Digital Communication System
Formats
DIGITAL-TO-ANALOG CONVERSION
Digital-to-analog conversion is the process of
changing one of the characteristics of an analog
signal based on the information in digital data.
Have to cover:
Aspects of Digital-to-Analog Conversion
Amplitude Shift Keying
Frequency Shift Keying
Phase Shift Keying
Quadrature Amplitude Modulation
Digital to Analog Conversion
Digital data needs to be carried on an
analog signal.
A carrier signal (frequency fc) performs
the function of transporting the digital
data in an analog waveform.
The analog carrier signal is manipulated
to uniquely identify the digital data
being carried.
Digital-to-analog conversion
Types of digital-to-analog conversion
Note: Bit rate, N, is the number of bits per
second (bps). Baud rate is the number of
signal elements per second (bauds).
In the analog transmission of digital data, the
signal or baud rate is less than
or equal to the bit rate.
S=Nx1/r bauds
Where r is the number of data bits per signal
element.
An analog signal carries 4 bits per signal element. If
1000 signal elements are sent per second, find the bit
rate.
Solution
In this case, r = 4, S = 1000, and N is unknown. We can
find the value of N from
An analog signal has a bit rate of 8000 bps and a baud
rate of 1000 baud. How many data elements are
carried by each signal element? How many signal
elements do we need?
Solution
In this example, S = 1000, N = 8000, and r and L are
unknown. We find first the value of r and then the value
of L.
Amplitude Shift Keying (ASK)
ASK is implemented by changing the
amplitude of a carrier signal to reflect
amplitude levels in the digital signal.
For example: a digital “1” could not affect the
signal, whereas a digital “0” would, by
making it zero.
The line encoding will determine the values of
the analog waveform to reflect the digital
data being carried.
Bandwidth of ASK
The bandwidth B of ASK is proportional
to the signal rate S.
B = (1+d)S
“d” is due to modulation and filtering,
lies between 0 and 1.
Binary amplitude shift keying
Implementation of binary ASK
We have an available bandwidth of 100 kHz which
spans from 200 to 300 kHz. What are the carrier
frequency and the bit rate if we modulated our data by
using ASK with d = 1?
Solution
The middle of the bandwidth is located at 250 kHz. This
means that our carrier frequency can be at fc = 250 kHz.
We can use the formula for bandwidth to find the bit rate
(with d = 1 and r = 1).
In data communications, we normally use full-duplex
links with communication in both directions. We need
to divide the bandwidth into two with two carrier
frequencies, as shown in Figure 5.5. The figure shows
the positions of two carrier frequencies and the
bandwidths. The available bandwidth for each
direction is now 50 kHz, which leaves us with a data
rate of 25 kbps in each direction.
Bandwidth of full-duplex ASK used in Example 5.4
Frequency Shift Keying
The digital data stream changes the
frequency of the carrier signal, fc.
For example, a “1” could be
represented by f1=fc +∆f, and a “0”
could be represented by f2=fc-∆f.
Binary frequency shift keying
Bandwidth of FSK
If the difference between the two
frequencies (f1 and f2) is 2∆f, then the
required BW B will be:
B = (1+d)xS +2∆f
We have an available bandwidth of 100 kHz which
spans from 200 to 300 kHz. What should be the carrier
frequency and the bit rate if we modulated our data by
using FSK with d = 1?
Solution
This problem is similar to Example 5.3, but we are
modulating by using FSK. The midpoint of the band is at
250 kHz. We choose 2Δf to be 50 kHz; this means
Coherent and Non Coherent
In a non-coherent FSK scheme, when
we change from one frequency to the
other, we do not adhere to the current
phase of the signal.
In coherent FSK, the switch from one
frequency signal to the other only
occurs at the same phase in the signal.
Multi level FSK
Similarly to ASK, FSK can use multiple
bits per signal element.
That means we need to provision for
multiple frequencies, each one to
represent a group of data bits.
The bandwidth for FSK can be higher
B = (1+d)xS + (L-1)/2∆f = LxS
Bandwidth of MFSK used in Example 5.6
We need to send data 3 bits at a time at a bit rate of 3
Mbps. The carrier frequency is 10 MHz. Calculate the
number of levels (different frequencies), the baud rate,
and the bandwidth.
Solution
We can have L = 23 = 8. The baud rate is S = 3 Mbps/3 =
1 Mbaud. This means that the carrier frequencies must be
1 MHz apart (2Δf = 1 MHz). The bandwidth is B = 8 ×
1M = 8M. Figure 5.8 shows the allocation of frequencies
and bandwidth.
