04

Convention used:
X n X n*  X n  An2  Bn2
2

If X n  An  jBn then X n*  X  n  An  jBn

Parseval’s Power Theorem:
1 2 1
T0 T0 T0 T0
P v (t ) dt  v(t )v * (t )dt

But according to complex Fourier series,

 1  j n 0 t
C e T0 T0
j n 0 t
v(t )  n where
Cn  v (t ) e dt
n  
Therefore,
*
  j n 0 t 

v (t )    Cn e
*
   C *  j n 0 t
ne
n  n
1   * j n  t
v(t )  Cn e
T0 T
P  dt
0

0
  
Thus,
 
P C C   C * 2
n n n
n   n  

Fig. A one-sided plot of spectral amplitude of periodic waveform.

Parseval’s theorem:
Let v(t) be a real signal and V() is its Fourier transform.

j  t
V ( )   v(t ) e dt ................................ (1)


1 jt
v(t ) 
2  V () e

d ................................ (2)

If v(t ) is real, then v * (t )  v(t ) and from equation (1),

1
 *
 j  t 
V ( )    v(t ) e
*
dt 
  
 V ()

Thus V ( )  V ( ) ................................ (3)

*

is the necessary condition for v(t) to be real.


 v1 (t )v2 (t )   v1 (t )v2 (t ) e
j  t
dt


1  
 
jt j  t
    V1 ( ) e d  v2 (t ) e dt

2   
1

 (   ) t 
V1 ( )v2 (t ) e
j
 dt  d
2 
Therefore,

 v1 (t )v2 (t ) 
1

2 
V1 ( )V2 (   )d ............................ (4)
Equation (4) is called the frequency convolution theorem. Also written as
1
v1 (t )v2 (t )  V1 ( ) *V2 ( )
2
That is,
 
j  t 1
 v (t )v (t ) e

1 2 dt 
2 
 V ( )V (   )d
1 2

Setting  = 0, we get,
 
1
1 2

v (t )v (t ) dt 
2  V ( )V ( )d

1 2

By changing the dummy variable of integration,

 
1


v1 (t )v2 (t ) dt 
2  V ()V ()d ............................ (5)

1 2

Equation (5) is called Parseval’s relation.

Setting v1(t) = v2(t) = v(t) in equation (5),
 

v (t )  dt
2

1
 V ( )V ( )d
2 
Therefore,
 

 v (t )  dt
2

1
 V () d ................................ (6)
2


2 

2
Equation (6) is Parseval’s identity or Parseval’s theorem.

E  v(t ) dt ..................................... (7)
2


T /2
1
T  T 
S  lim
2
v (t ) dt ..................................... (8)
T / 2

Normalized power in a Fourier expansion:

 2 n t

v(t )  C0   Cn cos  n  ...................... (9)
n 1  T0 
Considering the two typical terms of the Fourier expansion (fundamental and first harmonic),
 2 t   4 t 

v (t )  C1 cos
/
 1   C2 cos
  2  ..... (10)
 T0   T0 
To calculate the normalized power S/ of v/(t), we must square v/(t) and evaluate,
T /2
1
S   [v (t )] dt ..................................... (11)
/ / 2

T0 T / 2
However the two cosine functions in equation (10) are orthogonal. Hence,
C12 C22
S 
/
 ..................................... (12)
2 2

Cn2
S C  2
0 ..................................... (13)
n 1 2
Also,

An2  Bn2
S  A  2
0  ..................................... (14)
n 1 2 n 1 2
j 2 n t / T0
But in the complex Fourier representation the terms are of the form Vne .


V e
j 2 n t / T0
v(t )  n ..................................... (15)
n  
Where
T0 / 2
1 j 2  n t
Vn 
T0  v(t ) e
T0 / 2
dt ..................................... (16)

j 2 n t / T0  j 2 n t / T0
Vne Vne  VnVn  VnVn* ................... (17)
The total normalized power is
n  
S V V
n  
n n
*
................... (18)

3
 Fig. A two-sided power spedctrum
Sn is the power associated with each spectral component.

Power Spectral Density (PSD):

Fig. The sum S(f ) of the normalized power in all

spectral components from f = –  to f.

The normalized power at the frequency f in a range df,

dS ( f ) dS ( f )
dS ( f )  df ................... (19) and G( f )  ................... (20)
df df
The power in the range df at f is G(f)d(f).
The power in the positive – frequency range f1 to f2 is
f2

S ( f1  f  f 2 )   G( f )df ................... (21)

f1
The power in the negative – frequency range – f2 to – f1 is
 f1 f2
 f1

S ( f 2  f   f1 )   G( f )df ........ (22) and S ( f1  f  f 2 )   G( f )df   G( f )df ........... (23)

 f2
 f2 f1
Thus

G( f )   Vn  ( f  n f 0 ) ................... (24)
2

n  

Assignment:
Show that the two cosine functions cos nx and cos mx are orthogonal, where n and m are
integers. That is, find cos nxcos mx .