ELG3175 Introduction to

Communication Systems

Fourier transform,
Parseval’s theoren,
Autocorrelation and
Spectral Densities
Fourier Transform of a Periodic
Signal
• A periodic signal can be expressed as a complex
exponential Fourier series.
• If x(t) is a periodic signal with period T, then :
∞ n
j 2π
∑X
t
x(t ) = T
ne
n = −∞

• Its Fourier Transform is given by:

 ∞ j 2π t 
n
 ∞ 
 j 2π
n 
t ∞
X ( f ) = F  ∑ X ne T
 = ∑ X nF e T
 = ∑ X nδ ( f − nf o )
n = −∞  n = −∞   n = −∞

Lecture 4
Example
x(t)

A
… …
0.25 0.5 0.75 t

-A

∞
2 A j 4πnt ∞
 2 A  j 4πnt
x(t ) = ∑
n = −∞ jπn
e = ∑ − j
n = −∞  πn 
e
n is odd n is odd
∞
 2A 
X(f ) = ∑
n = −∞
− j

δ ( f − 2n )
πn 
n is odd

Lecture 4
 Example Continued

|X(f)|
2A/π 2A/π

2A/3π 2A/3π
2A/5π 2A/5π

-10 -6 -2 2 6 10 f

Lecture 4
Energy of a periodic signal
• If x(t) is periodic with period T, the energy on one
period is:

∫
E p = | x(t ) | 2 dt
T

• The energy on N periods is EN = NEp.

• The average normalized energy is E = lim NE p = ∞
N →∞

• Therefore periodic signals are never energy signals.

Lecture 4
Average normalized power of
periodic signals
• The power of x(t) on one period is :
1
Pp =
T ∫
| x(t ) | 2 dt
T
• And it’s power on N periods is :

 N iT 
  1
∑
1 1 1
PNp =
NT ∫
| x(t ) | 2 dt = 
N  i =1 T ∫ | x(t ) | 2 dt  = × NPp = Pp
 N
NT  (i −1)T 
• It’s average normalized power is therefore P = lim PNp = Pp
N →∞
• Therefore, for a periodic signal, its average normalized
power is equal to the power over one period.

Lecture 4
Hermetian symmetry of Fourier
Transform when x(t) is real
*
 ∞
− j 2πft

X ( f ) =  ∫ x(t )e
*
dt 
−∞ 
∞
j 2πft
∫
*
= x (t ) e dt
−∞
∞
j 2πft
= ∫ x (t ) e dt
−∞
∞
− j 2π ( − f ) t
= ∫ x (t ) e dt
−∞
= X (− f )

Lecture 4
Parseval’s Theorem
• Let us assume that x(t) is an energy signal.
• Its average normalized energy is :
∞

∫
2
E= x(t ) dt
−∞∞
= ∫
−∞
x(t ) x * (t )dt
∞∞ 
∫ ∫
∞ ∞
=  X ( f )e j 2πft
df  x * (t )dt
∫ ∫
2 2

− ∞ − ∞
 x(t ) dt = X ( f ) df

∞ ∞ −∞ −∞

∫ X(f ) ∫ x
* − j 2π ( − f )t
= (t )e dtdf
−∞ −∞
∞ ∞

∫ ∫
2
= X ( f ) X * ( f )df = X ( f ) df
−∞ −∞

Lecture 4
Example
∞

∫
2
sinc (t )dt = ?
−∞
∞ ∞

∫ ∫
2 2
sinc (t )dt = | sinc (t ) | dt
−∞ −∞
∞

∫
2
= | Π ( f ) | df
−∞
1/ 2
= ∫ df
−1 / 2
= 1

Lecture 4
Autocorrelation function of energy
signals
• The autocorrelation function is a measure of similarity
between a signal and itself delayed by τ.
• The function is given by :
∞
ϕ x (τ ) = ∫ x(t ) x* (t − τ )dt
−∞

• We note that
∞ ∞

∫ ∫
2
ϕ x ( 0) = x(t ) x * (t )dt = x(t ) dt = E
−∞ −∞

• We also note that ϕ x (τ ) = x(τ ) * x * (−τ )

Lecture 4
Energy spectral density

• Let x(t) be an energy signal and let us define G(f) =

|X(f)|2
• The inverse Fourier transform of G(f) is
∞ ∞

∫ G ( f )e ∫
j 2πfτ * j 2πfτ
df = X ( f ) X ( f ) e df
−∞ −∞
∞ ∞

∫ ∫
− j 2πft * j 2πfτ
= x (t ) e dtX ( f )e df
− ∞− ∞
∞ ∞

∫ x(t ) ∫ X ( f )e j 2πfτ e − j 2πft dfdt

*
=
−∞ −∞
∞ ∞

∫ ∫
* j 2πfτ j 2πf ( − t )
= x (t ) X ( f )e e dfdt
−∞ −∞
∞

∫
*
= x (t ) x (t − τ )dt
−∞

Lecture 4
Energy spectral density

• The energy spectral density G(f) = F{ϕ(τ)}.

• Example
– Find the autocorrelation function of Π(t).

Lecture 4
Energy spectral density of output
of LTI system
• If an energy signal with ESD Gx(f) is input to an LTI
system with impulse response H(f), then the output,
y(t), is also an energy signal with ESD Gy(f) =
Gx(f)|H(f)|2.

Lecture 4
ESD describes how the energy of
the signal is distributed throughout
the spectrum2
|H(f)|

|X(f)|2 1

-f c fc f
|Y(f)|2
|X(fc)|2 Ey = 2|X(f)|2∆f
∆f is Hz, therefore
|X(f)|2 is in J/Hz
-f c fc f

Lecture 4
Autocorrelation Function of Power
Signals
• We denote the autocorrelation of power signals as Rx(τ).
• Keeping in mind that for the energy signal
autocorrelation, ϕ(0)=Ex, then for the power signal
autocorrelation function Rx(0) should equal Px.
• Therefore

T
1
∫
*
Rx (τ ) = lim x (t ) x (t − τ )dt
T → ∞ 2T
−T

Lecture 4
Power Spectral Density

• Power signals can be described by their PSD.

• If x(t) is a power signal, then its PSD is denoted as
Sx(f), where Sx(f) = F{Rx(τ)}.
• The PSD describes how the signal’s power is distributed
throughout its spectrum.
• If x(t) is a power signal and is input to an LTI system,
then the output, y(t), is also a power signal with PSD
Sy(f) = Sx(f)|H(f)|2.

Lecture 4