Delta Modulation

EE 442 – Spring 2017

Lecture 10

Voice Input


Encoder Out

Decoder Out

1
 Key Attributes About Delta Modulation

1. Delta modulation (DM) is the simplest method for analog-to-digital

conversion (ADC).

2. Delta modulation uses 1-bit per sampling period (TS) – it is a 1-bit ADC

3. Delta modulation requires a sampling rate much greater than the

Nyquist rate (commonly four or five times the Nyquist rate).

4. DM is closely related to DPCM.

5. In DM we use a first-order predictor (one time delay TS is the predictor).

6. DM uses very simple hardware and is low cost for that reason.

7. The transmitted output is a binary stream of  pulses at fS. It gives

a stepwise approximation m(t) to m(t).

2
 Delta Modulation
Transmitter:
d(t) Sampler
m(t) + + dq[k]
 -
To the
channel
Comparator

m(t) Dq[k]
Integrator

(accumulator)
 time
dq[k]
-

step size
 m(t)
m(t) or m(t)

Slope
Overload

m(t)

TS time

3
 Delta Modulation Waveforms

Start-up interval
Slope Overload Noise


Band-Limited Analog Signal m(t)

Integrator Output mq(t)

Granular Noise
time

dq(t)


-
time
dq[k] 0 0 0 1 0 1 0 1 1 1 1 1 0 1 0 1 1 1 1 1 1 0 1 0 . . .

4
 Delta Demodulation

dq[k]
mq(t) 
m(t)
From the Low-pass
Integrator
channel filter
(accumulator)

Now that is really simple!

5
 Adaptive Delta Modulation

We can address the slope overload error problem with adaptive DM.
Of course, it does add more complexity to its implementation.

-3
Staircase with ADM 
2
m(t)
5

3 Staircase with DM
mq(t)
2

1 0 1 1 1 1 0 1 0 1 0 1 0 1 0 ...

6
 Comparing PCM, DPCM and DM

Parameters PCM DPCM DM

Number of 4, 8, 16, Typically 4 to 6 bits < 1 bit
bits/sample and so on. PCM
Bandwidth Highest required Lower than PCM Lowest
Step Size Fixed Fixed Fixed
Sampling Rate 8 kHz 8 kHz 64 to 128 kHz
Bit Rate 56 and 64 kbps 32 to 48 kbps 64 to 128 kbps
Quantization Depends upon q Slope overload and Slope overload and
Error granular noise granular noise

7
 Slope-Overload Distortion and Granular Noise

Slope

Delta modulation transmits the derivative of signal m(t). Suppose the

message signal is the sinusoidal signal Am cos(2fmt), then

 dm(t ) 
 dt   2 f m Am  2 B
  max
where fm < B. Slope overload means mq(t) can’t follow m(t). We then
have the condition, 2fmAm < fS or B Am < fS .

8
 Quantization Noise in DM
m(t)

m(t) is a sinusoidal waveform

2
Nq   
2

 = m(t) - m(t)


 TS

9
 Worked Example for DM
Problem:
A Delta modulated system is designed to operate at five times the Nyquist
Rate. The signal bandwidth B at its input port is 3 kHz and the quantized
step  is 250 millivolts (0.25 volt). For this problem we assume a 2 kHz
sinusoidal input – Find the maximum amplitude Am of this 2 kHz tone.

Solution:
We know that B = 3 kHz, fm = 2 kHz and  = 250 mV.

The Nyquist rate is 3,000  2 = 6,000 Hz. So five time the Nyquist rate is
30,000 Hz = fS.

Using the relationship given at the bottom of slide 8 allows us to write

  f S  0.25  30, 000

Am    0.60 volt
2 f m 2 (2, 000)

10
 Maximum Information Rate in Communications
Recall from prior lecture:
Basic relationship in digital communications:

A maximum of 2B independent elements of information per

second can be transmitted, error-free, over a noiseless channel
of bandwidth B Hz.

It is related to the sampling theorem:

Remember the sampling theorem states that a low-pass signal g(t) of
bandwidth B Hz can be fully recovered from uniform samples taken at
the rate of 2B samples per second.

The sampling theorem is important in signal analysis, digital signal

processing and transmission because it allows us to replace an analog
signal with a discrete sequence of numbers (i.e., digital signal).

11
 Comparing PCM With DM
Problem:
A one kilohertz (1 kHz) signal m(t) is sampled at 8 kHz with 12-bit encoding
for PCM transmission.
(a) How many bits are transmitted per second in in PCM? What is the
bandwidth required in this case?
(b) Now switch to using DM with 8 kHz sampling. How many bits are
transmitted per second using DM? What is the bandwidth required in
using DM?

Solution:
We know that the signal frequency is fm = 1 kHz and the sampling rate is
8 kHz.
(a) For PCM we have 8,000 samples per second and 12 bits per sample;
which equals 96,000 bits/second. The bandwidth is one-half of this
giving 48,000 Hz.
(b) Now for DM we have 1 bit per sample at 8,000 samples per second.
Thus, we have 8,000 bits per second and a bandwidth of 4,000 Hz.

12