UNIT IV

MULTIMEDIA SYSTEM DESIGN & MULTIMEDIA FILE HANDLING

Multimedia basics − Multimedia applications − Multimedia system architecture − Evolving

technologies for multimedia − Defining objects for multimedia systems − Multimedia data
interface standards − Multimedia databases. Compression and decompression − Data and file
format standards − Multimedia I/O technologies − Digital voice and audio − Video image and
animation − Full motion video − Storage and retrieval technologies.

Multimedia Basics:
Multimedia is a combination of text, graphic art, and sound, animation and video elements.
The IBM dictionary of computing describes multimedia as "comprehensive material, presented
in a combination of text, graphics, video, animation and sound. Any system that is capable of
presenting multimedia, is called a multimedia system"
A multimedia application accepts input from the user by means of a keyboard, voice or pointing
device.
Multimedia applications involve using multimedia technology for business, education and
entertainment.
Multimedia is now available on standard computer platforms. It is the best way to gain attention
of users and is widely used in many fields as follows:
 Business - In any business enterprise, multimedia exists in the form of advertisements,
presentations, video conferencing, voice mail, etc.
 Schools - Multimedia tools for learning are widely used these days. People of all ages
learn easily and quickly when they presented information with the visual treat.
 Home PCs equipped with CD-ROMs and game machines hooked up with TV screens
have brought home entertainment to new levels. These multimedia titles viewed at
home would probably be available on the multimedia highway soon.
 Public places - Interactive maps at public places like alone terminal.
 Virtual Reality (V R) - This technology helps us feel a virtual reality effect is very
popular.
Multimedia Elements:
High-impact multimedia applications, such as presentations, training and messaging, require
the use of moving images such as video and image animation, as well as sound (from the video
images as well as overlaid sound by a narrator) intermixed with document images and graphical
text displays. Multimedia applications require dynamic handling of data consisting of a mix of
text, voice, audio components, Video components and image animation. Integrated multimedia
applications allow the user to cut sections of all or any of these components and paste them in
a new document or in another application such as an animated sequence of events, a desktop
publishing system, or a spreadsheet.
The components that fall under our definition of multimedia are:
Data elements for Multimedia Systems
Facsimile: Facsimile transmissions were the first practical means of transmitting document
images over telephone lines. The basic technology, now widely used, has evolved to allow
higher scanning density for better-quality fax.
Document images: Document images are used for storing business documents that must be
retained for long periods of time or may need to be accessed by a large number of people.
Providing multimedia access to such documents removes the need for making several copies
of the original for storage or distribution.
Photographic images: Photographic images are used for a wide range of applications such as
employee records for instant identification at a security desk, real estate systems with
photographs of houses in the database containing the description of houses, medical case
histories, and so on.
Geographic information systems map (GIS): Map created in a GIS system are being used
wildly for natural resources and wild life management as well as urban planning. These systems
store the geographical information of the map along with a database containing information
relating highlighted map elements with statistical or item information such as wild life statistics
or details of the floors and rooms and workers in an office building.
Voice commands and voice synthesis: Voice commands and voice synthesis are used for
hands-free operations of a computer program. Voice synthesis is used for presenting the results
of an action to the user in a synthesized voice. Applications such as a patient monitoring system
in a surgical theatre will be prime beneficiaries of these capabilities. Voice commands allow
the user to direct computer operation by spoken commands
Audio message: Annotated voice mail already uses audio or voice message as attachments to
memos and documents such as maintenance manuals.
Video messages: Video messages are being used in a manner similar to voice mail.
Full-motion stored and live video: Full-motion video started out as a very useful idea for on-
line training and maintenance manuals. The capability to use full-motion stored video for
electronic mail or live video for presentations and video conferencing are important
evolutionary steps. Three-dimensional video techniques are being adapted to create the concept
of virtual reality.
Holographic images: All of the technologies so for essentially present a flat view of
information. Holographic images extend the concept of virtual reality by allowing the user to
get “inside” a part, such as, an engine and view its operation from the inside.
Fractals: Fractals started as a technology in the 1980s but have received serious attention only
recently. This technology is based on synthesizing and storing algorithms that describes the
information.
4.2 MULTIMEDIA APPLICATIONS
The first widely used application of multimedia is document image management. It is primarily
intended for scanning documents and retaining their images.
Another application is image processing. It is also known as Image recognition. It is intended
for recognizing objects by analyzing their raster images. Applications that present a view of
generic multimedia applications are:
1. Document Imaging
The fundamental concepts of storage, compression and decompression, and display
technologies used for multimedia systems were developed for document image management.
Organizations such as insurance agencies, law offices, country and state governments, and the
federal government manage large volumes of documents.
Document image technology is adopted by Department of Defence for applications ranging
from military personnel records to maintenance manuals and high-speed printing systems.
Almost all document image system use workflows that are customized for the purpose for
