Continuous-Time Fourier Transform

You might also like

You are on page 1of 68

Continuous-Time

Fourier Transform

Content
Introduction
Fourier

Integral
Fourier Transform
Properties of Fourier Transform
Convolution
Parsevals Theorem

Continuous-Time
Fourier Transform
Introduction

The Topic
Periodic

Discrete
Time

Fourier
Fourier
Series
Series

Discrete
Discrete
Fourier
Fourier
Transform
Transform

Aperiodic

Continuous
Time

Continuous
Continuous
Fourier
Fourier
Transform
Transform

Fourier
Fourier
Transform
Transform

Review of Fourier Series


Deal

with continuous-time periodic signals.


Discrete frequency spectra.
AAPeriodic
PeriodicSignal
Signal

f(t)
t
T

2T

3T

Two Forms for Fourier Series

Sinusoidal
aa00
22nt
22nt
nt
nt
ff((tt))
a
cos

b
sin
ann cos

bnn sin

Form
22 n 1
TT
TT
n 1
n 1

2 T /2
a0
f (t )dt

T
/
2
T

Complex
Form:

ff((tt))
ccnnee
nn

jnjn0t t
0

n 1

2 T /2
an
f (t ) cos n0tdt
T T / 2
2 T /2
bn
f (t ) sin n0tdt
T T / 2

1
cn
T

T /2

T / 2

f (t )e jn0t dt

How to Deal with Aperiodic Signal?


AAPeriodic
PeriodicSignal
Signal

f(t)
t
T

If T, what happens?

Continuous-Time
Fourier Transform
Fourier Integral

Fourier Integral
fT (t )

c e

1
cn
T

jn0t

n T

2 n
1

2 n
1

T /2

f T ( ) e

T / 2

T /2

T / 2

T /2

T / 2

jn0

d e

T /2

T / 2

fT (t )e jn0t dt

jn0t

fT ()e jn0 d 0 e jn0t

fT ()e jn0 d e jn0t

fT ()e j d e jt d

2
T

1 0

T 2

Let 0

2
T

T d 0

Fourier Integral
1
f (t )

f ()e

jt

d e d

F(j )

1
j t
f (t )
F ( j)e d

F ( j) f (t )e

j t

dt

Synthesis
Analysis

Fourier Series vs. Fourier Integral


Fourier
Series:

jnjn0t t
0
ff((tt))
c
e
c
e
n
n

Period Function

nn

11 TT/ /22
ccnn
ffTT((tt))eejnjn0t0tdt
dt
TT TT/ /22

Fourier
Integral:

Discrete Spectra

11
ff((tt)) FF((jj))eejjt tdd
22

FF((jj)) ff((tt))eejjt tdt


dt

Non-Period
Function

Continuous Spectra

Continuous-Time
Fourier Transform
Fourier Transform

Fourier Transform Pair


Inverse Fourier Transform:

11
ff ((tt)) FF((jj))eejjttdd
22

Synthesis

Fourier Transform:

FF((jj)) ff ((tt))ee

j
jt t

dt
dt

Analysis

Existence of the Fourier Transform

Sufficient Condition:
f(t) is absolutely integrable, i.e.,

|
f
(
t
)
|
dt

|
f
(
t
)
|
dt

Continuous Spectra

F ( j) f (t )e jt dt

F ( j) FR ( j) jFI ( j)

| F ( j) | e
Magnitude

j ( )
Phase

FI(j)

|
)

j
|F(
()

FR(j)

Example
1

-1

f(t)
t

1 j t
F ( j) f (t )e dt e dt
e

1
j
j j
2 sin
j
(e e )

j t

j t

F(
) )
F(

Example
-1

11
00

-1-1
f(t)
33
22

|F(
)|)|
|F(

33
22

11

-10
-10

-5-5

00

55

00
-10
-10
1
44

10
10

j t

arg[F(
)])]
arg[F(

1 j t
F ( j) f (t )e dt e dt
e

1
j
22
j j
2 sin
j
00
(e e ) -10-10 -5-5 00 55 1010

-5-5

j t

00

55

10
10

Example
f(t)

et
t

F ( j) f (t )e

j t

e
0

( j ) t

dt e t e jt dt
0

1
dt
j

Example
f(t)

