SAMPLING and PCM

Telecommunication I

EE419

Fall 2019

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 1 / 26

Table of Contents

1 Sampling Theorem

2 Signal Reconstruction

3 The Treachery of Aliasing

4 Practical Sampling

5 Pulse Modulation

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 2 / 26

Sampling Theorem

A signal g (t) whose spectrum is band-limited to B Hz can be

reconstructed exactly from its samples taken uniformly at a rate
R > 2B Hz.
Sampling g (t) at a rate of fs Hz can be achieved mathematically by
multiplying g (t) by an impulse train δTs (t)
X
g (t) = g (t) δTs (t) = g (nTs ) δ(t − nTs )
n

∞
1 X
G (f ) = G (f − nfs )
Ts n=−∞

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 3 / 26

Sampling Theorem

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 4 / 26

Signal Reconstruction: The Interpolation Formula

The process of reconstructing g (t) from its samples is also known as

interpolation.
If sample rate 1/Ts is greater than 2B, shifted copies of spectrum do not
overlap, so low pass filtering recovers original signal.
The ideal interpolating filter transfer function is
f
H(f ) = Ts rect( ), h(t) = 2BTs sinc(2πBt)
2B
1
Then if Ts = 2B
h(t) = sinc(2πBt)
∞
X
g (t) = h(t) ∗ g (t) = g (nTs ) sinc(2πB(t − nTs))
n=−∞

The last equation is the interpolation formula.

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 5 / 26

Signal Reconstruction: The Interpolation Formula
Ideal interpolation represents a signal as sum of shifted sincs.

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 6 / 26

Practical Signal Reconstruction
In practice, the signal is sampled at a higher rate than the Nyquist rate.
This yields G (f ) consisting of repetition of G (f ) with a finite band gap
between successive cycles. We can now use a low-pass filter with a gradual
cutoff characteristic instead of an ideal one. From the Paley-Wiener
criterion, it is impossible to realize a filter with zero band. However, by
increasing the sampling rate, the required filter can be closely
approximated with a smaller time delay.

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 7 / 26

The Treachery of Aliasing

The sampling theorem was proved on the assumption that the signal g (t)
is band-limited. However, All practical signals are time-limited which
means they are non-band-limited or have infinite bandwidth. In this case,
the spectral overlap in G (f ) is a constant feature, regardless of fs .

Note the spectra at frequency fs /2 = 1/2Ts Hz. This frequency is called

folding frequency.

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 8 / 26

The Treachery of Aliasing

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 9 / 26

The Antialiasing Filter

The solution to the aliasing problem is to suppress higher frequency before

the signal is sampled. This way we lose only the component beyond the
folding frequency fs /2.
These higher components now cannot re-apper to corrupt components
with frequencies below the folding frequency.
The suppression is accomplished by using a low-pass filter with a steep
cutoff. The filter is called Anti-Aliasing filter

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 10 / 26

The Antialiasing Filter

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 11 / 26

Practical Sampling

In practice, the sampling is obtained by multiplying a signal g (t) by a train

of pulses with finite width.
∞
X
g (t) = g (t) pTs (t) = g (nTs ) p(t − nTs )
n=−∞

Is it possible to reconstruct g (t) from the sampled signal g (t)?

∞
X 2π
pTs = C0 + Cn cos(nωs t + θn ), ωs =
Ts
n=1

∞
X
g (t) = C0 g (t) + Cn g (t)cos(nωs t + θn )
n=1

Clearly, the signal g (t) can be recovered by low-pass filtering of g (t),

provided fs > 2B.
Telecommunication I (EE419) SAMPLING and PCM Fall 2019 12 / 26
Practical Sampling

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 13 / 26

Maximum Information Rate

Two Pieces of Information per Second per Hertz

A maximum of 2B independent pieces of information per second

can be transmitted, errorfree, over a noiseless channel of band-
width B Hz.
The result follows from sampling theorem: assuming no noise, a
channel of bandwidth B Hz can transmit a signal of bandwidth
B Hz errorfree. But a signal of bandwidth B can be reconstructed
from its Nyquist samples, which are at a rate of 2B Hz

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 14 / 26

Pulse Modulation of Signals

In many cases, bandwidth of communication link is much greater than

signal bandwidth.
The signal can be transmitted using short pulses with low duty cycle:
• Pulse amplitude modulation: width fixed, amplitude varies
• Pulse width modulation: position fixed, width varies
• Pulse position modulation: width fixed, position varies
All three methods can be used with time-division multiplexing to carry
multiple signals over a single channel

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 15 / 26

PAM, PWM, PPM: Amplitude, Width, Position

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 16 / 26

Advantages of Digital Communication

• Analog communication (baseband and modulated) is subject to noise.

• Pulse modulations (PAM, PWM, PPM) represent analog signals by
analog variations in pulses and are also sunbject to noise.
• Long distance communication requires repeaters, which amplify signal
and noise. Each link adds noise.
• Digital communication suppresses noise by regenerating signal.
• Digital hardware implementation is flexible and permits the use of
microprocessors, digital switching, and large-scale integrated circuits.
• Digital signal storage is relatively easy and inexpensive. It also has the
ability to search and select information from distant electronic
database.