which they are being used. The workflow defines the sequence for scanning images,
performing data entry based on the contents of the Images, indexing them and storing them on
optical media.
Document Image Hardware requirements:
Real-time image decompression and display place an important role on image processing
hardware. Image decompression and display supports 4 to 8 planes. 4 planes provide 16 colors
and 8 planes provide 256 colors. The image planes are also called bit planes, because, they are
addressed by a bit in a byte. Images must be processed at the rate of tens to hundreds of pixels
per nano-second. For high-resolution images, processing of the order of 10 pixels/ ns is enough
for monochrome still images. Gray-scale images consist of pixels that have shades of gray
ranging from 16 to 256. Color images feature color hues instead of shades of gray. Most high-
resolution monitors support 16 to 256 colors display capability. The number of colors that can
be depicted depends on the number of bits used to define the palette.
2.Image processing and Image Recognition
Image processing involves image recognition, Image enhancement, image synthesis, and image
reconstruction. An image processing system may actually alter the contents of the image itself.
Image processing systems employ the compression and decompression techniques, a wide
range of algorithm for object recognition, comparing images of objects with pre-defined
objects, extrapolating finer details to view edges more clearly, gray-scale balancing and gray-
scale and color adjustments.
Let us briefly overview the various aspects of image processing and recognition.
Image enhancement:
Most image display systems feature some level of image adjustment. Increasing the sensitivity
and contrast makes the picture darker by making borderline pixels black or increasing the gray-
scale level of pixels.
Capabilities built in the compression boards might include the following:
 Image calibration: The overall image density is calibrated, and the image pixels are
adjusted to a predefined level.
 Real time alignment: The image is aligned in real-time for skewing caused by improper
feeding of paper.
 Gray-Scale normalization: The overall gray level of an image or picture is evaluated to
determine if it is skewed in one direction and if it needs correction.
 RGB hue intensity adjustment: Too much color makes a picture garish and fuzzy.
Automatic hue intensity adjustment brings the hue intensity within pre-defined ranges.
 Color Separation: A Picture with very little color contrast can be dull and may not bring
out the details. The hardware used can detect and adjust the range of color separation.
 Frame averaging: The intensity level of the frame is averaged to overcome the effects
of very dark or very light areas by adjusting the middle tones
Image Animation
Computers-created or scanned images can be displayed sequentially at controlled display
speeds to provide image animation that simulates real processes. The basic concept of
displaying successive images at short intervals to give the perception of motion is being used
successfully in designing moving parts such as automobile engines.
Image annotation
Image annotation can be performed in one of two ways: as a text file stored along with the
image or as a small image stored with the original image. The annotation is overlayed over the
original image for display purposes. It requires tracking multiple image components associated
with a single page, decompressing all of them, and ensuring correct spatial alignment they are
overlayed.
Optical Character Recognition
Data entry is the most expensive component of data processing, because it requires extensive
clerical staff work to enter data. Automating data entry, both typed and handwritten, is a
significant application that can provide high returns. Optical Character Recognition (OCR)
technology is used for data entry by scanning typed or printed words in a form.
Initially, people used dedicated OCR scanners. NOW, OCR Technology is available in
software. OCR technology, used as a means of data entry, may be used for capturing entire
paragraphs of text. The capturing text is almost always entered as a field in a database or in an
editable document
Handwriting recognition
Research for Handwriting recognition was performed for CAD / CAM systems for command
recognition. Pen-based systems are designed to allow the user to write commands on an
electronic tablet.
Handwriting recognition engines use complex algorithms designed to capture data in real time
as it is being input or from an image displayed in a window, depending on the application. Two
factors are important for handwriting recognition. They are the strokes or shapes being entered,
and the velocity of input or the vectoring that is taking place. The strokes are parsed and
processed by a shape recognizer that tries to determine the geometry and topology of the
strokes. It attempts to compare it to existing shapes, such as predefined characters. The stroke
is compare with the prototype character set until a match is found or all pre-defined prototypes
have been checked without a match. Multimedia system will use handwriting recognition as
another means of user input.
Non-Textual Image Recognition
Image recognition is a major technology component in designing, medical and manufacturing
fields. Let us review the basic concepts of image recognition architecture.
For example, a general Image recognition system, - the Image Understanding Architecture has
the design which calls for three processing layers.
(i) 512 x 512 array of custom pixel processors that extract basic features such as lines
and object boundaries.
(ii) The features of an object extracted by the first layer are tracked by the DSP array,
and that information is fed into 512-M byte RAM.
(iii) At the highest level, sophisticated AI algorithms perform the difficult task of object
and scene recognition.
3.Full motion Digital video Applications
Full motion video has applications in the games industry and training, as well as the business
world. Full motion video is the most complex and most demanding component of multimedia
applications.
For business applications, some core requirements are needed.
(i) Full-motion video clips should be sharable but should have only one sharable copy.
(ii) It should be possible to attach full-motion video clips to other documents such as
memos, chapter text, presentation, and so on.

