EE 321

Introduction to Communication
CHAPTER 3
ANALYSIS and TRANSMISSION of
SIGNALS

EE 321: Fuad Alsaadi, KAU 0

 3.1 Aperiodic Signal Representation by Fourier Integral

The Fourier transform G  f  of the signal g t  is defined as

 

G  f    g t    g t e  j 2  ftdt (Fourier Transform)


The signal g t  can be recovered from G  f  by

G  f   

g t    1
G  f e j 2  ftdf (inverse Fourier Transform)


We write
g t   G  f 
This is called a Fourier -Transform Pair

EE 321: Fuad Alsaadi, KAU 1

 3.1 Aperiodic Signal Representation by Fourier Integral

G  f  is in general complex and can be written (in polar form) as

j g  f 
G f   G f  e
 G  f  is called the “spectrum” of g t 
 G  f  is called the “magnitude or amplitude spectrum” of g t 
 g  f   G  f  is called the “phase spectrum” of g t 
 Example:
1 1  j tan1 2  f 
G f    e
1  j 2 f 2
1  4 f 2

Therefore,
1
G f   and g  f   G  f    tan1 2 f 
1  4 2 f 2

EE 321: Fuad Alsaadi, KAU 2

 3.1 Aperiodic Signal Representation by Fourier Integral
Example 3.1
Find the Fourier transform for x t   e at u t , a  0
Solution
  
X f    x t e  j 2  ft
dt   e u t e
at  j 2  ft
dt   e ate  j 2  ftdt
  0

 a  j 2  f t 

  a  j 2f t
  e  1 1  j tan1 2  f /a 
 e dt      e
0
  a  j 2 f  0 a  j 2 f a 2  4 2 f 2
1 g t 
Magnitude Spectrum: X  f  
a 2  4 2 f 2 1

e u t 
Phase Spectrum: x  f   X  f    tan1 2 f / a 
at

Fourier -Transform Pair : 1/a

G f 
1 

e at u t   ,a  0 2

a  j 2 f 
f
g  f   G  f 

2

EE 321: Fuad Alsaadi, KAU 3

Symmetry Property of the Fourier Transform

If g t  is a real function of t , then

G f   G   f  (conjuate-symmetric)
Therefore,
G f   G  f  (amplitude spectrum is an even function)
g f   g  f  (phase spectrum is an odd function)

If g t  is an imaginary, then
G f   G   f  (conjuate-antisymmetric)
Therefore,
G f   G  f  (amplitude spectrum is an even function)
g f   g  f  (phase spectrum is an even function)

EE 321: Fuad Alsaadi, KAU 4

 Linearity of Fourier Transform
If g1 t   G1  f  and g 2 t   G2  f 
Then, for all constants a1 and a2 , we have
a1g1 t   a2g2 t   a1G1  f   a2G2  f 

In other words,
   
 a1g1 t   a2g2 t   a1 g1 t   a2  g2 t   

EE 321: Fuad Alsaadi, KAU 5

3.2 Fourier Transforms of Some Useful Signals
Unit Impulse :

 

  t     t e  j 2  ftdt  e  j 2  ft   1
   t 0
 t   1

g t    t  G f 
1 1

0 t f

  t  has the same content at all frequencies

 It has infinite bandwidth

EE 321: Fuad Alsaadi, KAU 6

3.2 Fourier Transforms of Some Useful Signals
DC Signal :

 

 1
 f      f e j 2  ftdf  1


1   f 

g t   1 G f    f 

1 1

0
t 0
f

 g t   1 has content only at dc, i.e. at f  0

 It has “zero” bandwidth!!

EE 321: Fuad Alsaadi, KAU 7

3.2 Fourier Transforms of Some Useful Signals
Complex Exponential :

 

  f  f0      f  f0 e j 2  ftdf  e
1 j 2  f0t



   f  f0 
j 0t j 2  f0t
e e

g t   e  G  f     f  f0 
j 2  f0t

1

f0
f

 g t   e 0  e 0 has content only at frequency f0

j t j 2f t

 A complex exponential signal is a single-frequency signal

 It has “zero” bandwidth!!