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 17 / 26

Pulse-Code Modulation PCM

PCM is a method of converting analog signals into digital signals (A/D

conversion).
In PCM, a signal value is represented by a sequence of pulses (digits).
The value of pulses is chosen from L values.
Such a signal is known as L-ary digital signal.
Usually PCM uses only two pulse values, therefore L-ary signal is encoded
to a binary digital signal.
A binary digit is called a bit.

If m bits are used, then m signal values can be represented.

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 18 / 26

Quantization
In PCM, The signal samples are quantized, where Quantization is
rounding off the sample value to one of the closest permissible levels (L).
Thus, the signal is digitized with samples of L values.

The amplitudes of m(t) lie in (−mp ,mp ) range, which is partitioned into L
subintervals.
Note that mp is not a parameter of the signal m(t)
Telecommunication I (EE419) SAMPLING and PCM Fall 2019 19 / 26
PCM Tradeoff
• Signal bandwidth determines minimum sample rate.
• Desired signal fidelity determines precision of reproduced signal
• Signals can be quantized using digital-to-analog converter (DAC)

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 20 / 26

Quantization Error

Uniform quantization with L levels of a signal with peak amplitude mp has

maximum quantization error of ∆ν/
∆ν mp
max error =  = L

and mean square quantizing error q˜

∆ν/2 mp
Z
1
q2 = q 2 dq =
∆ν −∆ν/2 L
mp
The power of quantization error, Nq = L
ˆ = m(t) + q(t)
PCM signal, m(t)
m (t) 
The SNR, Signal-to-Noise Ratio, So
N = Nq = L mm(t)

p

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 21 / 26

Transmission Bandwidth and Quantization Error
What bandwidth is needed to transmit a PCM encoded signal?
Example: suppose that we want a maximum error 0.5% mp for a 3 kHz
signal which is sampled at 33% higher than the Nyquist rate.
∆ν m .
= Lp =  mp ⇒ L =  < 
L = , and m =  bits per sample

At Nyquist sample rate

RN =  x  = Hz

The actual sampling rate

RA =  x . = Hz

we need 8000 x 8 = 64000 bits/sec.

A maximum of two Pieces of Information per Second per Hertz
For binary PCM, one bit per symbol, we need 64000/2 = 32000 Hz.
Telecommunication I (EE419) SAMPLING and PCM Fall 2019 22 / 26
PCM SNR

The signal-to-noise ratio is

average signal power
SNR = average noise power

For uniform quantization noise,

average signal power ≈ amp2
quantization error ≈ 31 (mp /L)2
SNR ≈ cL2 = c22m

where m is the number of bits in the PCM sample, so L = m . c is a

constant.
SNR grows exponentially with the number of bits.
If we measure SNR in dB,
SNRdB =  log (c m ) =  [log (c) + m log  = (α + m)dB
where α =  log c.
Increasing m by one bit improves SNR by 6 dB! One bit quadruples SNR.

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 23 / 26

Logarithmic Units
In communications we often measure ratios using logarithms.
The bel (B) is the log of a ratio. More useful is the decibel (dB):
a a
b in dB is 10 log10 b

Examples: 2 ⇐⇒ 3.01 ≈ 3 dB, 5 ⇐⇒ 4.77 ≈ 5 dB

Why measure in dB?
• Some sensors (human eyes, ears) respond to logarithm of signal
power.
• Many transmission media have attenuation that is exponential in
length. Thus the signal loss in dB is proportional to length.
• Calculating how much power is needed in a communications system
requires a link budget, which is additive in dB.
rcv power (dBm) = xmit power (dBm) + gains (dB) - losses (dB)
• Since dB measures ratio, we must specify a reference value for 0 dB.
• dbW: 0 dB = 1 W
• dBm: 0 dB = 1 mW
Telecommunication I (EE419) SAMPLING and PCM Fall 2019 24 / 26
PCM in the Bell System

Starting in the 1920s, long distance telephone links used frequency division
multiplexing. (FDM requires amplifiers, built using vacuum tubes.)
A cable with bandwidth 3 MHz can support (in principle) 1000 3 kHz voice
channels. But 1000 filters, modulators, and demodulators are needed.
Local exchanges communicated by trunk lines. Each copper pair carried
one voice conversation.
Using PCM, multiple connections could be time division multiplexed.
The Bell System settled on 1.544 Mbit/s (by experimentation).
8000 x (24 x 8 + 1) = 8000 x 193 = 1,544,000

This TDM signal is called digital signal level 1 (DS1).

This T-1 carrier system uses the same copper that was used for voice!
PCM is credited to Bernard Oliver and Claude Shannon (patent 2,801,281,
1946)

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 25 / 26

T1 Carrier System

Telecommunication I (EE419) SAMPLING and PCM Fall 2019 26 / 26