Figure: Use of Full-Motion Video in Multimedia Applications.

The following features should be available:
a) Features, of a VCR metaphor, such as, rewind, fast-forward, play, and search.
b) Ability to move and resize the window displaying the video clip.
c) Ability to view the same clip on a variety of display terminal types with varying
resolution capabilities without the need for storing multiple copies in different form.
d) Ability to adjust the contrast and brightness of the video clip.
e) Ability to adjust the volume of the associated sound.
f) It should enable the users to place their own indexing marks to locate segments in video
clip.
4.A Universal Multimedia Application
It is an application that works on universal data type. This means that the application
manipulates datatypes that can be combined in a document, displayed on a screen, or printed,
with no special manipulations that the user needs to perform. The application is truly distributed
in nature.
An important consideration for such a universal application is the methodology for
dissemination of the information on a network.

Figure: Universal Multimedia Desktop

Figure describes the user screen for a universal multimedia application. In this screen, mix of
windows for displaying still video and document images, a video conference window with a
live session in progress, a remote live desktop, and a couple of other windows for applications
such as electronic mail and desktop publishing.
To maintain all of these windows requires a substantial amount of CPU power. Digital Signal
Processing assistance is needed to manage the multiple simultaneous decompressions for
JPEG, MPEG and windows applications.
Full-Motion Video Messages
In addition to textual messages, electronic mail capability allows embedding of voice messages
and video messages. Video messages may consist of video snapshots or live video with full-
motion picture and sound.
Two technological concepts at play in the implementation of full motion video messages;
(i) The storage and transmitted of a very large volume of data at a high rate.
(ii) Decompression of that data to present a continuous play back.

Audio and Video Indexing.

Indexing is an important and complex subject for multimedia design. Marking a position is
called Indexing. Audio and video indexing are used in full-motion video in a manner similar
to any video sequence, i.e., just as it would in a home movie, taped performance and so on.
The needs of the application must be a strong consideration for the type of indexing provided
with the system.
Key points for indexing of stored video clips:
 Indexing is useful only if the video is stored. Unless live video is stored, indexing
information is lost since the video cannot be repeated.
 When sound and video are decompressed and managed separately, synchronization is
very important.
 Depending on the application, indexing information must be maintained separately for
sound and video components of a video clip.
4.3 Multimedia Systems Architecture
Multimedia encompasses a large variety of technologies and integration of multiple
architectures interacting in real time. All of these multimedia capabilities must integrate with
the standard user interfaces such as Microsoft Windows.
The following figure describes the architecture of a multimedia workstation environment.

