Multimedia Systems and

Applications

Lecture 2 - Digitization Process

Introduction to Multimedia 1
Digital Technology
 Everyday, we encounter many values that
change continuously, for example, the voltage of
the electricity that lights up our room varies
continuously over time. These are also known as
analogue signals.

 However, modern computers are built to deal

with entities in completely different way. These
are known as digital computers because they
work with digits.

Introduction to Multimedia 2
Digital Technology
 Because of this, when using
a computer to process
continuous signals, we first
need to find a way to
represent them so that the
computer is able to handle
them. Usually, this is a
digital representation, i.e.,
we use a series of numbers
to denote the continuous
signals.

Introduction to Multimedia 3
Digital Technology
 Then, we have to convert the continuous signal into the
digital representation. This process is known as
digitization.
 The first step in the digitization process is sampling
which takes samples of the continuous signal. The
number of samples taken during a time period is known
as sampling rate.
 The second step is known as quantization where we
restrict the value of the samples to a fixed set of levels.

Introduction to Multimedia 4
Digital Technology

Introduction to Multimedia 5
Sound (Wave) Representation

Introduction to Multimedia 6
 Frequency
 Frequency is expressed in cycles per second - 1 cycle = 1 hertz (Hz)
the higher the frequency, the higher the pitch of the sound.
 For example, a violin = 2000 Hz, and a brass band = 100 Hz. A high
quality system will produce sounds in the range 20Hz to 22,050Hz
(Human range).

Introduction to Multimedia 7
A Low Frequency Sound Signal

Introduction to Multimedia 8
A High Frequency Sound Signal

Introduction to Multimedia 9
Amplitude
 The amplitude of an analogue audio signal is measured in
decibels (dB).

Introduction to Multimedia 10
Amplitude

Introduction to Multimedia 11
Amplitude

Introduction to Multimedia 12
Human Sound

Introduction to Multimedia 13
 Digitizing Sound
(Analogue to Digital Conversion)

Introduction to Multimedia 14
Sampling
 Readings of the analogue voltage are taken at uniformly
spaced time intervals - this is called the sampling rate:
the number of samples taken per second.

 In theory, the sampling rate must be at least twice the

highest frequency in the range of analogue voltage
values.

 Sampling rate = 2*f max ; f : Frequency. Nyquist Rule

 For Human sound : 20Hz to 22,050Hz

We take 2* 22,050 = 44,100 sample/sec

Introduction to Multimedia 15
Sampling the Low Frequency
Signal

Introduction to Multimedia 16
 Sampling the High Frequency
Signal

Clearly, the sampling rate for the high frequency signal is

inadequate - it does not 'catch' every peak and trough - the rate
would have to be increased. 17
Introduction to Multimedia
Quantization
 After sampling has been completed, it must be
remembered that the sampled data is still
analogue in form.
 It must therefore now be turned into digital
data. This is achieved using a process called
Quantization.
 Quantization is the process of converting
analogue values to digital values.

Introduction to Multimedia 18
Quantization

Introduction to Multimedia 19
Aliasing and Quantization Error
 If we under sample, i.e., taking less samples than as
required by Nyquist sampling theorem, some of the
frequency components will be mistakenly converted into
other frequencies. This is known as aliasing.
 On the other hand, if we use too few levels to represent
each sample value, there will be large amount of error
for each sample.
 This is known as quantization error. These errors can be
thought of as noise on the signal.

Introduction to Multimedia 20
Quantization Interval

 When 8 bits are used, there are 28 = 256 bands,

and also, therefore, 256 quantization intervals.

 The fewer the number of analogue bands, the

wider each band must be, and thus the greater
the range of analogue values which are
translated to a single digital value.

Introduction to Multimedia 21
Quantization Interval
 It is better to use a larger number of bits for each
quantization interval.
 For instance, when 16 bits are used, the number of
bands/quantization intervals is 216 = 65,536. Thus less
coarse (and therefore more accurate) than 8-bit
samples.
 16-bit sampling is in fact the norm for sound digitization.

 The data rate of a stream of digitized data is calculated

as:
sampling frequency x bits per sample / 8, expressed
in bytes/sec.
Introduction to Multimedia 22
Sampling and Sampling Rate
 To record everything the human ear can possibly
hear, we need to be sampling at a bit over
44kHz, or something like 44,100Hz, or 44.1kHz,
which is what audio compact disks (CDs) are.

 By sampling at this rate, we effectively remove

higher frequencies. For human listeners that's
not a problem (since we cannot hear higher
beyond 22kHz anyway).

