ECE 3614: Introduction to

Communications Systems
Lecture 14: Introduction to Digital
Communications

Lingjia Liu, Ph.D.

Associate Professor
Electrical and Computer Engineering
Virginia Tech, Blacksburg, VA
 Overview

● To date we have assumed that the communication systems that we

analyze are analog
● Today we will briefly introduce the concept of digital communications.
● Specifically we will examine digital communications of analog signals
that requires
– Sampling
– Quantization
– Waveform encoding
● We will examine sampling today and will examine quantization and
waveform encoding in the next lecture
● Reading
– Section 5.1-5.4
 Communications

o Definition: Communications is the transfer of information at one

time or location to another time or location

Information Destination
Source
Transmitter Channel Receiver or ‘sink’

Can be Converts information Transports

into appropriate Makes best
analog and corrupts guess as to
or digital. waveform signal
Can include analog-to- what the
Almost transmitted
always digital conversion,
modulation and signal was
baseband
signal waveform coding. Can
be baseband or
bandpass
 Transmitting Information

● The basic function of a communication system is to transfer the source

information from source to sink (destination)
● The transmitter converts the message signal to a format suitable for
transmission
● The signal sent can be
– Analog or digital
– Baseband or bandpass
● So far in this class we have considered an analog message which modulates
a sinusoidal carrier (bandpass)
● We will now focus on converting an analog signal to a digital signal for
digital modulation of either a pulse stream (baseband) or a sinusoidal carrier
(bandpass)
 Analog vs. Digital

● Analog Communications
– The message signal can take on an infinite number of possible values
– Directly uses an analog information source as the message to be sent
● Digital Communications
– The message signal must be one of a small number of discrete
messages
– Must convert analog signals into a sequence of discrete messages
– If the number of possible messages is 2, the system is binary and the
messages are termed binary digits or bits.
 Example

● Analog system that transmits a ● An example digital system

voltage value between -2V and transmits either +1V or -1V
+2V.

Analog Message Signal Digital Message Signal

 Baseband vs. Bandpass

● The transmitter’s job is to convert the information source into a

waveform suitable for transmission.
● The resulting transmit signal can either be “baseband” or “bandpass”
– Baseband - frequency content is primarily around DC
§ signal is typically a series of modulated pulses
– Bandpass – frequency content is primarily around some center
frequency fc >> 0
§ signal is typically a modulated sinusoid
 Why Digital Communications?

● Any noise introduces distortion to an analog signal. Since a digital

receiver need only distinguish between a finite number of waveforms it is
possible to recover digital information without corruption a large
percentage of time.
● Many signal processing techniques are available to improve system
performance: source coding, channel (error-correction) coding,
equalization, encryption
● Digital ICs are inexpensive to manufacture. A single chip can be mass
produced at low cost, no mater how complex
● Digital communications allows integration of voice, video, and data on a
single system
● Digital communications systems provide a more flexible tradeoff between
bandwidth efficiency and energy efficiency than analog communications
 Example Revisited

● When noise is added to the ● When noise is added to the

signal, all of the values are still signal, the resulting values are
valid not valid, thus we can correct
them.
How do we eliminate the noise?
 Limitation of Digital

● Analog system can naturally represent all of the message signal

values
● Digital systems cannot represent all possible input values, thus,
if the message is analog then all of the information cannot be
transmitted
● The process of converting an analog input signal to a digital
signal is termed analog-to-digital conversion and is in general a
lossy process
Block Diagram of Digital Communications
System
Analog Analog
Input Output
Signal Signal

Sample Digital D/A Converter

Output
Data
Quantize Source
Digital Decoder
Input Source
Data Encoder Decryption

Encryption Channel
Decoder
Channel
Encoder Equalization

Modulator Channel Demodulator

 Modulation
2
1.5

1.5

1
1

0.5
0.5

0 -0.5

-1
-0.5

Unmodulated pulse stream

-1.5 Unmodulated sinusoid
-2
-1 0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
Time
Time

