Delta or Differential PCM (DPCM)

• During transmission of any sampled signal, the

information contained in the present & next
sample does not vary a lot as the samples of
signals are highly correlated.
• The adjacent samples almost carry the same
information with a small difference when these
neighbouring samples are encoded by PCM
system, then the resulting encoded signal contains
some redundant information.
• The common bits of the adjacent samples are
known as redundant bits.
• By reducing these redundant bits, the total no. of
bits transmitted can be reduced.
• If the redundancy is eliminated then the overall
bit rate will decrease & no. of bits required to
transmit one sample will also be reduced. This
type of digital pulse modulation technique is
known as DPCM.
• Hence in DPCM only the difference between 2
consecutive samples are quantized & encoded for
transmission.
• The basic principle of DPCM is shown in Fig.-A.
Here, at each sampling time, say k-th sampling
time, transmit the difference between the sample
+
--

Fig.-A :- The basic principle of DPCM

 value m(k) & the sample value m(k-1).
• If such changes are transmitted, then simply by
adding up (accumulating) these changes we shall
generate at the receiver a waveform identical in
form to the message signal m(t).
• Thus by transmitting the difference signal m(k) -
m(k-1) by PCM, fewer quantization levels are
required as the difference signal is going to varry
within a smaller range of voltage extremes.
• Hence encoding in DPCM contains less no. of
bits in comparison to PCM.
• In a practical DPCM system the difference
between 2 successive sample would first be
quantized & then encoded into a binary bit stream
before transmission & decoded at the receiver.
For simplicity the encoder & decoder are not
shown in Fig.-A.
• Thus, the DPCM can operate at approximately
one-half of the bit rate of PCM with a consequent
saving of spectrum space or BW.
• Generally the predictor stores the past differences
& predict the next required increment using some
algorithm.
• For audio, DPCM encoding saves nearly 25%
bits compared to PCM for same audio quality.
• In adaptive DPCM (ADPCM) the predictor
coefficients are updated using an adaptation
scheme & LMS algorithm. For audio coding,
ADPCM is popular like PCM & gives nearly
50% savings of bits.
• Thus, the difference signal is much less in
amplitude than the actual sample & hence less
number of quantization levels are needed.
• So, the number of bits per code is reduced,
resulting in a reduced bit-rate.
• Thus the BW required in this case is less than the
PCM system.
• The disadvantage of DPCM is that the modulator
& demodulator circuits are more complicated
than those in PCM.

• NB:- Go through sufficient numbers of numerical from

different text books.
 Delta Modulation (DM)
• In PCM, all the bits are transmitted for which the
BW requirement & signalling rates are high. To
avoid these problems delta modulation is used.
• It is a DPCM scheme in which the difference signal
is encoded into a single bit either a ‘0’ or ‘1’.
• Hence DM transmits only one bit/sample & it
transmits either an 1 or 0 if it is greater or less than
the previous sample value.
• In DM, if the present sample value (amplitude) is
smaller than the previous sample, the logic zero is
transmitted.
• But, if the amplitude level of the present sample
is larger than the amplitude of the previous
sample, then logic 1 is transmitted.
• Thus in DM the step size, S is fixed. The
difference between the i/p signal m(t) & staircase
approximation, mq(t) is confined to two levels,
i.e., (+S) & (-S).
• If the difference is +ve, the approximated signal is
increased by one step S. if the difference is –ve,
then the approximated signal is reduced by one
step S.
• Block diagram of DM is shown in Fig.-B.
 +1

-1

+
--

FIG. B (i): Block diagram of Delta Modulator circuit.

• The pulse generator produces a pulse train, Pi(t)
of +ve pulses. The modulator multiplies Pi(t) with
the polarity ( either +1 or -1) of ∆(t) (the o/p of
the difference amplifier).
• In DM, magnitude of ∆(t) plays no role in
deciding Po(t) assuming that the probability of
∆(t) being exactly zero is zero.
• Fig.-C explains the operation of a DM.
• The waveform Po(t) is transmitted. At the
receiver, the quantizer takes a decision whether
the received pulse is + ve or –ve. The LPF
smoothens the o/p of the integrator to give a
waveform m~ (t ) which is similar to m(t).
FIG. B (ii): Block diagram of DM system
FIG. C: Waveforms in the DM system
• The up-down counter increments or decrements its
count by 1 at each edge of the clock (either leading
or trailing edge).
• Sample & hold ckt samples & holds the signal till a
new signal arrives. The difference signal, Δ(t) is
quantized & encoded into a single bit (+S as ‘1’ or –S
as ‘0’).
• Due to transmission of only one bit for each sample,
the signaling rate & channel BW requirement is very
small in comparison to PCM. Hence the transmitter
& receiver ckts are very simple in DM.
• DM is only suitable when mq(t) closely follows
message signal m(t).
• But, DM is practically not preferred due to certain
limitations like slope overload error & Granular
noise as shown in Fig.-D.
 +ve slope overload problem

