Viet Nam National University Ho Chi Minh City

University of Science
Faculty of Electronics & Telecommunications

Chapter 4:
Noise in Communication Systems

Dang Le Khoa

Email: dlkhoa@hcmus.edu.vn
 Outline
⚫ Properties of Signals
– Periodic Waveforms
– DC Value, Power
– Energy and Power Waveforms
– Signal-to-noise Ratio
– dB, dBm
⚫ Properties of Noise
– Random Variables
– Cumulative density function (CDF)
– Probability density function (PDF)
– Stationarity and Ergodicity
– Binomial distribution
– Gaussian Distribution
Faculty of Electronics & Telecommunications
 Properties of Signals & Noise
➢ In communication systems, the received waveform is
usually categorized into two parts:

Signal: Noise:
The desired part containing the The undesired part
information.

➢Properties of waveforms include:

• DC value, • Phase spectrum,
• Root-mean-square (rms) value, • Power spectral density,
• Normalized power, • Bandwidth
• Magnitude spectrum, • ………………..
3
 Where does noise come from?
⚫ External sources: e.g., atmospheric, galactic noise, interference;
⚫ Internal sources: shot noise and thermal noise (generated by
communication devices themselves).
– Thermal noise caused by the rapid and random motion of electrons
within a conductor due to thermal agitation.
It has a Gaussian distribution with zero mean
– Shot noise: the electrons are discrete and are not moving in a
continuous steady flow, so the current is randomly fluctuating.
It has a Gaussian distribution with zero mean.
– The Gaussian distribution follows from the central limit theorem.

Faculty of Electronics & Telecommunications

 Periodic Waveforms

➢ Definition
A waveform w(t) is periodic with period T0 . A sinusoidal waveform
of frequency f0 = 1/T0 Hertz is periodic

➢ Theorem: If the waveform involved is periodic, the time average

operator can be reduced to

where T0 is the period of the waveform and a is an arbitrary real constant, which
may be taken to be zero.

5
 DC Value

➢ Definition: The DC (direct “current”) value of a waveform

w(t) is given by its time average, w(t). Thus,

6
 Power

➢ Definition.
Let v(t) denote the voltage across a set of circuit terminals,
and let i(t) denote the current into the terminal, as shown .
The instantaneous power (incremental work divided by
incremental time) associated with the circuit is given by:
p(t) = v(t)i(t)
the instantaneous power flows into the circuit when p(t) is
positive and flows out of the circuit when p(t) is negative.
➢ The average power is:
7
 Evaluation of DC Value
➢ A 120V , 60 Hz fluorescent lamp wired in a high power factor configuration.
Assume the voltage and current are both sinusoids and in phase ( unity power
factor)

DC Value of this waveform is:

Vdc = v(t ) = V cos  0t Voltage
1 T0 / 2
=
T0 
−T0 / 2
V cos  0t  dt = 0

Where, Current
 0 = 2 / T0 , and
f 0 = 1 / T0 = 60Hz
Similarly,
I dc = 0 8
Instantenous Power
 Evaluation of Power

The instantaneous power is:

p(t ) = (V cos  0t )  (I cos  0t )
= 1 / 2 VI (1 + 2 cos  0t )
The Average power is:
P = 1/ 2 VI (1 + 2 cos 0t )
T0
VI
= −T 0
2
(1 + 2 cos 0t ) dt
2T0 2

VI
=
2
The Maximum power is: Pmax=VI
9
 RMS Value
➢ Definition: The root-mean-square (rms) value of w(t) is:

V
Wrms = V cos(ot ) =
2
➢ Rms value of a sinusoidal: 2

➢ Theorem:
If a load is resistive (i.e., with unity power factor), the average
power is:

where R is the value of the resistive load.

10
 Normalized Power
➢ In the concept of Normalized Power, R is assumed to be 1Ω,
although it may be another value in the actual circuit.
➢ Another way of expressing this concept is to say that the
power is given on a per-ohm basis.
➢ It can also be realized that the square root of the normalized
power is the rms value.
Definition. The average normalized power is given by:
Where w(t) is the voltage or current waveform


1
P = lim  g (t )dt
2
T → T
−

11
 Energy and Power Waveforms

➢ Definition: w(t) is a power waveform if and only if the normalized

average power P is finite and nonzero (0 < P < ∞).

➢ Definition: The total normalized energy is


E= 
−
g 2 (t )dt

➢ Definition: w(t) is an energy waveform if and only if the total

normalized energy is finite and nonzero (0 < E < ∞).

