0% found this document useful (0 votes)
19 views

Chap 3

The document discusses spectrum representation of signals. It describes how any signal can be represented as a sum of sinusoids and defines the spectrum as a plot of the signal components. It also discusses the spectra of periodic, multiplicative and modulated signals. The Fourier series is introduced as a way to obtain the frequency components of continuous periodic signals.

Uploaded by

김민성
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
19 views

Chap 3

The document discusses spectrum representation of signals. It describes how any signal can be represented as a sum of sinusoids and defines the spectrum as a plot of the signal components. It also discusses the spectra of periodic, multiplicative and modulated signals. The Fourier series is introduced as a way to obtain the frequency components of continuous periodic signals.

Uploaded by

김민성
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 17

7

Chap. 3: Spectrum Representation

• Spectrum : A graphical representation of the frequency content of a signal.

• We will show that



N
x(t) = A0 + Ak cos(2πfk t + ϕk )
k=1

N
= X0 + Re{ Xk · ej2πfk t }, where Xk = Ak · ejϕk
k=1
⇒ (Complicated sinusoidal waveforms can be constructed with a sum of
sinusoidal signals.)
• We will construct the spectrum of x(t)

3.1 Spectrum of sinusoidal signals

• A continuous signal:

N
x(t) = A0 + Ak cos(2πfk t + ϕk )
k=1
∑N
Ak jϕk j2πfk t Ak −jϕk −j2πfk t
= A0 + { e ·e + e ·e }
k=1
2 2
∑N
1 1
, X0 + { Xk · ej2πfk t + Xk∗ · e−j2πfk t }
k=1
2 2

where Xk = Ak · ejϕk

• Set of frequency components :


1 1 1 1
{(fk , Xk )} = {(0, X0 ), (f1 , X1 ), (−f1 , X1∗ ), (f2 , X2 ), (−f2 , X2∗ ),
2 2 2 2
1 1 ∗
. . . , ( XN , fN ), ( XN , −fN )}
2 2
8

• Spectrum : Plot of {(Xk , fk )}

Fig. 1. Plot of Spectrum {(Xk , fk )}

• Exercise: x(t) = 10 + 14 cos(200πt − π/3) + 8 cos(500πt + π/2)


9

• Example : Brain Signals

Fig. 2. Brain signals

Fig. 3. Spectrums of brain signals

⇒ It is much easier to analyze brain signals with spectrum.


10

3.2 Spectrums of multiplicative signals

• Many signals are created by multiplications of sinusoid signals.


- Audio effect such as combat sounds
- Music sounds
- Amplitude modulation
• Multiplication of sinusoids
- Multiplication of sinusoids can be decomposed into sums of sinusoids.

Ex: x(t) = sin(10πt) · cos(πt)


1 j10πt 1 1
= {e − e−j10πt } · {ejπt + e−jπt } = {ej11πt + e−j9πt − ej9πt − e−j11πt }
2j 2 4j
1 1
= {(ej11πt − e−j11πt ) + (ej9πt − e−j9πt )} = {2j · sin(11πt) + 2j sin(9πt)}
4j 4j
( )
1 1 1 π 1 π
= sin(11πt) + sin(9πt) = cos(11πt − ) + cos(9πt − )
2 2 2 2 2 2

• Spectrum of Beat signals

x(t) = cos(2πf1 t) + cos(2πf2 t) = Re{ej2πf1 t } + Re{ej2πf2 t } (f1 < f2 )

= Re{ej2πf1 t + ej2πf2 t } = Re{ej2π(fc −f△ )t + ej2π(fc +f△ )t }


1 1
where fc = (f1 + f2 ) : center freq., f△ = (f2 − f1 ) : deviation freq.
2 2
= Re{ej2πfc t · (e−j2πf△ t + ej2πf△ t )}

= Re{ej2πfc t · 2 cos(2πf△ t)}

= 2 cos(2πf△ t) · cos(2πfc t)

Fig. 4. Spectrum of beat signal


11

• Waveforms of a Beat signal


Ex) x(t) = 2 cos(2π · (20)t) · cos(2π · (200)t)
⇒ fc = 200 Hz, f△ = 20 Hz
⇒ f1 = 180 Hz, f2 = 220 Hz

Fig. 5. Analysis of Beat Signal

• Amplitude modulation
x(t) = v(t) · cos(2πfc t)
|{z} | {z }
signal to be transmitted carrier signal

Ex) x(t) = (5 + 2 cos(40πt)) · cos(400πt)

Fig. 6. Spectrum of amplitude modulated signal


12

3.3 Periodic Waveforms

• Periodic signals

N
x(t) = A0 + Ak cos(2π · k · f0 t + ϕk ), f0 : Fundamental freq.
k=1
⇒ A periodic signal whose period is To (= 1/fo ) can be represented by the
sum of sinusoids whose frequencies are multiples of the fundamental freq.
f0 .
• Spectrum of periodic signals

N
x(t) = A0 + Ak cos(2πk · f0 · t + ϕk )
k=1
∑N
= X0 + Re{ Xk · ej2πkf0 t }
k=1

where X0 = A0 , Xk = Ak ejϕk = |Xk |ej ∠Xk

Fig. 7. Spectrum of a periodic signal.

