Communications Engineering - ELEC 341

CHAPTER 3

ANALYSIS AND TRANSMISSION

OF SIGNALS
Lecture 3: Outline
1. Aperiodic Signal Representation By Fourier Integral
2. Transforms of Some Useful Functions
3. Properties of the Fourier Transform
4. Signal Transmission Through Linear System
5. Ideal Versus Practical Filters
6. Signal Distortion Over a Communication Channel
7. Signal Energy and Energy Spectral Density
8. Signal Power and Power Spectral Density
9. Numerical Computation of Fourier Transform

2
3.1 Aperiodic Signal Representation By Fourier Integral
 Fourier series decomposition is only applicable to periodic functions.
 Non-periodic functions can be analysed via the concept of Fourier
transform.
 Finite-energy functions are nonperiodic.
 Often (but not necessarily) nonperiodic functions have a finite support,
or at least most of the energy is concentrated within a finite time period

Fig. 3.1:
Construction of a periodic signal
by periodic extension of aperiodic
signal g(t).

3
 Aperiodic Signal Representation By Fourier Integral
 Periodic signals have line spectra (discontinuous spectra)
 They can be developed in Fourier series.
T0 / 2 
1 1
g 
 j 2nf0t  j 2nf0t
Dn  T0 (t )e dt Dn  g (t ) e dt
T0 T0 / 2 T0 

Fig 3.2:
Change in the Fourier spectrum
when the period T0 in Fig. 3.2 a
is doubled.

4
 2.1 Size of Signal
 In the limit as T0 → ∞, Δf = 1/ T0 → 0 we have gT0 (t )  g (t )
 Therefore 
g (t )  lim gT0 (t )  lim  G(nf )e( j 2nf )t f
T0  f 0
n  
 G(f) is Fourier transform of continuous time function of w

g (t )   G( f )e jwt dt


 Fourier coefficients Dn are the samples of G(f) at intervals of f0:

1
Dn  G (nf 0 )
T0

Figure 3.3
The Fourier series becomes
the Fourier integral in the
limit as T0 →∞.

5
Fourier transform and continuous spectrum

 Consider signals whose energy E   g (t ) dt is finite.
g
2


 Which means that the signal is concentrated to a relatively
short time period (time-limited).
 Fourier transform for this kind of energy signal is defined as



 jwt
G( f )  F [ g (t )]  g (t ) e dt


G(f) is the spectrum of signal g(t).

 Most non-periodic signals have continuous spectra

 => Fourier transform is used (instead of Fourier series) .
Fourier transform and continuous spectrum
 The spectrum has the following properties:
1. G(f) is a complex function. |G(f)| is the amplitude spectrum and
argG(f) is the phase spectrum. 

2. The value of G at f=0 equals the net area of g(t) : G(0)   g (t )dt

*
3. If g(t) is real => G(-f) = G (f) (Hermitian symmetry), i.e,
| G( f ) | G( f ) arg G( f )   arg G( f )

 The time function is obtained from by using the inverse Fourier

transform:

g (t )  F 1[ f ]   G( f )e j 2ft df


7
Example 3.1
 Find the Fourier transform of g (t )  eatu(t ) where a  0

 The FT of g(t):
G(f)  FTg (t ) 
1 1  j tan 1 ( 2af )
a0 G( f )  e
a  jw a  2f 
2 2

Magnitude: Phase:
1
G( f )   g ( f )   tan 1 ( 2af )
a 2  2f 
2
Existence of the Fourier Transform
 In example 3.1, if a < 0 then the integral of eatu(t ) does
not converge, FT of g (t )  eatu(t ) when a  0 does not exit


 Existence of FT g(t) should satisfy Dirichlet condition



 g (t ) dt  


Recall | e j 2ft | 1
Hence 
G( f )   g (t ) dt


9
FT of Useful Functions

3.6 rectangular pulse

3.7 Triangular pulse

Sinc pulse: sinc(x)=sin(πx)/ πx

10
Example: Rectangular pulse
 The notation  (t /  )is used for rectangular pulse. It is defined as
A t  / 2
 (t /  )  
0 t  / 2
Consider the signal g (t )  A (t /  ) . Its Fourier transform is
 /2
A

 j 2ft
G( f )  Ae dt  sin f  A sinc f
 / 2
f

 It can be seen that the spectrum of the rectangular pulse corresponds

to the envelope of the spectrum of the rectangular pulse train (see
previous example).
 It can be also noted that most of the spectral content is located in the
frequency band of f  1 / 
 This means that the spectrum of a narrow pulse is wide.

11
Example: Rectangular pulse train
Coefficients of the Fourier series can be calculated as follows:
T0 / 2  /2
1 1
  Ae
 j 2nf0t  j 2nf0t
Dn  v(t )e dt  dt
T0 T0 / 2
T0  /2


A
 j 2nf 0T0

e  jnf0  e jnf0 
A sin nf 0

T0 nf 0
A
 sincnf 0
T0

Where,

12
Example: Dirac pulse
 Find the FT of unit impulse signal δ(t)

G( f )  FT [ (t )]    (t )e  j 2ft dt  e  j 2f 0  1


13
FT of sign function: sgn(t)
1 t  0

 Sign function: sgn( t )  0 t  0
 1 t  0


 
sgn( t )  lim e  atu(t )  e atu(t )  FT [sgn( t )]
a 0

 1 1  1
FT [sgn( t )]  lim   
 jf
a 0 a  j 2f a  j 2f
 

14
Time Frequency Duality
Near symmetry between direct and inverse Fourier transforms
Time-shifting property Frequency-shifting property
g (t  t0 )  G( f )e j 2ft0 g (t )e j 2f0t  G( f  f0 )

15