1
1
|F(j)|
|F(j)|

=2
=2

0.5
0.5
0
0

2
2

e
0

arg[F(j)]
arg[F(j)]

F ( j) f (t )e

-10
-10

j t

0
0

-2

( j-2) t

et

-5
-5

0
0

5
5

t 1010

dt e t e jt dt
0

1
dt
j
-10
-10

-5
-5

0
0

5
5

10
10

Continuous-Time
Fourier Transform
Properties of
Fourier Transform

Notation
FF[[ ff ((tt)]
)] FF(( jj
))

FF [[FF(( jj
)]
)] ff ((tt))
--11

ff ((tt))

FF(( jj
))
FF

Linearity
aa11ff11((tt))aa22 ff22((tt))

aa11FF11(( jj
))aa22FF22(( jj
))
FF

e
v
e
l
v
e
l
s
e
r
s
u
r
o
u
y
o
y
y
b
y
b
d
e
d
v
e
PPrroov

Time Scaling
11

ff ((at

at))

FF jj
||aa|| aa
FF

e
v
e
l
v
e
l
s
e
r
s
u
r
o
u
y
o
y
y
b
y
b
d
e
d
v
e
PPrroov

Time Reversal
ff ((tt))

FF jj

FF

Pf) F [ f (t )] f (t )e jt dt t f (t )e jt dt

j t

f
(
t
)
e
d (t )
f (t )e d (t ) t

j t

f (t )e dt
j t

f (t )e jt dt

f (t )e jt dt F ( j)

Time Shifting
ff ((tt tt00))

FF jj
ee

j
jtt00

FF

Pf) F [ f (t t )] f (t t )e jt dt t f (t t )e jt dt
0
0
0

t t0

j t 0

j t 0

t t 0

f (t )e j( t t0 ) d (t t0 )

f (t )e jt dt

j t
f (t )e jt dt F ( j)e 0

Frequency Shifting (Modulation)


j
j0

ff ((tt))ee

FF jj((

00))

Pf)
F [ f (t )e

FF

j 0 t

] f (t )e j0t e jt dt

f (t )e j ( 0 ) t dt

F j ( 0 )

Symmetry Property
FF[[FF(( jtjt)])] 22ff ((
))
Pf)

2f (t ) F ( j)e jt d

2f (t ) F ( j)e jt d

Interchange symbols and t

2f () F ( jt )e jt dt F [ F ( jt )]

Fourier Transform for


Real Functions
If f(t) is a real function, and F(j) = FR(j) + jFI(j)
F(j) = F*(j)

F ( j) f (t )e

jt

dt

F * ( j) f (t )e dt F ( j)

jt

Fourier Transform for


Real Functions
If f(t) is a real function, and F(j) = FR(j) + jFI(j)
F(j) = F*(j)
FR(j) is even, and FI(j) is odd.

F R( j ) = F R(j ) F (I j ) = F (I j )

Magnitude spectrum |F(j)| is even, and


phase spectrum () is odd.

Fourier Transform for


Real Functions
If f(t) is real and even
F(j) is real
Pf)
Even

f (t ) f (t )

If f(t) is real and odd


F(j) is pure imaginary
Pf)
Odd

F ( j) F ( j)

Real

F ( j) F * ( j)
F ( j) F * ( j)

f (t ) f (t )
F ( j) F ( j)

Real

F ( j) F * ( j)
F ( j) F * ( j)

Example:
F [ f (t )] F ( j)
Sol)

F [ f (t ) cos 0t ] ?