EE 321: Fuad Alsaadi, KAU 8

3.2 Fourier Transforms of Some Useful Signals
Sinusoid :
 1 j  t  1 

 cos 0t     e 0  e 0
 2
j t
 j t
   e 0   e 0 
 2    j t 


 
0  2 f0 1
   f  f0     f
 f0 
2 
1
cos 0t  cos 2 f0t     f  f0     f  f0 
2 
g t   cos 0t G f 
1 / 2 1 / 2
t
0 0
f
f0 f0
 g t   cos 0t has content only at frequency f   f0
i.e. a real sinusoid has the same content (power) at both frequency sides
 A sinusoid is a single-frequency real signal
EE 321: Fuad Alsaadi, KAU 9
3.2 Fourier Transforms of Some Useful Signals
Rectangular Pulse :
 /2
  t   t   j 2  ft  e  j 2  ft 
   
  /2

    e dt   /2 e dt   j 2 f 

 j 2  ft

       /2
e j 2  f  /2  e  j 2  f  /2 sin  f  
    sinc  f  
j 2 f f 

 t /     sinc  f  

G  f    sinc  f  
g t    t /   
1
1 

 0  t f
 
2 2 null-to-null bandwidth  1 /  Hz

EE 321: Fuad Alsaadi, KAU 10

3.2 Fourier Transforms of Some Useful Signals
Triangular Pulse :
  t  

 t   j 2  ft   t   j 2  ft
        e dt     e dt
           
 
0  t   j 2  ft  t   j 2  ft

  1  e 
dt   1  e dt
 
  
 0 
  
 t  j 2  ft
0  t
 e dt  
 j 2  ft
e dt   e  j 2  ftdt
   0 

  sinc  f  
2

 t /     sinc2  f  

EE 321: Fuad Alsaadi, KAU 11

3.2 Fourier Transforms of Some Useful Signals
g t  G f 
 t  1 g t  G f 
u t  1
2
  f   1 / 2 j  f   t /    sinc  f  
sgn t  1 / jf   t /    sinc2  f  
1 / t j sgn  f  sinc Wt  1
W
 f /W 
1  f  sinc2 Wt  1
W
 f /W 
  f  f0  2 cos  f  
j 2  f0t
e ,
cos t /    t /  
cos 2 f0t 1   f  f     f  f   1  2 f  2
2  0 0 

sin 2 f0t 1 
 f  f0     f  f0 
2 j  
cos Wt / 2 2  f   2 f 
 cos     

 2W   W 
1  Wt / 
2
1 W
e atu t , a  0
a  j 2 f  

eatu t , a  0
1   t  mT  0
fs    f  nf , s
fs  1 / T0
m  n 
a  j 2 f
1
te atu t , a  0
a  j 2 f 
2

a t 2a
Note:
e ,a  0
a 2  2 f 
2
A table similar to this will be provided
e t
2
/2  2
,  0  2e  
2 2
/2 in all exams
2
 2 2  f  /2
te t 2 /2  2
,  0 j  2 f e
3

EE 321: Fuad Alsaadi, KAU 12

3.3 Some Properties of the Fourier Transform
Linearity : a1g1(t )  a2g2 (t )  a1G1( f )  a2G2 ( f )
Complex conjugate : g * (t )  G * (f )
Time Scaling :  
g(t )  1 /  G  f /  ,   0 real
Time / Frequency Reversal : g(t )  G f 
 j 2  ft0
Time Shift : g(t  t0 )  e G ( f ), t0 real
j 2  f0t
Frequency translation : e g(t )  G ( f  f0 ), f0 real
Amplitude modulation : g(t )cos 2 f0t  21 G ( f  f0 )  G ( f  f0 ) , f0 real
Amplitude modulation : g(t )sin 2 f0t  21j G ( f  f0 )  G ( f  f0 ) , f0 real
Amplitude modulation : g(t )cos 2 f0t     21 G ( f  f0 )e j  G ( f  f0 )e  j  , f0 and  real
 
