Double Sideband Suppressed Carrier Modulation (DSB-SC)

A DSB-SC signals is an AM signals that has the form:

𝑠(𝑡) = 𝐴𝑐 𝑚(𝑡)cos⁡(2𝜋𝑓𝑐 𝑡)

where fc >> w , w is the baseband signal’s bandwidth .

The spectrum of 𝑠(𝑡) is:

𝐴𝑐
𝑆(𝑓) = [𝑀(𝑓 − 𝑓𝑐 ) + 𝑀(𝑓 + 𝑓𝑐 )]
2

Remarks:

1. No impulses are present in the

spectrum at⁡± fc and, hence, no carriers
is transmitted.

2. The transmission B.W of s(t)=2w.

(same of AM).

3. power efficiency
𝑝𝑜𝑤𝑒𝑟⁡𝑖𝑛⁡𝑡ℎ𝑒⁡𝑠𝑖𝑑𝑒⁡𝑏𝑎𝑛𝑑𝑠
= =100%.
𝑡𝑜𝑡𝑎𝑙⁡𝑡𝑟𝑎𝑛𝑠𝑚𝑖𝑡𝑡𝑒𝑑⁡𝑝𝑜𝑤𝑒𝑟

This is a power efficient modulation scheme.

4. Coherent detector is required to extract m(t) from s(t), as we shall demonstrate shortly.

5. Envelope detection cannot be used.

Computer simulation: The next figure shows a DSB-SC signal when 𝑚(𝑡) = cos⁡(2𝜋𝑡)
and 𝑐(𝑡) = cos⁡(2𝜋(10)𝑡). You can easily see that m(t) cannot be recovered using
envelope detection.

1
Demodulation of a DSB-SC signals

A DSB-SC signal is demodulated using what is known as coherent demodulation. This

means that the modulated signal s(t) is multiplied by
a locally generated signal at the receiver which has
the same frequency and phase as the carrier c(t) at
transmitting side.

a. Perfect coherent demodulation.

Let c(t)= Accos2πfct , c\(t)= A c\ cos(2πfct)

Mixing the received signal with the version of the

carrier at the receiving side, we get

v(t)=s(t) A c\ cos2πfct =Ac A c\ m(t)cos22πfct

2
 Ac⁡A⁡c\⁡
= m(t)[1+ cos 2 (2πfct)]
2

Ac⁡A⁡c\⁡ Ac⁡A⁡c\⁡
= m(t)+⁡ m(t) cos 2(2πfct)
2 2

The first term on the RHS is proportional to⁡𝑚(𝑡), while the second term is a DSB signal
modulated on a carrier with frequency 2𝑓𝑐 . The high frequency component can be
eliminated using the LPF. The output is

𝐴𝑐 𝐴′𝑐
𝑦(𝑡) = 𝑚(𝑡)
2
Therefore, 𝑚(𝑡) has been recovered from 𝑠(𝑡) without distortion, i.e., a distortion less
system.

b. Effect of carrier noncoherence on demodulated signal (to be covered in class)

Here we consider two cases.

Case 1: A constant phase difference between c(t) and c\(t)

Let c(t) = Accos2πfct , c\(t) = A c\ cos(2πfct+Ø)

We use the demodulator considered above

v(t)= Acm(t)cos2πfct . A c\ cos(2πfct+Ø)

Ac⁡A⁡c\⁡
= m(t)[ cos (4πfct+Ø)+cos Ø]
2

Ac⁡A⁡c\⁡ Ac⁡A⁡c\⁡
= m(t) cos (4πfct+Ø)+⁡ m(t) cos Ø
2 2

high frequency term low frequency term

The output of the low pass filter is:

Ac⁡A⁡c\⁡
y(t) =⁡ m(t) cos Ø
2

𝜋
For 0 < Ø < 2 , 0 < cos Ø < 1, y(t) suffer from an attenuation due to Ø.

𝜋
However, for Ø= 2 , cos Ø = 0 and y(t) = 0, signal disappears. The disappearance of a
message component at the demodulator output is called quadrature null effect.

