Unit 5 Shashi
Unit 5 Shashi
• Uniform quantizer: A uniform quantizer is that type of quantizer in which the ‘step size’
remains same throughout the input range.
• Non-uniform quantizer: A non-uniform quantizer is that type of quantizer in which the
‘step size’ varies according to the input signal values.
Quantization Electrical representation of
binary digits
Uniform Quantizer:
• There are two types of uniform quantizer as
under:
I. Symmetric quantizer of the midtread type
II. Symmetric quantizer of the midrise type
• The quantizer characteristic can also be
midtread or midrise type. Figure 7(a) shows the
input-output characteristic of a uniform
quantizer of the midtread type, which is so
called because the origin lies in the middle of a
tread of the staircase like graph.
here, q represents the total number of digital levels of a q-level quantizer. For example if
v=4 bits, the total number of levels will be,
q 24 16 levels 7
Each sample is converted to ‘v’ binary bits. i.e., Number of bits per sample=v.
We know that, number of samples per second= f S
Therefore, Number of bits per second is expressed as
(Number of bits per second) = (Number of bits per samples) x (Number of samples per
second)= v bits per sample x f S samples per second. 8
As a matter of fact, the number of bits per second is known as signaling rate of PCM and
is denoted by ‘r’ i.e., Signaling rate in PCM, r=v f S ; where f S 2 f m
Transmission Bandwidth in a PCM System
• Since bandwidth needed for PCM transmission is given by half of the signaling rate.
Transmission Bandwidth in PCM,
1
BW r 9
2
• But, r vf S
1
therefore BW vf s
2 10
vf m
• Here, Vnoise
2
is taken as the mean square value of noise voltage. Since, here noise is defined
by random variable ‘ ’ and PDF f therefore, its mean square value is given as,
Mean square value= E ε 2 =ε 2 =Vnoise
2
22
• We know that the mean square value of a random variable ‘X’ is expressed as,
X E X
2 2
x 2 f X x dx 23
E 2 f d
2
Here 24
Quantization Noise/Error in PCM
• Using equation (19), above equation may be written as
2 3 2
3 3
1 1 1 2 2 1 3
3
E
2 2
d 25
3 2 3 3 3 8 8
2
• Simplifying, we get
2
E 2 26
12
• Now, using equation (22), the mean square value of noise 2voltage would be
2
Vnoise mean square value 27
12
• Also, if load resistance, R= 1 ohm, then the noise power is normalized i.e.,
2
Vnoise 2 12 2
Noise power (normalized) 28
1 1 12
• Hence, finally, we write
Normalized noise power or Quantization noise power
2
or Quantization error (in terms of power)= 29
12
Signal to Quantization Noise Ratio for Linear
Quantization
• We know that in a PCM system for linear quantization the signal to quantization noise
ratio is given as,
S Normalized signal power
= 30
N Normalized noise power
12
S Normalized signal power
=
Therefore, N 2 12 31
We know that the number of bits ‘v’ and quantization levels are related as,
q 2v 32
Let us assume that input x nTs to a linear quantizer has continuous amplitude in the range
, xmax to xmax . Therefore, total amplitude range = x max - -x max = 2 x max 33
Signal to Quantization Noise Ratio for Linear
Quantization
• Now, the step size will be
2 x max
= 34
q
• Here, substituting the value of q from equation (32) in equation (34), we get
2 x max
= 35
2v
• Now substituting this value in equation (31), we get,
S Normalized signal power
= 2 36
N 2 x max 1
2v 12 .
• Let normalized signal power be denoted as ‘P’.
• Then, This is the required relation for
S P 3P
= 2 2 .22 v signal to quantization noise ratio for
N 4 xmax 1 xmax
2v
linear quantization in a PCM system.
2 12
Signal to Quantization Noise Ratio for Linear
Quantization
Hence, signal to quantization noise ratio:
S 3P 2 v
= 2 .2 37
N xmax
• This expression shows that signal to noise power ratio of quantizer increases
exponentially with increasing bits per sample.