Figure: Multimedia Workstation Architecture.

In this diagram, the right side shows the new architectural entities required for supporting
multimedia applications. The add-on multimedia devices and peripherals include scanners,
video cameras, VCRs and sound equipment along with their associated device controllers and
encoding hardware (such as DVI- JPEG- or MPEG-enabled boards). For each of these special
device, a software device driver is need to provide the interface from an application to the
device, The GUI require control extensions to support applications such as full motion video.
High Resolution Graphics Display
The various graphics standards such as MCA, GGA and XGA have demonstrated the
increasing demands for higher resolutions for GUIs. Combined graphics and imaging
applications require functionality at three levels. They are provided by three classes of single-
monitor architecture.
(i) VGA mixing: In VGA mixing, the image acquisition memory serves as the display
source memory, thereby fixing its position and size on screen.
(ii) VGA mixing with scaling: Use of scalar ICs allows sizing and positioning of images in
pre-defined windows; resizing the window causes the image to be retrieved again.
(iii) Dual-buffered VGA/Mixing/Scaling: Double buffer schemes maintain the original
images in a decompression buffer and the resized image in a display buffer.
The IMA Architectural Framework
The Interactive Multimedia Association has a task group to define the architectural framework
for multimedia to provide interoperability. The task group has concentrated on the desktops
and the servers. Desktop focus is to define the interchange formats. This format allows
multimedia objects be displayed on work station.
The architectural approach taken by IMA is based on defining interfaces to a multimedia
interface bus. This bus would be the interface between systems and multimedia sources. It
provides streaming I/O services, including filters and translators Figure describes the
generalized architectural approach

Figure: Multimedia Architecture Based on Interface Bus.

Network Architecture for Multimedia Systems:
Multimedia systems need special networks. Because large volumes of images and video
messages are being transmitted.
Asynchronous Transfer Mode technology (ATM) simplifies transfers across LANs and WANs.
Task based Multi-level networking
Higher classes of service require more expensive components in the workstations as well as in
the servers supporting the workstation applications. Rather than impose this cost on all work
stations, an alternate approach is to adjust the class of service to the specific requirement for
the user. This approach is to adjust the class of services according to the type of data being
handled at a time also. We call this approach task-based multilevel networking.
High speed server to server Links
Duplication: It is the process of duplicating an object that the user can manipulate. There is no
requirement for the duplicated object to remain synchronized with the source (or master)
object.
Replication: Replication is defined as the process of maintaining two or more copies of the
same object in a network that periodically re-synchronize to provide the user faster and more
reliable access to the data Replication is a complex process.
Networking Standards
The two well-known networking standards are Ethernet and token ring.
ATM and FDDI are the two technologies which we are going to discuss in detail.
ATM: ATM is an acronym for Asynchronous Transfer Mode. Its topology was originally
designed for broadband applications in public networks. ATM is a method of multiplexing and
relaying (cell-switching) 53 byte cells. (48 bytes of user information and 5 bits of header
information).
Cell Switching: It is a form of fast packet switching based on the use of cells.
Cells: Short, fixed length packets are called cells.
ATM provides high capacity, low-latency switching fabric for data. It is independent of
protocol and distances. ATM effectively manage a mix of data types, including text data, voice,
images and full motion video. ATM was proposed as a means of transmitting multimedia
applications over asynchronous networks.
FDDI: FDDI is an acronym of Fiber Distributed Data Interface. This FDDI network is an
excellent candidate to act as the hub in a network configuration, or as a backbone that
interconnects different types of LANs. FDDI presents a potential for standardization for high
speed networks. The ANSI standard for FDDI allows large-distance networking. It can be used
as high-performance backbone networks to complement and extend current LANs.
4.4 Evolving Technologies for Multimedia Systems
Multimedia applications use a number of technologies generated for both commercial business
application as well as the video game industry.
Let us review some of these technologies in this section.
Hypermedia documents
Hypermedia documents are documents which have text, embedded or linked multimedia
objects such as image, audio, hologram, or full-motion video.
Hypertext
Hypertext systems allow authors to link information together, create information paths through
a large volume of related text in documents. It also allows to annotate existing text, and append
notes. It allows fast and easy searching and reading of selected excerpts.
Hypermedia
It is an extension of hypertext. In that, we can include texts, any kind of information that can
be stored in electronic storage, such as audio, animated video, graphics or full-motion video.
Hypermedia documents used for electronic mail and work flow applications provide a rich
functionality for exchanging a variety of information types. The hypermedia document is a
definition of a document and a set of pointers to help locate the various elements of the
document on the network.
Hyperspeech
Multimedia stimulated the development of general-purpose speech interfaces. Speech synthesis
and speech recognition are fundamental requirement for hyperspeech systems. Speech
recognition is nothing but converting the analog speech into a computer action and into ASCII
text. Speech-recognition systems cannot segment a stream of sounds without breaks into
meaningful units. The user must speak in a stilted fashion. He should make sure to interpose
silence between each word.