3
This highlights the importance of maintaining synchronism between the transmitting and
receiving carrier signals c\(t) and c(t).

Case 2: Constant frequency difference between c(t) and c\(t) (to be done at home
as an exercise)

Let c(t)= Accos2πfct ,C\(t)= A c\ cos(2πfc+∆f)t

In an analysis similar to case a, we get

v(t)= Acm(t)cos2πfct . A c\ cos(2πfc+∆f)t

Ac⁡A⁡c\⁡
= m(t)[ cos (4πfct+2π∆ft)+cos 2π∆ft]
2

After low pass filtering,

Ac⁡A⁡c\⁡
y(t)= m(t) cos 2π∆ft
2

So the demodulated signals appears as if double side band modulated on a carrier with
magnitude ∆f. As can be observed, this is not a distortionless transmission.

Example: Let m(t)=cos2π(1000)t and let ∆f=100Hz (to be done at home as an exercise)

From the analysis in case 2 above,

Ac⁡A⁡c\⁡
y(t)=⁡ cos2π(1000)t cos2π(100)t
2

Ac⁡A⁡c\⁡
= [ cos 2π(1100)t+ cos 2π(900)t]
4

The original message was a signal with a frequency of =1000Hz, while the output
consists of a signal two frequencies at f1=1100Hz and f2=900Hz.

Distortion

Exercise: Use Matlab to plot both m(t) and y(t) and see the distortion caused by the lack
of synchronization between the transmitting and receiving oscillators.

Generation of DSB-SC

a. Product modulator : It multiplies the message signal m(t) with the carrier c(t).
This technique is usually applicable when low power levels are possible and over a
limited carrier frequency range.

4
b .Ring modulator: (to be covered in class)

consider the scheme shown in the

figure.

Let c(t)>> m(t). Here the carrier c(t)

control the behavior of the diodes .

During the positive half cycle of c(t),

c(t) > 0, and D1 and D2 are ON while
D3 and D4 are OFF.

y(t) = m(t)

and the circuit appears like this

During the negative half cycle of

c(t), c(t) < 0 and D3 and D4 are ON
while and D1 and D2 are OFF

y(t) = -m(t)

and the circuit appears like this

5
So m(t) is multiplied by +1 during the +ve half cycle of c(t) and m(t) multiplied by -1
during the -ve half cycle of c(t). Mathematically, y(t) behaves as if multiplied by the
1
switching function gp(t) where gp(t) is the square periodic function with period Tc=𝑓𝑐,
where fc the period of c(t). By expanding gp(t) in a Fourier series, we get

4 4 4
y(t) = m(t)[𝜋 cos2πfct -3𝜋 cos 3(2πfct)+ 5𝜋 cos5(2πfct)]⁡

4 4 4
=m(t)𝜋 cos2πfct -m(t)3𝜋 cos 3(2πfct)+ m(t) 5𝜋 cos5(2πfct)

When y(t) passes through the BPF, the only component that appears at the output is the
desired DSB-SC signal, which is
4
s(t)= m(t)𝜋 cos2πfct

C. Nonlinear characteristic (to be done at home as an excecise)

Consider the scheme shown in the figure

6
Let the non linear characteristic be of the form

y(t)=a0x(t)+a1x3(t)

Let x(t)= Acos2πfct+m(t), (m(t) is the message signals)

y= a0( Acos2πfct+m(t))+a1 (Acos2πfct+m(t))3

=
a0 Acos2πfct+ a0m(t)+ a1A3cos32πfct+ a1 m(t)3+3 a1A2 m(t)cos22πfct

+3A a1cos2πfct

After some algebraic manipulations, a DSB-SC term appear in x(t) along with other
undesirable terms. The band pass filter will admit the desired signal, which is
3(A)2 a1
s(t)= m(t) cos(2)2πfct,
2

Note that the carrier frequency=2fc in this case.