Introduction to Multimedia 23
Sampling and Sampling Rate
 For example, telephone samples at 8kHz,
which effectively means that any frequency
higher than 4kHz is wiped out; which is precisely
why on the phone we sometimes cannot tell the
difference between certain letters (like "f" and
"s"), and why we need to spell our names (s-as-
in-Solid, and f-as-in-fast), and the listener still
manages to make mistakes when writing it
down.

Introduction to Multimedia 24
Sample Size
 Another issue we have to care about when sampling
sound is how big to make each sample. Basically, how
many distinct values can a wave can take:
* We want to record 0 dB – 60 dB (Normal
Conversation)
60/8 = 7.5  8 bits needed
* To record from 60 dB – 140 dB (Machines and
Music)
140/8 =17.5  16 or 24 bits
 Most sound is encoded at 8-bit or 16-bit. Some (very
rare) high end sound cards have 32-bit samples. Audio
CD samples are 16-bit.

Introduction to Multimedia 25
Sample Size
 For example, a 22kHz stereo sound, with 16-
bit sampling will need to be sampled at 44.1kHz
or 44,100 times per second. Each sample will be
16-bits, and we'll have 2 channels (one left, and
one right---for stereo). Thus, we'll have
44,100 sample/sec * 2 bytes/sample * 2
channels = 1,411,200 bits/s
= 176,400 bytes/s
=173 kbyte/sec.

Introduction to Multimedia 26
Image Sampling & Quantization

Introduction to Multimedia 27
Basic Concepts

 To create a digital image, we need to

convert continuous sensed data into digital
form.

 This involves two processes: sampling and

quantization

Introduction to Multimedia 28
RGB Channels

Introduction to Multimedia 29
 Continuous Image

 This figure shows a

continuous image, f (x, y),
that we want to convert to
digital form.
 To convert it to digital
form, we have to sample
the function in both
coordinates and in
amplitude.

Introduction to Multimedia 30
Image Sampling and Quantization

 Digitizing the coordinate values is called

sampling.

 Digitizing the amplitude values is called

quantization.

Introduction to Multimedia 31
 This is a plot of amplitute (gray level) values of the
continuous image along the line segment A B
 To sample this function, we take equally spaced samples
along line AB.
 Location of each sample is given by a vertical tick mark
in the bottom part of the figure.

Introduction to Multimedia 32
 However, the values of the samples still span
(vertically) a continuous range of gray‑level
values.
 In order to form a digital function, the gray‑level
values also must be converted (quantized) into
discrete quantities.
Introduction to Multimedia 33
Quantization

 The right side of the above figure shows the

gray‑level scale divided into eight discrete levels,
ranging from black to white.
 The vertical tick marks indicate the specific value
assigned to each of eight gray levels.
 The continuous gray levels are quantized simply
by assigning one of the eight discrete gray levels
to each sample.

Introduction to Multimedia 34
 The assignment is made depending on the
vertical proximity of a sample to a vertical tick
mark.
Introduction to Multimedia 35
Representing Digital Images

 The values of the coordinates at the origin are

(x,y) = (0,0).
 The next coordinate values along the first row
are (x,y) = (0,1).
 The notation (0,1) is used to signify the 2nd
sample along the 1st row.

Introduction to Multimedia 36
Coordinate convention used to represent digital images

Introduction to Multimedia 37
A digital image of size M x N

Introduction to Multimedia 38
 Representing Digital Images

 The number of bits required to store a digitised

image is
b=MxNxk
where M & N are the number of rows and
columns, respectively and k is the number of bits
/pixel.
 The number of gray levels is an integer power of
2:
L = 2k where k =1,2,…
 It is common practice to refer to the image as a
“k-bit image”
Introduction to Multimedia 39
Introduction to Multimedia 40
Image Sampling

 Dense sampling will produce a high resolution

image in which there are many pixels, each of
which represents of a small part of the scene.

 Coarse sampling, will produce a low resolution

image in which there are a few pixels, each of
which represents of a relatively large part of the
scene.

Introduction to Multimedia 41
Introduction to Multimedia 42
High Data Volume of
Multimedia Information

Speech 8000 samples/s 8Kbytes/s

CD Audio 44,100 samples/s, 2 176Kbytes/s

bytes/sample

Satellite 180X180 km^2 600MB/image

Imagery 30m^2 resolution (60MB
compressed)
NTSC Video 30fps, 640X480 30Mbytes/s
pixels, 3bytes/pixel (2-8 Mbits/s
compressed)

Introduction to Multimedia 43