1
f (t ) cos 0t f (t )(e j0t e j0t )
2
1
1
j 0 t
F [ f (t ) cos 0t ] F [ f (t )e ] F [ f (t )e j0t ]
2
2
1
1
F [ j ( 0 )] F [ j ( 0 )]
2
2

Example:
1

d/2

f(t)=wd(t)cos0t

wd(t)
t

d/2

d/2

d/2

2 d
Wd ( j) F [ wd (t )] e dt sin

d / 2
2
d
d
sin ( 0 ) sin ( 0 )
2
2

F ( j) F [ wd (t ) cos 0t ]
0
0
d /2

jt

1.5
1.5

d=2
d=2
=5

0 =5
0

1
1

F(j)
F(j)

Example:

0.5
0.5
0
0

-0.5
-0.5 -60
-60

d/2

-40
-40

-20
-20

wd(t)
t

d/2

d/2

0
0

20
20

40
40

60
60

f(t)=wd(t)cos0t

d/2

2 d
Wd ( j) F [ wd (t )] e dt sin

d / 2
2
d
d
sin ( 0 ) sin ( 0 )
2
2

F ( j) F [ wd (t ) cos 0t ]
0
0
d /2

jt

Example:
sin at
f (t )
t
Sol)

wd(t)
t

d/2

d/2

F ( j) ?

2 d
Wd ( j) sin

2
2 td
F [Wd ( jt )] F sin
2wd ()
2
t
0 | a |
sin at
F [ f (t )] F
w2 a ()

t
1 | a |

Fourier Transform of f(t)


ff ((tt))

ff ((tt)) 00

FF jj
and
and tlim
lim

FF

ff ''((tt))

jj
FF(( jj
))
FF

Pf) F [ f ' (t )] f ' (t )e jt dt

f (t )e

j t

jF ( j)

j f (t )e jt dt

Fourier Transform of f (t)


(n)

ff ((tt))

ff ((tt)) 00

FF jj
and
and tlim
lim

FF

nn
ff ((tt))

(
j

)
( j) FF(( jj
))
((nn))

FF

e
v
e
l
v
e
l
s
e
r
s
u
r
o
u
y
o
y
y
b
y
b
d
e
d
v
e
o
v
PPrro

Fourier Transform of f (t)


(n)

ff ((tt))

ff ((tt)) 00

FF jj
and
and tlim
lim

FF

nn
ff ((tt))

(
j

)
( j) FF(( jj
))
((nn))

FF

e
v
e
l
v
e
l
s
e
r
s
u
r
o
u
y
o
y
y
b
y
b
d
e
d
v
e
o
v
PPrro

Fourier Transform of Integral

ff((tt))

FF jj and
and ff((tt))dt
dt FF00 00
FF

tt
11

FF ff((xx))dx
FF jj
dx

jj

Let (t )

f ( x)dx

lim (t ) 0
t

F [' (t )] F [ f (t )] F ( j) j ( j)
1
( j)
F ( j)
j

The Derivative of Fourier Transform

dF
jj

dF
FF[[jtf

jtf ((tt)])]

dd
FF

Pf)

F ( j) f (t )e jt dt

dF ( j) d
j t
jt

f (t )e dt f (t )
e dt

d
d

[ jtf (t )]e jt dt F [ jtf (t )]

Continuous-Time
Fourier Transform
Convolution

Basic Concept
fi(t)

Linear
Linear System
System

fi(t)=a1 fi1(t) + a2 fi2(t)

fo(t)=L[fi(t)]

fo(t)=L[a1 fi1(t) + a2 fi2(t)]

A linear system satisfies fo(t) = a1L[fi1(t)] + a2L[fi2(t)]

= a1fo1(t) + a2fo2(t)

Basic Concept
fi(t)

Time
Time Invariant
Invariant
System
System

fi(t +t0)
fi(t t0)
fi(t)
fi(t+t0)

fi(tt0)

t
t
t

fo(t)
fo(t + t0)
fo(t t0)
fo(t)
fo(t+t0)
fo(tt0)

t
t
t

Basic Concept
fi(t)

Causal
Causal
System
System

fo(t)

A causal system satisfies


fi(t) = 0 for t < t0
fo(t) = 0 for t < t0

Which
Which of
of the
the following
following
systems
systems are
are causal?
causal?