12
 Example
⚫ Find the energies of the signals shown in Figure below

2 2 2
1 1
E =  sin (t )dt =
2
 dt −  cos 2tdt =  + 0 = 
0
2 0
2 0

Faculty of Electronics & Telecommunications

 Example
⚫ Find the power of a sinusoid C cos (0t +  )
+
1
P = lim  C 2 cos 2 (0t +  )dt
T → T
−
T
1 0 2
=  C cos 2 (0t +  )dt
T0 0
T0
C2
=
2T0  1 + cos(2 t + 2 ) dt
0
0

C2 T0 T0

=   dt +  cos(20t + 2 )dt 
2T0  0 0 
C2 C2
= T0 + 0 =
2T0 2

Faculty of Electronics & Telecommunications

 Example
⚫ Find the energie and the power of e−t

 2 

 ( e ) dt =  ( e )dt = 
−t −2 t
E=
− −
T /2 2 T /2
P = lim
1
T → T  e −t
dt ( )
= lim
1
T → T  e −2 t
dt =  ( )
−T /2 −T /2

Faculty of Electronics & Telecommunications

 Energy and Power Waveforms
➢ If a waveform is classified as either one of these types, it
cannot be of the other type.
➢ If w(t) has finite energy, the power averaged over infinite time
is zero.
➢ If the power (averaged over infinite time) is finite, the energy
if infinite.
➢ However, mathematical functions can be found that have both
infinite energy and infinite power and, consequently, cannot be
classified into either of these two categories. (w(t) = e-t).
➢ Physically realizable waveforms are of the energy type.
– We can find a finite power for these!!

16
 Decibel
➢ A base 10 logarithmic measure of power ratios.
➢ The ratio of the power level at the output of a circuit
compared with that at the input is often specified by
the decibel gain instead of the actual ratio.
➢ Decibel measure can be defined in 3 ways
Decibel Gain
Decibel signal-to-noise ratio
Mill watt Decibel or dBm

➢ Definition: Decibel Gain of a circuit is:

17
 Decibel Gain

➢ If resistive loads are involved,

Definition of dB may be reduced to,

or

18
 Decibel Signal-to-noise Ratio (SNR)
➢ Definition. The decibel signal-to-noise ratio (S/R, SNR) is:

Where, Signal Power (S) =

And, Noise Power (N) =

So, definition is equivalent to

19
 Decibel with Mili watt Reference (dBm)

➢ Definition. The decibel power level with respect to 1 mW is:

= 30 + 10 log (Actual Power Level (watts)

• Here the “m” in the dBm denotes a milliwatt reference.
• When a 1-W reference level is used, the decibel level is
denoted dBW;
• when a 1-kW reference level is used, the decibel level is
denoted dBk.
E.g.: If an antenna receives a signal power of 0.3W, what is the
received power level in dBm?
dBm = 30 + 10xlog(0.3) = 30 + 10x(-0.523)3 = 24.77 dBm
20
Why probability in Communications?
⚫ Modeling effects of noise
– Quantization
– Channel
– Thermal
⚫ What happens when noise and signal are filtered, mixed, etc?
⚫ Making the “best” decision at the receiver
 Signals
⚫ Two types of signals
– Deterministic – know everything with complete certainty
– Random – highly uncertain, perturbed with noise
⚫ Which contains the most information? Information content is determined
from the amount of uncertainty and unpredictability. There is no
information in deterministic signals

Information = Uncertainty

Let x(t) be a radio broadcast. How useful is it if x(t) is known? Noise is ubiquitous.
2.5
y[n]
2 x[n]

1.5

x(t) y(t) 0.5

(t)
-0.5

-1

-1.5

-2

-2.5
0 100 200 300 400 500 600 700 800 900 1000
 Need for Probabilistic Analysis
⚫ Consider a server process
– e.g. internet packet switcher, HDTV frame decoder, bank teller line,
instant messenger video display, IP phone, multitasking operating
system, hard disk drive controller, etc., etc.

