whitepaper_mpeg-if_understanding_mpeg4

Understanding MPEG-4:
Technologies, Advantages, and Markets

An MPEGIF White Paper

MPEG-4: The Unified Framework of High-Performance,
Open International Standards Enabling New Services and
Greater Quality for Video, Audio, and Multimedia

Building on the enormous success of the MPEG-1 and MPEG-2
standards, MPEG-4 offers new audio and video codecs with twice
the performance of MPEG-2, enabling new services ranging from
HDTV satellite broadcasting to mobile video and games on
handheld devices.
Understanding MPEG-4:
In addition to excellent audio and video codecs, MPEG-4 offers a
comprehensive unifying framework of tools to combine them with
Technologies, Advantages, and Markets
2D and 3D graphics, animation, text, and other objects in interactive
rich media experiences. An MPEGIF White Paper
In this paper, you will learn about:

§ What MPEG-4 Actually Is

§ Why Manufacturers and System Operators Have Chosen MPEG-4

§ Applications and Markets for MPEG-4

§ How MPEG-4 Relates to Other Standards

§ What's in the MPEG-4 Standard and How It Evolves

§ MPEGIF and the MP4 Logo Program

§ Answers To Common Questions on MPEG-4:

- Licensing

- How MPEG-4 Relates to Other Codecs

- Interactive TV and MHP

MPEG-4: The Unified Framework of High-Performance,
The MPEG Industry Forum
http://www.mpegif.org
Open International Standards Enabling New Services
+1 510-744-4025 phone
and Greater Quality for Video, Audio, and Multimedia
+1 510-608-5917 fax

Copyright © 2005, The MPEG Industry Forum
ALL RIGHTS RESERVED

Document Number mp-in-40182

Copies of this document may be freely downloaded from MPEGIF web site at
http://www.mpegif.org

For further reading please refer to the MPEGIF resources page at
http://www.mpegif.org/resources.php.

This document presents an overview of the technology and business landscape -
the applications, advantages and benefits of MPEG-4 - The Media Standard.
Comments are welcome at: papers@mpegif.org

Editor: Robert Bleidt, Fraunhofer IIS

Contributors: Glenn Bulycz, Apple
Farzin Aghdasi, Pelco
Mikael Bourges-Sevenier, Mindego

This document is based on the original 2002 white paper produced by Martin
Jacklin from contributions by Tim Schaaff of Apple, Sebastian Moeritz of dicas
digital image coding, Yuval Fisher of Envivio, U. Georg Ohler of Fraunhofer IIS,
David Price and Fadi Malak of Harmonic, Jennifer Toton of iVAST, Rob Koenen
of InterTrust, Jan van der Meer of Philips, Olivier Avaro of France Telecom,
Professor Klaus Diepold of the Technical University of Munich, and other
members of the MPEGIF Board of Directors and Marketing Workgroup.

Trademarks used herein are the property of their mark holders and are used for
illustrative purposes only. MPEGIF thanks Apple Computer, Envivio, Inc.,
Harmonic, NTT DoCoMo, Panasonic, Pelco, Samsung, Streamcrest Associates,
Sony, Tandberg TV, and others for the use of the photographs and illustrations
they have provided.

Table of Contents
What Is MPEG-4? .......................................................................................... 1

A Unified Framework for Advanced Multimedia .......................................1

Excellent Conventional Codecs .................................................................1

Framework for Rich Interactive Media ......................................................3

Why Have Manufacturers And Operators Chosen MPEG-4?.......................... 5

Excellent Performance ...............................................................................5

Open, Collaborative Development to Select the Best Technologies...........6

Competitive but Compatible Implementations..........................................6

Lack of Strategic Control by a Supplier......................................................8

Public, Known Development Roadmap.....................................................9

Encode Once, Play Anywhere..................................................................10

Flexible Integration with Transport Networks..........................................10

Established Terms and Venues for Patent Licensing .................................10

MPEG-4 Markets and Applications .............................................................. 12

Television Broadcasting ...........................................................................12

IP-based Television Distribution..............................................................12

Portable Gaming .....................................................................................13

Mobile Communication and Entertainment............................................13

Internet Streaming...................................................................................13

Packaged Media.......................................................................................14

Video Conferencing ................................................................................14

Home Networking and PVRs..................................................................14

Digital Still Cameras and Convergence Devices.......................................15

Satellite Radio .........................................................................................15

Security................................................................................................... 15

MPEG-4 and Related Standards ................................................................... 17

ISO/IEC’s Moving Picture Experts Group (MPEG).................................... 17

MPEG-1................................................................................................. 17

MPEG-2................................................................................................. 17

MPEG-4................................................................................................. 17

MPEG-7 and MPEG-21......................................................................... 18

MPEG-7................................................................................................. 18

MPEG-21............................................................................................... 19

Related Standards........................................................................................ 19

Internet Engineering Task Force (IETF).................................................. 19

3rd Generation Partnership Project (3GPP and 3GPP2) ......................... 20

Internet Streaming Media Alliance (ISMA) ............................................. 20

The DVB Project .................................................................................... 20

The MPEGIF Interoperability and Qualification Programs .......................... 22

MPEG-4 Technical Overview ....................................................................... 24

What are the parts of the MPEG-4 standard?.......................................... 24

What are profiles and levels?.................................................................... 27

MPEG-4’s Rich Multimedia Framework ................................................. 28

The Importance of Interoperability ......................................................... 29

Responsible Upgrades in MPEG-4.......................................................... 30

Clarifying Common Questions..................................................................... 32

Who licenses MPEG-4 technology? ........................................................ 32

What is the role of MPEGIF in licensing?............................................... 33

With the release of Part 10 AVC, is Part 2 video coding obsolete? ........... 33

What is the relationship between MPEG-4 Visual and the DivX codec?. 34

Is Microsoft Windows Media an MPEG-4 codec?................................... 34

The future is all downloadable software codecs, why do we need a standard?
................................................................................................................34

Is MPEG-4 based on QuickTime? ...........................................................34

I read a benchmark of MPEG-4 where it did poorly, how can you claim it
is higher performance?.............................................................................35

How does MPEG-4 Compare to Other Internet Media Formats?............35

How Will MPEG-4 Be Used in Interactive TV? ......................................37

What is the Difference between MPEG-4 and MHP?..............................38

How Does MPEG-4 Compare to SMIL and SVG?..................................40

MPEG-4’s Textual Format: XMT ............................................................41

The MPEG Industry Forum ......................................................................... 43

Join the forum.........................................................................................44

Help Drive Success ..................................................................................44

whitepaper_mpeg-if_understanding_mpeg4

What Is MPEG-4?

A Unified Framework for Advanced Multimedia
MPEG-4 is a family of open international standards that provide tools for the
delivery of multimedia. These tools include both excellent codecs for
compressing conventional audio and video, and those that form a framework for
rich multimedia – combinations of audio, video, graphics, and interactive
features.

Excellent Conventional Codecs
MPEG-4’s conventional audio and video codecs provide the highest quality and
compression efficiency available today, and have been adopted as the foundation
of many new media products and services.

MPEG-4’s latest video codec is AVC, the Advanced Video Codec. Also
identically standardized as ITU H.264, the AVC codec represents the latest
developments in video coding, offering a typical compression rate half that of
MPEG-2 for similar perceived quality. This dramatic improvement has led to
AVC becoming the new standard for video transmission, being employed in
most new video products and services where quality and compression efficiency
are paramount. New HDTV satellite broadcasting and DSL video services will
use AVC, as will the Sony PlayStation Portable and Apple QuickTime 7 player.
AVC will also be used in video broadcasting to mobile handsets using the
DVB-H, DMB, and MediaFlo systems, and specified in the HD-DVD and
BluRay high-definition optical disc standards.

AVC is intended to be practical when implemented on the latest generation of
high-performance hardware and processors. For applications where hardware cost
or power considerations make implementing AVC difficult, MPEG-4 offers the
Simple and Advanced Simple Profile codecs. These codecs offer good
performance while using less complex encoder and decoder architectures. They
are commonly used for 3G wireless videophony, digital still camera or
convergence devices, and for security or intranet video applications.

These codecs are usually coupled with AAC, MPEG-4’s family of
general-purpose audio codecs. The core AAC codec offers excellent quality at
stereo bitrates above 128 Kb/s. Compatible extensions to AAC, the HE-AAC

Understanding MPEG-4: Technologies, Advantages, and Markets 1

and HE-AAC v2 codecs, improve its quality at lower bitrates, while maintaining
compatibility with existing AAC decoders.

AAC is being used not only in television broadcasting, where it is the codec for
Japan’s ISDB digital TV system, but also in music players and distribution
services, such as Apple’s iPod and iTunes.