Figure: Progression of Resolution of Television Pictures

HDTV and UDTV
HDTV is an acronym of High-Definition Television. The broadcasting standards such as
NTSC, PAL, SECAM, NHK have an idea of bringing the world together on a single high-
definition Television broadcasting standard. The Japanese broadcasting services developed a
1125-line, along MUSE system. A competing standard in the U.S. changed direction from
analog to digital technology: A 1125-line digital HDTV has been developed and is being
commercialized. NHK of Japan is trying to leapfrog the digital technology to develop ultra-
definition television (digital UDTV) featuring approximately 3000 lines.
3-D Technologies and Holography
Three-dimensional technologies are concerned with two areas: pointing devices and displays.
3-D pointing devices are essential to manipulate object in a 3-D display system. 3-D displays
are achieved using holography techniques. The techniques developed for holography have been
adapted for direct computer use.
Fuzzy Logic
Fuzzy logic is logic which is used for low-level process controllers. Use of fuzzy logic in
multimedia chips is the key to the emerging graphical interfaces of the future. It is expected to
become an integral part of multimedia hardware. Fuzzy logic has mathematical principles.
Hence, the application of multimedia can benefit those principles.
Digital Signal Processing
Digital Signal Processing are used in applications such as digital servos in hard disk drives,
and fax/modems. DSP technology is used in Digital wireless communications, such as personal
communication networks (pens), wireless local area networks and digital cordless phones.
DSP Architectures and Applications
A typical DSP operating system architecture would contain the following subsystems:
Memory Management: DSP architectures provide dynamic allocation of arrays from multiple
segments, including RAM, SRAM and DRAM.
Hardware-Interrupt handling: A DSP operating system must be designed to minimize
hardware-interrupt latency to ensure fast response to real time events for applications, such as
servo systems.
Multitasking: DSPs need real-time kernels that provide pre-emptive multitasking and user-
defined and dynamic task prioritization
Intertask Synchronization and Communication: Mechanisms for intertask communication
include message queues, semaphores, shared memory, and quick response event flags.
Multiple timer services: The ability for the developer to set system clock interrupt managed
timers to control and synchronize tasks is needed for most real-time applications.
Device-Independent I/O: DSP operating system should supports
(i) Asynchronous data stream
(ii) Synchronous message passing.
Use of DSP' s has evolved from traditional general purpose digital signal processors to
application-specific and customizable DSPs. DSPs were conceived as math engines with a
system architecture that was like that of a mini-computer with an array processor.