Bandwidth of MFSK used in Example 5.6
Phase Shift Keyeing
We vary the phase shift of the carrier
signal to represent digital data.
The bandwidth requirement, B is:
B = (1+d)xS
PSK is much more robust than ASK as it
is not that vulnerable to noise, which
changes amplitude of the signal.
Binary phase shift keying
Implementation of BASK
Quadrature PSK
To increase the bit rate, we can code 2 or
more bits onto one signal element.
In QPSK, we parallelize the bit stream so that
every two incoming bits are split up and PSK
a carrier frequency. One carrier frequency is
phase shifted 90o from the other - in
quadrature.
The two PSKed signals are then added to
produce one of 4 signal elements. L = 4 here.
QPSK and its implementation
Find the bandwidth for a signal transmitting at 12
Mbps for QPSK. The value of d = 0.
Solution
For QPSK, 2 bits is carried by one signal element. This
means that r = 2. So the signal rate (baud rate) is S = N ×
(1/r) = 6 Mbaud. With a value of d = 0, we have B = S = 6
MHz.
Constellation Diagrams
A constellation diagram helps us to
define the amplitude and phase of a
signal when we are using two carriers,
one in quadrature of the other.
The X-axis represents the in-phase
carrier and the Y-axis represents
quadrature carrier.
Concept of a constellation diagram
Show the constellation diagrams for an ASK (OOK),
BPSK, and QPSK signals.
Solution
Figure shows the three constellation diagrams.
Note: Quadrature amplitude modulation
is a combination of ASK and PSK.
Constellation diagrams for some QAMs
MODEM components
• Modulator/demodulator
– Basic operations modulation and pulse shaping Tx/Rx filter pair
of demodulation
modulator/demodulator
– Complex notations for modulation and
demodulation map
bits ↔ symbols
– Carrier recovery and timing recovery
equalisation (distorting
map
This lecture: bits ↔ symbols channel)
fs
Recall that to transmit at a rate fs requires at least baseband bandwidth of 2
Can you see why do we want to group several bits into a symbol?
Bits to Symbols
• The bit stream to be transmitted is serial to parallel multiplexed onto a stream of
symbols with q bits per symbol (discrete 2q levels)
• Example for q = 2 bits per symbol (4-ary modulation): symbol period Ts is twice
of bit period Tb
1
bit stream 1 0 0 1 1 1 0 1 0 0 time
0
z(k)
(1,1)
(1,0)
symbol stream
(0,1)
time
(0,0)
• Symbol rate is half of bit rate; symbol stream is then pulse shaped ... (what
happens to required bandwidth?) 41
Mapping to Constellation
Pattern
• It is typical practice to describe a symbol x(k) by a point in constellation diagram,
i.e. its in-phase and quadrature components, x i (k) and x q (k)
• Example for a case of q = 2 bits per symbol (QPSK): xq(k)
(0,1) (0,0)
• From the constellation pattern, the values x i (k) and
x q (k) of symbol x(k) are determined xi(k)
• There is a one-to-one relationship between
symbol
set (constellation diagram) and modulation signal
set (actually transmitted modulated signal) (1,1) (1,0)
(0,1) (0,0)
• In the receiver, the constellation point and x^ (k)
the transmitted symbol value is determined from
therefore
the received signal sample xˆ(k)
(1,1) (1,0)
Phase Shift Keying (PSK)
• In PSK, carrier phase used to carry symbol information, and modulation signal
set:
q
si(t) = A cos(ωct + φi(t)), 0 ≤ t ≤ Ts, 1≤i ≤M =2
where Ts: symbol period, A: constant carrier amplitude, M : number of
symbol points in constellation diagram
• “Phase” carries symbol information, namely to transmit i-th symbol value (point),
signal s(t) = si(t) is sent, note:
s(t) = A cos(ωct+φi(t)) = A cos(φi(t)) · cos(ωct)+ (−A sin(φ (t))) · sin(ωct)
...,, ...,, i
inphse symbol x i (t) quadrature symbol x q (t)
• Recall previously in slide 29, we say transmitted signal is
s(t) = xi(t) cos(ωct) + xq(t) sin(ωct)
Binary Phase Shift Keying (BPSK)
•One bit per symbol, note the mapping from bits to symbols in constellation
diagram, where quadrature branch is not used