The low-bitrate quality improvements of the HE-AAC codec has enabled new
digital music broadcasting services, such as XM satellite radio and the music
download services of mobile carriers KDDI and Orange. The latest version,
HE-AAC v2, incorporating MPEG-4’s Parametric Stereo tool, has just been
standardized by 3GPP for future music streaming services.

Codec Features Typical Applications and Users

AVC – Advanced Video Highest-Performance video HDTV Broadcasting – DirecTV, bSkyb,
Codec codec for demanding Premiere
(MPEG-4 Part 10) applications Mobile Multimedia – DMB, DVB-H,
MediaFlo systems
Internet Video – Apple QuickTime
Gaming – Sony PSP UMD Disc

SP - Simple Profile, High-Performance video codec 3G Wireless Videophony – DoCoMo,
ASP - Advanced Simple with scalability and error Hutchinson-Whampoa
Profile resilience features Intranet Video – Envivio, vBrick
(MPEG-4 Part 2) Internet Video – Apple QuickTime
Digital Cameras – Panasonic, Samsung,
Sanyo

AAC – Advanced Audio High-Performance audio codec Portable Music – Apple iPod, iTunes
Codec for excellent quality at moderate DTV Broadcasting – ISDB (Japan)
bitrates

HE-AAC – High-Performance audio codec Satellite Radio – XM Radio
High-Efficiency AAC for superior quality at bitrates Mobile Handset Download – KDDI,
below 48 Kb/s Orange

HE-AAC v2 Highest-Performance audio Mobile Handset Streaming - 3GPP
(HE-AAC + Parametric codec for excellent quality at
Stereo) bitrates below 48 Kb/s

Table 1. Popular MPEG-4 Audio and Video Codecs

2 The MPEG Industry Forum

Framework for Rich Interactive Media
MPEG-4’s rich media tools include codecs for combining audio and video with
text, still images, animation, and both 2D and 3D vector graphics into
interactive and personalized media experiences. MPEG-4 includes both a
scripting language for simple interaction and a Java variant, MPEG-J for more
elaborate programming.

While video and audio are coded with MPEG-4’s excellent conventional codecs,
graphics, text and synthetic objects have their own coding, rather than forcing
them into pixels or waveforms. This makes representation more efficient and
handling them much more flexible. Additionally, MPEG-4 offers revolutionary
synthetic content tools, like structured audio (a language for describing sound
generation), animated faces and bodies, 2D and 3D meshes, and vector-based
graphics. This allows separate interaction with each object and sharp rendering at
all bandwidths, e.g. subtitles that can be turned on and off and remain sharp
even at high compression rates.

Figure 1. MPEG-4’s rich media tools allow the tight integration of
graphics, stored video, and a live presentation in this educational
application. Courtesy Envivio, Inc.

Because MPEG-4’s rich media tools may be object-based, it is possible to
construct multimedia scenes which revolutionize the possibilities of interactive
media. Authors can allow end-users to interact with objects in the scene: to
change the color of a car to see how it will look, to tag a player on the field and
watch their moves, or to personalize an enhanced video program. MPEG-4 also


opens up new revenue opportunities through the integration of back-office
systems including transaction and e-commerce systems.

With MPEG-4, advanced interactive programming can be authored to
seamlessly integrate audio and video with 2D and 3D objects, animation, and
interactivity. For example, a viewer can navigate a sporting event’s course from a
3D map, select information about aspects of the program, listen to commentary
within a picture-in-picture window, and watch sponsored advertising – all within
a single MPEG-4 stream supporting multiple media objects.

MPEG-4 also allows the same interactive programming to be used across
different delivery channels. The same interactive program can be used on a DVD
or delivered across a broadband network, something that was previously
impossible.

Truly a “framework of the future”, MPEG-4’s rich media tools enable the
creation of content that is difficult or impossible to achieve with the proprietary
tools available today. These tools are already in use for distributing complex
corporate and educational multimedia presentations that require tight
synchronization of video and graphics content, with entertainment applications
expected as the creative community moves towards interactive content.


Why Have Manufacturers And
Operators Chosen MPEG-4?

Excellent Performance
A codec’s compression efficiency, or the bitrate required for a given perceived
quality level, is one of its most important features. MPEG-4’s latest audio and
video codecs, AAC and AVC, provide outstanding compression efficiency,
equaling or exceeding the performance of any codecs now available.

MPEG is constantly evaluating new algorithms and technology to see if they
offer a practical and significant improvement over existing MPEG standards.
Successful techniques are then incorporated as compatible extensions or
improvements to existing standards, or if necessary are issued as complete new
standards. An example of the latter is AVC, which is sufficiently different from
MPEG-4’s Part 2 video that it required development of Part 10 of MPEG-4.

Where possible, however, MPEG’s philosophy is to extend rather than to replace.
An example is the continual improvement in audio coding efficiency enabled by
successive versions of the AAC codec. AAC was a significant improvement over
MPEG-1’s Layer 3 codec (MP3), but MPEG has extended AAC through the
techniques in the HE-AAC and Parametric Stereo codecs, while preserving
compatibility with existing AAC decoders. The upcoming MPEG Surround
codec will improve coding efficiency even further, offering a four to six times
improvement in bitrate over MP3 for multi-channel content.

350
5.1 Surround
300
250 Stereo
200
Kb/s
150
100
50
0
MP3 AAC HE-AAC MPEG
Surround

Figure 2. The evolution of MPEG Audio codecs has produced a
5 to1 reduction in bitrate for similar perceived quality.


Open, Collaborative Development to
Select the Best Technologies
Key to this evolution in performance is MPEG-4’s open and collaborative
development process, where new techniques are proposed and evaluated not on
business considerations, but strictly on their technical merit in a careful,
thorough, and objective set of experiments and tests.

This development, like all international standards, is carried out in a public
process by delegations from many countries, representing their top scientists and
engineers. In MPEG’s case, this development is done by the Motion Picture
Experts Group, a working group of the ISO and IEC standards organizations,
and typically involves hundreds of companies from tens of countries.

Once developed and proposed, new standards go through careful review and
approval to insure that they are completely and unambiguously specified, so they
can be implemented solely by following the standard. Additionally, test results,
frameworks, and the source code to reference software implementations are made
available so anyone can see how the codec was implemented, study why design
choices were made, and understand how the codec operates.

Competitive but Compatible Implementations
With MPEG-4, manufacturers and technology developers collaborate to create
the standard, but they compete to implement it. In MPEG-4, the bitstream
format and decoder operations are carefully specified to allow complete
interoperability, but manufacturers are not restricted in how they design
encoders and decoders internally.

Decoder manufacturers, particularly those building decoders in hardware chips
or devices, can take comfort in the knowledge that the bitstream format will not
change. Encoder manufacturers can use their proprietary techniques and
experience to continually improve coding efficiency and quality, knowing that
their output bitstreams will be rendered by all decoders.

These same principles apply to MPEG-2, where history has shown them to be
very worthwhile. Early MPEG-2 encoders required a 6 Mb/s bitrate for
acceptable picture quality, while today’s encoders can operate at 2 to 2.5 Mb/s,
saving 60-65% of the original bandwidth costs, without any changes to the
installed base of MPEG-2 decoders. A similar experience curve is expected for
MPEG-4.


For users and operators, this competition offers several benefits. Besides potential
bandwidth savings as encoders improve, they can choose between multiple
suppliers for each component in their signal chain – encoders, servers,
transmission equipment, and decoders – with different prices, performance levels,
and business terms. Once chosen, they are not locked-in to a specific supplier,
but can supplement or replace their components with MPEG-4 equipment from
another supplier secure in the knowledge that they will operate together.

Harmonic’s view of improvements in the coding
efficiency of their MPEG-2 encoders over time,
expressed as the bit rate required for “broadcast
quality” with moderately difficult source material

Figure 3. Improvements in coding efficiency are facilitated by MPEG-based competition. Courtesy Harmonic.


9

1st commercial
8
encoder
Bit rate required for
7 constant broadcast
quality
Bit rate (Mbit/s)

6

5

4
Size reduction
3 System 3000

2
New generation
Evo5K
1

0
1994 1995 1996 1997 1998 1999 2000 2001 2002

Figure 4. Improvements in coding efficiency are facilitated by MPEG-based competition. Courtesy Tandberg TV.

Lack of Strategic Control by a Supplier
In addition to the benefits of intense day-to-day competition among
manufacturers to best implement the standard, MPEG-4 also benefits from the
lack of strategic control by a supplier, as is common with proprietary codecs.