Baseband signal Bandpass signal

2
1.5

Amplitude modulated 1.5

Amplitude modulated
1
pulse stream 1
sinusoid
0.5
0.5

0 -0.5

-1
-0.5
-1.5

-1 -2
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time Time
 Baseband Communications

● There are multiple baseband communications

techniques that modulate a pulse stream using the
message signal
– Pulse Amplitude Modulation (PAM)
Pulse modulated
– Pulse Width Modulation (PWM) by analog or
digital signal
– Pulse Position Modulation (PPM)
– Pulse Code Modulation (PCM) Requires conversion to
digital signal
prior to modulation

● However, PCM is the most common

 Analog Information

● Regardless of the type of system (analog or digital; bandpass or

baseband) the original information source can be either analog or
digital
● Traditional communication systems focused on the transfer of
analog information
– Examples: voice or video
● If the system uses continuous pulse modulation, the analog
information signal must be sampled (made discrete in time)
● If the system is digital, the analog information signal must be
sampled and quantized (made discrete in time and amplitude)
 Digital Information

● With the rise of the internet, very often the ‘source’ of information
is simply a computer which inherently uses digital information
● Such digital information fits naturally with a digital
communication system
● No analog-to-digital conversion is necessary
● There may be conversion from binary to M-ary information within
the digital communication system
Basic Structure of PAM/PPM/PWM

Input Signal

Makes signal
Sampling discrete in time

Transmitter

Maps samples
Pulse to
waveforms
Modulation

Transmitted Signal
 Basic Structure of PCM

Input Signal

Makes signal
Sampling discrete in time

Transmitter Makes signal

Quantization discrete in
Amplitude and
converts to bits
Map Quant.
Levels to bits & Maps bits to
waveforms
Pulse
Modulation

Transmitted Signal
PAM / PWM / PPM vs PCM
o PAM/PWM/PPM are systems where the information signal is typically
discrete in time but not necessarily in amplitude (thus not truly digital)
o Infinite number of waveforms can be sent
o Useful for time multiplexing multiple signals
o Noise readily degrades information
o Not particularly common
o PAM is the first step in PCM, thus is useful for study of PCM
o PCM are systems where the information must be discrete in time and
amplitude
o Finite number of waveforms can be sent (i.e., digital)
o Requires both sampling and quantization
o Can be made more robust to noise
o Both require sampling, thus we study sampling first
 Digital Representation of Analog
Signals

● Analog signals (e.g. voice, video) are continuous in time and amplitude:

0.8

0.6

0.4

0.2

-0.2

-0.4

-0.6

-0.8

-1
0 2 4 6 8 10
 Digital Representation of Analog
Signals

● Sampling analog signals makes them discrete in time:

0.8

0.6

0.4

0.2

-0.2

-0.4

-0.6

-0.8

-1
0 2 4 6 8 10
 Digital Representation of Analog
Signals

● Quantization of sampled analog signals makes the samples discrete in

amplitude:

2.5

2 The number of
1.5
discrete amplitude
levels is directly
1
related to the
0.5
number of bits we
0
are willing to use to
-0.5 represent each
-1 sample.
-1.5

-2 Thus, we trade-off
-2.5
bit rate and fidelity
0 2 4 6 8 10
 Digital Representation of Analog
Signals

● If done properly, sampling introduces no distortion into the signal

● Quantization does introduce distortion
– There is a tradeoff between distortion and bandwidth requirements
– More bits per sample means less distortion
– Fewer bits per sample means lower bandwidth requirement
● We consider sampling today
● We will discuss quantization in the next lecture
 The Sampling Theorem

● The impulse sampling (also called ideal sampling) of a signal is modeled

as ∞

() ()
ws t = w t ∑ δ t − nTs
! n=−∞
( )
original !##"## $
signal impulse
train
∞
= ∑ w (nT )δ (t − nT )
s s
n=−∞

– The train of impulse functions select sample values at regular intervals.

– How often do we have to sample to retrieve the original information? (i.e.,
how large can Ts be?)
 The Sampling Theorem (cntd.)
¥ ¥
ws (t ) = w(t ) å d (t - nTs ) = å w(nT )d (t - nT )
s s
n = -¥ n = -¥

● The train of impulse functions select sample values at regular intervals.