Carriers recovery for coherent demodulation (will not be covered)

We consider briefly two circuits that are used to extract the carriers fc from the incoming
DSB-SC signal. we recall that demodulation of DSB-SC signal requires the availability
of a signal with the same frequency and phase as the carrier c(t) at the transmitter

a. Squaring loop :
The basic elements of squaring loop are shown in the figure below. The incoming signal
is the DSB-SC signal:

s(t)=Acm(t)cos2πfct.

In the figure, we mark five signals that appear at the output of the five blocks. In
summary these signals are:

7
 𝐴𝑐⁡
1- (Acm(t)cos2πfct)2=( 𝑚(𝑡))2(1+cos2wct)
2
𝐴𝑐⁡ 𝐴𝑐⁡
=( 𝑚(𝑡))2+( 𝑚(𝑡))2 (cos2wct)
2 2

Low pass term band pass around 2fc

𝐴𝑐⁡
2- ( 𝑚(𝑡))2 (cos2wct) = Kcos2wct (when the BPF is of narrow B.W)
2
3- Kcos2wct (The limiter removes any variation in amplitude but keeps the
frequency unchanged). The frequency divider produces a signal Kcoswct.
4- (Acm(t)cos2πfct) Kcoswct
= AcKm(t)cos2wct
𝐴𝑐 (𝐾) 𝐴𝑐 (𝐾)
m(t)+⁡ m(t) cos2wct
2 2

(𝐾)
5- 𝐴𝑐 m(t)
2

Therefore, demodulation has been achieved, even though the receiver does not
have a copy of the carrier, but was able to generate its own version of the carrier
via this loop.

8
Costas Loop:

The VCO: is an oscillator that produces a signal whose frequency is proportional to the
voltage ec(t).

When ec(t) = 0, the frequency of the oscillator is called the free running frequency. Let
this frequency = fc (the incoming carrier frequency)

When there is a phase difference Ø, we have

V1(t)=⁡ AcA0m(t) cos (wct) cos(wct+ Ø)

𝐴𝑐𝐴0
V2(t)= m(t)cos Ø
2

V4(t)=Low pass {A0Acm(t)cos(wct)sin(wct+ Ø)}

𝐴𝑐𝐴0
V4(t)=⁡ m(t)sin Ø after LPF 2
2

𝐴𝑐𝐴0
e(t)=⁡ m(t)2 (sinØ)(cosØ)
2

9
 𝐴𝑐𝐴0
= m(t)2 sin2Ø
24

When the B.W of LPF3 is very narrow, the output can be approximated as:

ec(t)= c0sin2Ø

This is the feedback signal that is applied to the VCO. Ideally, when Ø=0, ec(t)=0 and
VCO frequency (and phase) are equal to the frequency of the input signal s(t).

If the phase difference Ø between the incoming s(t) and the VCO output increases, then
ec(t) increases forcing the frequency of the VCO to decrease so that it remains in
synchronism with the input phase. (Recall that the frequency of the VCO decreases if its
input voltage increases; the slope of the VCO characteristics is negative).

10
 Single Sideband Modulation

In this type of modulation, only one of the two sidebands of DSB-SC is retained while
the other sideband is suppressed. This mean that B.W of the SSB signal is one half that
of DSB-SC. The saving in the bandwidth comes at the expense of increasing modulation
complexity.

The time-domain representation of a SSB signal is

s (t)= Acm(t)coswct±Ac 𝑚
̂ (t) sinwct

𝑚
̂ (t) : Hilbert transform of m(t) obtained by passing m(t) through a 90-degree phase
shifter.

-sign: upper sideband is retained

+sign: lower sideband is retained

Generation of SSB: Filtering Method (Frequency Discrimination Method)

A DSB-SC signal X(t)=2Acm(t) coswct is generated first. A band pass filter with
appropriate B.W and center frequency is used to pass the desired side band only

11
The band pass filter must satisfy two conditions:

a. The pass band of the filter must occupy the same frequency range as the desired
sideband.
b. The width of the transition band of the filter separating the pass band and the stop
band must be at least 1% of the center frequency of the filter. i.e., 0.01f0≤∆f.
This is sort of a rule of thumb for realizable filters on the relationship between the
transition band and the center frequency.