The specifications and tools needed to implement MPEG-4 are available to
anyone, and there are no restrictions on how or where it is used, or limits on
what hardware or software can be employed. MPEG-4 is not tied to any
particular application program or service offering, and it may be used with any
DRM system, operating system, software platform, or hardware chip, without
any bundling, limitations, or conditions being placed on the user.

The business failure of an MPEG-4 supplier does not put the standard in
jeopardy, and users may shift new purchases of MPEG-4 equipment to other
suppliers without losing their investment in equipment or software already
purchased. Conversely, the success of a MPEG-4 supplier does not mean research
and development efforts on the core functionality of the codec will be shifted to
product extensions into other areas, as is common when a proprietary standard
matures.


A user’s risk is also reduced through the stability of the standard. There are no
time limits on the use of the MPEG-4 standard, and non-discriminatory patent
licenses for MPEG-4 have guaranteed terms and renewals, along with strict
limits on changes in royalties.

For MPEG-4 manufacturers, there are no restrictions, favoritism, or delays in
accessing specifications as they are developed, and no requirements for
certification or approval from another party before it is produced. (MPEGIF
does operate voluntary interoperability and conformance programs, but these are
not limits on what a manufacturer can do.)

Public, Known Development Roadmap
Since MPEG-4 development occurs in a public, international forum within
established procedures, its development process is known and can be monitored
by manufacturers and operators to aid their planning and development. Indeed,
the breadth of the MPEG-4 standard offers both excellent performance today
and a long-term framework for the future, without the switching costs or
technology locks incurred with proprietary formats.

Consumption
Creation
Contribution Distribution

TV

Playout
video
Production Storage terrestrial
PC

Encode
audio interaction Packaged satellite
DVD

2D images animation MPEG-4 wireless
server
CellPhone
PDA

3D images BIFS IPMP Cable
DRM
Game Console

text Retail
Radio

Figure 5. The MPEG-4 ecosystem liberates multimedia for
delivery across any network to any user of any device


Encode Once, Play Anywhere
MPEG-4 is truly a multi-platform standard. It operates as easily on embedded
hardware as it does in software, and it is designed to be independent of any
transport medium, working as well over wireless networks as on legacy MPEG-2
transport streams. This means content providers can avoid duplicate production
for different delivery media by encoding it once in MPEG-4, instead of in several
delivery formats.

Flexible Integration with Transport Networks
MPEG-4 is designed to be transport independent, operating over both IP-based
and traditional MPEG-2 transport stream networks. Thus, it can be easily
transmitted over the Internet to PCs, or over MPEG-2 cable or satellite networks
to set-top boxes. MPEG, MPEGIF, IETF, DVB, ISMA, and other organizations
have developed standards for the carriage of MPEG-4 content

Further, MPEG-4 has been designed from the beginning to operate over lossy
networks such as mobile 3G or WiFi, and incorporates scalability, error
concealment, and error recovery techniques in its codecs so that program quality
is preserved when changes in bandwidth or signal dropouts occur.

Popular transport protocols for MPEG-4 include:

The MPEG-4 File Format. This is the standardized way to store
MPEG-4 content as computer files, and includes MPEG-4’s hinting
mechanism to pre-packetize content to increase video server efficiency.

The IETF RTP Protocols. This is a family of real-time streaming
protocols standardized by Internet RFCs that allow MPEG-4 to be
streamed over IP networks.

MPEG-4 over MPEG-2 Transport Streams. This allows MPEG-4
content to be combined with MPEG-2 programming, if desired, and
carried over existing cable TV or satellite networks and equipment.

Audio-only Protocols. MPEG-4 also includes several protocols such as
LATM/LOAS, ADIF, and ADTS that are popular for carrying
audio-only programming.

Established Terms and Venues for Patent Licensing
Audio and video coding is a field where patents have been asserted and licensed
for decades, and many organizations have substantial IP portfolios that apply to


media codecs. The use of most MPEG codecs has involved the payment of
royalties through patent pools to patent holders. With MPEG-4’s patent pool
organizations, a manufacturer or operator can be assured that he has a license to
essential patents from the major IP holders with fixed, renewable prices and
conditions.

Though patent pools and royalties have been attacked as a disadvantage of
MPEG-4, it is actually an advantage compared to the patent uncertainty
surrounding proprietary or “royalty free” media codecs, where patents from other
organizations could be later asserted.

Though there is always a chance that an unknown inventor may someday assert a
patent against an MPEG-4 manufacturer or user, it is more likely the inventor
will join an established pool where the costs of licensing and royalty collection
can be shared with many other licensors. His royalties would be distributed as a
share of the pool’s, and a user’s royalties would be unchanged.


MPEG-4 Markets and Applications

Television Broadcasting
MPEG-2 is the current standard for digital television production and
distribution, but doesn’t offer enough compression for transmitting the hundreds
of channels cable and satellite TV consumers will expect in high-definition. As
more programming moves to HDTV, cable and particularly satellite distribution
systems will begin using more efficient compression codecs such as AVC. Already,
major satellite TV operators such as DirecTV, bSkyb, and Premiere have
announced new high-definition services using AVC.

IP-based Television Distribution
With the rise in availability of DSL and other broadband IP networks, IP-based
television distribution is now an alternative to cable, satellite, or over-the-air
broadcasting. Because IP TV is inherently two-way, improved interactive and
video-on-demand services are easily provided. But the limited bandwidth of
many of these networks requires high-efficiency coding such as AVC to allow for
multiple sets to be served over a single connection. This is also a natural
application for AVC.

Thousands of AVC encoders have already been deployed in the field by both
established and new encoder suppliers and a number of system operators are
undertaking extensive trials with a view to launching services soon, most notably
UK-based Video Networks Ltd., which has recently launched the first full
revenue-bearing service based on AVC.

Figure 6. Sony’s PlayStation Portable uses MPEG-4 AAC and
AVC. Courtesy Sony.


Portable Gaming
Gaming devices are typically closed systems, where
manufacturers are free to use whatever proprietary
technologies meet their needs. Notably, Sony’s new
portable gaming platform, the PlayStation Portable,
uses MPEG-4 Simple Profile to view video stored
on memory cards, and AVC for video playback
from its internal UMD optical disc drive, along
with AAC audio encoding.

Mobile Communication and
Entertainment
Figure 7. 3G Mobile Handset with
MPEG-4 Simple Profile video has been a part of MPEG-4 Two-Way Video Capability.
international 3G mobile standards since Courtesy NTT DoCoMo

videophone service was introduced in 2001. Today,
mobile operators such as NTT DoCoMo and
Hutchinson-Whampoa have deployed millions of
MPEG-4 handsets to enable two-way video calls or
watch video programming over 3G networks.

Additionally, MPEG-4 HE-AAC audio coding has been employed in several
music download services, such from operators KDDI and Orange, and HE-AAC
v2 with the Parametric Stereo tool has been selected by 3GPP as a new standard
for music streaming.

The next phase of mobile entertainment is broadcasting to handsets, as opposed
to point-to-point connections, to reduce the bandwidth and expense required.
MPEG-4’s AVC is a part of all three of the major systems being deployed for this
service: DVB-H, DMB, and MediaFlo.

Internet Streaming
MPEG-4 was designed from the start for streaming, and it is now featured in
systems from several manufacturers. Apple supports MPEG-4 Simple Profile
video and AAC audio in its QuickTime platform and now includes support for
AVC in QuickTime 7. Real Networks supports decoding of MPEG-4 content,
and the popular DivX codec is also MPEG-4 compliant. A number of MPEG-4
vendors offer plug-ins for Microsoft’s Windows Media Player that enable users to
also watch MPEG-4 content in this player.


Apple supports MPEG-4 natively in their
QuickTime Streaming Server and Darwin
Streaming Servers, available for free
download or as source code, and in its new
QuickTime 7 encoder and player. At the
time of writing, QuickTime 6 has been
downloaded one billion times since its
introduction in summer 2002. Real
Networks also supports MPEG-4 on its
Helix servers.

Figure 8. iPod portable music player using
MPEG-4 AAC audio coding. Packaged Media
Courtesy Apple.
As industry consortia prepare standards for
the next generation of optical discs for video
distribution, the performance of AVC has led
to its adoption in both of the leading disc
formats. AVC is a codec specified for both
the HD-DVD and Blu-Ray standards.

Video Conferencing
Video conference terminals are a natural application for MPEG-4’s AVC codec,
and recently market leaders Polycom, Tandberg, and Sony have all introduced
AVC support in their products. AVC will enable increased video quality over the
same connection compared to the earlier H.261 and H.263 standards, or a
half-bandwidth connection may be used for similar quality.