S1(t) S2(t)
Q
bit 1 bit 0
I
Tb Tb
• Modulation signal set si(t) = A cos(ωct + φi), i = 1, 2
bit 0 or symbol 1: φ1 = 0
bit 1 or symbol 2: φ2 = π
Phase separation:
π
Quadrature Phase Shift Keying
(QPSK)
• Two bits per symbol with a minimum phase
(0,1) xq(k)
separation of 2π
• A QPSK constellation diagram:
(0,0)
(A “different” one shown in slide 42)
(1,1) xi(k)
• Modulation signal set
si(t) = A cos(ωct + φi), 1≤i ≤4 (1,0)
S1(t) S2(t) S3(t) S4(t)
Ts Ts Ts Ts
bit (0,0): φ1 = 0 bit (0,1): φ2 = bit (1,1): φ3 = π bit (1,0): φ4 =
π/2 3π/2
Amplitude Shift Keying (ASK)
• Pure ASK: carrier amplitude is used to carry symbol information
• An example of 4-ASK with constellation diagram and modulation signal set si(t) =
Ai cos(ωct), 1 ≤ i ≤ 4
Q
10 11 01 00
I
Ts Ts Ts Ts
(1,0): A1 (1,1): A2 (0,1): A3 (0,0): A4
•Note quadrature branch is not used, pure ASK rarely used itself as amplitude
can easily be distorted by channel
Combined ASK / PSK
•PSK and ASK can be combined. Here is an example of 4-ary or 4-PAM (pulse
amplitude modulation) with constellation pattern and transmitted signal
s(t):
x (k) (00) (01) (11) (00) (10)
q
(1,1) (1,0) (0,1) (0,0)
xi(k)
−→ time t
2 amplitude levels and
phase shift of π are
combined to represent 4-ary
symbols
√ 47 √
•Note in M -ary or M -
PAM, quadrature
component is not used, a
Quadrature Amplitude Modulation
(QAM)
• QAM: combines features of PSK and ASK, uses both I and Q components, and is
bandwidth very efficient
xq(k)
• An example of (squared) 16-QAM:
• Note for squared M -QAM,
√ I and
Q branches are both M -ary (of
previous slide) x i (k)
• Depending on the channel quality, 64-QAM, 128-QAM, or
256-QAM
are possible
• Why high-order QAM particularly bandwidth efficient? and what is penalty paid?
48
Gray Mapping
• Gray coding: adjacent constellation points only differ in a single bit (minimum
Hamming distance)
xq(k) (0111) (0110)
Gray code d
(0000) (0001) (0011) (0010)
x^(
k)
(0100) (0101) (0111) (0110)
(1100) (1101) (1111) (1110)
(k) Non−gray cod ed (1111) (0110)
x^(
xi k)
Symobl (0110) was sent but received sample in neighbor
(1000) (1001) (1011) (1010) region due to noise
•If noise or distortions
cause misclassification
49 in
the receiver, Gray
coding can
minimise the bit error rate
Eye Diagram — Perfect
Channel
• By now we have discussed all components of MODEM, and we turn to “channel” again
• We are looking at stacked 2 symbol period intervals of the demodulated signal xˆi(t) in a QPSK
scheme (xˆi(t) is BPSK):
2
1.5
0.5
x (t)
0
i
−0.5
−1
−1.5
−2
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
time /symbol periods
• This is called an eye diagram; ideal sampling of xˆi(k) will sample the crossing points xˆi(t) = ±1
−→ clock/timing recovery (τ ≈ 0.85Ts or tk = kTs + 0.85Ts)
Eye Diagram — Noisy
Channel
• With channel noise at 10 dB SNR, the eye diagram looks
different: 2
1.5
0.5
0
(t)
xi
−0.5
−1
−1.5
−2
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
time /symbol periods
• As long as the sampling points can be clearly determined and the eye is “open”,
x̂i (k) will correctly resemble x i (k)
• At higher noise levels, misclassifications can occur if the eye is “closed”
Eye Diagram — Distorting
Channel
1
• Non-ideal channel with an impulse response c(t) = δ(t) − 2
· δ(t − T s/4), where T s
is the symbol
period: 1
0.8
0.6
0.4
0.2
x (t)
0
i
−0.2
−0.4
−0.6
−0.8
−1
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
time /symbol periods
• The eye diagram is distorted; this intersymbol interference together with
noise effect will make the eye completely closed, leading to misclassification
Intersymbol Interference (ISI)
• Combined impulse response of an ideal pulse shaping filter of regular zero
with ideal channel gc (t) = δ(t) and non-ideal channel g (t) 1δ(t − t /4):
crossings c = δ(t) − 2 s
1
0.8
txrx filter * channel
0.6
0.4
0.2
−0.2
0 1 2 3 4 5 6 7 8
time / symbol periods
