIGITAL DATA TRANSMISSION
3
Two types of Digital Data Transmission:
1) Base-Band data transmission:
Uses low frequency carrier signal to transmit the data
2) Band-Pass data transmission:
Uses high frequency carrier signal to transmit the data
BASE-BAND DATA TRANSMISSION 13
• Base-Band data transmission = Line coding
• The binary data an electrical signal to transmit
them in the channel
some of the common baseband signaling techniques
used
• Non-Return to Zero (NRZ)
• Unipolar Return to Zero (Unipolar-RZ)
• Bi-Polar Return to Zero (Bi-polar RZ)
• Return to Zero Alternate Mark Inversion (RZ-AMI)
• Non-Return to Zero – Mark (NRZ-Mark)
• splitphase Manchester coding (Biphase)
Line Coding
On the channel, we might want to send binary numbers directly.
The resulting bit patterns on the channel might create a static voltage, which is not
desired.
Use line code to eliminate the average static voltage.
- Save power
- Save bandwidth (possibly)
Line coding: Technique to convert binary data to digital signal.
In the transmission of digital information over wire or optical fiber systems, line
coding (also known as modulation/data translation coding) is the method by which 1’s
& 0’s are represented as transmitted waveforms.
Reasons For Using Line Coding
Spectrum shaping and relocation without modulation or filtering. This is
important in telephone line applications, e.g, where the transfer characteristic
has heavy attenuation below 300 Hz.
Error detection capabilities.
Bandwidth usage; the possibility of transmitting at a higher rate
than other schemes over the same bandwidth.
Reasons For Using Line Coding
DC component can be eliminated; this allows AC
(capacitor or transformer) coupling between stages (as
in telephone lines )
Properties of a Line Code:
Self–Synchronization : There is enough timing information built
into the code so that bit synchronizers can extract the
timing or clock signal.
A long series of binary 1’s or 0’s should not cause a problem
in time recovery
Properties of a Line Code:
Low Probability of Bit Error : Receivers can be designed that will recover the binary
data with a low probability of bit error when the input data is corrupted by noise or
ISI.
A Spectrum that is Suitable for the Channel:
In addition, the signal bandwidth needs to be
sufficiently small compared to the channel bandwidth, so
that ISI will not be a problem.
Properties of a Line Code:
Transmission Bandwidth: This should be as small as possible
Error Detection Capability: It should be possible to implement this feature easily by the
addition of channel encoders and decoders, or the feature should be incorporated
into the line code
Line coding schemes
Unipolar encoding
Unipolar encoding uses only one voltage level ( use only one polarity +ve or
-ve).
Unipolar signaling: 1 = +A volt, 0 = 0 volt
Two problems:
A dc component
Lack of synchronization : If data contain long sequence of 0’s or 1’s, there is no
transition in the signal during this duration that can alert the receiver to
synchronization problem.
The receiver receives a continuous voltage and determines how many bits are
sent by relaying on its clock (bit-duration), which may not be synchronized with
the sender clock.
Polar Encoding
Polar encoding uses two voltage levels (positive and negative).
By using two levels , the average voltage level is reduced and
the dc component problem may be alleviated( if balance)
Polar signaling: 1 = +A volt, 0 = -A volt
Non-Return to Zero (NRZ) 29
• The “1” is represented by some level
• The “0” is represented by the opposite for complete bit period
• Non-return to zero means the signal switched from one level to another
without taking the zero value at any time during transmission.
1 is transmitted by +A for period=Tb and 0 is transmitted by –A or 0 volt
for entire bit period Tb
NRZ: NRZ-L
The level of the signal is dependent upon the state of the bit. A positive voltage
means 0, while negative means 1.
Has lack of synchronization , when the data contain a long stream of 0s or 1s.
NRZ: NRZ-I
In NRZ-I , the signal is inverted if a 1 is encountered. It is the
transition between a +ve and a –ve voltage , not the voltage
itself.
A 0 bit is represented by no change
NRZ-I is superior to NRZ-L due to synchronization each time a 1 bit
is encountered.
RZ encoding
To ensure synchronization, there must be a signal change
(transition) for each bit
RZ uses three values +ve, zero and –ve
The signal changes during each bit
A1-bit is represented by transition from +ve to zero
A 0-bit by –ve to zero
UNIPOLAR RETURN TO ZERO (UNIPOLAR RZ) 33
• Binary “1”: some level that is half the width of the signal
• Binary “0” : absence of the pulse
BIPOLAR RETURN TO ZERO (BIPOLAR RZ) 34
• Binary “1” : some level that is half the width of the signal
• Binary “0”: a pulse that is half width the signal but with the
opposite sign
Bipolar signaling: 1 = +A or –A, 0 = 0 volt
(Also called the alternate mark inversion – AMI)
Bi phase: Manchester encoding
The transition at the middle of the bit is used for both synchronization and bit
representation.