12
Two remarks should be considered when generating a SSB signal.

1. Ideal filter do not exist in practice meaning that a complete elimination of the
undesired side band is not possible. The consequence of this is that either part of the
undesired side band is passed or the desired one will be highly attenuated. SSB
modulation is suitable for signals with low frequency components that are not rich in
terms of their power content.

2. The width of the transition band of the filter should be at most twice the lowest
frequency components of the message signal so that a reasonable separation of the two
side band is possible.

13
Generation of SSB Signal: Phase Discrimination Method .

The method is based on the time –domain representation of the SSB signal

s (t)= Acm(t)coswct±Ac 𝑚
̂ (t) sinwct

14
Two- stage generation of SSB signal (will not be covered in class)

When the conditions on the filter cannot be met in a single-stage SSB system, a two-stage
scheme is used instead where less stringent conditions on the filters can be imposed. The
block diagram illustrates this procedure.

m1(t) is the base band signal with a gap in its spectrum extending over (0, fI).

15
 y1(t): is a DSB-SC signal on a carrier frequency f1.

BPF1 selects the upper side band of y1(t). The parameters of the filter are f01 (center
frequency) and the transition band length ∆f1.

We must maintain that

∆f1 ≥ 0.01 f01 and ∆f1 ≤ 2 fI

s1(t) is a single side band signal. The frequency gap of this signal extends over (0, f1+ fL).
The second modulator views this signal as the baseband signal to be modulated on a
carrier with frequency f2.

The second modulator generates a DSB signal. The second BPF with center frequency f02
and transition band ∆f2 selects the upper side band. Again, we maintain that

∆f2 ≥ 0.01 f02 and ∆f2 ≤ 2 (f1+fI)

16
Demodulation of SSB: Time-Domain Analysis

A SSB signal can be demodulated using coherent demodulation (oscillator at the receiver
should have the same frequency and phase as those of transmitter carrier ) as shown in
the figure:

Let the received signal be the upper single sideband

𝑠(𝑡) = ⁡ 𝐴𝐶 𝑚(𝑡)𝑐𝑜𝑠𝜔𝑐 𝑡 − 𝐴𝐶 𝑚
̂(𝑡)𝑠𝑖𝑛𝜔𝑐 𝑡⁡

At the receiver,⁡𝑠(𝑡) is mixed with the carrier signal. The result is

𝑣(𝑡) = ⁡𝑠(𝑡)𝐴𝐶 ⁡́𝑐𝑜𝑠𝜔𝑐 𝑡

⁡⁡⁡⁡⁡⁡⁡⁡⁡= ⁡ 𝐴𝐶 ⁡́⁡[𝐴𝐶 𝑚(𝑡)𝑐𝑜𝑠𝜔𝑐 𝑡 − 𝐴𝐶 𝑚

̂(𝑡)𝑠𝑖𝑛𝜔𝑐 𝑡]𝑐𝑜𝑠𝜔𝑐 𝑡

⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡= ⁡ 𝐴𝐶 𝐴𝐶 ⁡́𝑚(𝑡)𝑐𝑜𝑠2𝜔𝑐 𝑡 − 𝐴𝐶 𝐴𝐶 ⁡́𝑚

̂(𝑡)⁡𝑠𝑖𝑛𝜔𝑐 𝑡⁡𝑐𝑜𝑠𝜔𝑐 𝑡

𝐴𝐶 𝐴𝐶 ⁡́ 𝐴𝐶 𝐴𝐶 ⁡́ 𝐴𝐶 𝐴𝐶 ⁡́
⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡= ⁡ 𝑚(𝑡) + 𝑚(𝑡)𝑐𝑜𝑠2𝜔𝑐 𝑡 − ̂ (𝑡)⁡𝑠𝑖𝑛2𝜔𝑐 𝑡
𝑚
2 2 2