Time convolution : g1(t ) * g2 (t )  G1( f )G2 ( f )
Frequency convolution : g1(t )g2 (t )  G1  f   G2  f 
Duality (Symmetry) : G (t )  g(f )
dn
 
n
Time differentiation : g (t )  j 2 f G (f )
dt n
dn
j 2t  g(t )  df n G(f )
n
Frequency differentiation :
t 1 1
Time integration :  g (  )d  
j 2 f
G ( f ) 
2
G (0)( f )

EE 321: Fuad Alsaadi, KAU 13

EE 321: Fuad Alsaadi, KAU 14
3.3 Some Properties of the Fourier Transform

Time-Frequency Duality

 g t e  j 2  ft
dt


g t  G f 

 G  f e j 2  ft
df


EE 321: Fuad Alsaadi, KAU 15

3.3 Some Properties of the Fourier Transform
Duality Property :
If g t   G  f 
Then,
G t   g f    
i.e.  G t     1 G  f  
 
 replace t with f
Example :


 1  

Find   2

1  t 
 

Solution :
a t 2a
From tables, e 
a  2 f 
2
2



 1  2  f
Then,    e


 1  t 2




EE 321: Fuad Alsaadi, KAU 16

3.3 Some Properties of the Fourier Transform
Time - Scaling Property :
If g t   G  f 
1  f  1
Then, g at   G  , a  0
a a 

i.e.  g at   a
 
 g t 

   replace f with f /a

Time - Reversal Property :

a  1 : g t   G f 
Example :

For a  0, Find  eat u t  
Solution :
1
From tables, e at u t  
a  j 2 f
1
Then, 
 eat u t   
a  j 2 f

EE 321: Fuad Alsaadi, KAU 17

 Significance of Time-Scaling Property

 t 

Π ( t ) ⇔ sin c ( f )

t f
 1 1 1 1
2 2

 2t 
Π ( 2t ) ⇔
1
2
 f 
sin c  
2 
t f
 1 1 2 2
4 4

 t / 2
t
Π   ⇔ 2sin c ( 2 f )
2

t f
2 2  1 1
2 2

EE 321: Fuad Alsaadi, KAU 18

Reciprocity of Signal Duration and Its Bandwidth

g t    t /   G  f    sinc  f  

 
t f
  1 1
 
2 2  
2/

EE 321: Fuad Alsaadi, KAU 19

3.3 Some Properties of the Fourier Transform

Complex - Conjugation Property :

If g t   G  f 
Then, g  t   G  f 

EE 321: Fuad Alsaadi, KAU 20

3.3 Some Properties of the Fourier Transform
Time - Shifting Property :
If g t   G  f 
Then, g t  t0   G  f e
 j 2  ft0

i.e.  g t  t0   e   j 2  ft0
 
 g t 
Example :
Find  e a t t0

Solution :

From tables, e
a t

2a
a  2 f 
2
2
Then, 
 e
a t t0
 
2a
a 2  2 f 
2
e
 j 2  ft0

EE 321: Fuad Alsaadi, KAU 21

 Amplitude Modulation
Frequency - Shifing Property :
If g t   G  f 
Then, g t e  G  f  f0 
j 2  f0t


i.e  g t e
j 2  f0t
   g t  replace f by f  f0

Amplitude Modulation Property :

1
g t  cos 2 f0t   G  f  f0   G  f  f0 
2 
1 
g t  sin 2 f0t   G  f  f0   G  f  f0 