It has not Dc component
It achieves the same level of synchronization as RZ but with only two levels and
less B-W
Binary “1”: a positive pulse half width the signal followed by a negative pulse
• Binary “0”: a negative pulse half width the signal followed by a positive pulse
Bipolar AMI (Alternate Mark Inversion)
Encoding
Bit 0 represents by zero voltage
Bit 1s are represented by alternating +ve and –ve voltages
In bipolar encoding AMI , we use three levels: positive, zero, and negative as RZ.
Has DC component
AMI has a lack of synchronization when there is a stream of sequential zeros.
Return to Zero Alternate Mark Inversion (RZ-AMI) 37
• Binary “1” : a pulse alternating in sign
• Binary “0”: the absence of the pulse
Non-Return to Zero – Mark (NRZ-Mark) 38
• Also known as differential encoding
• Binary “1” : The change of the level
– High to low
– Low to high
• Binary “0” : No change in the level
39
Bit Rate
• The number of bits transferred between Example – Bit rate calculation40
• Suppose that we have a binary data source
devices per second
• If each bit is represented by a pulse of width that generates bits. Each bit is represented
Tb, then the bit rate by a pulse of width Tb = 0.1 mSec
• Calculate the bit rate for the source
• Solution:
Rb = 1/ Tb bits/sec Rb = 1/Tb =1/ 0.1* 10 – =10000 bits/sec
3
Suppose we have an image frame of size 200x200 pixels. Each pixel is represented by three
primary colors red, green and blue (RGB). Each one of these colors is represented by 8 bits, if
we transmit 1000 frames in 5 seconds what is the bit rate for this image?
We have a total size of 200x200 = 40000 pixels
• Each pixel has three colors, RGB that each of them has 8 bits.
– 3 x 8 = 24 bits ( for each pixel with RGB)
• Therefore, for the whole image we have a total size of 24 x 40000 = 960000 bits
• Since we have 1000 frames in 5 seconds, then the total number of bits transmitted
will be 1000 x 960000 = 960000000 bits in 5 seconds
• Bit rate = 96000000/5 = 192000000 bits/second
Baud rate (Symbol rate) • Symbols transmission rather than bits transmission:
• The number of symbols transmitted per second
Reduce the spectrum of the transmitted signal.
through the communication channel. • Symbol transmission rather than bit transmission
• The symbol rate is related to the bit rate by the when the transmission bandwidth is limited
following equation: •
• Rb = bit rate e.g. we need to transmit a data at high rate and the
• Rs = symbol rate bit duration Tb is very small; to overcome this problem
• N = Number of bits per symbol we take a group of more than one bit, say 2, therefore :
Rs = Rb / N Tb f o = 1/Tb
2Tb f = 1/ 2Tb = 1/ 2 f o
4Tb f =1/ 4Tb = 1/4f o
A binary data source transmits binary data, the bit duration is 1μsec, Suppose we want to
transmit symbols rather than bits, if each symbol is represented by four bits. what is the
symbol rate?
• Each bit is represented by a pulse of duration 1μ second, hence the bit rate Rb = 1/ 1
*10 -6 = 1000000 bits/sec
Therefore, the symbol rate will be
Rs= Rb/ N= 1000000/ 4= 250000 symbols/sec
SIGNAL REPRESENTATION AND
BASEBAND PULSE TRANSMISSION
Introduction
Digital data – broad spectrum; low-frequency components;
Transmission channel bandwidth – should accommodate the essential
frequency content of the data stream
Channel is dispersive
channel is noisy – control over additive white noise (old problem..)
received signal pulses are affected by adjacent symbols (new problem) – intersymbol
interference (ISI); major source of interference;
Distorted pulse shape (new problem) - channel requires control over pulse shape
Main issue to be discussed:
Detection of digital pulses corrupted by the effect of the channel
We know the shape of the transmitted pulse
Device to be used – matched filter
Matched Filter
Basic task – detecting transmitted pulses at the front end of the receiver (corrupted by
noise)
Receiver model
• The function of the receiver is
to detect the pulse g(t) in an
optimum manner, providing
that the shape of the pulse is
known and the distortion is due
to effects of noise
• To optimize the design of a filter
x(t ) g t w(t ), 0t T
The filter input x(t) is: so as to minimize the effects of
noise at the filter output in
(4.1)
where some statistical sense.
T is an arbitrary observation interval
g(t) is a binary symbol 1 or 0
w(t) is a sample function of white noise, zero mean, psd N0/2
46/
46
Designing the filter
y (t ) g o (t ) n(t )
Since we assume the filter is (4.2)
linear its output can be
described as: instantaneous
power in the
where output signal
g0(t) is the recovered signal
| g o (T ) |2
n(t) produced noise 2
(4.3)
This is equal to maximizing
E[ n (t )]
the peak signal-to-noise
average
ratio: output noise
SO we have to define the impulse response of the filter h(t) in
such a way that the signal-to-noise ratio (4.3) is maximized.