The low pass filter admits only the first terms. The output is:
𝐴𝐶 𝐴𝐶 ⁡́
𝑦(𝑡) = ⁡ 2
𝑚(𝑡)

The following steps demonstrate the demodulation process viewed in the frequency
domain .
𝐴𝐶 ⁡́ 𝐴𝐶 ⁡́
𝑉(𝑓) = ⁡ 2
𝑆(𝑓 − 𝑓𝐶 ) + ⁡ 2
𝑆(𝑓 + 𝑓𝐶 )

𝐴𝐶 ⁡́ 𝐴𝐶 ⁡́
𝑌(𝑓) = ⁡𝐿𝑜𝑤⁡𝑝𝑎𝑠𝑠⁡ {⁡ 2
𝑆(𝑓 − 𝑓𝐶 ) + ⁡ 2
𝑆(𝑓 + 𝑓𝐶 )⁡}

17
Demodulation of SSB signal : Why one side band is enough ?: A frequency-domain
perspective

18
Demodulation of SSB : Coherent Demodulation

a. perfect coherent

when ⁡⁡⁡⁡𝑐(𝑡) = 𝐴𝑐 cos 𝜔𝑐 𝑡⁡⁡, 𝑐́ (𝑡) = 𝐴́𝑐 cos 𝜔𝑐 𝑡 ,

we have perfect coherence and

𝐴𝑐 𝐴́𝑐
𝑦(𝑡) = 𝑚(𝑡)
2

as was derived earlier

b. Constant phase difference

The local oscillator takes the form

𝑐́ (𝑡) = 𝐴́𝑐 cos(𝜔𝑐 𝑡 + ∅);

̂ (𝑡) sin 𝜔𝑐 𝑡]𝐴́𝑐 cos(𝜔𝑐 𝑡 + ∅)

⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡𝑣(𝑡) = [𝐴𝑐 𝑚(𝑡) cos 𝜔𝑐 𝑡 − 𝐴𝑐 𝑚

⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡= 𝐴𝑐 𝐴́𝑐 𝑚(𝑡) cos 𝜔𝑐 𝑡 cos(𝜔𝑐 𝑡 + ∅) − 𝐴𝑐 𝐴́𝑐 𝑚

̂ (𝑡) sin 𝜔𝑐 𝑡 cos(𝜔𝑐 𝑡 + ∅)

𝐴𝑐 𝐴́𝑐 𝐴𝑐 𝐴́𝑐
⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡= 𝑚(𝑡) cos(2𝜔𝑐 𝑡 + ∅) + 𝑚(𝑡) cos(∅)
2 2

𝐴𝑐 𝐴́𝑐 𝐴𝑐 𝐴́𝑐
⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡− 𝑚
̂ (𝑡) cos(2𝜔𝑐 𝑡 + ∅) − 𝑚
̂ (𝑡) sin(∅)
2 2
𝐴𝑐 𝐴́𝑐 𝐴𝑐 𝐴́𝑐
 ⁡𝑦(𝑡) = 2
𝑚(𝑡) cos(∅) − 2
̂ (𝑡) sin(∅)
𝑚

Note that there is a distortion due to the appearance of the Hilbert transform of the
message signal at the output.

c. 𝑐́ (𝑡) = 𝐴́𝑐 cos ⁡2𝜋(𝑓𝑐 + ∆𝑓)𝑡; Constant frequency shift

̂ (𝑡) sin 𝜔𝑐 𝑡]𝐴́𝑐 cos ⁡2𝜋(𝑓𝑐 + ∆𝑓)𝑡

𝑣(𝑡) = [𝐴𝑐 𝑚(𝑡) cos 𝜔𝑐 𝑡 − 𝐴𝑐 𝑚

19
 𝐴𝑐 𝐴́𝑐
= 2
𝑚(𝑡)[cos(2𝜔𝑐 + ∆𝜔)𝑡 + cos 2𝜋∆𝑓𝑡]