Granular noise
-ve slope overload problem problem
FIG. D: Limitations of the DM system
• When the message signal suddenly
increases/decreases, then the quantization (DM)
signal (mq(t)) is unable to follow the message signal
due to larger slope variations & fixed step size, S,
hence slope overload error occurs as shown in Fig.-D
(a) & (b).
• To avoid slope overload error, the step size S can be
increased, i.e., the following condition must be
satisfied:- Sfs ≥ |dm(t)/dt|max = ωmAm, if, m(t) =
Amcosωmt .
• Again, Granular noise occurs (Fig.-D ©) when the
step size is too large compared to small/slow
variations in the i/p signal m(t). Hence, the DM o/p is
a square wave & recovered as DC, though m(t) is not
DC. To overcome this problem, the step size, S must
be reduced.
 Adaptive Delta Modulation (ADM)
• In DM, a large step size is required to decrease
slope overload error & a small step-size is
required to reduce granular noise.
• Hence ADM is the modification to overcome
these 2 problems of DM by suitably changing the
step size, i.e., in ADM the step size varies as per
the situation/demand.
• Thus, slope-overload can be eliminated if the step
size is increased/decreased in such a way that the
magnitude of the slope of m ~ (t ) becomes greater
than the magnitude of the slope of m(t).
• When the message signal variations are less than
the step size, the step size may be reduced to take
care of the situation to avoid the granular noise
problem.
• Thus, ADM is a DM scheme where the step size
is automatically varied depending on the
amplitude characteristic of the analog i/p, m(t).
• Figs.-E & F show the block diagrams of ADM
ckts.
• In digital processor one accumulator is present
which produces steps to increase or decrease the
step size according to the following algorithm:-
S(k) = |S(k-1)|e(k) + S0e(k-1) …(i)
 +

--

Fig. –E (i): A complete Block diagram of an ADM.

FIG. E(ii): Block diagram of an adaptive Delta Modulator ckt.
Fig.–F (i): A complete Block diagram of an ADM.
• Where S(k) is the step size at the k-th sample =
nS0, e(k) is the o/p of the DM, S0 is the basic/unit
step size which will be increased by nS0 (where n
= 1,2,3,…).
• In the slope-overload case, the o/p of the RC
integrator is either of a large +ve or –ve value.
• The gain of the amplifier depends on the gain
control i/p, the square law device, which is
always a large +ve value for the slope-overload
case.
• When the signal variations are within the step
size, P0 (t) is a sequence of alternate +ve &-ve
pulses & the RC integrator o/p in this case is 0,
 A

FIG. F(ii): Block diagram of an adaptive Delta Modulator.

 hence, the gain control i/p of the variable gain
amplifier is also zero. Thus, the gain of the
variable gain amplifier decreases, resulting in a
reduced step size to take care of the situation to
avoid the granular noise problem.
• On the receiver side, the o/p of the quantizer is
fed to a variable gain amplifier controlled by via
RC integrator & a square law device.
• Thus, an adaptive adjustment of the step size is
obtained at the receiver to get an undistorted
reception of the transmitted signal.
• In DM signal error will be more but, in ADM
signal error is less.
• In DM up/down counter is used whereas digital
processor is used in ADM.
• Step size is fixed in DM but, step size is adaptive
or variable in ADM.
• In ADM slope overload error & granular noise is
very very less.

• NB:- Go through sufficient numericals from

various text books ?
 Assignment –Module-I
• Q1. Explain in detail the DM with circuit
diagram ?
• Q2. Explain in detail the DPCM with circuit
diagram?
• Q3. Explain in detail the ADM with circuit
diagram ?
• Q4. Why line coding is used ? Classify all types
of line coding used generally ?
• Q5. Explain in detail about companding process
with its characteristics ?
• (6) Explain in detail the typical Digital
Communication System with block diagram ?
• (7) Explain briefly how, multiple band-limited
signals can be transmitted simultaneously using
sampling/TDM process ?

• Summit the assignment for 05 marks of your

Assignment.
Thank you.