Let us assume that:
- G(f) - FT of the signal g(t);
- H(f) – frequency response of the filter
then: FT of the output signal g0(t)= H(f).G(f), or
g o (t ) H ( f )G ( f ) exp( j 2 ft )df (4.4)
sampled at time t=T and no noise 2
| g o (T ) | H ( f )G ( f ) exp( j 2 ft ) df
2
(4.5)
Next step is to add the noise.
What we know is that the power spectral density of white noise is:
N0
SN ( f ) | H ( f ) |2 (4.6)
2
Thus the average power of the output noise n(t) is:
E [ n (t )]
2
S N ( f )df
N 0
2
| H ( f ) |2 d f
( 4 .7 )
Substituting 4.5 &4.7 into 4.3,
2
H ( f )G ( f ) exp( j 2 ft ) df
(4.8)
N0
2
| H ( f ) | df
2
So, given the function G(f), the problem is reduced to finding such an H(f) that would maximize η.
We use Schwartz inequality which states that for two complex functions, satisfying the
conditions:
| 2 ( x) |2 dx
| ( x) | dx
2
1
The following is true
2
1 ( x)2 ( x)dx | 1 ( x) | dx | 2 ( x) |2 dx
2
(4.9)
and equality holds if:
1 ( x) k2* ( x) (4.10)
In our case this inequality will have the form:
2
H ( f )G ( f ) exp( j 2 fT )df | H ( f ) |2 df
| G ( f ) |2 df
(4.11)
and we can re-write the equation for the peak signal-to-noise ratio as:
2
N0
| G ( f ) |2 df (4.12)
Remarks:
The right hand side of this equation does not depend on H(f).
It depends only on:
signal energy
noise power spectral density
Max value is for:
2
N0
| G ( f ) |2 d f ( 4 .1 3)
Let us denote the optimum value of H(f) by
Hopt(f). To find it we use equation 4.10:
H opt ( f ) kG* ( f ) exp( j 2 fT ) (4.14)
The result: Except for a scaling coefficient k exp(-
2πfT), the frequency response of the optimum
filter is the same as the complex conjugate of the
FT of the input signal.
Definition of the filter functions:
In the frequency domain:
knowing the input signal we can define the
frequency response of the filter (in the
frequency domain) as the FT of its complex
conjugate.
In the time domain…
Take inverse FT of Hopt(f):
hopt (t ) k G* ( f ) exp[ j 2 f (T t )]df (4.15)
and keeping in mind that for real signals G*(f) = G(-f):
hopt (t ) k
G ( f ) e x p [ j 2 f ( T t )] d f
k G ( f ) e x p [ j 2 f ( T t )] d f
k g (T t )
( 4 .1 6 )
So,
in the time domain it turns out that the impulse response of the
filter, except for a scaling factor k, is a time-reversed and
delayed function of the input signal
This means it is “matched” to the input signal, that is why
this type of time-invariant linear filters is known as
“matched filter”
NOTE: The only assumption for the channel noise was
that it is stationary, white, with psd N0/2.
Properties of Matched Filters
Property 1:
A filter matched to a pulse signal g(t) of duration T is characterized
by an impulse response that is time-reversed and delayed version of
the input g(t):
Time domain: hopt(t) = k . g(T-t)
Frequency Domain: Hopt(f) = kG*(f)exp(-j2πfT)
Property 2:
A matched filter is uniquely defined by the waveform of the pulse
but for the:
- time delay T
- scaling factor k
Property 3:
Go ( f ) H opt ( f )G ( f )
kG* ( f )G ( f ) exp( j 2 fT )
The peak signal-to-noise
k | G ( f ) |2 exp( j 2 fT )
ratio of the matched filter
depends only on the ratio
(4.17)
of the signal energy to the
power spectral density of using the inverse FT
the white noise at the filter
input. g o (T ) G ( f ) exp( j 2 fT )df
0
k
2
| G ( f ) | df
The integral of the
squared magnitude
spectrum of a pulse signal
with respect to frequency
E
g 2 (t )dt G ( f ) |2 df
is equal to the signal
energy E (Rayleigh so g0 (T ) k E (4.18)
Theorem) so substituting
in the previous formula
we get:
H opt ( f ) kG* ( f ) exp( j 2 fT )
(4.14)
After substitution of (4.14)
into (4.7) we get the E[n (t )] S N ( f )df
2
expression for the
average output noise
power as: N0
2
| H ( f ) |2 df
(4.7)
we finally get the following expression:
2
k N0
E[n (t )]
2 2
| G ( f ) | df
2
k N0 E / 2
2
(4.19)
and
(kE ) 2 2E
max 2 (4.20)
(k N 0 E / 2) N 0