𝐴𝑐 𝐴́𝑐
− 2
̂ (𝑡)[sin(2𝜔𝑐
𝑚 + ∆𝜔)𝑡 + sin 2𝜋∆𝑓𝑡]

𝐴𝑐 𝐴́ 𝑐 𝐴𝑐 𝐴́𝑐
 𝑦(𝑡) = 2
𝑚(𝑡) cos 2𝜋∆𝑓𝑡 + 2
̂ (𝑡) sin 2𝜋∆𝑓𝑡
𝑚

Once again we have distortion and m(t) appears as if single sideband modulated on a
carrier frequency = ∆ f.

Example :

Let m( t ) = ⁡𝑐𝑜𝑠2𝜋(1000)𝑡⁡⁡, ∆𝑓 =⁡ 100Hz and let s ( t ) be an upper sideband signal.

Then ,
𝐴𝑐 𝐴́𝑐 𝐴𝑐 𝐴́𝑐
𝑦(𝑡) = ⁡ 𝑐𝑜𝑠2𝜋(1000)𝑡⁡𝑐𝑜𝑠2𝜋(100)𝑡 + 𝑠𝑖𝑛2𝜋(1000)𝑡⁡⁡𝑠𝑖𝑛2𝜋(100)𝑡
2 2

cos(𝑎 − 𝑏) = cos 𝑎 cos 𝑏 + sin 𝑎 sin 𝑏

𝑦(𝑡) = cos 2𝜋(900)𝑡 ≠ cos 2𝜋(1000)𝑡

 Distortion

So, a message component with f = 1000Hz appears as a 900Hz component at the

demodulator output.

Again , distortion results as a result of failing to synchronize the transmitter and receiver
carrier frequencies.

20
Vestigial sideband (VSB) modulation : (Not included and will not be covered)

 This type of modulation finds applications in the transmission of video signal .

 Unlike the audio signal , the video signal is rich in low frequency components
around the zero frequency .
 The B.W of a video signal is about 5MHz.
 If a video signal is to be transmitted using DSB, it requires a 10 MHz B.W ;
too large .
 If a video signal is to be transmitted using SSB (B.W = 5MHz) distortion will
results due to the inability to suppress one of the sidebands completely using
practical filters.
 A compromise between DSB and SSB was proposed called vestigial sideband
modulation .
 Here, a DSB-SC signal is first generated The DSB is applied to a band pass
filter (called a vestigial filter ) that has an asymmetrical frequency response
about ( -+fc).
 The filter allows one of the sidebands to pass almost without attenuation ,
while a trace or a vestige of the second sideband is allowed to pass (most of
the second sideband is attenuated )
 A typical spectral density of an audio and a video signal is shown below.

21
Generation of a VSB : Filtering method

Let Hv(f) be the transfer function of the vestigial filter. We need to find a condition on the
characteristic of the filter such that the demodulated signal at the receiver is proportional
to the message signal. Now we proceed to find such a condition.

𝑥(𝑡) = 𝐴𝑐 𝑚(𝑡) cos 𝜔𝑐 𝑡 ; A DSB-SC signal

𝑆(𝑓) = ⁡𝑋(𝑓)𝐻𝑣 (𝑓) ; The Fourier transform of the filter output.

𝐴𝑐
= {𝑀(𝑓 − 𝑓𝑐 ) + 𝑀(𝑓 + 𝑓𝑐 )}𝐻𝑣 (𝑓) ; VSB signal
2

The objective is to specify a condition on Hv(f) such that V0(t) is an exact replica of m(t) .

𝑦(𝑡) = 𝐴́𝑐 𝑠(𝑡) cos 𝜔𝑐 𝑡

𝐴́𝑐
𝑌(𝑓) = {𝑆(𝑓 − 𝑓𝑐 ) + 𝑆(𝑓 + 𝑓𝑐 )}
2

𝐴𝑐 𝐴́𝑐
= {𝑀(𝑓 − 2𝑓𝑐 ) + 𝑀(𝑓)}𝐻𝑣 (𝑓 − 𝑓𝑐 )
4

𝐴𝑐 𝐴́𝑐
+ {𝑀(𝑓 + 2𝑓𝑐 ) + 𝑀(𝑓)}𝐻𝑣 (𝑓 + 𝑓𝑐 )
4

The LPF will eliminate the high frequency component and retains only the low frequency
terms.

