Digital modulation

In digital modulation, modulation is applied to a carrier by changing the

amplitude, frequency or phase of the latter. What characterizes digital
modulation is that the modulating signal (the message) is digital in nature. A
signal resulting from digital modulation is an RF signal. Digital phase
modulation is called phase-shift keying (PSK) or phase-shift modulation. In the
simplest case, the carrier is modulated by a binary signal. We then speak of
'binary phase-shift keying' (BPSK).
Pourquoi la modulation numérique?

Why digital modulation?

Digital modulation is at the heart of the vast majority of modern

telecommunications systems. Several reasons explain the choice of digital
modulation, including its greater capacity and its robustness to noise.

Modern telecommunications systems, in addition to using increasingly complex

digital modulation techniques (ie QPSK, 256QAM, 4096QAM, etc.), also take
advantage of multiplexing techniques (ie TDMA, CDMA, OFDM, etc.) .
Multiplexing allows, among other things, a more efficient use of a transmission
channel by allowing several devices to use it at the same time.
Bandwidth

Regardless of the transmission medium (optical, electrical, electromagnetic),

bandwidth is in demand. Digital modulation, thanks to its modulation
techniques, makes it possible to transmit more information, for the same
bandwidth, than what analog modulation techniques allow.

A more complex digital modulation technique can transmit more bits per
second than a less complex digital modulation technique. There is therefore a
trade-off between the complexities of the system and its capacity. In recent
years, microprocessors, DPS and FPGAs have improved in performance, while
being smaller and cheaper. These technological advances have contributed to
the improvement of systems using digital modulation.
Robustness to noise

An analog modulation is not very robust to noise while a digital modulation can
be exposed to a certain intensity of noise before displaying transmission errors.
This difference is explained by the very nature of digital modulation, that is to
say, by the fact that the digitally modulated signal take one state among a finite
number of states. Thus, the receiver will succeed in re-establishing a digital
signal disturbed by noise (up to a certain noise level). ASK-FSK-PSK

ASK-FSK-PSK.jpg

A carrier is always sinusoidal in shape. The modulation of this carrier consists in

encoding information by changing one or more parameters of the latter so that
it carries the information. The three parameters that can be changed are:

A: the amplitude
f: frequency
φ: phase
y (t) = A sin (2 π ft + φ)

When the amplitude, frequency, or phase of a sinusoidal carrier is changed by

an analog signal, we are talking about AM, FM, and PM modulations,
respectively. On the other hand, when these three parameters are modified by
a digital signal, they are:

Amplitude Shift Keying (ASK) or Amplitude Shift Keying (MDA)

 Frequency Shift Keying (FSK) or Frequency Shift Keying (MDF)
Phase Shift Keying (PSK) or phase shift modulation (MDP)

For example, when a carrier is phase shift modulated by a binary signal, the
carrier can take only two phase values. In short, the carrier can take a finite
number of phases (2) which means that it is a digital modulation.
PSK
Digital phase modulation is called phase-shift keying (PSK) or phase-shift
modulation. In the simplest case, the carrier is modulated by a binary signal.
We then speak of 'binary phase-shift keying' (BPSK). When the modulating
signal has more than one level, for example 4, the modulated carrier has four
possible phases (4PSK). When the modulating signal has 3 bits per symbol (8
levels), the carrier has 8 possible phases (8PSK). As the number of bits per
symbol increases, it becomes difficult for a demodulator to distinguish the
symbols since they are very close on the constellation. PSK modulation only
involves a phase change of the carrier which gives a circular pattern on the
constellation. QAM modulation uses a matrix scheme which allows more bits
per symbol to be transmitted under the same conditions (same noise, same
bandwidth).
I/Q
500px-Unit circle.svg.png

I means in phase and corresponds to a phase of 0 rad. Q means quadrature and

corresponds to a phase of π / 2 rad (90 °).

A sinusoidal signal, when phase shifted by 90 ° becomes a cosine signal. So it's

fair to say that a sine as well as a cosine are quadrature signals.

I / Q data represents the coordinates of a polar system (amplitude and phase of

a carrier). The I / Q representation is equivalent in every way to the polar
representation. It is only the form that changes. [1]
The I / Q representation is useful in telecommunications because it allows to
simply represent the data as it is injected into most QAM modulators. Indeed, it
is technically very difficult to change the phase and amplitude of an RF carrier
while it is simpler to handle two signals, one in phase (I) and the other in
quadrature (Q)

This article titled I / Q Data for Dummies provides excellent visual aids that can
facilitate understanding of spinning vectors as well as I / Q data.
QAM
Schematic diagram of a QAM modulator.

QAM modulation is intimately linked to I / Q data. Truth be told, a QAM

modulator is made up of two ASK modulators using quadrature carriers. A QAM
signal is the signal resulting from the sum of two signals modulated in ASK by
carriers phase shifted by 90 °.
Modulation error
QAM4.png error

A QAM modulator is made up of several electronic circuits which can induce

modulation errors. This means that the symbol transmitted, or received, is not
necessarily in the expected location. The difference between the received
symbol and the reference symbol gives a vector which is the error vector. The
ratio, in dB, of the average signal power to the average error power is called
the Modulation Error Ratio (MER). This measurement is often thought of as the
signal to noise ratio of a digitally modulated signal.