Basic Concept
fi(t)

fi(t)
t0
t0
t0

fi(t)
fi(t)

fo(t)

Causal
Causal
System
System
t
t
t

fo(t)
t0
t0
t0

fo(t)
fo(t)

t
t
t

Unit Impulse Response


(t)

h(t)=L[(t)]

LTI
LTI
System
System

f(t)

L[f(t)]=?

Facts: f ()(t )d f (t )()d


L[ f (t )] L

f (t )

f ()(t )d f () L[(t )]d

f ()h(t )d

Convolution

Unit Impulse Response


(t)

h(t)=L[(t)]

LTI
LTI
System
System

f(t)

L[f(t)]=?

L[ f (t )] f (t ) * h(t )

Facts: f ()(t )d f (t )()d


L[ f (t )] L

f ()(t )d

f ()h(t )d

f (t )

f () L[(t )]d

Convolution

Unit Impulse Response


Impulse
Impulse Response
Response
LTI
LTI System
System

f(t)

h(t)
h(t)

f(t)*h(t)

Convolution Definition
The convolution of two functions f1(t) and
f2(t) is defined as:

f (t ) f1 () f 2 (t )d

f1 (t ) * f 2 (t )

Properties of Convolution
ff11((tt))** ff22((tt)) ff22((tt))** ff11((tt))

f1 (t ) * f 2 (t ) f1 () f 2 (t )d

f1 () f 2 (t )d

f1 (t ) f 2 [t (t )]d (t )
f1 (t ) f 2 ()d

f1 (t ) f 2 ()d f 2 (t ) * f1 (t )

Properties of Convolution
ff11((tt))** ff22((tt)) ff22((tt))** ff11((tt))
f(t)

Impulse
ImpulseResponse
Response
LTI
LTISystem
System

f(t)*h(t)

h(t)
h(t)

h(t)

Impulse
ImpulseResponse
Response
LTI
LTISystem
System

f(t)
f(t)

h(t)*f(t)

Properties of Convolution
[[ ff11((tt))** ff22((tt)])]** ff33((tt)) ff11((tt))**[[ ff22((tt))** ff33((tt)])]

e
s
r
u
o
y
y
b
e
v
o
Pr

The
The following
following two
two
systems
systems are
are identical
identical

Properties of Convolution
[[ ff11((tt))** ff22((tt)])]** ff33((tt)) ff11((tt))**[[ ff22((tt))** ff33((tt)])]

hh11(t)
(t)

hh22(t)
(t)

hh33(t)
(t)

hh22(t)
(t)

hh33(t)
(t)

hh11(t)
(t)

Properties of Convolution
ff ((tt))**((tt)) ff ((tt)) f(t)

f (t ) * (t ) f ()(t )d

f (t )()d

f (t )

(t)
(t)

f(t)

Properties of Convolution
ff ((tt))**((tt)) ff ((tt)) f(t)

(t)
(t)

ff ((tt))**((tt TT)) ff ((tt TT))

f (t ) * (t T ) f ()(t T )d

f (t T )()d

f (t T )

f(t)

Properties of Convolution
ff ((tt))**((tt TT)) ff ((tt TT))
(tT)

f(t T)

f(t)
0

f (t)
0

f (t)
0

System
System function
function (tT)
(tT) serves
serves as
as an
an
ideal
delay
or
a
copier.
ideal
delay
or
a
copier.
Properties of
Convolution
ff ((tt))**((tt TT)) ff ((tt TT))
(tT)

f(t T)

f(t)
0

f (t)
0

f (t)
0

Properties of Convolution
ff11((tt))** ff22((tt))