22
 𝐴𝑐 𝐴́𝑐
𝑉𝑜(𝑓) = {𝐻𝑣 (𝑓 − 𝑓𝑐 ) + 𝐻𝑣 (𝑓 + 𝑓𝑐 )}𝑀(𝑓)
4

In order for V0(f) to be proportional to M( f ) , we require that

𝐻𝑣 (𝑓 − 𝑓𝑐 ) + 𝐻𝑣 (𝑓 + 𝑓𝑐 ) = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 = 2𝐻𝑣 (𝑓𝑐 )

When this condition is imposed on the filter, the output becomes

𝐴𝑐 𝐴́𝑐
𝑉𝑜 (𝑓) = 𝐻𝑣 (𝑓𝑐 )𝑀(𝑓)
2

𝐴𝑐 𝐴́𝑐
𝑣𝑜 (𝑡) = 𝐻𝑣 (𝑓𝑐 )𝑚(𝑡)
2

23
 Two remarks :

1. B.W = 𝑊 + 𝑓𝑣 ; 𝑓𝑣 is the size of the vestige .

2. VSB can be demodulated using coherent demodulation .

Generation of VSB: phase discrimination method

The time-domain representation of a VSB signal is

𝑠(𝑡) = ⁡ 𝐴𝑐 𝑚(𝑡) cos 𝜔𝑐 𝑡 ∓ 𝐴𝑐 𝑚𝑠 (𝑡) sin 𝜔𝑐 𝑡

Where ms(t) is the response of a vestigial filter (in the base band spectrum) to the message m(t).
Using the time-domain representation ,

the following scheme may be used

to generate a VSB signal

The – sign means that most of the upper sideband is admitted

+ sign means that most of the lower sideband is admitted

24
The transfer function HS(f) of the low pass filter is related to the band pass characteristic
by:

𝐻𝑠 (𝑓) = 𝐿𝑜𝑤⁡𝑝𝑎𝑠𝑠⁡{𝐻𝑣 (𝑓 + 𝑓𝑐 ) − 𝐻𝑣 (𝑓 − 𝑓𝑐 )}

Coherent Detection of VSB : Time Domain Analysis

Let the received VSB signal be given as:

𝑠(𝑡) = 𝐴𝑐 𝑚(𝑡) cos 𝜔𝑐 𝑡 − 𝐴𝑐 𝑚𝑠 (𝑡) sin 𝜔𝑐 𝑡

This signal is mixed with a version of the

transmitted carrier of the same phase and frequency.

𝑣(𝑡) = 𝑠(𝑡)𝐴́𝑐 cos 2𝜋𝑓𝑐 𝑡

= 𝐴𝑐 𝐴́𝑐 [𝑚(𝑡) cos 𝜔𝑐 𝑡 − 𝑚𝑠 (𝑡) sin 𝜔𝑐 𝑡] cos 𝜔𝑐 𝑡

= 𝐴𝑐 𝐴́𝑐 𝑚(𝑡) cos2 𝜔𝑐 𝑡 − 𝐴𝑐 𝐴́𝑐 𝑚𝑠 (𝑡) sin 𝜔𝑐 𝑡 cos 𝜔𝑐 𝑡

𝐴𝑐 𝐴́𝑐 𝐴𝑐 𝐴́𝑐 𝐴𝑐 𝐴́𝑐

= 𝑚(𝑡) + 𝑚(𝑡) cos 2𝜔𝑐 𝑡 − 𝑚𝑠 (𝑡) sin 2𝜔𝑐 𝑡
2 2 2

25
The low pass filter admits only the low pass component, which is nothing but a scaled
version of the message signal.

𝐴𝑐 𝐴́𝑐
𝑦(𝑡) = 𝑚(𝑡)
2

Envelope Detection of VSB + Carrier :