This equation is used to calculate the MER:

MER (d B) = 10 log 10 ⁡(P signal P error) {\ displaystyle \ mathrm {MER (dB)} =

10 \ log _ {10} \ left ({P _ {\ mathrm {signal}} \ over P_ {\ mathrm {error}}} \
right)} {\ displaystyle \ mathrm {MER (dB)} = 10 \ log _ {10} \ left ({P _ {\
mathrm {signal}} \ over P _ {\ mathrm { error}}} \ right)}

A small MER implies a large cloud of points over the constellation. That is to
say, significant modulation errors which is bad!
A large MER implies a small cloud of points over the constellation. That is,
low modulation errors which is good!

For more information on MER, see the presentation by Ron Hranac, Engineer at
Cisco.
Filtering

Filtering plays an important role in the modulation as well as the transmission

of a digitally modulated RF signal. The following document written by Keysight
is an excellent introduction to so-called pulse shapping filters.

Digital Modulation Webinar by Frederic BIS, Signal Analysis Specialist at

Keysight Technologies (formerly Agilent)

Error detection and correction

Error detection and correction are techniques that allow the transmission of
digital data over imperfect transmission media (with noise).
Error detection

In general, error detection is the transmission of additional data allowing the

receiver to determine whether there is an error during transmission or not.
When an error occurs and an error detection code is used, the message will be
rejected and a retransmission will be requested or not (depending on the
communication protocol.
Redundancy

The simplest error detection method is to transmit the same data successively.
In Example 1 of the table above, one byte is transmitted twice and these two
bytes are received without problem. In example 2, a byte is transmitted twice
without the reception having received two different bytes, which indicates a
transmission error.

Another method of error detection is called goal parity. Suppose the goal is the
transmission of seven bits plus the parity bit. The parity bit can be defined as
being equal to zero if the sum of the other bits is even and to one otherwise.
We speak of even parity (3rd column of the table below).
Parity

A second convention can be taken, the parity bit is then defined as being equal
to one if the sum of the other bits is even and to zero otherwise. The result
corresponds to the fourth column of the table below.

We must not confuse parity of a number, and the fact that it is even or odd (in
the mathematical sense of the term). The binary number 00000011 (3 in
decimal number) is odd (not divisible by 2) but of even parity (even number of
bits at 1).
An error is detected when the parity bit of the data received does not
correspond to the number of bits equal to 1. This method has the advantage of
being simple, except that it only allows the detection of an odd number of
errors.

CRC
In telecommunications, computers and some digital devices, a cyclic
redundancy check or CRC (Cyclic Redundancy Check) is a technique for
detecting transmission errors. CRCs are evaluated (sampled) before and after
transmission or transfer, then compared to make sure the data is exactly the
same. The most commonly used CRC calculations are designed so that errors of
certain types can always be detected, such as those due, for example, to
interference during transmission.

CRC functions are found in various software such as those dedicated to backup,
data capture (sampling) as well as in digital signal transmission devices and
devices: DVB, MPEG-2 TS, DAB, etc.
Error correction

While an error detection code helps identify transmission errors, error

correction codes help detect as well as correct errors. The complexity of the
error correcting code will determine the total number of erroneous bits that
can be corrected.

There are several error correction codes, but the most famous are:

Hamming code
Golay code
BCH code
Reed-Solomon code
Reed-solomon

The Reed-Solomon code is used among others for CD, DVB protocol (European
digital television), DOCSIS, ADSL.
The number of bits that can be corrected depends on the complexity of the
code.

Concept (approximation)

Imagine a block of 3 numbers that we want to transmit: 02 09 12

Let's add two numbers of information redundancy.

The first is the sum of the 3 numbers: 02 + 09 + 12 = 23
The second is the weighted sum of the 3 numbers, each one is multiplied by its
rank: 02 × 1 + 09 × 2 + 12 × 3 = 56
At the encoder output, the block to be transmitted is: 02 09 12 23 56

Following a disturbance, the receiver receives: 02 13 12 23 56

From the data received, the decoder calculates:

Its simple sum: 02 + 13 + 12 = 27
Its weighted sum: 02 × 1 + 13 × 2 + 12 × 3 = 64

The difference between the simple sum calculated (27) and that received (23)
indicates the value of the error: 4 (27-23 = 4)
The difference between the calculated weighted sum (64) and that received
(56), itself divided by the value of the error indicates the position where the
error is found: 2 ((64-56) / 4 = 2) .
We must therefore remove 4 from the number of row 2.

The original block is therefore 02 (13-4 = 09) 12 23 56

During a transmission without disturbance, the differences of the simple sums
and the weighted sums are zero.