Using a Fourier Series representation of the impulse train:
¥ 1 ¥ jnwst 2p
å d (t - nTs ) = åe ,w s =
n = -¥ Ts n =-¥ Ts
● Rewriting, we have:
¥
1 jnw st
ws (t ) = w(t ) å e
n =-¥ Ts
 The Sampling Theorem (cntd.)

● Taking the Fourier Transform of signals:

1 ì ¥ jnwst ü
Ws ( f ) = W ( f ) * F í å e ý
Ts în =-¥ þ
∞
1 jnω t
( )
= W f ∗∑F e s
Ts n=−∞
{ }
1 ¥ w
Ws ( f ) = W ( f )* å d ( f - nf s ), f s = s
Ts n =-¥ 2p

¥
1
Ws ( f ) = å W ( f - nf )
Note: This also follows from
s the fact that the Fourier
Ts n =-¥
Transform of an impulse train
is simply an impulse train.
 Sampling Theorem

Original Spectrum

-fo fo f

Sampled Spectrum

-fs -fo fo fs f
 Sampling Theorem

● Let w(t) be a bandlimited signal with Fourier Transform

W ( f ) = 0, for f > B

● w(t) can be perfectly reconstructed from uniformly spaced samples,

provided those samples are taken at a rate f s ³ 2 B
– 2B is called the Nyquist Rate
– If fs<2B, aliasing results.
– If the signal is not strictly bandlimited, then it must be passed through
lowpass filter before sampling to practically limit its bandwidth
 Recovering the Original Signal

● Sampled signal: 1 ¥
Ws ( f ) = å W ( f - nf s )
Ts n =-¥
● Apply lowpass filter to recover original signal

Ws ( f ) Ideal Lowpass Filter Y( f )

w/bandwidth B
 Recovering the Original Signal
Sampled Spectrum Ws(f)

-fs -fo fo fs f

Filter, H(f)

-B B

Filtered Signal Spectrum, W(f)

-fo fo f
 Example

Original Spectrum Time Signal

 Example: System

● Simple sampling and reconstruction

Input Signal

Microphone

Sampling Low pass

fs Filter Output Signal
B = fs/2

Sampled spectrum
 Example: fs = 32kHz

No Aliasing Perfect Reconstruction

 Example: fs = 2kHz

Substantial Aliasing Imperfect Reconstruction

 Practical Sampling Rates

● Speech:
– Telephone quality speech has a bandwidth of 4 kHz
– Most digital telephone systems sample at 8000
samples/sec
● Audio:
– The highest frequency the human ear can hear is
approximately 15 kHz
– CDs sample at rate 44,100 samples/sec
● Video:
– The human eye requires samples at a rate of at least 20
frames/sec to achieve smooth motion
 Summary

● Today we have introduced the concept of digital communications

● Digital communication systems are capable of transmitting
analog information signals but require analog-to-digital
conversion
– Sampling
– Quantization
– Waveform mapping
● We focused on sampling first. Next class we will discuss
quantization and waveform mapping.
 Appendix

Alternate view of the Sampling Theorem

Another View of the Sampling
Theorem
æ f ö
W ( f ) = Ws ( f )P ç ÷
è 2B ø

ì æ f öü
w(t ) = ws (t ) * Á-1 íP ç ÷ý
î è 2 B øþ
= ws (t ) * sinc ( 2Bt )
æ ¥ ö
= ç å w ( nTs ) d ( t - nTs ) ÷ * sinc ( 2Bt )
è n =-¥ ø
¥
= å w ( nT ) sinc ( 2Bt - n2BT )
n =-¥
s s

¥
æ t ö
= å w ( s) ç
nT sinc
T
- n ÷
n =-¥ è s ø
 Time-Domain View of the Sampling
Theorem

Original Signal
Reconstructed Signal

Samples
Weighted Sinc-functions

sin(x)/x is also referred to as the Sampling Function

Ideal Reconstruction

Sinc functions provide ideal

reconstruction of values between
samples