• For non-ideal channel, the combined Tx-filter – channel – Rx filter has lost the
property of a Nyquist system, no longer has regular zero crossings at symbol
spacing
Dispersive Channel
• Recall that zero ISI is achieved if combined Tx and Rx filters is a Nyquist system
• But this is only true if the channel is ideal ⇒ GTx(f )Gc(f )GRx(f ) = GTxGRx(f )
• If Gc(f ) is non-ideal, GTx(f )Gc(f )GRx(f ) will not be a Nyquist system; example of
a distorting channel:
Amplitude Spectrum (dB)
transmission
bandwidth
f
• Dispersive channel is caused by: (i) a restricted bandwidth (channel bandwidth is
insufficient for the required transmission rate);
or (ii) multipath distorting
• Equalisation is needed for overcoming this channel distortion (next lecture)
Summary
• Mapping bit stream to symbol stream
• Symbol constellation and modulation signal set
• Modulation schemes: PSK, ASK, PAM (Q-ary)
and QAM
• Gray code
• Eye diagram, effects of noise
• Effects of dispersive channel
M-ary Digital Carrier Mo
► M-ary send many bits in a single symbol
► Many encoding schemes use different features of the pulses
► Amplitude: PAM
► Phase: PSK, QPSK, OQPSK
► Frequency: FSK, MFSK, GMFSK
► Many Frequencies: FDM, OFDM
Based on lecture notes from John Gill
M-ary Signalling
► There are many ways to send multiple bits of information with each
pulse, or symbol.
► Send information by different amplitudes of a common pulse, or by
sending different pulses
�
y(t) = akpk (t − kTb)
k
where ak is chosen from a set of more than two values (i.e., not just
±1). Both k and pk (t) may be complex.
► Some methods use a common pulse p(t) with different amplitudes
► Amplitude shift keying (AKS) where ak is real
► Quadrature amplitude modulation (QAM) where ak is complex
► Phase shift keying (PSK) where ak is complex, and unit magnitude
► Other methods use different waveforms and amplitudes
► Frequency shift keying (FSK)
► Orthogonal frequency domain multiplexing (OFDM)
Amplitude Shift Keying w
One widely used encoding of two bits into four levels is
−3 message bits 00
−1 message bits 01
ak =
+1 message bits 11
+3 message bits 10
This is used in ISDN, a digital home and office network connection
since replaced by DSL.
ASK used in
► Ethernet 100Base-T (PAM-3), 1000Base-T (PAM-5), 10GBase-T
(PAM-16)
► GDDR6X memory (used in the RTX 3090) with PAM-4.
► ATSC Digital TV, (PAM-8)
Constellation Plo
► For all of the methods that use a common pulse p(t) we can
characterize them by plotting the coefficients ak
► In general ak can be complex, so the plot is in the complex domain.
► Some examples so far:
OOK Polar PAM-5
► These are all on the real axis.
Constellation Plo
► Noise blurs the constellation causing decoding errors
OOK Polar PAM-5
Decision Boundary
Binary Phase Shift Keyin
► Binary PSK is the same as polar ASK.
Carrier
Message
Transmited Signal
► Usually use an integer number of cycles of an offset frequency, and
modulate that up to the carrier
► Phase changes at zero crossing to minimize bandwidth
► Detection generally must be synchronous, since the envelope is
M-ary Phase Shift Keying
► Binary phase shift keying is the same as polar signal
► Only two phase shifts are used, ±π
► We can add more symbols by using additional phase
shifts
PSK-8
PSK-2 PSK-4
► The waveforms for M symbols are
pk (t) = cos(2πfct/Tb + 2πk/M )
► This is a coefficient ak = ej2πk/M for the complex pulse ej2πfct/T b.
► Generally need a synchronous receiver
PSK Constellation Examples
These are all used in telephone modems:
Quadrature AM (Q
► Another way to fill in the space is to use linear combinations of sines
and cosines
pk (t) = ak cos(2πfct/Tb) + bk sin(2πfct/Tb)
The cosine term is along the real axis, and the sine term along the
imaginary axis
► The constellations look like
QAM-4 QAM-16
► Here the symbols are spaced rectilinearly and evenly, many other options
► M-bits / symbol
► Common for cable TV, where QAM-64 or QAM-256 are used.