Customers arrive Queue,

at random times Length L Satisfied customer

Server:
1 customer
Rejected customer, per  seconds
Queue full
 Relative Frequency
⚫ nA – number of elements in a set, e.g. the number of times an
event occurs in N trials
⚫ Probability is related to the relative frequency
⚫ For N small, fraction varies a lot; usually gets better as N
increases n
f ( A) = A Relative Frequency
n
n 
P ( A ) = lim  A  Probability
n →
 n 
0  P ( A)  1
P ( A) = 0 Never Occurs
P ( A) = 1 Always Occurs
 Random Variables
⚫ Definition: A real-valued random variable (RV) is a real-
valued function defined on the events of the probability
system

Event RV P(x)
Value
E P(x)
A 3 0.2
B 1
A B -2 0.5
D 0.5
C C 0 0.1
D -1 0.2
-2 -1 0 3 x
 Cumulative Density Function

⚫ The cumulative density function (CDF) of the RV, x,

is given by Fx(a)=Px(x<a)

P(x) Fx(a)
1 1
0.5
0.2 0.1 0.2 0.5

-2 -1 0 3 x -2 -1 0 3 a
 Probability Density Function
⚫ The probability density function(PDF) of the RV x is
given by f(x)
⚫ Shows how probability is distributed across the axis

dFx ( a ) dPx ( x  a )
fx ( x ) = =
da a = x da a=x

fx(x)
1
0.5
0.2 0.1 0.2

-2 -1 0 3 x
 Types of Distributions
⚫ Discrete-M discrete values at x1, x2, x3,. . . , xm
⚫ Continuous- Can take on any value in an defined interval
fx(x) Fx(a)
1 1
0.5
DISCRETE
0.2 0.1 0.2 0.5

-2 -1 0 3 x -2 -1 0 3 a
fx(x) Fx(a) Continuous
1 1

0.5 0.5

-1 0 1 x -1 0 1 x
 PDF Properties
⚫ fx(x) is nonnegative, fx(x) > 0
⚫ The total probability adds up to one

 f x ( x ) dx = Fx (  ) = 1
−

fx(x) Fx(a)
CDF
2 PDF 1

-1 0 1 -1 1
 Calculating Probability
⚫ To calculate the probability for a range of values
Px ( a  x  b ) = Px ( x  b ) − Px ( x  a )
= Fx ( b ) − Fx ( a )
b +
= lim  f x ( x ) dx
 →0 a +

AREA= F(b)- F(a) F(b)

fx(x) 2
F(a)

-1 0 a b 1 -1 a b 1
 Discrete Random Variables
• Summations are used instead of integrals for discrete RV.
• Discrete events are represented by using DELTA
functions.

If x is discretely distributed and xi represents a discrete event

M
f ( x) =  P( xi ) ( x - xi )
i =1
L
F( a) =  P( xi )
i =1
PDF and CDF of a Triangular Wave
⚫ Calculate Probability that the amplitude of a triangle wave is
greater than 1 Volt, if A=2.
⚫ Sweep a narrow window across the waveform and measure the
relative frequency of occurrence of different voltages.

fx(x)
s(t)
A A fx(x)

1/2A
-A -A
-A 0 A
 PDF and CDF of a Triangular Wave

⚫ Calculate Probability that the amplitude of a triangle wave is

greater than 1 Volt, if A=2.
 1 1
PV ( v  1) =  fV ( v ) dv =  dv =
2

1 1 4 4
3 1
PV ( v  1) = FV (  ) − FV (1) = 1 − =
4 4

FV(v)
fV(v) 1
3/4
1/4

-2 0 1 2 -2 0 1 2
 PDF and CDF of a Triangular Wave
⚫ Calculate Probability that the amplitude of a triangle wave is in
the range [0.5,1] v, if A=2.

1 1
PV ( 0.5  v  1) =  fV ( v ) dv = 
1 1
dv =
0.5 0.5 4 8
1
PV ( 0.5  v  1) = FV (1) − FV ( 0.5 ) =
8

FV(v)
1
fV(v) 3/4
5/8
1/4

-2 0 1 2 -2 0 1 2
 PDF and CDF of a Square Wave
⚫ Calculate Probability that the amplitude of a square wave is at
+A.
⚫ Sketch PDF and CDF

s(t)
A fx(x)

-A 0 A
-A
 PDF and CDF of a Square Wave
⚫ Calculate Probability that the amplitude of a square wave is at
+A. 1/4
⚫ Sketch PDF and CDF