Home Networking and PVRs
Though traditional standard-definition cable and satellite TV is likely to remain
MPEG-2 based, given the huge installed base of MPEG-2 receivers, the
consumer’s home is moving from islands of digital video, as with current set-top
boxes and personal video recorders, to interconnected home networks. In the
future, consumers will be able to seamlessly switch from watching programming
on their traditional TV to viewing it on a phone or PDA device, or downloading
it to their car. Given the bandwidth constraints of home wireless networks and
devices, content will likely be transcoded to new codecs such as MPEG-4 AVC
when sent around the home. Standards consortia such as DVB and DLNA have
adopted AVC for this purpose, as well as for future IP network delivery.


Digital Still Cameras and Convergence Devices
Digital still cameras now include “movie
modes” for capture of short video sequences,
and with the new affordability of
high-capacity flash memory, it is possible to
build camera-like Mobile Content
Convergence Devices that include the
functions of a still camera, camcorder, and
music player in one device. Given its
compression efficiency and multi-platform
Figure 9. Digital Convergence Device using
support, plus its freedom from platform MPEG-4 for flash-memory based recording.
bundling requirements, MPEG-4 is an ideal Courtesy Panasonic.
fit for these devices, and manufacturers
Panasonic, Sanyo, Samsung, Audiovox, and
Archos, among others, have all introduced
portable convergence devices using
MPEG-4.

Satellite Radio
MPEG-4’s HE-AAC audio codec, as well as its AAC and BSAC audio codecs,
have been employed in several systems for satellite radio and multimedia
broadcasting. XM Radio employs HE-AAC coding, while Digital Radio
Mondiale employs HE-AAC as well as the MPEG-4 speech codecs CELP and
HVXC. Korea’s DMB system will employ AAC and BSAC audio coding and
AVC video.

Security
Traditional video surveillance
systems have employed time-lapse
video recorders and more recently
small Digital Video Recorders using
some variant of motion JPEG or
wavelet compression. These devices
often must limit the video
resolution and frame rate to provide
a reasonable recording time, and
Figure 10. Combination CDMA2000 mobile handset
they require proprietary video
and receiver for Satellite DMB service using MPEG-4
players or browser plug-ins for users AVC and AAC. Courtesy Samsung.


to view stored content. Additionally, users desire better resolution and frame
rates to examine security alarms or events.

Thus, several security video firms have begun offering improved video systems
with recording of MPEG-4 video at full D1 resolution and frame rates of 25 or
30 frames per second. MPEG-4 coding dramatically reduces the storage costs: A
typical installation with 3000 cameras might use 800 TB of storage to support
15 days of recording. MPEG-4’s interoperability also allows users to combine
equipment from different manufacturers in their systems and still be able to
export the video in a universally readable format.

Figure 11. MPEG-4 coding enables this casino to store weeks of video with better quality for forensic
analysis. Courtesy Pelco.


MPEG-4 and Related Standards

ISO/IEC’s Moving Picture Experts Group (MPEG)
The Emmy Award-winning MPEG committee has built the foundations of
digital content delivery with its highly successful standards since 1991 with the
release of the MPEG-1 standard. Today, close to a billion devices support one of
MPEG’s standards.

MPEG-1
MPEG-1 was originally intended to enable video to be compressed for storage
on optical disks of the era, and led to the Video CD format, hugely popular in
Asia for video content distribution. MPEG-1 was also used in early
video-on-demand systems and interactive media and has become almost
universally supported by PC media players. MPEG-1 audio offers Layer 2 coding,
used in the DAB digital radio standard, and Layer 3 coding, the MP3 format.

MPEG-2
MPEG-2 is the audiovisual standard most widely used for entertainment video
applications. MPEG-2 enables digital television and DVDs, with hundreds of
millions of MPEG-2 decoders deployed in digital satellite and cable set-top
boxes, DVD players and PCs. It is a more powerful format than MPEG-1,
capable of achieving higher compression ratios and supporting interlaced video.
MPEG–2 video decoding and encoding are more CPU-intensive than for
MPEG-1.

Virtually every image you see on television today, even on an analog receiver, has
at some point been coded and decoded in MPEG-2.

MPEG-4
The subject of this paper, MPEG-4 is the successor standard to MPEG-2,
extending its application to IP-based and lossy networks, to rich multimedia
presentations, and to objects and interactivity. MPEG-4 also has improvements
in the core audio and video codecs over MPEG-2, notably the AVC, AAC, and
HE-AAC codecs which have enabled so many new products and services.


Relationship of the MPEG Media Standards
What is it?
“IT” Metadata
Content or Essence MPEG-7
Title, Author, Dates,
AVC, AAC Dialogue, Thumbnails, Melody
MPEG-4
MPEG-2
How can I use it?
Rights
Video, Audio,
Graphics, Interactivity MPEG-21
Digital Item Identification
Rights Expression Language

Figure 12. The Metadata and Rights Standards MPEG-7 and MPEG-21 are designed to
complement coding standards MPEG-1, -2, and -4. Courtesy Streamcrest Associates.

MPEG-7 and MPEG-21
The related standards MPEG-7 and MPEG-21 are additional toolsets which
extend the functionality of MPEG and interface tightly with MPEG-4 to create
new content management features. MPEG has taken care that MPEG-4
integrates well with MPEG-7 and MPEG-21. MPEG-7 descriptions and
metadata can be carried as MPEG-4 streams, and MPEG-21’s specifications are
being written to complement MPEG-4’s content representation.

MPEG-7
MPEG-7 is a recently finalized standard for description of multimedia content.
It will be used for indexing, cataloging, advanced search tools, program selection,
smart reasoning about content and more. The standard comprises syntax and
semantics of multimedia descriptors and descriptor schemes. MPEG-7 is an
important standard because it allows the management, search and retrieval of
ever-growing amounts of content stored locally, on-line, and in broadcasts.

For example, an overheard song can be captured by a mobile handset and
converted locally to an MPEG-7 description. This description can then be
transmitted to a search engine which offers a download of the song. Another
example is facilitating complex editing tasks based on rich, hierarchical
descriptions of the raw footage. A third example is broadcasting MPEG-7
descriptions along with TV content, allowing TVs and Personal Video Recorders
(PVRs) to autonomously choose programs based on user preference. MPEG-4


has a built-in MPEG-7 data type, allowing the close integration of MPEG-7
descriptions and MPEG-4 content.

MPEG-21
MPEG-21 is an emerging standard with the goal of describing a “big picture” of
how different elements to build an infrastructure for the delivery and
consumption of multimedia content – existing or under development – work
together. The MPEG-21 world consists of Users that interact with Digital Items.
A Digital Item can be anything from an elemental piece of content (a single
picture, a sound track) to a complete collection of audiovisual works. A User can
be anyone who deals with a Digital Item, from producers to vendors to
end-users.

Interestingly, all Users are “equal” in MPEG-21, in the sense that they all have
their rights and interests in Digital Items, and they all need to be able to express
those. For example: usage information is valuable content in itself; an end-user
will want control over its utilization. A driving force behind MPEG-21 is the
notion that the digital revolution gives every consumer the chance to play new
roles in the multimedia food chain. While MPEG-21 has lofty goals, it has very
practical implementations.

MPEG-21 includes a universal declaration of multimedia content, a language
facilitating the dynamic adaptation of content to delivery network and
consumption devices, and various tools for making Digital Rights Management
more interoperable.

MPEG-21 is about managing content and access to content. Even with fully
interoperable coding there are still additional steps to insure that all the features
of different networks work together. MPEG-21 is a framework which allows
interoperability and portability of content.

Related Standards

Internet Engineering Task Force (IETF)
The Internet Engineering Task Force (IETF) is a large, open international
community of network designers, operators, vendors, and researchers concerned
with the evolution of the Internet architecture and the smooth operation of the
Internet. The IETF addresses transport/session protocols for streaming media.
Their work relevant to MPEG-4 includes audio-video elementary stream


payloads, generic MPEG-4 payload formats, the Real Time Protocol (RTP) - a
transport protocol for real-time applications, an RTP profile for audio and video
conferences with minimal control, and the Real Time Streaming Protocol
(RTSP).

3rd Generation Partnership Project (3GPP and 3GPP2)
rd rd
The 3 Generation Partnership Project (3GPP) defines standards for 3
generation mobile networks and services evolving from GSM-based systems.
3GPP2 does the same for CDMA-based systems. Both give much attention to
mobile multimedia.

In their Wireless terminal specification, 3GPP and 3GPP2 use MPEG-4 Simple
Visual profile for video, MPEG-4 file format for multimedia messaging, and
RTP/RTSP for streaming protocols and control. 3GPP has also recently adopted
HE-AAC v2 for music delivery.

