You are on page 1of 27

1

Lecture 2. Deterministic Signal


Analysis

Fourier Transform
Energy Spectrum, Power Spectrum and Signal Bandwidth
Signal Transmission through a Linear System
Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

Signals in Time Domain


A signal is a set of data or information, which can be
represented as a function of time: s (t )
Deterministic signal is a signal whose physical description
is known completely, either in a mathematical form or a graphical
form.

Signal Energy:

Es | s (t ) |2 dt

Signal Power:

1 T /2
Ps lim | s (t ) |2 dt
T T T / 2

Signal Classification:
Continuous-time vs. Discrete-time signal
Periodic signal vs. Aperiodic signal

Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

Signals in Frequency Domain


Time domain

Frequency domain

S( f )

s (t )

s t e j 2 ft dt

cos(2f0t)
t

Periodic signal with period 1/f0:

f0

1/ f 0

sn f 0 0 s(t )e j 2 nf t dt
0

sn e j 2 nf0t

Lin Dai (City University of Hong Kong)

nf0
EE3008 Principles of Communications

f
Lecture 2

Fourier Transform

Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

Fourier Transform
Given a time domain signal s(t), its Fourier transform is defined as follows.

Fourier transform:

S( f )

s t e j 2 ft dt

The time domain signal s(t) can be expressed by S(f) using an inverse transform.
Inverse Fourier transform:

s (t )

S f e j 2 ft df

(Fourier) spectrum of s(t):

S( f )

Magnitude spectrum of s(t): S ( f )


Lin Dai (City University of Hong Kong)

s (t ) S ( f )

EE3008 Principles of Communications

Lecture 2

Example 1: Spectrum of Unit Impulse (t)

1/

(t) is a unit impulse, which is zero everywhere except


at t=0, and has unit area.

(t )dt 1

and

s (t ) (t )

(t)

t 0
(t )
0 t0

S( f )

j 2 ft

(
t
)
e
dt

area =1

s(t)

S(f)

Lin Dai (City University of Hong Kong)

0
f

EE3008 Principles of Communications

Lecture 2

Example 2: Spectrum of Constant Signal


s (t )
A

s (t ) A
t

S ( f ) A e j 2 ft dt

S(f)

S (0) A e j 2 0t dt A 1dt

A
0

S ( f ) A e j 2 ft dt 0 for f 0

S ( f ) A ( f )
Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

Example 3: Spectrum of Sinusoidal Signal


s(t)

s (t ) cos(2 f 0t )

S( f )

S(f)

1/2
-f0

j 2 ft
cos
2

f
t

e
dt
0

1
2

(e j 2 f0t e j 2 f0t ) e j 2 ft dt

f0

1
2

j 2 ( f f 0 ) t

dt 12

j 2 ( f f 0 ) t
e
dt

12 ( ( f f 0 ) ( f f 0 ))
Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

Example 4: Spectrum of s(t)cos(2f0t)


x(t ) s(t ) cos(2 f 0t )

X(f )

s (t ) cos(2 f 0t ) e j 2 ft dt

1
2

s (t ) 12 (e j 2 f0t e j 2 f0t ) e j 2 ft dt

s (t ) e

j 2 ( f f 0 ) t

dt 12

s (t ) e j 2 ( f f0 )t dt 12 [ S ( f f 0 ) S ( f f 0 )]

S(f)

X(f)

Lin Dai (City University of Hong Kong)

-f0

EE3008 Principles of Communications

f0

f
Lecture 2

10

Example 5: Spectrum of Single Rectangular Pulse


s(t)
A
-/2

/2

A
s (t )
0

/ 2 t / 2
otherwise

S(f)

/2

S( f ) A

A
-2/

-1/

sinc( x)

sin( x)
x

/2

1/

2/

j 2 ft

e j f e j f
dt A
j 2 f

sin( f )
A sinc( f )
f

sinc function is an even, oscillating function with a decreasing magnitude.

It has unit peak at x=0, and zero crossing points at x= non-zero integers.

Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

11

Properties of Fourier Transform

S1 ( f ) S2 ( f )

Linearity

s1 (t ) s2 (t )

S1 ( f ) S2 ( f )

Convolution

S (t )

s ( f )

Duality

s (t )

S ( f ) e j 2 f

Time shift

s (t )e j 2 f0t

S ( f f0 )

Frequency shift

s (t ) cos(2 f 0t )

s (at )

s1 (t ) s2 (t )

1
2

(for any real a 0 )


Lin Dai (City University of Hong Kong)

[ S ( f f 0 ) S ( f f 0 )]

Modulation

1 f
S
a a

Time scale

EE3008 Principles of Communications

Lecture 2

12

Review Examples 2 & 4


s (t ) (t )

Duality: S (t ) s ( f )

S( f ) 1

S( f ) ( f )

s (t ) 1

Modulation:

x(t ) s(t ) cos(2 f 0t )

s1 (t ) cos(2 f 0t )

X ( f ) S ( f ) ( ( f f 0 ) ( f f 0 ))
1
2

12 [ S ( f f 0 ) S ( f f 0 )]

1
2

[ S ( f f 0 ) S ( f f 0 )]

Convolution:

s1 (t ) s2 (t )
Lin Dai (City University of Hong Kong)

S1 ( f ) S2 ( f )

EE3008 Principles of Communications

Lecture 2

13

Example 6: Spectrum of Impulse Train


s (t ) n (t nT0 )

se

j 2 nf 0t

f0

1
T0

T0

2T0

S ( f ) n sn ( f nf 0 )

1
sn
T0

T0

s (t )e

j 2 nf 0t

1/ f 0

dt f 0 0

(t )e j 2 nf t dt f 0
0

S(f) f 0 n ( f nf 0 )

f0

f f0

Lin Dai (City University of Hong Kong)

2f0

EE3008 Principles of Communications

f
Lecture 2

14

Example 7: Spectrum of Periodic Signal

For periodic signal s(t) with period T0, define sT0 (t ) as


s (t )
sT0 (t )
0

T0 / 2 t T0 / 2
otherwise

s (t ) n sT0 (t nT0 ) sT0 (t ) n (t nT0 )

S ( f ) ST0 ( f ) f 0 n ( f nf 0 ) f 0 n ST0 (nf 0 ) ( f nf 0 )

Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

15

Example 8: Spectrum of Sampled Signal


x(t ) s (t ) n (t nT0 )

X ( f ) S ( f ) f 0 n ( f nf 0 ) f 0 n S ( f nf 0 )

s(t)

x(t)
t

S(f)

X(f)

Lin Dai (City University of Hong Kong)

-2f0

-f0

EE3008 Principles of Communications

f0

2f0
Lecture 2

16

Energy Spectrum, Power Spectrum


and Signal Bandwidth

Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

17

Energy-type Signal and Power-type Signal


Energy-type Signal: A signal is an energy-type signal if and
only if its energy is positive and finite.
s(t) is an energy-type signal if and only if 0 Es

| s (t ) |2 dt .

Power-type Signal: A signal is a power-type signal if and only if


its power is positive and finite.
1 T /2
s(t) is a power-type signal if and only if 0 Ps lim | s (t ) |2 dt .
T T T /2

How to determine if a signal is an energy-type signal or


a power-type signal from the frequency domain?
Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

18

Energy and Energy Spectrum


Energy of energy-type signal s(t):

*
j 2 ft

s
(
t
)
S
(
f
)
e
df dt

Es | s (t ) | dt
s (t ) s (t )dt

2
*
*
j 2 ft

dt df S ( f ) S ( f )df
S ( f ) s (t )e
S ( f ) df

U s ( f )df

Parsevals Theorem:

Es | s (t ) | dt

S ( f ) df

Energy spectrum: U s ( f ) | S ( f ) |2
Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

19

Example 9: Energy Spectrum of Single Rectangular


Pulse
s(t)
A
-/2

/2

A
s (t )
0

/ 2 t / 2
otherwise

S(f)
A

Fourier spectrum:

S ( f ) A sinc( f )

-2/

-1/

1/

2/

2/

Us(f)

Energy spectrum:

A2

U s ( f ) | S ( f ) |2 A2 2sinc 2 ( f )
Lin Dai (City University of Hong Kong)