4.5 Defining Objects for Multimedia Systems

The basic data types of object using in multimedia include text, image, audio, holograms and
full-motion video.
Text
It is the simplest of data types and requires the least amount of storage. Text is the base element
of a relational database. It is also the basic building of a document. The major attributes of text
include paragraph styling, character styling, font families and sizes, and relative location in a
document
Hypertext
It is an application of indexing text to provide a rapid search of specific text strings in one or
more documents. It is an integral component of hypermedia documents. A hypermedia
document is the basic complex object of which text is a sub object. Sub-objects include images,
sound and full motion video. A hypermedia document always has text and has one or more
other types of sub-objects
Images
Image object is an object that is represented in graphics or encoded form. Image object is a
subobject of the hypermedia document object. In this object, there is no direct relationship
between successive representations in time. The image object includes all data types that are
not coded text. It does not have a temporal property associated with them. The data types such
as document images, facsimile systems, fractals, bitmaps, meta files, and still pictures or still
video frames are grouped together.

Figure: Image Hierarchy

Non-Visible
This type of images are not stored as images. But they are displayed as images. Example:
Pressure gauges, and temperature gauges.
Abstract
Abstract images are computer-generated images based on some arithmetic calculations. They
are really not images that ever existed as real-world objects. Example of these images is
fractals.
Audio and Voice
Stored-Audio and Video objects contain compressed audio information. This can consist of
music, speech, telephone conversation and voice commands. An Audio object needs to store
information about the sound clip. Information here means length of the sound clip, its
compression algorithm, playback characteristics, and any annotations associated with the
original clip.
Full Motion and Live Video
Full motion video refers to pre-stored video clips. Live video refers to live and it must be
processed while it is being captured by the camera. From a storage perspective, we should have
the information about the coding algorithm used for compression. It need decoding also. From
a processing perspective, video should be presented to user with smooth and there should not
be any unexpected breaks.
Hence, video object and its associated audio object must be transferred over the network to the
decompression unit. It should be then played at the fixed rate specified for it. For successful
playback of compressed video, there are number of technologies. They are database storage,
network media and protocols, decompression engines and display engines.

4.6 Multimedia Data Interface Standards

File Formats for Multimedia Systems:
1. Device-independent Bitmap (DIB): This file format contains bitmap, color, and color
palette information.
2. RIFF device Independent Bitmap (RDIB): Resource Interchange File Format (RIFF) is
the standard file format defined for Microsoft Windows and OS/2. It allows a more
complex set of bit maps than can be handled by DIB.
3. Musical Instrument Digital interface (MIDI): This is the interface standard for file
transfer between a computer and a musical instrument such as a digital piano. It is also,
used for full-motion video and voice-mail messaging systems. It has the advantage of
ready availability of MIDI device controller boards for personal computers.
4. RIFF Musical Instrument Digital Interface: A MIDI format within a RIFF envelope
provides a more complex interface.
5. Palette File Format (PAL): An interface that allows defining a palette of 1 to 256
colours in a representation as RGB values.
 6. Rich Text Format (RTF) This file format allows embedding graphics and other file
formats within a document. This format is used by products such as Lotus Notes. This
format is also the basis for the use of OLE.
7. Waveform Audio File Format (WAVE): A digital file representation of digital audio.
8. Windows Metafile Format (WMF): This is a vector graphic format used by Microsoft
Windows as an interchange format.
9. Multimedia Movie Format (MMM): This is a format used for digital video animation.
10. Apple's Movie Format: This format was defined as the standard for file exchange by
Quick Time enabled systems.
11. Digital Video Command Set (DVCS): This is the set of digital video commands
simulating VCR controls.
12. Digital Video Media Control Interface (DV-MCI): Microsoft's high level control
interface for VCR controls, including play, rewind, record and so on.
13. Vendor - Independent Messaging (VIM): Developed by a consortium of Vendors
providing a standardized format for cross-product messages.
14. Apple's Audio Interchange File Format: Apple's standard file format for compressed
audio and voice data.
15. SDTS GIS Standard: The Spatial Data Transfer Standard (SDTS) is designed to provide
a common storage format for geographic and cartographic data.
Video Processing Standards
Intels DVI
DVI is an acronym of Digital Video Interface. DVI standard is to provide a processor
independent specification for a video interface. That video interface should accommodate most
compression algorithms for fast multimedia displays. An example of custom-designed chip
which supports DVI is Intel's i750 B. This chip is designed for enhancing low-end, software
based PC video.
Advantages of the DVI Chip
(i) It can operate software video processing real time.
(ii) It can share the processing with the host CPU.
(iii) It can handle additional vector-quantization-type algorithms in conjunction with host
processing. DVI silicon chip relies on a programmable video processor. It gives
potential to DVI chips to run a range of compression algorithms.
Apple’s Quick Time
Quick Time standard is developed by Apple Computer. It is designed to Support multimedia
applications. It is integrated with the operating system. Quick time refers to both the extensions
to the Mac Operating system and to the compression/decompression functionality of the
environment. Quick Time is designed to be the graphics standard for time-based graphic data
types.
Quick Time's definition has been extended to include
(i) System Software.
(ii) File Formats.
(iii) Compression / decompression algorithms.
 (iv) Human Interface Standards.

