Formatting and

Baseband Modulation
Lecture 2
Formatting and transmission of
Baseband signals

2
7-bit ASCII code

3
Message, character and symbols

4
Formatting the analog
information
• Sampling Theorem:
The process of sampling popularly uses sample-
and-hold circuit, and the output of sampling
process is Pulse Amplitude Modulated (PAM)
signal.
Nyquist Criteria: i.e.
This allows the signal to get reconstructed
after passing through an LPF.

Nyquist Rate:

5
Impulse sampling

6
Sampling Theorem: fs>2fm and fs
< 2fm

7
Aliasing due to under-sampling
Solution:
1) Use anti-
aliasing
Filter
2) Use
sampling
frequency

8
fs < 2fm i.e. Aliasing

9
Natural Sampling

10
Quantization of signal

a) Signal
b) Natural-
sampled
signal
c) Quantized
signal
d) Sample and
Hold

11
Errors
• Due to quantization process
A) Quantization noise
B) Quantization saturation
C) Timing jitter

• Channel effect
A) Chanel Noise
B) Inter-symbol Interference

12
Quantization error
The quantization error
variance i.e. average
Quantization noise power

Peak power of analog signal

normalized to 1 Ω resistor is

Peak signal power to avg.

quantization noise ration is

SNR is function of
square of L

13
Pulse Code Modulation

14
Uniform and non-uniform
quantization
Statistically, in
speech signal low
amplitude signals have
highest probability
than the high amplitude
signals.
Thus, if it quantized
with uniform quantizer,
the receiver may face
huge quantization
noise.
Hence, non-uniform
quantizers are
preferred.

15
Uniform and non-uniform
quantization

16
Non-uniform Quantization
A Law and µ Law

µ Law
A Law
Segmented Companding

• E.g. for quantization levels of 256 thus 8-bit

codewords,
1) Divide the binary numbers in equal halves i.e.
8 on upper side and 8 on lower side
2) First bit of each codeword represents the
polarity of quantization intervals
3) Next three bits represents the segment number
4) Last four bits indicates quantization interval
within the segment
19
Time division multiplexing

20
Time division multiplexing
a) All the signal sources
are having same data
rate
b) D) Channel A has 4
times more data rate
than other three
sources

21
Self Study

• T1 carrier system i.e. PCM system for 24 voice

channels multiplexing, block diagram and
description

22
Baseband transmission waveform
representation
PCM:
 Line coding-various PCM
codes
Primary categories:
1) Non-return-to zero
(NRZ): used in
magnetic tape
recording
2) Return-to-zero (RZ)
3) Phase encoded: in
optical
communication
NRZ-M (NRZ-Mark
4) Multilevel ) :
binary
Change is indicated by
1
NRZ-S (NRZ-Space) :
Change is indicated by
Bi-ϕ-L: Manchester
0
coding
Bi-ϕ-M: mark
Bi-ϕ-S: Space
Delay Modulation:
Miler Coding
Parameters of line
coding
1) DC component: eliminating dc energy
helps in ac coupling
2) Self-Clocking: inherent synchronization
aids in clock recovery
3) Error detection: error can be detected
without introducing error detection
coding
4) Bandwidth compression: multilevel coding
helps in transmitting more information
per unit bandwidth
5) Differential encoding: the signal
inversion happening due to error pro
conditions, this is helpful in many
applications
Spectral densities of various
PCM waveforms
Bits per PCM word
PCM Word size:
Let, L=no of Quant. Levels= 2^ no of bits needed to
represent the levels
If ‘p’ is part of Vpp and ‘e’ is
quantization error, then

Since quantization error can not

be more than q/2, then

= bits per PCM word

Source coding: Differential Pulse
Code Modulation
d[k]= 𝑚 𝑘 − 𝑚ƴ 𝑞 𝑘

The output will be

𝑑𝑞 𝑘 = d[k]+𝑞 𝑘
Where, 𝑞 𝑘 is
quantization error.
𝑚𝑞 𝑘 is the quantized
version of m[k].

The SNR improvement:

𝑃𝑜𝑤𝑒𝑟 𝑜𝑓 𝐷𝑃𝐶𝑀
𝐺𝑝 = 10𝑙𝑜𝑔10
𝑃𝑜𝑤𝑒𝑟 𝑜𝑓 𝑃𝐶𝑀
In practice SNR
improvement of 5 to 25
dB is recorded in
literature

28
Delta Modulation
DM is special case of
DPCM
𝑚𝑞 [𝑘] = 𝑚𝑞 [𝑘 −1] +𝑑𝑞 [𝑘]

mq (t) cannot follow m(t)

29
Practical implementation of DM
a) DM
Transmitter
b) DM receiver
c) Message
signal and
integrator
output
d) Delta
modulated
pulses
e) Modulation
error

When 𝑚ƴ 𝑞 (𝑡) cannot follow m(t), slope overload condition occur, this can be avoided by choosing step size
correctly.
When input is relatively constant amplitude, the reconstructed signal has variations and is not truly input
signal. This is called Granular noise.
30
Distortion in DM

𝐴 = 𝜋𝑟 2

1 2𝜋𝑓𝐴 Where, fs is sampling frequency , f is

𝑓𝑠 = ≥
𝑇𝑠 ∆ the frequency of input signal, A is it’s
amplitude and Δ is step size
31
Adaptive Delta Modulation
To overcome the slope overload
distortion and granular noise the
step size should be made adaptive to
the variations in the input signal.
If the variations in the input
signal are small the step size
should be small on the other hand if
input varies over a large dynamic
range step size should be large.

32
Multi-level signaling

Eight level
signaling

Two level
signaling
M-ary pulse modulation
• PAM, PPM, PWM (or PDM) are like
amplitude, phase and frequency
modulation.
• PCM need large bandwidth; to reduce
it, one can use Multi-level signaling.
• If data rate, R= bits per second
• Instead of transmitting each pulse
from PCM, partition the data in k-
groups. i.e. R/k symbols per second.
• Then M= 2^k for M-ary PCM
• The reduction in bandwidth comes with
increased power requirements to
indentify the symbols at receiver end