2j  
1
g t  cos 2 f0t     G  f  f0 e j   G  f  f0 e  j  
2 

EE 321: Fuad Alsaadi, KAU 22

 Amplitude Modulation
To sketch a signal g t  cos 2 f0t , we observe that,
 g t  when cos 2 f t   1

g t  cos 2 f0t   
0

g t  when cos 2 f t   1

 0

Therefore,
 g t  cos 2 f0t  touches g t  when cos 2 f0t  is at its positive peaks
 g t  cos 2 f0t  touches -g t  when cos 2 f0t  is at its negative peaks
g t 
G f 

G f 

t f
g  f 

g t  cos 2 f0t 

t f
f0 f0

EE 321: Fuad Alsaadi, KAU 23

 Lowpass Signals
A signal g t  is said to be “lowpass” if its magnitude spectrum
G  f  is zero or negligible above a certain frequency B Hz, where
B is "not very far from" f  0.
The “bandwidth” of this signal is B Hz.

G f 

f
B

EE 321: Fuad Alsaadi, KAU 24

 Bandpass Signals
A signal gbp t  is said to be “bandpass” if its magnitude spectrum
Gbp  f  is zero or negligible except in an interval of width 2B Hz
around a “center frequency” f0 . The “bandwidth” of this bandpass
signal is 2B Hz.
Usually, B  f0
Gbp  f 
2B

f
f0

EE 321: Fuad Alsaadi, KAU 25

 Bandpass Signal Representation
A real-valued bandpass signal gbp t  with “center frequency” f0 and
“bandwidth” 2B Hz can be represented as
gbp t   gc t  cos 2 f0t  gs t  sin 2 f0t

 E t  cos 2 f0t   t  
where gc t  and gs t  are low-pass signals, each with a bandwidth B Hz
and G f  bp 2B
E t   g t   g t   (real) envelope of gbp t 
2
c
2
s

 t   tan1 gs t  / gc t  f

  f0
E t 
gbp t 

EE 321: Fuad Alsaadi, KAU 26

 Fourier Transform of a Periodic Signal
Consider a periodic signal g t  with period T , 0

and let Gn be the n th complex Fourier series coefficient.


Then, g t    jn 2  f0t
Gne
n 
    
and, G  f      Gne
n 
jn 2  f0 
t
   Gn  e 0   Gn   f  nf0 
 n 
j 2 nf t

n 
 
Example :
Find the Fourier transform of the periodic impulse train function

T t     t  kT  0
0
k 
Solution :
    1   
    t  kT0       1
e    f  nf 
jn 2  f0t

k   T0 n   T0 n 

0

EE 321: Fuad Alsaadi, KAU 27

 Computing Fourier Series Coefficients
using Fourier Transform
Consider a periodic signal v t  with period T0 and fundamental
frequency f0  1 / T0 , and let Vn be the n th complex
Fourier series coefficient.
Also, let
vT t   a single period of v t 
0

 
VT  f    vT t  .
0 0

Then,
1
Vn  VT nf0 
T0 0

EE 321: Fuad Alsaadi, KAU 28

 Convolution
The convolution of the two signals g1 t  and g2 t  is defined as

g1 t   g2 t  
 g1 t    g2  d 


 g1   g2 t   d 


g 1  


g 2  


g1 t   


t
EE 321: Fuad Alsaadi, KAU 29
 Convolution

g1 t 

w1  b1  a1 g1 t   g 2 t 
a1 t
b1
w 2  w1  w 2
g 2 t  t
a1  a 2 b1  b2
w2  b2  a2
a2 t
b2

