Fourier Series
[p. 3] Fourier series definition
[p. 3] Complex exponential expansion
[p. 6] Fourier analysis equation
[p. 8] Examples
06/09/10
Fourier Series
Most signals may be represented as Sinusoids with
DIFFERENT Frequencies
N
x(t ) =
j 2 kf0t
a
e
+ A0
k
k =
06/09/10
x (t) =
ake
j 2 k f0 t
k =
06/09/10
SYNTHESIS
Easy
Given (k,Ak,k) create
x(t)
Synthesis can be
HARD
Synthesize speech so that it
sounds good
06/09/10
ANALYSIS
Hard
Given x(t), extract
(k,Ak,k)
How many
harmonics?
Need algorithm for
computer
STRATEGY: x(t) ak
x (t ) =
ak e
j 2 k f 0 t
k =
ak = T1
0
T0
j0k t
x(t)e
dt
2 f 0
6
T0
(DC component)
Property: a k = ak*
06/09/10
Example:
06/09/10
06/09/10
Example:
x(t)
x (t ) = sin 2 (3 t )
06/09/10
10
06/09/10
11
Example:
x(t)
x ( t ) = sin 3 ( 3 t )
06/09/10
12
06/09/10
13
Example:
x(t)
A periodic signal x(t) is described over one period [0, T0], by the equation
t , 0 t tc
x(t ) =
0, tc t T0 ,
(a) Sketch the periodic function x(t) for the specific case tc=T0/2
(b) Determine the D.C. coefficient of the Fourier Series, a0.
06/09/10
14
x(t ) =
k =
ak e j 30 kt
1/(4 + 2 jk ) for k = 3, 2, 1, 0
ak =
0 for | k |> 3
(a) Sketch the two-sided spectrum of this signal. Label all complex
amplitudes in polar form.
(b) Determine the fundamental frequency (in Hz) and the fundamental
period (in secs.) of this signal.
06/09/10
15
x(t ) =
ae
k =
jk 2 f 0t
jk 2 f 0t
k
k =
y (t ) =
06/09/10
be
16
x(t ) =
ae
k =
jk 2 f 0t
y (t ) =
be
jk 2 f1t
k =
Question 2: How do we compute bks ?
06/09/10
17
x(t ) =
k =
ak e jk 2 f0t
y (t ) =
be
k =
jk 2 f1t
06/09/10
18
Summary
Basic FS Coefficient Properties
x(t)
y(t)
ak
bk
06/09/10
x(t) real
==> ak=a-k*
ak real
ak imaginary
19
Summary
Basic FS Coefficient Properties
x(t)
y(t)
ak
bk
Property
Signal
Linearity
Ax(t) +By(t)
x ( t t0 )
Time-shift
Time reversal
Conjugation
Parsevals relation:
06/09/10
x ( t )
x * (t )
Coefficients
Aak +Bbk
ak e
a k
a* k
jk 0t0
2
1
2
Px = x ( t ) dt = ak
TT
k
20
Example
1
t
-1
06/09/10
21
22
06/09/10
23
x(t ) =
T0
j 2 kf0 kt
j0kt
1
a
e
;
a
=
x
(
t
)
e
dt
k
k
T0
k =
2 f 0
ak = T10
= T10
T0
x(t )e jk0 t dt
T0
0
T0
T0
Real part: R
06/09/10
Imaginary part: I
24
25
06/09/10
26
06/09/10
27
end
>a=0, ,0,0
06/09/10
28
ak =
T0
1
T0
x (t)e
j ( 2 / T0 ) k t
dt
29
30
1 2
2
06/09/10
31
1 2
2
y ( t ) = + cos(50 t 2 ) +
sin(150 t ) + ...
2
3
2
2
sin(250 t ) +
sin(350 t )
+
5
7
06/09/10
32
06/09/10
33
Gibbs Phenomenon
Convergence at DISCONTINUITY of x(t)
There is always an overshoot
9% for the Square Wave case
06/09/10
34