QAM Constellation Mea
QAM-16 Experimental Measurement
►Note the ordering. Adjacent symbols differ by only one
bit. This is called a Gray code.
►Errors due to noise will mostly be between adjacent
symbols
►These will result in single bit errors that are easily
correctable (next class).
Binary Frequency Shift Ke
► Binary FSK uses two frequencies for 1 and
zero.
Carrier
Message
Transmited Signal
► Usually integer numbers of cycles of each offset frequency, so that
they are orthogonal
► Easy to receive, can be done with filters and an envelope detector
(see this week’s lab!). Does not need to be synchronous.
M-ary Frequency Shift Key
► For M-ary FSK we use several frequencies
pk (t) = cos(2π(fc + k∆f )t)
where ∆f is the spacing in frequency.
► Decoding is easy if (∆f )Tb = 1, so that we have integer numbers of
cycles of the cosine (or sine).
This is a modulation index m = 1, meaning the frequencies are
multiples of one cycle.
►These are all orthogonal
►The receiver can use an array of filters, one for each frequency,
followed by enveloped detectors, just as in the binary case.
Minimum Shift Keying
► If we want to get the most bits/symbol across a channel for FSK,
we’d like the frequencies to be as close together as possible
► The minimum duration orthogonal signals are multiples of half a cycle
Tb Tb Tb Tb
1/2 1 3/2 2
These are all orthogonal. It only works because we are
limiting the waveforms to only cosines.
►These FSK frequencies are Minimum Shift Keying,
MSK
►The modulation index is m = 0.5, and all the
frequencies are multiples of that for one
half cycle
►At this point the receiver has to be more
Gaussian Minimum Shift Ke
► FSK can be thought of as a time varying frequency waveform
y(t) = cos(2π(fc + f (t))t)
where f (t) looks like this:
f (t)
k f
t
► The abrupt frequency changes cause sidelobes, and interfere
with adjacent channels
Gaussian Minimum Shift Ke
► To minimize this, the f (t) waveform is first convolved with a Gaussian
waveform to smooth it
fg (t)
k f
t
► If this is used with Minimum Shift Keying, the results is Gaussian
Minimum Shift Keying or GMSK
► This is used in GSM cell telephones, with a modulation index of
m = 0.3, even less than MSK!
► The results is ISI, and mitigation strategies and error correction are
required to make it work.
Orthogonal Frequency Domain M
► So far we have been trying to get as many symbols per second
through the channel, each encoded for as many bits as possible
► Another approach is to go for very large number of frequencies per
symbol, so that the symbol rate can be much lower
► If we have a sequence ck of bits or QAM encoded symbols, we can
modulate (FSK encode) one sample as
dk (t) = ckej2πk∆ft
Each sample ck is transmitted at its own FSK frequency k∆f
► Each frequency we can think of as an independent subchannel
f
BW f
OFDM
► We can send data on all of them at once by adding the signals up
► If we combine all of the samples in an interval (frame) we get
N
�−1
d(t) = ckej2πk∆ft
k=0
► If there are N samples, then there are N output FSK frequencies
► In practice d(t) is sampled, and this operation is computed with an
inverse FFT.
► Very effectively uses the entire available spectrum
► Different channels will have different noise and propagation
properties
► We can equalize each independently, and allocate more or less bits
to each
OFDM Encode
The transmitter looks like this
N Channels
with Symbol
Rate R/(N*M)
Input Bit Stream
Rate R
1 Symbol Rate
R/(N*M) N
Demux M- Frequency
M bits bit Blocks
NxN iFFT Channels
into N D/A
Channels
TB t
NM Bits in Frame TBNM
N Bit Streams QAM-M Encoders M
Rate R/N bits into one
Complex number
OFDM Decoder
The receiver basically just inverts the operations of the transmitter
Output Bit Stream
Rate R
1 Symbol Rate
R/(N*M) N
Frequency M bits
Mux N
Channels
M-bit Block
NxN
A/D FFT TB t
NM Bits in Frame
TBNM
OFDM
► OFDM has many practical advantages
► The many different frequency channels are resistant to channel
variations
► The much lower symbol rate makes timing and pulse shaping much
easier
► The number of frequencies can be anywhere from 64 to 8k or more
► The constellation encoder may be anything from BPSK through
QAM-256
► That can be a lot of bits per symbol!
► OFDM is widely used
► Digital Satelite TV : DVB-T, and others
► WiFi : 802.11a/g/n and WiMAX
► Digital Radio : Digital Radio Modiale (DRM)
► ADSL
