Friday, 25 April

2025

DIGITAL CARRIER
MODULATION
Lecture 25 Session 4

Anirban
Decision Regions of Binary FSK
▪ Observation vector
R2 Decision boundary
Message
point
R1
Message point

▪ The receiver decides in favor of s1 if the observation vector r

falls inside region R1. This occurs when r1 > r2

▪ When r1 < r2 , r falls inside region R2 and the receiver decides

in favor of s2
2
 Probability of Error for Binary FSK
▪ Given that s1 is transmitted,
and
▪ Since the condition r1 < r2 corresponds to the receiver making a
decision in favor of symbol s2, the conditional probability of error given
s1 is transmitted is given by

▪ Define a new random variable

▪ Since n1 and n2 are i.i.d with
▪ Thus, n is also Gaussian with

3
 Pe for BFSK (cont’d)
▪ By symmetry

▪ Since the two signals are equally likely to be transmitted,

the average probability of error for coherent binary FSK is

3 dB worse than BPSK

To achieve the same Pe, BFSK needs 3dB more

transmission power than BPSK

4
So now can you take a guess
how the MFSK will look like in the
signal space??

Ohh!! You will just need a M-

dimensional signal space
corresponding to M orthogonal
carrier signals of frequency
𝒇𝒊 = 𝒇𝟏 + (𝒊 − 𝟏)𝜹𝒇
Where 𝜹𝒇 is the frequency difference between two consecutive carriers
And 𝒊 = 𝟏, 𝟐, … . , 𝑴
5
Did you know??
For MFSK, frequency deviation, ∆𝒇 = (𝑴 − 𝟏)𝜹𝒇
𝟏
If 𝜹𝒇 = or half of the sampling frequency then
𝟐𝑻𝒃
it is known as Minimum Shift Keying.

Bandwidth of FSK or MFSK, as per Carson’s rule is

𝟐(∆𝒇 + 𝑩)
𝑴+𝟏 𝟏
For MSK, it evaluates as where 𝑩 = is the
𝑻𝒃 𝑻𝒃
Bandwidth of the message signal.
If the message signal is Gaussian filtered and then MSK is done then it is known
As Gaussian MSK or GMSK. It finds application in GSM Cellular Network or 2G.

If the FSK or MFSK signal maintains a continuous phase during transition from
One frequency to other then it is known as Continuous Phase FSK or CP-FSK.
It is important to maintain the continuous phase to reduce the power
contained in side lobes. MSK is an example of CP-FSK.
SRM University - AP 6
 Binary ASK

▪ Modulation
2𝐸 𝑻𝒃
“1” 𝜑 𝑡 = 𝐴𝑐𝑜𝑠 𝜔𝑐 𝑡 = cos 𝜔𝑐 𝑡 = 𝐸𝜓1 (𝑡) 𝑬=𝑨
𝑇𝑏 𝟐
“0” 𝜑 𝑡 = 0 = 0 𝜓1 (𝑡)

▪ Average energy per bit 𝟐

𝝍𝟏 𝒕 = 𝒄𝒐𝒔 𝝎𝒄 𝒕
𝑻𝒃

0 1 1 0 1 0 0 1
0 1

0 𝜓1 (𝑡)
𝐸

7
SRM University - AP
 Binary ASK
▪ Modulation 0 1 1 0 1 0 0 1

“1”
“0”
▪ Average energy per bit (On-off signaling)

▪ Decision Region Region R2 Region R1

s2 s1

SRM University - AP
Probability of Error for Binary ASK
Region R2 Region R1

s2 s1

9
 Probability of Error for Binary ASK

▪ Average probability of error is

Identical to that of coherent binary FSK

10
 M – ary ASK (Signal Space)
Disadvantage: The power requirement highly increases with increase in M

0 𝟐𝑨𝒄𝒐𝒔 𝝎𝒄 𝒕

𝑨𝒄𝒐𝒔 𝝎𝒄 𝒕 (𝑴 − 𝟏)𝑨𝒄𝒐𝒔 𝝎𝒄 𝒕

𝜓1 (𝑡)
𝟐 𝑬
(𝑴 − 𝟏) 𝑬
𝑬

𝑻𝒃 𝟐
𝑬=𝑨 𝝍𝟏 𝒕 = 𝒄𝒐𝒔 𝝎𝒄 𝒕
𝟐 𝑻𝒃

Each waveform is carrying log 2 𝑀 bits

SRM University - AP 3
Binary PSK
0 1 1 0 1 0 0 1
▪ Modulation

“1”

“0”

▪ , bit duration
▪ : carrier frequency, chosen to be for some fixed
integer or fc  1/ Tb
▪ : transmitted signal energy per bit, i.e.