Movie Toolbox

Image Compression
Manager

Component Manager

Figure: QuickTime Architecture

Figure Shows the components in the Quick Time Architecture. Quick Time adjust
automatically to the hardware being used by the user. MPEG is another competing standard
which is comparatively higher-end, hardware-assisted standard. It can produce better
resolutions at faster rates.
Microsoft’s AVI
AVI is an acronym for Audio Video Interleave Standard. It is similar to Apple's Quick Time.
It offers low-cost, low-resolution video processing for the average desktop user. It is a layered
product. AVI is scalable. It allows users to set parameter such as window size, frame rate,
quality and compression algorithm through a number of dialog boxes. AVI-compatible
hardware allows enhancing performance through hardware-accelerated compression
algorithms such as DVI and MPEG. AVI supports several compression algorithms.

4.7 Multimedia Databases

Images, sounds and movies can be stored, retrieved and played by many databases. In future,
multimedia databases will become a main source of interaction between users and multimedia
elements.
Multimedia storage and retrieval
Multimedia storage is characterized by a number of considerations. They are:
(i) massive storage volumes.
(ii) large object sizes.
(iii) multiple related objects.
(iv) temporal requirements for retrieval.
Massive Data Volumes
A single multimedia document may be a combination of different media Hence indexing of
documents, films and tapes is more complex. Locating massive data volumes requires
searching through massive storage files. Locating and indexing systems can be understood only
by a few key staff personnel. Hence it requires a major organizational eff0l1 to ensure that they
are returned in proper sequence to their original storage location.
Storage Technologies
There are two major mass storage technologies used currently for storage of multimedia
documents.
(i) Optical disk storage systems.
(ii) High-speed magnetic storage.

Advantages of Optical disk storage systems:

(i) Managing a few optical disk platters in a juke box is much simpler than managing a large
magnetic disk farm. (ii) Optical disk storage is excellent storage system for off line archival of
old and infrequently referenced documents for significant periods of time.
Multimedia object storage
Multimedia object storage in an optical medium serves its original purpose, only if it can be
located fast and automatically. A key issue here is random keyed Access t6 various components
of hypermedia database record. Optical media provides very dense storage. Speed of retrieval
is another consideration. Retrieval speed is a direct result of the storage latency, size of the data
relative to display resolution, transmission media and speed, and decompression efficiency.
Indexing is important for fast retrieval of information. Indexing can be at multiple levels.
Multimedia document retrieval
The simplest form of identifying a multimedia document is by storage platter identification and
its relative position on the platter (file number). These objects can then be grouped using a
database in folders (replicating the concept of paper storage in file folders) or within complex
objects representing hypermedia documents. The capability to access objects using identifiers
stored in a database requires capability in the database to perform the required multimedia
object directory functions. Another important application for sound and full motion video is
the ability to clip parts of it and combine them with another set. Indexing of sound and full-
motion video is the subject of intense debate and a number of approaches have been used.
Database Management Systems for Multimedia Systems
Since most multimedia applications are based primarily on communications technologies, such
as electronic mail, the database system must be fully distributed. A number of database storge
choices are available. The choices available are:
 Extending the existing relational database management systems, (RDBMSs) to support
the various objects for multimedia as binary objects.
  Extending RDBMSs beyond basis binary objects to the concepts of inheritance and
classes. RDBMSs supporting these. features provide extensions for object-
programming front ends and/or C++ support.
 Converting to a full-fledged object oriented database that supports the standard SQL
language.
 Converting the database and the application to an object-oriented database and using
an object-oriented language, or an object-enabled SQL for development.
Multimedia applications combine numerical and textual data, graphics from GUI front-ends,
CAD/CAM systems and GIS applications, still video, audio and full-motion video with
recorded audio and annotated voice components. Relational databases, the dominant database
paradigm, have lacked the ability to support multimedia databases. Key limitations of relational
database systems for implementing multimedia applications stem from two areas: the relational
data model and the relational computational model. RDBMSs have been designed to manage
only tabular alphanumeric forms of data (along with some additional data types stored in binary
form such as dates).
RDBMS Extensions for Multimedia
Binary Large Object (BLOB) is a data type which has been adapted by most of the leading
relational databases. BLOBs are used for objects such as images or other binary data types. The
relational database is extended to access these BLOBs to present the user 'with a complete' data
set. Extended relational databases provide a gradual migration path to a more object-oriented
environment. Relational database tables include location information for the BLOBs which
may be stored outside the database on separate image or video servers. Relational databases
have the strength of rigorous set management for maintaining the integrity of the database.
Object-Oriented Databases for Multimedia
In object databases, data remains in RMS or flat files. Object databases can provide the fastest
route to multimedia support. Object programming embodies the principles of reusable code
and modularity. This will ease future maintenance of these databases. Object database
capabilities such as message passing, extensibility, and the support of hierarchical structures,
are important for multimedia systems. We can develop the application fastest class definitions.
ODBMSs are extensible. They allow incremental changes to the database applications.
Extensibility: Extensibility means that the set of operations, structures and constraints that are
available to operations are not fixed, and developers can define new operations, which can then
be added as needed to their application.
Object-oriented software technology has three important concepts. They are:
Encapsulation: It is the ability to deal with software entities as units that interact in pre-defined
and controllable manner, and where the control routines are integral with entity.
Association: It is the ability to define a software entity in terms of its differences from another
entity.
Classification: It is the ability to represent with a single software entity a number of data items
that all have the same behaviour and the same state attributes.
Object orientation helps to organize the software in a more, modular and re-usable manner.
Encapsulation allows for the development of open systems where one part of the application
does not need to know the functioning of other part. It also provides autonomy; Autonomy
means we can interface to a variety of external programs can be built in one class of objects
and the storage of the data in another class of objects.
Database Organization for Multimedia Applications
Data organization for multimedia systems has some key issues. They are:
1. Data independence.
2. Common distributed database architecture.
3. Distributed database servers.
4. Multimedia object management.
Data Independence
Flexible access by a number of databases requires that the data be independent from the
application so that future applications can access the data without constraints related to a
previous application.
Key features of data independent designs are:
1. Storage design in independent of specific applications.
2. Explicit data definitions are independent of application program.
3. Users need not know data formats or physical storage structures.
4. Integrity assurance in independent of application programs.
5. Recovery in independent of application programs.
Distributed Data servers
Distributed database servers are a dedicated resource on a network accessible to a number of
applications. The database server is built for growth and enhancement, and the network
provides the opportunity for the growth of applications and distributed access to the data.
Multimedia Object Management
The object management system must be capable of indexing, grouping and storing multimedia
objects in distributed hierarchical optional storage systems, and accessing these objects on or
keyed basis. The design of the object management system should be capable indexing objects
in such a manner that there is no need to maintain multiple storage copies.
Multimedia transactions are very complex transactions. We define a multimedia transaction as
the sequence of events that starts when a user makes a request to display, edit, or print a hyper
media document. The transaction is complete when the user releases the hypermedia document
and stores back the edited versions or discards the copy in memory (including virtual memory)
or local storage.