Internet Streaming Media Alliance (ISMA)
The Internet Streaming Media Alliance creates a set of vertical specifications for
Internet Streaming. ISMA has chosen specific MPEG-4 Audio and Visual
Profiles and Levels (see “What Are Profiles and Levels?”), and augmented these
with IETF transport specifications to create cross-vendor interoperability for
video on the Internet. ISMA uses the MPEG-4 file format for file storage, and
the IETF protocols RTP and RTSP for streaming protocols and control.

There are two ISMA specifications: ISMA 1.0 is mature, conformance testing is
in place, and ISMA-compliant implementations are available. It uses MPEG-4
Simple Visual and Advanced Simple profiles for video, and MPEG-4 High
Quality Audio Profile for audio.

ISMA 2.0 updates the codec suite to use Advanced Video Coding (MPEG-4 Part
10 AVC or H.264) Baseline, Main, and High video profiles, and High-Efficiency
AAC (HE-AAC) audio. Implementations are emerging and conformance testing
is available.

The DVB Project
DVB’s main transmission standards, DVB-S for Satellite, DVB-C for Cable and
DVB-T for terrestrial, are used worldwide for television transmission and are the
basis for many alternative standards. While DVB has been historically based on


MPEG-2 transport streams and video coding, DVB has recently adopted
MPEG-4 AVC and HE-AAC codecs for future systems. MPEG-4’s rich media
framework is also closely related to DVB’s Multimedia Home Platform (MHP)
standard for digital interactive television middleware.


The MPEGIF Interoperability and
Qualification Programs

MPEGIF operates two programs to aid members in establishing interoperability
of their MPEG-4 implementations.

The Interoperability Program has been active since 2000 and currently conducts
three test rounds per year. It is intended as an informal, confidential program
where members can test their initial designs and implementations, and resolve
interoperability problems.

The program organizes and executes interoperability tests around MPEG
technologies, such as audio codecs, video codecs, file format, transport, and
other systems elements. The activities of the program are worked out by
consensus to the benefit of all members. For example, new features and test
points are regularly added based on the interest of the companies who participate
in the tests.

The main work of the group is done through conference calls and a private
reflector. Typically, calls are held every two weeks during an active test round.
Calls may be held less frequently or not at all during the periods between active
test rounds.

The Logo Qualification Program was established in 2004 and offers members a
procedure by which companies may qualify encoder & decoder products and
earn the right to label their products with a logo provided by MPEGIF.
MPEGIF also operates a website with a searchable directory of Qualified
Products. The program is designed to meet the following goals:

Promote “MPEG-4 Brand Awareness”

Provide marketing value for MPEGIF members who participate in the
program.


Encourage participation in the MPEGIF Interoperability Working
Group, and in the Forum in general.

Provide an environment for companies to discover and resolve
interoperability problems.

Encourage growth of the industry by building confidence that products
will interoperate.

The Logo Qualification Program is tightly linked to the MPEGIF
Interoperability Testing program, and in fact participating companies are
required to complete at least one round of testing in the Interoperability WG
before submitting their product for qualification. An additional verification step
is also required:

Decoder companies must verify that the product correctly decodes all
streams in the relevant Qualification Stream Set. This set is a
combination of published conformance streams and streams posted to
and tested within the Interoperability WG.

Encoder companies must obtain Qualification Endorsements from three
other companies with qualified decoder products. The endorsement is a
statement from the decoder company that a relevant stream set generated
by the encoder company was decodable without problems.

Note that this is essentially a Self-Qualification process. In order to complete the
qualification procedure, a company will sign an agreement stating that the
testing criteria have been met. The program stops short of full “Certification,”
which normally would require an independent third party to administer
certification tests and make judgments on whether the product passes or fails. In
this way MPEGIF is providing a framework and encouraging a base level of
testing, however the details of testing and verification are the responsibility of the
individual company, not that of MPEGIF.

The program will initially focus on MPEG-4 Audio and Video
encoders/decoders, but may eventually be expanded to include other MPEG
standards. Support for Profile/Level combinations is designed to be flexible in
that new combinations may be added at any time based on interest from the
membership, and also based on the evolution of testing in the Interoperability
WG. Initially, the program includes AVC Main, Baseline, and Extended visual
and High Efficiency AAC audio profiles.

The Qualified Products Directory may be searched, and additional information
on the program found at http://logo.mpegif.org.


MPEG-4 Technical Overview
This section presents an overview of the technical advantages of MPEG-4 and
some of the latest information on extensions and the MPEG-4 upgrade path. For
a much more detailed technical overview of MPEG-4 the official “MPEG-4
Overview” is recommended. Please visit http://www.mpegif.org/resources.php
for access to this document.

What are the Parts of the MPEG-4 Standard?
MPEG-4 consists of closely interrelated but distinct individual Parts, that can be
individually implemented (e.g., MPEG-4 Audio can stand alone) or combined
with other parts.

The fundamentals of MPEG-4 are described by the parts on Systems (part 1),
Visual (part 2) and Audio (part 3), along with the new part 10 on Advanced
Video Coding. DMIF (Delivery Multimedia Integration Framework, part 6)
defines an interface between an application and the network or storage.
Conformance (part 4) defines how to test an MPEG-4 implementation, and part
5 offers a significant body of example Reference Software, that can be used to
start implementing the standard.

Part 7 of MPEG-4 defines an optimized video encoder (in addition to the
Reference Software, which is a correct, but not necessarily optimal
implementation of the standard)

Other parts in MPEG-4 include:

Part 8: Transport is in principle not defined in the standard, but Part 8
defines how to map MPEG-4 streams onto IP transport.

Part 9: Reference Hardware Description: Phase 1 - Hardware
Accelerators, Phase 2 - Optimized Reference Software integration
through Virtual Sockets

Part 11: Scene Description (BIFS) and Application Engine (MPEG-J)

Part 12: ISO Base Media File Format

Part 13 : IPMP Extensions

Part 14 : MP4 File Format (based on part 12)

Part 15 : AVC File Format (also based on part 12)


Part 16 : AFX (Animation Framework eXtension)

Part 17: Streaming Text Format

Part 18: Font Compression and Streaming

Part 19: Synthesized Texture Streaming

Part 20: Lightweight Application Scene Representation (Laser)

Part 21: MPEG-J Graphical Framework eXtension (GFX)

Part 22: Open Font Format

MPEG-4’s Framework of Tools
Visual Audio Interactivity “Transport”

Rectangular Video Audio Codecs MPEG-J / GFX .mp4
Simple Profile AAC JavaScript IPMP (DRM)
Tools
HE-AAC BIFS LATM, ATDS
Advanced Parametric Stereo (Audio-only)
LASER
Simple Profile
Tools MPEG Surround
RTP/RTSP/SDP
Natural

AVC / H.264 BSAC Authoring
MPEG-2 Transport
Twin VQ XMT Stream
Video
Scalability HILN
Audio Scalability
Video Objects
Speech Codecs
Sprite
CELP
Shape
HVXC
Texture/Stills
Graphics Synthetic Audio
Synthetic

2D / 3D Vector Text To Speech
Graphics
Structured
Face/Body © 2004, 2005 Streamcrest Associates
Audio
Animation This slide is an overview and is not meant to be technically precise.
Audio Rendering Note that RTP is defined by IETF and MPEG-2 TS in MPEG-2. As MPEG-4
AFX Tools
is “transport-agnostic” other transport standards can be used as well.

Figure 13. This is merely one way to classify MPEG-4’s toolset. Courtesy Streamcrest Associates.


objects
11. Scene Description
presentation
20. LASeR 21. GFX

5. Reference SW
4. Conformance

2. 10. 3.
decoding
Visual AVC Audio

1. Systems demux & buffer

6. DMIF transport interface

Transport not in standard

Note: there currently are 21 parts in the
bits
MPEG-4 Standard. Not all are depicted

Figure 14. The parts of MPEG-4. The arrows represent the flow of bits through the MPEG-4 system.


What are Profiles and Levels?

MPEG-4 device Scene
Graph
0,1
Profiles
0,1
Scal. Etc. 0,1 Comp.
Simple 2D
Synth speech 0,1 0,1 Complete
2D
0,1 Audio

Audio
Profiles
Core
core Compl. 2D
hybrid etc.
FA Simple 2D Object Descriptor
Compl. Profile
Visual Graphics
Profiles Profiles Main personal

Media Profiles MPEG-J Profiles

Figure 15. Profiles and Levels

MPEG-4 consists of a large number of tools, not all of which are useful in any
given application. In order to allow different market segments to select subsets of
tools, MPEG-4 contains profiles, which are simply groups of tools. For example,
the MPEG-4 Advanced Simple visual profile contains ¼ pel motion
compensation, B-frames, and global motion vectors, but it does not contain
shape coded video.