EE 321: Fuad Alsaadi, KAU 30

 3.3 Some Properties of the Fourier Transform
Time Convolution :
If g1 t   G1  f 
g2 t   G2  f 
Then, g1 t   g2 t   G1  f G2  f 

    
i.e.  g1 t   g2 t    g1 t   g2 t  
Frequency Convolution :
If g1 t   G1  f   If g1 t  has bandwidth B1 and

g2 t   G2  f   g2 t  has bandwidth B2

Then, g1 t  g2 t   G1  f   G2  f    g1 t  g2 t  has bandwidth B1  B2

 If g t  has bandwidth B

   
i.e.  g1 t  g2 t    g1 t    g2 t     g n t  has bandwidth nB

EE 321: Fuad Alsaadi, KAU 31

Properties of the Fourier Transform
Time Differentiation :
If g t   G  f 
dg t 
Then,  j 2 fG  f 
dt

 d n 

 
Repeat application of this proprty yields   n g t    j 2 f   g t   
n

 
dt
 

Example :
Use time differentiation to find   t /    
Solution :
d 2  t /   1 2 1
dt 2


 t    

  t  

 t   
1  j 2f  4
         f  
2
j 2f   2
j 2  f   t /   e  2  e    sin

2
 sin  f  
1     sin c 2 f 

Then,   t /     sin 2
  f     
 
  
  f 
2
 f 
EE 321: Fuad Alsaadi, KAU 32
Properties of the Fourier Transform
Time Integration :
If g t   G  f 
t 1 1
Then,  g  d   G  f   G  0   f 
 j 2 f 2

  1
 1
  
t
i.e.  g  d    g t     g t     f 
 j 2 f 2   0
Example :
Use time integration to find  u t   
Solution :

    1
  1
 
t
 u t      d     t      t     f 
 j 2 f 2   0
1 1
   f 
j 2 f 2

EE 321: Fuad Alsaadi, KAU 33

 Fourier Transform and DC Value
Consider a signal g t  with Fourier transform G  f  .
What is the dc value of g t  ?
Is it G 0 ?

EE 321: Fuad Alsaadi, KAU 34

 Classification of Systems
A system  with input signal x t  and output y t  is described by
an input-output relationship of the form

y t    x t  
e.g.
 
Amplifier:  x t   Gx t 

  x  d 

t
Integrator:  x t  

Differentiator:  x t   dx t  / dt

Square-Law Device:  x t   x t 

2

 
d 2y t  dy t 
 
Dynamic System:  x t   solution of
dt 2
b
dt
 c  x t 

EE 321: Fuad Alsaadi, KAU 35

 Classification of Systems
Linearity :
A system  is linear if
x 1 t 
   
 1x 1 t   2x 2 t   1  x 1 t   2  x 2 t    1 
for all 1, 2 , x 1 t , x 2 t  x 2 t 

For a linear system, it makes no diffrence if we 2


make the linear combination before or after
x 1 t  
the system 1
x 2 t  
2

EE 321: Fuad Alsaadi, KAU 36

 Classification of Systems
Time - Invariant :
A system  is called time-invariant if a time shift (delay or advance)
in the input signal causes the same time shift in the output signal
Thus, for a continuous-time system, the system is time-invariant if

 
 x t     y t    for any real value of 

EE 321: Fuad Alsaadi, KAU 37

 Classification of Systems
Causality :
A system  is causal if current ouput of the system depends
on only current or past input (i.e., it is a realistic system)

Equivalent Definition :
A system  is causal if current input does not produce past
output

EE 321: Fuad Alsaadi, KAU 38

Linear Time-Invariant (LTI) Systems
Linear Time - Invariant (LTI) Systems :
A system is called linear time-invariant (LTI) if it is linear and also time-invariant
Impulse Response :
The impulse response h t  of an LTI system  is the system response to a unit
impulse input, i.e.
 