This type of modulation takes the form :

𝑠(𝑡) = 𝑐𝑎𝑟𝑟𝑖𝑒𝑟 + 𝑉𝑆𝐵

𝑠(𝑡) = 𝐴𝑐 cos 𝜔𝑐 𝑡 + 𝐴𝑐 𝛽𝑚(𝑡) cos 𝜔𝑐 𝑡 ∓ 𝐴𝑐 𝛽𝑚𝑠 (𝑡) sin 𝜔𝑐 𝑡

𝛽 is a scaling factor chosen to minimize envelope distortion. The addition of the carrier is
meant to simplify the demodulation of the video signal in practical TV systems and
avoids the complexity of coherent demodulation. It is also less expensive since a simple
envelope detector, of the type described in demodulating a normal AM signal, can be
used.

𝑠(𝑡) = (𝐴𝑐 + 𝐴𝑐 𝛽𝑚(𝑡)) cos 𝜔𝑐 𝑡 ∓ 𝐴𝑐 𝛽𝑚𝑠 (𝑡) sin 𝜔𝑐 𝑡

2
𝑠(𝑡) = √(𝐴𝑐 + 𝐴𝑐 𝛽𝑚(𝑡)) + 𝐴𝑐 2 𝛽2 𝑚𝑠 2 (𝑡)⁡⁡cos(𝜔𝑐 𝑡 + ⁡∅)

If s( t ) is applied to an envelope detector (which is insensitive to phase variations), the

output is

2
𝑦(𝑡) = √(𝐴𝑐 (1 + 𝛽𝑚(𝑡)) + 𝐴𝑐 2 𝛽2 𝑚𝑠 2 (𝑡)

If 𝐴𝑐 ≫ 𝛽𝑚(𝑡) , then

𝑦(𝑡) ≅ 𝐴𝑐 (1 + 𝛽𝑚(𝑡))

Hence , m(t) can be demodulated, almost without distortion, using simple envelope
detection techniques if the above condition is satisfied .

26
Example: A VSB is generated from the DSB-SC signal 2𝑚(𝑡) cos 𝜔𝑐 𝑡 . M(f) and Hv(f)
are shown below . Find the spectrum of the transmitted signal s(t).

+FC

27
Baseband Signal :

The input signal consists of five frequency components. It is represented as:

⁡𝑚(𝑡) = ⁡0.6 cos 2𝜋𝑓1 𝑡 + 1.6 cos 2𝜋𝑓2 𝑡

+2 cos 2𝜋𝑓3 𝑡 + 1.3 cos 2𝜋𝑓4 𝑡 + 0.8 cos 2𝜋𝑓5 𝑡

Transmitted signal :

The spectrum of the transmitted signal is:

𝑆(𝑓) = 𝐻𝑣 (𝑓)𝑀(𝑓 − 𝑓𝑐 ) + 𝐻𝑣 (𝑓)𝑀(𝑓 + 𝑓𝑐 )

If we perform the multiplication in the frequency domain and take the inverse Fourier
transform, we get the time domain representation of the transmitted signal.

𝑠(𝑡) = ⁡0.4 cos 2𝜋(𝑓𝑐 − 𝑓2 )𝑡 + 0.2 cos 2𝜋(𝑓𝑐 + 𝑓1 )𝑡

+0.4 cos 2𝜋(𝑓𝑐 + 𝑓1 )𝑡 + 1.2 cos 2𝜋(𝑓𝑐 + 𝑓2 )𝑡 + 2 cos 2𝜋(𝑓𝑐 + 𝑓3 )𝑡

+⁡1.3⁡ cos 2𝜋(𝑓𝑐 + 𝑓4 )𝑡 + 0.8⁡ cos 2𝜋(𝑓𝑐 + 𝑓5 )𝑡

28