Profiles allow users to choose from a variety of toolsets supporting just the
functionality they need. Profiles exist at a number of levels, which provide a way
to limit computational complexity, e.g. by specifying the bitrate, the maximum
number of objects in the scene, audio decoding “complexity units,” etc.

The concept of MPEG-2 Video Profiles has been extended to include the Visual,
Audio and Systems parts of the standard, so that all the tools can be
appropriately “subsetted” for a given application domain.


Common Video Profiles Common Audio Profiles

Profile Features Profile Features

Simple Visual Similar to MPEG-2 High-Quality Audio Includes the most
coding popular AAC object
type, AAC Low
Advanced Simple Adds support for Complexity, and the
Visual B-frames, Global CELP speech coder
Motion Compensation,
Interlace (at levels 4-5) Low Delay Audio A variant of the AAC
codec with ~20mS
Core Visual Adds support for binary delay, suitable for
shapes (video objects) high-quality
and B-frames conferences or
conversations
AVC Baseline Low Delay, Lower
Processor Load High Efficiency AAC Adds Spectral Band
Audio Replication tool to
AVC Main Supports Interlaced improve coding
video, B-Frames, efficiency at low
CABAC encoding bitrates

AVC Extended Includes Error HE-AAC v2 Audio Adds Parametric
Resilience Tools, Stereo Tool to further
B-Frames improve coding
efficiency at low
AVC High Supports High-quality,
bitrates
High-resolution
formats for Digital
Cinema
Table 2. Common MPEG-4 Profiles

MPEG-4’s Rich Multimedia Framework
The best way to understand MPEG-4’s new multimedia paradigm is by
comparing it to MPEG-2.

In the MPEG-2 world, content is created from various resources such as video,
graphics, and text. After it is “composited” into a plane of pixels, these are
encoded as if they all were video pixels. At the playback side, decoding is a
straightforward operation.

MPEG-2 is a static presentation engine: if one broadcaster is retransmitting
another broadcaster’s coverage of an event, the latter’s logo cannot be removed,
also, viewers may occasionally see the word “live” on the screen when a
broadcaster is showing third-party live footage from earlier in the day. You may
add graphic and textual elements to the final presentation, but you cannot delete
them.


The MPEG-4 paradigm turns this upside down. It is dynamic, where MPEG-2
is static. Different objects can be encoded and transmitted separately to the
decoder in their own elementary streams. The composition only takes places
after decoding instead of before encoding. This actually applies for visual objects
and audio alike, although the concept is a little easier to explain for visual
elements. In order to be able to do the composition, MPEG-4 includes a special
scene description language, called BIFS, for Binary Format for Scenes.

The BIFS language not only describes where and when the objects appear in the
scene, it can also describe behavior (make an object spin or make two videos do a
cross-fade) and even conditional behavior – objects doing things in response to
an event, usually user input. This makes the interactivity of MPEG-4 rich
multimedia possible. All the objects can be encoded with their own optimal
coding scheme – video is coded as video, text as text, graphics as graphics –
instead of treating all the pixels as moving video, which they often really aren’t.
For applications that need more complex logic in response to an event,
ECMAScript can be used or the Java language via MPEG-J.

Recently, MPEG has redefined the scene description based on W3C's SVG
instead of VRML (Virtual Reality Modeling Language) that BIFS was based on.
This lightweight scene representation - called Laser – is geared at 2D applications
on limited resources devices such as mobile handsets.

BIFS and Laser are descriptions of a scene to compose audio-visual objects. As all
scene descriptions, they only contain a limited number of features and
sometimes features specific to some applications. To allow content creators more
freedom in the composition and logic of their applications, a programming
language is necessary. The Java-based Graphical eXtension Framework (GFX)
provides a programmatic way to compose and to render audio-visual objects.
GFX was designed around mobile entertainment applications such as 3D games
enhanced with video.

As all the coders in MPEG-4 are optimized for the appropriate data types,
MPEG-4 includes efficient coders for audio, speech, video and even synthetic
content such as animated faces and bodies.

The Importance of Interoperability
Interoperability is the capability of products from different vendors to seamlessly
work together. Interoperability is the goal of standards like MPEG-4.

Competition thrives and consumers benefit when multiple vendors’ products
interoperate. The MPEG committee’s contribution to interoperability is to


publish the specification and conformance points, such as profiles and levels.
Other groups, notably the MPEG Industry Forum and the Internet Streaming
Media Alliance go farther. Both organizations sponsor interoperability programs
and certification processes that allow vendors to test their products before they
are marketed and consumers to recognize they are purchasing a product that will
interoperate with other MPEG-4 products. Customers get choice and quality;
vendors get secure customers and a larger market.

Responsible Upgrades in MPEG-4
MPEG-4 is a dynamic standard. The first parts were published in early 1999,
and work is ongoing. Changes to existing parts of the standard are always done
in a backward-compatible way, as MPEG does not want to render already
deployed systems non-conformant. This means that changes are usually done in
the form of additions, sometimes as an amendment to an existing part of the
standard, sometimes as a new part. There are basically two types of such
additions: those that add functionality that was not present before and those that
improve existing functionality.

Vid MP
eo EG
com -4
pro
pre
Bitrate

ssio gre
ns ss
cie
nce

Simple Advanced MPEG-4
Profile Simple AVC/H.264

1998 2000 2002
Figure 16. MPEG-4 - A predictable, responsible upgrade strategy

Examples of the first kind are support for fonts, synthesized textures and the
Animation Framework eXtension (AFX). Examples of the latter are MPEG-4
Advanced Video Coding (AVC), Lightweight Scene Representation (Laser), and
MPEG-J Graphical Framework eXtension (GFX).


MPEG-4 is a toolbox, as stated earlier. As the market moves, the requirements
for the toolbox evolve as well, and the standard development work follows these
requirements. Although not at the pace that some people believe, compression
technology is still progressing, both for audio and video coding. In order for
standards like MPEG-4 to be of most use to the market, interoperability and
stability need to be combined with solid performance. MPEG standards, when
they are issued, are always state-of-the-art, created through the collaboration of
the world’s best experts.

An example of this upgrade strategy is MPEG-4 Advanced Audio Coding (AAC).
While AAC offers excellent quality at bitrates of 64 Kb/s per channel and higher,
MPEG has extended it with a technique called Spectral Bandwidth Replication,
which gives spectacular bandwidth savings for applications like Internet audio
and digital broadcasting. MPEG-4 AAC with SBR, known as High Efficiency
AAC, or HE-AAC, can deliver high quality stereo audio at a mere 48 Kb/s. The
SBR extension is both forward and backward compatible: an existing MPEG-4
AAC decoder can decode a signal without using the technique, while a decoder
with SBR uses the extended signal to enhance the upper octave of the signal.

For digital audio broadcasting, MPEG-4 AAC is becoming the codec of choice.
Satellite-based XM Radio uses HE-AAC, as does terrestrial Digital Radio
Mondiale.

MPEG-4 Audio is also inherently scalable. If, for example, a transmission uses an
error-prone channel with limited bandwidth, an audio stream consisting of a
small base layer and a larger extension layer provides a robust solution. Strong
error protection on the base layer (adding only little overhead to the overall
bitrate) makes sure there is always a signal, even with difficult reception. The
extension layer (with little error protection) and base layer together give excellent
quality in normal conditions. Any errors lead only to a subtle degradation of
quality but never in a total interruption of the audio stream.


Clarifying Common Questions

Who licenses MPEG-4 technology?
The Motion Picture Experts Group of the international standards bodies ISO
and IEC develops and publishes MPEG standards. MPEG requires that
companies proposing technologies for MPEG standards to commit to licensing
their patents on Reasonable and Non-Discriminatory Terms and Conditions
(also called “RAND”). Other than the costs of publication, there are no fees for
using MPEG standards themselves, and neither MPEG nor MPEGIF is involved
in patent licensing.

MPEG MPEGIF Patent Pools Manufacturers

Develops MPEG Promotes MPEG License Patents Sell or license their
Standards standards Essential to MPEG implementations
Operates standards (systems, chips, or
Interoperability and software) of MPEG
conformance standards
programs
Table 3. Roles of Organizations in the IP Licensing Process

Patent pool organizations provide a single convenient point for licensing patents
that are essential to implementing one of the MPEG standards. They operate
independently of MPEG and MPEGIF, and each sets its own terms and royalties
for patent licensing. Currently, MPEGIF is aware of two firms that operate pools
for MPEG-4 patents: MPEG-LA and Via Licensing.