h t     t 
 t   h t 
Input - Output Relationship :
For an LTI system with impulse response h t , the output y t  for an input x t  is
y t   x t   h t 
- The frequency domain relationship between the input and output
is obtained by taking Fourier transform of both sides:
Y f   H f   X f 
H  f  is the Fourier transform of the impulse response h t , and is referred to as
the transfere function or the frequency response of the LTI system
j h  f 
In general, H  f  is complex and can be written as H  f   H  f  e
where H  f  is Amplitude response and h  f  is the Phase response
EE 321: Fuad Alsaadi, KAU 39
Linear Time-Invariant (LTI) Systems
Recall that h t  is the system response to impulse input  t ,
which is applied at t  0
Then,
a linear time-invariant system with impulse response h t  is causal if
h t   0 for all t  0
In this case, we have
y t   x t   h t 

 x t    h  d 


  x t    h  d 
0

EE 321: Fuad Alsaadi, KAU 40

 Signal Transmission Through a Linear System
Input signal Output signal
LTI System
x t  y t   h t   x t 
h t 
Time domain

Frequency domain X f  H f  Y f   H f X f 

Notes :
 If H  f  is zero or very small (negligible) at a certain frequency,
then this frequency will be disregarded (filtered out) from the output
 If H  f   1 at a certain frequency, this frequency will be amplified
 If H  f   1 at a certain frequency, this frequency will be attenuated
 An LTI system is usually called a “filter”
 The bandwidth of the filter is defined based on H  f  as in the
case of signals

EE 321: Fuad Alsaadi, KAU 41

 Signal Distortion During Transmission
j x  f 
X f   X f  e
H f   H f  e
j h  f 
x t  h t  y t 
 f 
y

j x  f  j h  f  j x  f h  f 
Y f   X f H f   X f  e H f  e  X f  H f  e  

Y f 

Y f   X f  H f 
y  f   x  f   h  f 
 The transmission of an input signal x t  through a system changes it
into the output signal y t 
 The amplitude of the specrum is changed to X  f  H  f 
 The phase of the spectrum is changed to x  f   h  f 
 This, in general, results in a change in the shape (distortion)
of the input signal
EE 321: Fuad Alsaadi, KAU 42
 Distortionless Transmission
 Transmission is said to be distortionless if the input and the output have
identical wave shapes within a multiplicative constant
 Thus, in distortionless transmission, the input x t  and the output y t 
satisfy the condition
y t   kx t  td  x t  h t  y t 
for some real gain/attenuation k and some delay td
 Taking the Fourier transform of both sides yields
 
Y  f    kx t  td   kX  f e
 j 2  ftd
 X  f ke
 j 2  ft

d

H f 

i.e. For a distortionless system,

H  f   ke
 j 2  ftd

H f   k
h  f   2 ftd
i.e. a distortionless system has
 a magnitude response that is constant for all frequencies
 a phase response that is a linear function of f

EE 321: Fuad Alsaadi, KAU 43

 Ideal vs. Practical Filters
 Ideal filters allow distortionless transmission of a certain band
of frequencies and suppress all the remaining frequencies
 The ideal low-pass filter allows all components below f  B Hz to pass
without distortion and suppresses all components above f  B Hz
Ideal Low-Pass Filter (LPF) Ideal High-Pass Filter (HPF)

H f  H f 

1 1

f f
B B B B
h  f   2 ftd
h  f   2 ftd

Ideal Band-Pass Filter (BPF)

H f 

f
f0 f0
h  f   2 ftd

EE 321: Fuad Alsaadi, KAU 44

 Ideal vs. Practical Filters
 Ideal filters are not physically realizable,
i.e. we cannot build them in the real world
 Why?
Because they are not causal, i.e. h t   0 for t  0
 e.g.
 f   j 2  ft
H LP  f     e
 2B 
d

 hLP t   2B sinc 2B t  td  
 H  f  is causal (physically realizable) if
 ln H  f 
 df   (Paley-Wiener condition) h t  1
1  2 f 
 2
2B
h t   0
 H  f  is not physically realizable if H  f   0
for t  0
over any finite band t
0 td
 H  f  is not physically realizable if H  f  is flat
over any finite band
 H  f  is not physically realizable if H  f  has an abrupt change
 H  f  can be zero at discrete frequencies and still be realizable