▪ The pair of signals differ only in a 180-degree phase shift

13
 Signal Space for BPSK
▪ There is one basis function

0 s1
s2

14
Probability of Error for Binary PSK

0 s1
s2

Anirban Friday, 25 April 2025

Probability of Error and the Distance
Between Signals
BPSK BFSK BASK

▪ In general, BASK and BFSK

needs 3 dB more bit
energy to achieve
same BER as BPSK

16
 Probability of Error for BPSK and FSK/ASK
0
10
-1
10
Probability of Bit Error ASK/FSK
-2
10
-3
10
PSK
-4
10
-5
10
3dB
-6
10
-7
10
0 2 4 6 8 10 12 14
Eb/No in Meixia
[dB]Tao @ SJTU 16
Summary of Pe for Different Binary
Modulations
Coherent PSK

Coherent ASK

Coherent FSK

Non-Coherent FSK

DPSK

18
Pe Plots for Different Binary Modulations
0
10
-1
10
ASK/FSK
-2
Probability of Bit Error 10 NC FSK
-3 BPSK(QPSK)
10
-4
10
-5 DPSK
10
-6
10

-7
10
0 2 4 6 8 10 12 14
Eb/No in [dB]
19
M-PSK

▪ Two orthogonal basis function is

needed for MPSK. Each Dot
represents a symbol in the signal
space representation

▪ The symbols maintain a constant

distance from the centre of the signal
space and lie on a circle.

▪ M symbols in MPSK are represented

BPSK 4-PSK or 8-PSK by M different phase angles which are
QPSK uniformly chosen as
2𝜋
𝜃𝑖 = 𝜃1 + 𝑀 (𝑖 − 1) where 𝑖 =
1,2, . . , 𝑀

20
 QPSK and OQPSK

▪ QPSK consists of two BPSK - one

modulated by 𝒄𝒐𝒔(𝝎𝒄 𝒕) and other by ▪ OQPSK is QPSK where 𝑸(𝒕) is shifted
𝒔𝒊𝒏(𝝎𝒄 𝒕) corresponding to the two by one bit period to reduce the abrupt
orthogonal basis function phase change which increases
spectral broadening.

21
 𝝅
-QPSK
𝟒
❖The symbols in this case are
alternately chosen from two
different constellation sets.
❖In the figure alongside, for the
blue I-Q waveforms during the
first symbol duration, the symbol
is chosen from one of the blue
dots shown in the constellation
diagram.
❖For the next orange set of I-Q
waveforms, the symbol is
chosen from one of the orange
dots shown in the constellation
and the process goes on like that
22
 Why are
QAM there so
➢ There are 16 dots
corresponding to 16 symbols
many dots?? that are depicted. Also there
are two orthogonal basis
function as in MPSK
➢ QAM differs from MPSK in the
fact that the symbols need not
be on a circle, i.e. does not
maintain constant distance
from the origin
➢ Advantage of MQAM over
MPSK is it provides better bit
error rate performance for the
same signal to noise ratio.

Where 𝒎𝟏 𝒕 = 𝒂𝒊 𝒑 𝒕 and 𝒎𝟐 𝒕 = 𝒃𝒊 𝒑 𝒕
are the baseband PAM signals where 𝒂𝒊 and
𝒒𝑸𝑨𝑴 𝒕 = 𝒂𝒊 𝒑 𝒕 𝐜𝐨𝐬 𝝎𝒄 𝒕 + 𝒃𝒊 𝒑 𝒕 𝐬𝐢𝐧(𝝎𝒄 𝒕) 𝒃𝒊 determines the position of the symbols
in the signal space

23
 Why are
QAM there so
➢ There are 16 dots
corresponding to 16 symbols
many dots?? that are depicted. Also there
are two orthogonal basis
function as in MPSK
➢ QAM differs from MPSK in the
fact that the symbols need not
be on a circle, i.e. does not
maintain constant distance
from the origin
➢ Advantage of MQAM over
MPSK is it provides better bit
error rate performance for the
same signal to noise ratio.

Where 𝒎𝟏 𝒕 = 𝒂𝒊 𝒑 𝒕 and 𝒎𝟐 𝒕 = 𝒃𝒊 𝒑 𝒕
are the baseband PAM signals where 𝒂𝒊 and
𝒒𝑸𝑨𝑴 𝒕 = 𝒂𝒊 𝒑 𝒕 𝐜𝐨𝐬 𝝎𝒄 𝒕 + 𝒃𝒊 𝒑 𝒕 𝐬𝐢𝐧(𝝎𝒄 𝒕) 𝒃𝒊 determines the position of the symbols
in the signal space

24
25