Now, if we assume that input x(t) is normalized, i.e.,
xmax 1 38
Am2
Here P and xmax Am
2
Substituting, these values in the equation 3, we get
Am2
3
S 3
22 22 v 22 v 1.5 22 v
N Am 2
Signal to Quantization noise ratio for the
Sinusodial Signal
Expressing signal to noise power ratio in dB, we get
S S
10 log10 1.5 2
2v
dB 10 log10
N N
10 log10 1.5 10 log10 22 v
1.76 2v 10 0.3
1.8 6v
Therefore, we have
S S
dB in PCM 10 log10 dB 1.8 6v For Sinusodial Signal
N N
Questions
Example 1. A Television signal having a bandwidth of 4.2 MHz is transmitted using binary
PCM system. Given that the number of quantization levels is 512. Determine:
1. Code word length
2. Transmission bandwidth
3. Final bit rate
4. Output signal to quantization noise ratio.
Solution: Given that television signal has the maximum frequency 4.2 MHz. So f m 4.2 MHz
1. The quantization level is 512. So 512 2v and v 9 bits . Thus the code word length is
9 bits.
2. Transmission bandwidth vf m 9 4.2 106 Hz 37.8 MHz
3. Final bit rate r vf s 9 2 4.2 106 bits/sec
4. Output signal to quantization noise ratio
S
dB 4.8 6v 4.8 6 9 4.8 54 58.8dB
N
Questions
Example 2: The information in an analog signal voltage waveform is to be transmitted over
a PCM system with an accuracy of ± 0.1% (full scale). The analog voltage waveform has a
bandwidth of 100 Hz and an amplitude range of - 10 to + 10 volts,
i. Find the minimum sampling rate required.
ii. Find the number of bits in each PCM word.
iii. Find minimum bit rate required in the PCM signal.
iv. Find the minimun absolute channel bandwidth required for the transmission of the
PCM signal. (WBTU, Kolkata, Sem. Exam., 2004-05)
Solution: Here an accuracy is given as ± 0.1%, This means that the quantization error must
be 0.1% or the maximum quantization error must be ± 0.1%. Thus,
max 0.1% 0.001
xmax 10 volts
Questions
Example 3: In a binary PCM system, the output signal to quantization noise ratio is to be
held to a minimum value of 40 dB. Determine the number of required levels, and find the
corresponding output signal-to-quantization-noise ratio. (Gate Examination-1997)
Solution: in a binary PCM system, q 2v where v is the number of binary digits. Since, the
output signal to quantization noise ratio is minimum, thus it is clear that the signal is
sinusoidal which has minimum signal-to-quantization-noise ratio in comparison with other.
Then, we have,
S
dB 1.76 10 log10 2 2v
1.76 20 log10 2 v
1.76 6.02v dB
N
Given that minimum signal-to-quantization-noise ratio is 40 dB.
Thus, 40 1.76 6.02v dB
v 6.35 bits 7 bits
Number of levels required is= q 2 2 128
v 7
Maximum quantization error = max
2 2xmax
Also the step size ‘ ’ is expressed as,
q
If x(t) is normalized, its maximum value i.e., xmax 1.
2 2 1
Therefore, we have, ; for v=4
q 16 8
Hence, quantization error is given from equation (1) as,
1 1
max
2 2 8 16
Robust Quantization: Necessity of Non Uniform
Quantization in a PCM System
• Hence, note that here the quantization error is (1/16)th part of the full voltage range. For
simplicity, we assume that full range voltage is 16 volts. Then maximum quantization
error will be 1 volt. However, for the low signal amplitudes like 2 volts, 3 volts etc., the
maximum quantization error of 1 volt is quite high i.e., about 30 to 50%.
• This means that for signal amplitudes which are close to 15 volts, 16 volts etc., the
maximum quantization error (which is same throughout the range) of 1 volt can be
considered to be small.
• In fact, this problem arises because of uniform quantization. Therefore non-uniform
quantization should be used in such cases.
• In other words, we can say that it is desirable that SNR, should remain essentially
constant over a wide range of input power levels. A quantizer that satisfies all these
requirements is known as a Robust Quantizer. Infact, such a robust performance can be
obtained by using a nonuniform quantization.
Necessity of Non Uniform Quantization for
Speech Signal
We know that speech and music signals are characterized by large crest factor. This means
that for such signals the ratio of peak to rms value is quite high.