⇒ Spectrum of a periodic signal has signal components at only k · fo (k =


0, ±1, ±2, . . . ), i.e., only the integer multiples of the fundamental frequency.
13

• Ex: Normal Brain signal

Fig. 8. Spectrum of brain signal and signals as adding frequency components


14

• Ex : Determine the fundamental freq. of the following spectrum.

13 ′ 7 ′ f′ f′
{f ′ , f , f } = {20, 26, 35} · ⇒ f0 =
10 4 20 20

• Ex : The following signal is NOT periodic (although it seems to be periodic).

√ √
x(t) = 2 cos(20πt) = 23 cos(20π 8t) + 25 cos(20π 27t)
15
16

3.4 Fourier Series (Spectrum of continuous periodic signals)

• Any periodic signal even with finite discontinuities can be represented by a


linear weighted sum of sinusoids of which frequencies are integer multiples
of fundamental frequency.


x(t) = x(t + mTo ) = Ao + Ak cos(2πkf0 t + ϕk )
k=1


= X0 + Re{ Xk ej2πkf0 t }, where To = 1/fo .
k=1

• Fourier Series : A way to obtain Xk for x(t) = x(t + mTo ).


∑N
1
x(t) = X0 + Re{ Xk ej2πfo t }, f0 =
k=1
T0
 ∫ T0

 2 −j2πkf0 t


 T0 0 x(t)e dt : k = 1, 2, 3, . . .
Xk = ∫ T



 1
 x(t)dt :k=0
T0 0


∑ ∫
1
* If we de denote x(t) = Xk e j2πkf0 t
, Xk = x(t)e−j2πkt/T0 dt.
k=−∞
T0 <T0 >
17

• Derivation of F.S

jm T2π t jl T2π t
ejn T0 t · e−jl T0 t dt
2π 2π
<e 0 ·e 0 >=
<T0 >
 0 : m ̸= l
= = T0 · δml
T :m=l
0


jk T2π t
x(t) · e−jk T0 t dt

< x(t) · e 0 >=
∫<T0 >∑
Xl ejl T0 t ) · e−jk T0 dt
2π 2π
= (
<T0 l

∑ ∫
ejl T0 t · e−jk T0 dt
2π 2π
= Xl
l=−∞ <T0 >

∑∞
= Xl · T0 · δlk
l=−∞

 T ·X :l =k
0 l
=
0 : l ̸= k

1
x(t)e−jk T0 t

⇒ Xk =
T0 <T0 >
18

• Ex : Square wave :


1 : 0 ≤ t < T20
x(t) =
 −1 : T0 ≤ t < T
2 0

∫ T0 /2 ∫ T0
2 2
- FS Coefficient : Xk = e−j2πkt/T0 dt + (−1)e−j2πkt/T0 dt
T0 0 T0 T0 /2

T0
2 e−j2πk( 2 )/T0 − e−j2πk(0)/T0
= { }+
T0 −j2πk/T0
T0
(−2) e−j2πk(T0 )/T0 − e−j2πk( 2 )/T0
{ }
T0 −j2πk/T0

e−jπk − 1 e−jπk − e−j2πk 2 − 2e=jπk


= + =
−jπk −jπk jπk


jπk : k = ±1, ±3, ±5, . . .
4
2{1 − (−1)k }
= =
jπk 0 : k = 0, ±2, ±4, ±6, . . .
19

- Spectrum for square wave

- Summing harmonic sinusoids


20

3.5 Fourier Transform

• Spectrum for continuous Non-periodic signal x(t) ̸= x(t + T0 ) .


∫ ∞
Fourier Tr. : X(ω) = x(t)e−jωt dt, ω = 2πf
−∞

↕ F.T
∫ ∞
1
Inverse Fourier Tr. : x(t) = X(ω)ejωt dω
2π −∞

 1 : |t| ≤ T
1
• Ex : Rectangular wave : x(t) =
 0 : |t| ≥ T
1

∫ ∞ ∫ T1
−jωt
Fourier. Tr. : X(ω) = x(t)e dt = e−jωt dt
−∞ −T1

sin(ωT1 )
=2 , 2T1 sinc(ωT1 )
ω

Fig. 9. Spectrum of rectangular wave.

−at
• Ex : Exponential wave : x(t) = e u(t), a>0
21

• Ex : Spectrum of square wave

• Observation
- Spectrum by Fourier series is discrete on freq. domain
- Spectrum by Fourier Transform is continuous on freq. domain
⇒ Discrete spectrum on freq. domain ↔ Periodic on time domain
⇒ Continuous spectrum on freq. domain ↔ Non-Periodic on time domain

- Fourer Series Xk is the discrete version of Fourier Transform X(ω)

1
⇒ Xk = X(ω)|ω=kω0 =k T2π
T0 0
22
23

You might also like