MPEG-4 Patent Pools

Section of Topic Patent Pools
standard

Part 1 Systems MPEG-LA

Part 2 Video MPEG-LA

Part 3 Audio Via Licensing

Part 10 Advanced Video MPEG-LA,
Via Licensing
Table 4. Current MPEG-4 Patent Pool Consortia


The exact details of the licensing model are outside the scope of this paper. Please
visit the MPEGIF patents page (http://www.mpegif.org/patents/) for links to the
latest information, including links to the licensing terms themselves.

These joint licensing schemes are not carried out on behalf of ISO, MPEG or
MPEGIF, nor are they, or do they need to be, officially blessed by any such
organization. There is no “authority” involved in licensing, it is a matter of
private companies working together to offer convenience to the market.

What is the role of MPEGIF in licensing?
The MPEG Industry Forum has written in its statutes that it shall not license
patents or determine licensing fees. It does not share in the license royalties.
MPEGIF has acted as a catalyst, promoting the use of patent pools. MPEGIF
has among its members licensors, licensees, and other entities which have an
interest in fair and reasonable licensing. MPEGIF continues to monitor terms
and conditions of licensing, and to offer forums where parties can exchange
information and opinions on licensing

With the release of Part 10 AVC, is Part 2 video coding
obsolete?
AVC is currently MPEG-4’s highest performance video codec, offering the lowest
bitrate for a given quality among any MPEG or proprietary codecs, including
MPEG-4 Part2.

MPEG-4 AVC makes use of the latest research in video coding. The coding
methods rely on the fact that computational power and memory have become
cheaper than 5 years ago, meaning that coding methods can be more complex
than could previously be accommodated in hardware and software
environments.

In applications where coding efficiency is paramount, such as HDTV satellite
broadcasting, the performance needed means Part 2 video, as well as MPEG-2
and other codecs, are not suitable. Part 2 will likely continue to be used in
applications where power consumption is critical or bandwidth is less so, as in
digital still cameras or webcams. Part 2 codecs will also remain important for
object-coded rich multimedia systems where AVC is not yet supported.

Additionally, one must remember that Part 2 video is a part of several deployed
standards, such as mobile videophony. The installed base of 3G mobile video
handsets means that Part 2 video encoders will continue to be improved and


future handsets will support Part 2 for interoperability. Standards continue to
flourish for many years after their introduction – with MPEG-2 we are likely just
now reaching the peak of annual production of MPEG-2 encoders and decoders,
a decade after its release.

What is the relationship between MPEG-4 Visual
and the DivX codec?
DivX5 is an implementation of MPEG-4 Advanced Simple Visual Profile. DivX
Networks is also working on file format compliance.

Is Microsoft Windows Media an MPEG-4 codec?
Microsoft was one of the first companies to deploy an MPEG-4 Video codec in
previous versions of its Windows Media platform. Explicit support for MPEG-4
was removed from Windows Media several years ago. It is unknown to what
extent the current version of Windows Media uses MPEG-4 concepts internally.
Some developers will know of Microsoft’s contribution to the MPEG-4
Reference Software, one of the two implementations of the part 2 MPEG-4
Visual standard that developers can download from ISO’s website (The other
implementation is from the European project ‘MoMuSys’).

The future is all downloadable software codecs, why do we
need a standard?
There are many environments in which downloading codecs is not possible. The
future is video and multimedia on many different devices, with very many totally
different uses. While the Internet is growing exponentially, and streaming media
and video on demand are poised to be large applications, the future is also about
wireless connectivity on mobile phone, PDA, or camera devices. In many of
these devices, available memory and power supply requires a hardware decoder.
Also, it is important to realize that standards such as MPEG-4 are not just about
stable decoders that can be implemented in hardware. They also allow
interoperability among all implementations, whether in software or hardware.

Is MPEG-4 based on QuickTime?
The file format of MPEG-4 (MP4) is based on the QuickTime architecture. The
rest of the MPEG-4 standard was developed independent of QuickTime.
QuickTime started supporting MPEG-4 with its version 6, which includes


Simple Visual profile and AAC. QuickTime 7 added support for MPEG4 Part
10 (AVC/H.264) on April 28, 2005.

I read a benchmark of MPEG-4 where it did poorly, how can
you claim it is higher performance?
One of the virtues of MPEG-4 is that it is an open standard that can be
implemented by anyone. While there are conformance and interoperability
programs run by MPEGIF and other organizations to insure that products work
together and correctly implement the standard, they are voluntary, unlike the
tight controls imposed by proprietary codec licensing. So it is possible to test an
immature or poor implementation of MPEG-4 against the latest proprietary
codec, or even a mature MPEG-2 codec, and have MPEG-4 fair poorly.

One of the advantages of MPEG-4 is that encoders from different manufacturers
will be of different quality, yet all will be compliant to the standard and decoded
by any decoder. This variety means a user or system manufacturer can pick an
implementation that has the cost and performance needed in his application,
and he can switch encoders without making changes to his installed base of
decoders or re-encoding existing content.

Benchmark tests are also complex to perform correctly because of the biases,
fatigue, and other psychological effects of the observers. A viewing jury in a test
conducted according to ITU R-500 sees a controlled set of clips that they rank
without knowing which codec is being used, and the test clips are carefully
chosen so there is a mix of motion, detail, and other features so the test
represents all types of scenes that would be encoded. In less formal benchmarks,
particularly those carried out by journalists or enthusiasts, these effects are often
not considered correctly.

How does MPEG-4 Compare to Other Internet Media Formats?
Multi-vendor support ensures market driven solutions. Standards like MPEG’s
have a potential for broad industry support. Proprietary solutions can only
succeed if they are adopted by large market segments, which has not happened
with existing technologies. The table below gives a comparison of MPEG-4
against most commonly used multimedia formats on the Internet today.


Table 5. Comparing MPEG-4 to Other Internet Media Formats

MPEG-4 Windows Media Real Flash

Audio/Video Standards based; Proprietary Proprietary, but Proprietary +
Codec multi-vendor supports proprietary Real
support. automatic and QuickTime
download of formats.
MPEG-4 plug-in.

Interactivity Highly interactive. Limited Yes, via SMIL. Highly
interactive.

Digital Rights Interfaces to Microsoft DRM Content access Content access
Management proprietary DRM. control control
More
interoperable
DRM under
development in
MPEG-4 and
MPEG-21

Real-time Yes Yes Yes No
stream control

Synchronization Audio, video and Tight Tight No
all other objects synchronization synchronization synchronization
can be tightly between audio between audio between scene
synchronized with and video and video and streams
high accuracy

Broadcast Yes, including A/V only Scene must be No
capable interactive unicast
features

Object model Video/audio and Audio/Video only Video/audio and Video/audio and
support rich 2D/3D mixed mixed media mixed media
media, synthetic through SMIL through
graphics. DRM on based protocol. proprietary
separate streams. No streaming of protocol.
mixed media.

Graphic Objects Yes No No Yes

Transport Support exists for HTTP, UDP, HTTP, RTP/RTSP, HTTP
HTTP, UDP, RTP/RTSP, mobile mobile
RTP/RTSP,
MPEG-2TS, mobile

PC, Set Top Box, Yes Yes Yes Yes
Wireless


How Will MPEG-4 Be Used in Interactive TV?
Along with significantly less bandwidth for the same quality, the native support
for interactivity is a key difference between MPEG-4 and the MPEG-2
technology broadly deployed in current digital television systems.

In every case, making Interactive TV work in an MPEG-2 based environment
means that operators need to adopt one or more proprietary solutions, or
solutions based on technologies not native to MPEG, and add them productively
to an MPEG-2 delivery environment. This has led to the emergence of several
proprietary add-on technologies competing for the business of ITV operators.

Each operator has a unique composite solution of technologies, usually
determined by their MPEG compression platform, their Conditional Access
System and their Middleware platform, e.g. OpenTV. This has led to the
emergence of several incompatible vertical solutions and markets. The problem
with vertical markets, not only in the business sense but also in the technology
sense, is that at the end of the day, end-users don’t benefit from them, and service
deployment is slowed-down. Several attempts to dissolve them into the
horizontal market have taken place and are meeting great resistance from this
sort of “economic gravity” which makes vertical markets inevitable without open
standards.

One very promising technology is the DVB standard for interactive TV APIs,
Multimedia Home Platform (MHP), described in more detail below. In the
United States, MHP has its equivalents in the Java-based Digital Applications
Software environment (DASE), an Advanced Television Systems Committee
(ATSC) activity, and in OCAP, the Open Cable Application Platform specified
by the OpenCable consortium, which is based on MHP.