Peak Value
Crest factor = Very high for speech and music. 1
rms Value
We know that the signal to noise ratio is given by,
3 22 v P
S
2
N
Expressing in decibles, the last expression becomes
S
N
dB 10 log10 3 2 2v
P 3
If we normalized the signal power i.e., if P=1, then the above equation becomes,
S
dB 4.8 6v dB 4
N
Here, power P is defined as
2
Vsignal x2 t
P
R R
Necessity of Non Uniform Quantization for
Speech Signal
2
Vsignal =Mean square value of signal voltage = x 2 t
x2 t
Hence, normalized power will be P [with R=1]
R 5
P x2 t
From equation (1), crest factor is given as,
6
xmax
or crest factor= since P x 2 t
P
3 2 2v
P
P 1
3 22 v P
P 1
9
Necessity of Non Uniform Quantization for
Speech Signal
• This equation illustrates that the signal to noise ratio for large crest factor signal (P << 1)
would be very very less than that of the calculated theoretical value. The theoretical value
is obtained for normalized power (P = 1) by equation (4).
• Important Notes: Therefore, such large crest use factor signals (i.e., speech and music)
should use non-uniform quantization to overcome the problem just discussed. Signal to
noise ratio reduces at low power levels (P << 1) just now we have observed by equation
(9). This means that at low signal levels, signal to noise ratio reduces i.e., noise increases.
• However the quantization noise is directly related to step size. Therefore, at low signal
levels (P <<1), noise can be kept low by keeping step size low.
• This means that at low signal levels, signal to noise ratio can be increased by decreasing
step size ‘’, This means that step size ‘ ’ should be varied according to the signal level
to keep signal to noise ratio at the required value. This is nothing but nonuniform
quantization. The non uniform quantization is achieved through companding.
Nonunifom Quantization
• If the quantizer characteristics is nonlinear and the step size is not constant instead if it is
variable, dependent on the amplitude of input signal then the quantization is known as
non uniform quantization.
• In non-uniform quantization, the step size is reduced with the reduction in signal level.
For weak signals (P << 1), the step size is small, therefore the quantization noise reduces,
to improve the signal to quantization noise ratio for weak signals.
• The step size is thus varied according to the signal level to keep the signal to noise ratio
adequately high. This is non uniform quantization. The non-uniform quantization is
practically achieved through a process called companding.
Companding
• As a matter of fact, companding is non uniform quantization. It is required to be
implemented to improve the signal to quantization noise ratio of weak signals. We know
that the quantization noise is given by
2
Nq
12
• This shows that in the uniform quantization, once the step size is fixed, the quantization
noise power remains constant. However, the signal power is not constant. It is
proportional to the square of signal amplitude.
• Hence signal power will be small for weak signals, but quantization noise power is
constant. Therefore, the signal to quantization noise for the weak signals is very poor.
This will affect the quality of signal.
• The remedy is to use companding. Companding is a term derived from two words i.e.,
compression and expansion as under:
Companding = Compressing + Expanding
Companding
• In pratice, it is difficult to implement the non-uniform quantization because it is not
known in advance about the changes in the signal level.
• Therefore, a particular method is used. The weak signals are amplified and strong signals
are attenuated before applying them to a uniform quantizer. This process is called as
compression and the block that provides it is called as a compressor.
• At the receiver exactly opposite is followed which is called expansion. The circuit used
for providing expansion is called as an expander.
• The compression of signal at the transmitter and expansion at the receiver is combined to
be called companding. The process of companding has been shown in the form of a block
diagram in figure 12.
A x xmax x
for 0 1
Z x 1 log e A xmax
xmax 1 log e A x xmax 1 x
for 1
1 log e A A xmax
The value of companding gain should be as high as possible. The effect of the companding
gain of 33.3 dB has been shown in figure 17. It shows that due to companding, the smallest
step size with companding is smaller than the step size of a uniform quantizer by a factor
equal to Gc = 32.
Companding
Difference between -law and A-law companding
The most important difference between the two types of compressors is that the A-law
compressor has a midriser at the origin whereas the - law compressor has a midtread at
the origin. Thus, the A-law compressor has no zero value.
Application, Advantages and Drawback of PCM
• Some of the applications of PCM may be listed as under:
i. With the advent of fibre optic cables, PCM is used in telephony.
ii. In space communication, space craft transmits signals to earth. Here, the transmitted power
is quite small (i.e., 10 or l5 W) and the distances are very large (i.e., a few million km).
However, due to the high noise immunity, only PCM systems can be used in such
applications.
• Following are the advantages of a PCM system:
i. PCM provides high noise immunity.
ii. Due to digital nature of the signal, we can place repeaters between the transmitter and the receivers.