Fx(x)
s(t) fx(x)
A

-A 0 A -A 0 A

-A
 Ensemble Averages

⚫ The expected value (or ensemble average) of y=h(x) is:


y = E  y  =  h ( x ) f x ( x ) dx
−

[•] =  [•] f x ( x ) dx
−

For Discrete distributions


y = [h( x)] = E  y  =  h(x ) f (x )
i x i
i =−
 Moments

⚫ The r th moment of RV x about x=xo is


( x − xo ) =  ( x − xo ) r f x ( x ) dx
r
−

MEAN is the first moment taken about x o =0


m = x =  x f x ( x ) dx
−

VARIANCE  2 is the second moment around the mean


 = ( x − x) =  ( x − x) 2 f x ( x ) dx
2 2
−

STANDARD DEVIATION  - the second moment around mean


=  = 2
 ( x − x ) 2 f x ( x ) dx
−

 2 = x 2 − ( x) 2
 Binomial distribution

⚫ This is the Binomial distribution,

n−k  n  k n−k
P[ Ak ] = K p q
k
=   p q = b(k ; n, p)
k 
n n!
K =   = Ckn =
k  k !(n − k )!
= n choose k is the Binomial coefficient
= Combinations of n things taken k at a time.
 Binomial distribution Example
⚫ Full redundancy error correction code:
– In a disaster, residents must hear a call to evacuate three times
before they will leave.
– No ack / nack possible (unidirectional message)
– 0.7 probability that a single message is not received by a given
resident.
– Message is repeated 10 times by radio
⚫ What is the probability a given resident will not evacuate?
 Binomial distribution Example
⚫ let Ec = A0A1A2 be the event the resident did not evacuate.
– Ai is the event that i messages were received.
– Now apply binomial distribution:

P[ E c ] = P[ A0 ] + P[ A1 ] + P[ A2 ]

10  10  10 

=  (0.3) 0 (0.7)10 +  (0.3)1 (0.7) 9 +  (0.3) 2 (0.7)8
0 1 2
= (1)(1)(0.0282) + (10)(.3)(0.0404) + (45)(.09)(0.0576)
= 0.383
 Uniform Distribution
⚫ The UNIFORM distribution is:
  2m − A 
0, x  fx(x)
  2 
1 A
f ( x) =  , x−m  A
A 2
  2m − A  1/A
0, x 
  2 
m x
MEAN is:
A
 m+ 1
x =  x f x ( x ) dx =  A
2
x =m
− m− A
2

VARIANCE  2 is:
A
 m+ 1 A2
 = ( x − x) =  ( x − x) f x ( x ) dx = 
2 2 2
A
2
( x − m) dx =
2
− m− A 12
2

Quantization noise is uniformly distributed

 Gaussian Distribution
⚫ Gaussian distribution also called Normal distribution is one of
the most common and important distributions
⚫ PDF

fx ( x ) =
1
e
−( x − m )
2
( )
2 2
m is the mean and is  variance
2 
 Gaussian CDF
⚫ Start with definition of CDF:

F (a) =
a
f x ( x ) dx =
1 a −( x − m )
2
( )dx
2 2

− 
2 −
e

⚫ Change variables, y=(m-x)/:

1 ( m−a )  − y2 ( 2 ) ( − ) dy
F (a) =  e
2 
1  − 2
2dy = Q  m − a 

= y
e  
2 ( m − a )    
 m−a  1  − 2 2
F (a) = Q   Q( z) =  e d
   2 z
Q function
 Q function vs Error function
1 ( m−a )  − y 2 ( 2)
F (a) = ( − ) dy
2 
e

1  − 2

2 ( m − a ) 
= e y 2dy

1  − 2
Change variables t = y / 2 F (a) = 
t
e dt
 ( m−a ) 2

 m−a 1  m−a
F (a) = Q   = erfc   Complementary Error Function
   2  2 
2  − 2
erfc ( z ) =  e d  = 1 − erf ( z ) Error function
 z
 Ideal Low-Pass Filtered White Noise
⚫ Suppose white noise is applied to an ideal low-pass filter such
that
⚫ Baseband

⚫ Passband

The bandwidth of the complex signal is B, then the bandwidth of

the transmitted signal is 2B.
 Signal-to-noise power ratio (SNR)
⚫ Energy-per-bit (Eb), Energy per-symbol (Es). The power
density is N0/2
⚫ Since noise n(t) has a uniform power spectral density N0/2,
the total noise power in bandwidth 2B is N = N0/2 × 2B =
N0B.
Pr
SNR =
N0 B
⚫ Convert to Es
Pr Es
SNR = =
N0 B N0 BTs

where Ts is the symbol period.

⚫ Let , SNR = Es/N0

Faculty of Electronics & Telecommunications [48]