FF11(( jj
))FF22(( jj
))
FF

e jt dt
f
(

)
f
(
t

)
d

1
2

f1 () f 2 (t )e jt dt d

F [ f1 (t ) * f 2 (t )]

f1 () F2 ( j)e j d

F2 ( j) f1 ()e

d F1 ( j) F2 ( j)

Time Domain

Frequency Domain

convolution

multiplication

Properties of Convolution
ff11((tt))** ff22((tt))

FF11(( jj
))FF22(( jj
))
FF

e jt dt
f
(

)
f
(
t

)
d

1
2

f1 () f 2 (t )e jt dt d

F [ f1 (t ) * f 2 (t )]

f1 () F2 ( j)e j d

F2 ( j) f1 ()e

d F1 ( j) F2 ( j)

Time Domain

Frequency Domain

convolution

multiplication

Properties of Convolution
ff11((tt))** ff22((tt))

FF11(( jj
))FF22(( jj
))
FF

f(t)

F(j)

Impulse
ImpulseResponse
Response
LTI
LTISystem
System

h(t)
h(t)

Impulse
ImpulseResponse
Response
LTI
LTISystem
System

H(j)
H(j)

f(t)*h(t)

F(j)H(j)

Time Domain

Frequency Domain

convolution

multiplication

Properties of Convolution
ff11((tt))** ff22((tt))

FF11(( jj
))FF22(( jj
))
FF

F(j)H1(j)H2(j)H3(j)

F(j)H1(j)

F(j)

H
H11(j)
(j)

H
H22(j)
(j)

H
H33(j)
(j)

F(j)H1(j)H2(j)

Properties of Convolution
ff11((tt))** ff22((tt))

FF11(( jj
))FF22(( jj
))
FF

Fi(j)

H(j)

Fo(j)

p 0 p

An Ideal Low-Pass Filter

Properties of Convolution
ff11((tt))** ff22((tt))

FF11(( jj
))FF22(( jj
))
FF

Fi(j)

H(j)

Fo(j)

p 0 p

An Ideal High-Pass Filter

Properties of Convolution
11
ff11((tt)) ff22((tt))

FF11(( jj))FF22[[ jj((


)])]dd
22
FF

11
ff11((tt)) ff22((tt))

FF11(( jj
))**FF22(( jj
))
22
FF

e
s
r
u
o
y
y
b
Prove

Time Domain
multiplication

Frequency Domain
convolution

Properties of Convolution
11
ff11((tt)) ff22((tt))

FF11(( jj))FF22[[ jj((


)])]dd
22
FF

11
ff11((tt)) ff22((tt))

FF11(( jj
))**FF22(( jj
))
22
FF

e
s
r
u
o
y
y
b
Prove

Continuous-Time
Fourier Transform
Parsevals Theorem

Properties of Convolution
11
[[ff11((tt))ff22((tt)])]dtdt 22FF11((jj))FF22[[jj]]dd

11
ff11((tt))ff22((tt))

FF11((jj))FF22[[jj(()])]dd
22
FF

1
[ f1 (t ) f 2 (t )]e dt 2 F1 ( j) F2 [ j ( )]d

=0
=0

j t

1
[ f1 (t ) f 2 (t )]dt 2 F1 ( j) F2 [ j ()]d

Properties of Convolution
11
[[ff11((tt))ff22((tt)])]dtdt 22FF11((jj))FF22[[jj]]dd

If f1(t) and f2(t) are real functions,


11
**
[
f
(
t
)
f
(
t
)]
dt

F
(
j

)
F
[ f11 (t ) f22 (t )]dt 22 F11 ( j) F22[[jj]]dd

f2(t) real

F2 [ j] F2*[ j]

Parsevals Theorem:
Energy Preserving

11
22
|| ff ((tt))|| dtdt 22 || FF(( jj))|| dd

22

F [ f * (t )] f * (t )e jt dt

f (t )e dt
jt

F * ( j)

| f (t ) | dt f (t ) f * (t )dt
2

1
1
2

F
(
j

)
F
*
[

)]
d

|
F
(
j

)
|
d

2
2

You might also like