Infact, the repeaters regenerate the received PCM signal. This can not be possible in analog systems.
Repeaters further reduce the effect of noise.
iii. We can store the PCM signal due to its digital nature.
iv. We can use various coding techniques so that only the desired person can decode the received signal.
• A PCM system has few drawbacks as under:
i. The encoding, decoding and quantizing circuitry of PCM is complex.
ii. PCM requires a large bandwidth as compared to the other systems.
Delta Modulation
Reason to use Delta Modulation
• We have observed in PCM that it transmits all the bits which are used to code a sample.
Hence, signaling rate and transmission channel bandwidth are quite large in PCM. To
overcome this problem, Delta Modulation is used.
Working Principle
• Delta modulation transmits only one bit per sample. Here, the present sample value is
compared with the previous sample value and this result whether the amplitude is
increased or decreased is transmitted.
• Input signal x(t) is approximated to step signal by the delta modulator. This step size is
kept fixed.
• The difference between the input signal x(t) and staircase approximated signal is confined
to two levels, i.e., + and - . Now, if the difference is positive, then approximated signal
is increased by one step, i.e., ‘ ’. If the difference is negative, then approximated signal is
reduced by ‘ ’.
Delta Modulation
• When the step is reduced, 0' is transmitted
and if the step is increased, 1' is transmitted.
Hence, for each sample, only one binary bit is
transmitted. Figure 18 shows the analog
signal x(t) and its staircase approximated
signal by the delta modulator.
Mathematical Expressions
Thus, the principle of delta modulation can be
explained with the help of few equations as
under:
The error between the sampled value of x(t) and
last approximated sample is given as,
Figure 18. Delta modulation waveform
e nTs =x nTs -xˆ nTs
where e nTs = error at present sample
x nTs = sampled signal of x(t).
Delta Modulation
x̂ nTs = last sample approximation of the staircase waveform.
If we assume u nTs as the present sample approximation of staircase output,
then, u n-1 Ts =xˆ nTs =last sample approximation of staircase waveform
Let us define a quantity b nTs in such a way that,
b nTs =Δsgn e nTs
This means that depending on the sign of error e nTs , the sign of step size is decided. In
other words, we can write
if x nTs xˆ nTs
b nTs =
if x nTs < xˆ nTs
Example 5: A delta modulator system is designed to operate at five times the Nyquist rate
for a signal having a bandwidth equal to 3 kHz bandwidth. Calculate the maximum
amplitude of a 2 kHz input sinusoidal signal for which the delta modulator does not have
slope over load. Given that the quantizing step size is 250 mV. Also, derive the formula that
you use. (U.P,SC. I.E.S. Engg. Examination-1999))
Delta Modulation
Solution: In last example, we have derived the relation for slope overload distortion which
will occur if,
Am
2 f m Ts
So, slope overload will not occur if, Am
2 f m Ts
The maximum frequency in the signal is, f m 3 kHz
Nyquist rate = 2 f m 6 kHz
Sampling frequency = 5 times Nyquist rate = f s 30 kHz
1 1
Hence, Sampling interval= Ts = = 3
seconds
fs 30×10
Given the step size = =250 mV
And the frequency of the signal = f m 2 kHz
Substitute all these values in the expression of not occurring slop overload distortion and we
get,
Am 0.6 volts
2 f m Ts
Adaptive Delta Modulation
Reason to use Adaptive Delta Modulation:
• To overcome the quantization errors due to slope overload and granular noise, the step
size ( ) is made adaptive to variations in the input signal x(t). Particularly in the steep
segment of the signal x(t), the step size is increased.
• Also, if the input is varying slowly, the step size is reduced. Then, this method is known
as Adaptive Delta Modulation (ADM). The adaptive delta modulators can take continuous
changes in step size or discrete changes in step size.
Transmitter Part:
• Figure 21 (a) shows the transmitter and 21 (b)
shows receiver of adaptive delta modulator.
• The logic for step size control is added in the
diagram.
• The step size increases or decreases according
to a specified rule depending on one bit
quantizer output.
Figure 21 (a). Transmitter of Adaptive Delta Modulator
Adaptive Delta Modulation
• As an example, if one bit quantizer output is
high (i.e. 1), then step size may be doubled for
next sample. If one bit quantizer output is low,
then step size may be reduced by one step.
• Figure 22 shows the staircase waveforms of
adaptive delta modulator and the sequence of
bits to be transmitted.