-2/

-1/

1/

EE3008 Principles of Communications

Lecture 2

20

Power and Power Spectrum


Power of power-type signal s(t):

s(t )
sT (t )
0

T / 2 t T / 2
otherwise

1 T /2
1 T /2
Ps lim | s (t ) |2 dt lim T /2 | sT (t ) |2 dt
T T
T T T / 2

1
1
2
2

lim | ST ( f ) | df lim | ST ( f ) | df Gs ( f )df

T T
T T

Power spectrum:

1
| ST ( f ) |2
T T
1 T /2
Gs ( f ) lim s (t ) s* (t )dt
T T T / 2
Gs ( f ) lim

Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

21

Example 10: Power Spectrum of Periodic Signal


For periodic signal s(t) with period T0: s (t ) n sn e j 2 nf0t

Fourier spectrum:
S ( f ) n sn ( f nf 0 )

Power spectrum:
1 T /2
1
Gs ( f ) lim s (t ) s* (t )dt
T T T / 2
T0
Gs ( f ) n sn ( f nf 0 )

s2

S(f)
s
s1 0 s1
-f0 0

T0 / 2

T0 / 2

s2

f0

j 2 nf 0
s (t ) s (t )dt n sn e

Gs ( f )
| s0 |2
| s1 |2 | s1 |2
| s2 |2
| s2 |2
-f0 0

Lin Dai (City University of Hong Kong)

f 0 1/ T0

f0

EE3008 Principles of Communications

f
Lecture 2

22

Signal Bandwidth
Bandwidth of signal s(t): the amount of positive frequency
spectrum that signal s(t) occupies.
S(f)
0

Bs

Effective Bandwidth: x% of the signals power (energy) are


included.

Gs(f)

Bs _ 90%

Bs _ 95%

90% power
95% power
Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

23

Signal Transmission through a


Linear System

Lin Dai (City University of Hong Kong)

EE3008 Principles of Communications

Lecture 2

24

Linear Time Invariant (LTI) System

Linear system: in the time domain, a linear system is described in


terms of its impulse response (the response of the system to a
unit impulse (t)).

Linear Time Invariant (LTI) system: the shape of the impulse


response is the same no matter when the unit impulse (t) is
applied to the system.

s(t)

S(f)

Impulse Response
h(t)
Transfer Function
H(f)

Lin Dai (City University of Hong Kong)

y (t ) s (t ) h(t ) s ( )h(t )d

Y ( f ) S( f ) H ( f )
EE3008 Principles of Communications

Lecture 2

25

Ideal Lowpass System

Transfer Function H(f) of an ideal lowpass system:


H(f)
1
-Bh

Bandwidth
Bh

For a baseband input signal with bandwidth Bs:


S(f)

if Bh Bs

H(f)

x
-Bs

Bh

Bs

1
-Bh

Lin Dai (City University of Hong Kong)

Y(f)

Bh

-Bs 0

EE3008 Principles of Communications

Bs

Lecture 2

26

Ideal Bandpass System

Transfer Function H(f) of an ideal bandpass system:


H(f)
1
0

f c 12 Bh fc
f c 12 Bh
Bandwidth
Bh

For a baseband input signal with bandwidth Bs:

S(f)

S(f)

Bs f

fc-Bs fc

If: 1) the center


frequency of S(f)
Y(f)
is shifted to fc
2) Bh 2Bs

H(f)

x
-Bs

fc+Bs f

Lin Dai (City University of Hong Kong)

=
0

f c 12 Bh

fc

f c 12 Bh

EE3008 Principles of Communications

fc-Bs fc fc+Bs f
Lecture 2

27

Baseband Channel and Bandpass Channel


Baseband channel

Bandpass channel

A baseband channel
efficiently passes frequency
components from dc (zero)
to the cutoff frequency Bh Hz.

A bandpass channel efficiently


passes frequency components
within a certain band, say,
between f c 12 Bh and f c 12 Bh Hz.

Examples: coaxial cable

Examples: EM wave, fibre

In this course, a baseband channel and a bandpass channel are modeled as an


ideal low-pass LTI system and an ideal bandpass LTI system, respectively.
H(f)
H(f)
1

1
-Bh

Bh

Lin Dai (City University of Hong Kong)

f c 12 Bh

fc

EE3008 Principles of Communications

f c 12 Bh

Lecture 2