EE 321: Fuad Alsaadi, KAU 45

 Ideal vs. Practical Filters
 We can approximate an ideal filter by truncating the negative-time side

of h t  to get h t 

h t   h t  u t 
 This corresponds to
 t
H f  0 td

f
B B

 As td increases, h t  gets closer to h t 

EE 321: Fuad Alsaadi, KAU 46

 Practical Filters
H f 
Butterworth Filters :
1
H f  
1

1  f / B 
2n
1
2
B  3dB cutoff frequency n 1

n  Filter order
2
3
n   16 8 4
H f  f
B
Ideal Butterworth, n = 4
1 h t 
1 Butterworth, n = 4 Ideal
2

f /B t
1
Butterworth, n = 4

h  f 
Ideal

EE 321: Fuad Alsaadi, KAU 47

3.7 Signal Energy and Energy Spectral Density
Parseval's Theorem for Energy Signals
For a signal g t  with spectrum G  f ,
 2  2
Eg   g t  dt   G  f  df
 

 This means that the enrergy of a signal can be computed in time

domain or in frequency domain
Example : Verify Parseval's theorem for the signal g t   e at u t , a  0
 2  1
Solution : We have, E g   g t  dt   e dt  2at
 0 2a
1
We can also determine E g from the signal spectrum G  f  
a  j 2 f

 2  1 1 
 1  2 f 
 1
E g   G  f  df   df  tan  
 

a  2 f 
 2 2
2  a  a 
2a
which verifies Parseval's theorem
EE 321: Fuad Alsaadi, KAU 48
 3.7.2 Energy Spectral Density (ESD)
The energy spectral density of a signal g t  is defined as
2
g f   G f 

The ernergy of a real signal g t  in a frequency band  f1, f2  is

f2 f2 2
E B  2   g  f df  2  G  f  df
f1 f1

EE 321: Fuad Alsaadi, KAU 49

 Signal Bandwidth
 The bandwidth of a signal g t  is the width of the positive-frequency
range over which the magnitude spectrum G  f  has a “significant content”
 The meaning of “significant content” depends on the application.
Therefore, there are many definitions for bandwidth
 Absolute Bandwidth : the range of frequenceis for which G  f   0

G f  G f  G f 

f f f
Babs  
Babs Babs

EE 321: Fuad Alsaadi, KAU 50

 Essential Bandwidth of a Signal
 The essential bandwidth of a signal g t  is the witdth of the frequency range
which contains x % of the total signal energy.
 x is chosen large enough so that most of the energy is included.
 This is also called the x%-energy bandwidth.
Example : Find the (95%) essential bandwidth of the signal g t   e atu t 
1 2 1
Solution : In this case, G  f   and the ESD is G  f  
a  j 2 f a 2  2 f 
2

B
B 2 B 1 1  
1  2  f  1 
1  2B 

E B   G  f  df   df  tan    tan  
  a    a 
a  2 f 
2
B B 2 2 a B
a
If the essential band is required to contain 95% of the signal energy
EB 1
Then,  0.95 where E g 
Eg 2a
0.95 1 
1  2B 
 0.95 2B  0.95 180 
EB   tan      tan   
2a a  a  2a a 
 2  
The essential bandwidth is
B  2a Hz
EE 321: Fuad Alsaadi, KAU 51
 Other Definitions of Bandwidth
(a) 3-dB Bandwidth = half-power bandwidth

(b) absolute bandwidth: the range of frequencies for which G f   0

(c) null-to-null or main-lobe bandwidth
G f 
(d) x-dB Bandwidth K

(e) x%-energy/power bandwidth

Baseband Signal
K/ 2

= Essential Bandwidth 10x /20 K

f
(a)
(c)

(d)

G f 
(b)

K
K/ 2 Bandpass Signal

10x /20 K
f
(a)
(c)

(d)

(b)

EE 321: Fuad Alsaadi, KAU 52