The mainstream of the Broadcast Industry likes Java, because unlike the host of
other proprietary and flavored web-standards based approaches (e.g.
MediaHighway, Liberate, OpenTV), it offers content creators and providers and
service operators a chance to ‘write once, run many times’ using the same
content, which is itself indispensable to creating a horizontal market.

This paper will not compare MPEG-4 to these technologies and
operator-specific or platform-specific architectures. We will give a few examples
of how the power of MPEG-4 can be easily added to complement the most
important standards in this area. As an example of a platform based on
procedural content, we examine ways in which MPEG-4 can add value to MHP.
Then in the other track of Interactive TV, away from Java, there is significant


interest in the W3C work on newer generations of its meta-languages based on
HTML, e.g. XML and XHTML. In fact the MHP platform supports both.

What is the Difference between MPEG-4 and MHP?
An often-asked question about MPEG-4 is how it relates to the Multimedia
Home Platform specification of DVB (Digital Video Broadcasting).

The first thing to understand is that there are two relevant groups of DVB
specifications. The first, DVB 1.0, is the transport foundation of the DVB family
of standards. This specification spells out how to implement DVB-compliant
MPEG streams. DVB is traditionally MPEG-2 based, but MPEG-4 is seen as a
logical evolution, and one which will be more efficient when DVB services are to
be delivered over IP. To this goal, DVB has included MPEG-4 Main Profile AVC
and HE-AAC in its latest revision.

The second DVB, also sometimes referred to as DVB 2.0, addresses the
Multimedia Home Platform and a variety of next generation delivery
applications, including Copy Protection and Copy Management and delivering
DVB services over IP. The Multimedia Home Platform (MHP) defines a generic
interface between interactive digital applications and the terminals on which
those applications execute. The MHP specification specifies how to download
applications and media content, typically delivered over a DVB compliant
transport stream, and optionally in the presence of a return channel.

MPEG-4 is a natural companion to MHP applications, with low bitrate video
and scene representation formats streamed or delivered over IP to set top boxes.
The application, interaction, and synchronization models of MPEG-4 allow
more dynamic content to be added to MHP-type of applications.

Because MPEG-4 can be carried by MPEG-2 transports we can achieve a very
fine-grain synchronization between the broadcast program and the MPEG-4
multimedia content. Integrating MHP with MPEG-4 can enable object-based
interactive digital television.

The combination of MHP and MPEG-4 provides the ability to develop very
flexible and rich interactive applications for the interactive broadcast domain.
The MPEG-4 features can be introduced smoothly and gradually, in a
backwards-compatible manner.

The MHP architecture is defined in terms of three layers: resources, system
software and applications. Typical MHP resources are those elements that can be


called upon by applications to perform certain functions, for example MPEG
processing, I/O, CPU, memory and graphics handling. The system software
presents a standardized abstract view of the resources of the platform to the
applications, thus enabling "platform independence.” An "application manager"
is provided to manage the interaction between these elements.

Generic Application Program Interfaces (APIs) are specified by DVB-MHP,
based around DVB-J, which includes the Java Virtual Machine (VM) as
originally specified by Sun Microsystems. MHP applications can only access the
resources of the platform via these specified APIs, a feature which guarantees the
stability of the platform and its robustness against "rogue" applications. These
APIs are specified by the DVB Technical Module (TAM) and are tested for
conformance with the MHP specification through the use of agreed test
applications.

MPEG-J(ava) offers a set of functionalities complementary to those offered by
DVB-J. MPEG-J in MPEG-4 is a set of Java APIs that may be present on a
MPEG4 terminal. MPEG-J applications (MPEGlets), which are sent as part of
the presentation, use the MPEG-J API's to control the capabilities of the
MPEG-J terminal. Java packages that must be in the terminal that supports
MPEG-J include java.lang, java.io, and java.util.

MHP's "DVB-J" consists of a generic Java platform that is similar to
PersonalJava 1.2a. It supports "Xlets" (but not applets), and has a very limited
subset of java.awt -- the widget set is removed. Additionally, it includes JMF 1.1,
Java TV 1.0, and a number of DVB-specific APIs for accessing TV-specific data
and controlling TV-specific functionality.


Figure 17. MHP reference architecture showing role of DVB-J

How Does MPEG-4 Compare to SMIL and SVG?
SMIL is the Synchronized Multimedia Integration Language of the W3C and
SVG is W3C’s Scalable Vector Graphics specification. A comparison of MPEG-4
and SMIL+SVG capabilities follows below. MPEG-4 provides a rich multimedia
experience, in which interactivity, streaming, and various mixed media, including
graphics objects, are combined seamlessly. SMIL and SVG, as currently proposed
for use by 3GPP, provide somewhat similar functionality, with notable
differences, as SMIL is more declarative in nature and MPEG-4 is more
procedural. The comparison mostly concerns MPEG-4 BIFS (the Binary Format
for Scenes) and the Object Descriptor framework of MPEG-4, which takes care
of the synchronization between the different objects.
Table 6. Comparing MPEG-4 to SMIL and SVG

Requirement MPEG-4 SMIL+SVG

Spatial and temporal composition of Very simple to very complex Only 2D
text, graphics, images and streamed composition. 2D and 3D composition
media (audio and visual streams) profiles

Flexible synchronization models of Yes Yes
different objects (co-start, co-end, …)

Broadcast-grade synchronization of all Yes No
objects on a rigid timeline (e.g. A and V)


Streaming scene description Yes No

Compression of scene description Yes No

Dynamic scenes (add/ remove objects, Yes No
etc)

Streamed animation of scene Yes No
components

Broadcast capable Yes No

DRM tightly coupled with scene (e.g. Yes No
can protect streams independently)

MPEG-4’s Textual Format: XMT
Originally, MPEG-4 only contained a binary scene description language. Later, it
became clear that it would be helpful to add a textual representation as well, in
the form of XMT, the eXtensible MPEG-4 Textual format. XMT is an
XML-based language, like SMIL. MPEG has been careful to build XMT as
compatible with SMIL as possible, to aid interoperability in media distribution.
Another goal was to build compatibility with the X3D specifications for
interactive 3D content (which is an XML extension of the older Virtual Reality
Modeling Language (VRML)). MPEG-4’s scene description model is based on
the textual VRML language, to which MPEG added streaming behavior, 2D
support and a binary representation for efficiency.

In the course of defining the textual format, MPEG-4 was also extended with the
flexible timing models that SMIL uses. The so-called “flextime” support was
added to the broadcast-type of time stamp-based, rigid MPEG-2 type of
synchronization.


SMIL Player
SMIL Parse
Compile
SVG VRML
XMT Browser

MPEG-7
MPEG-4
Representation MPEG-4
(e.g. mp4 file)
X3D Player

Figure 18. Relation of MPEG-4 XMT to other specifications.

The textual format and the binary format are largely dual representations of the
same information. In most situations, one would want to deliver scene
description information in binary form, as that is much more efficient. For
exchanging scenes between authors or storing scenes inside a single organization
in a way that is understood by multiple tools, a textual format is a useful tool. It
is easy to go from text to binary representation, and while the other way is just as
easy in theory, it is harder to do so in a ways that is meaningful for an author,
much like a decompiled program can be hard to read.


The MPEG Industry Forum

The MPEG Industry Forum represents more than 80 companies from diverse
industries evenly distributed across North America, Europe and Asia, addressing
MPEG-4 adoption issues that go beyond the charter of ISO/IEC MPEG.

MPEGIF is vital to the success of the MPEG-4 standard, since the work done by
MPEG is necessary but not sufficient. In its endeavors to promote wide adoption
of MPEG-4, MPEGIF picks up where MPEG stops.

The following is a list of MPEGIF's current activities:

• Promoting the emerging MPEG standards (MPEG-4, MPEG-7 and
MPEG-21), and serving as a single point of information on technology,
products and services for these standards;

• Carrying out interoperability tests, which lead to an ecosystem of
interoperable products. Over 30 companies have tested their products in
MPEGIF's MPEG-4 interop program;

• Developing and establishing an MPEG-4 Certification program, which
comes with the right to carry MPEGIF's MP4 logo;

• Organization of and participation in many trade show events - MPEGIF
has show floor presence together with some of its members, at shows
such as NAB and IBC. In many other shows we organize panels and
presentations.

• Organization of MPEG exhibitions and tutorials. MPEGIF has
organized several Workshops and Exhibitions on MPEG-4, such as in
Geneva (2000) and San Jose (June 2001 and June 2002);

• Establishing a forum for discussions that led to the formation of
independent patent pools for licensing MPEG-4 patents on fair terms.
MPEGIF's website has a wealth of information on the MPEG-4 standard and is
starting to collect information on MPEG-7 and MPEG-21. It has many links to
external resources, and is updated daily with latest relevant news and press
releases. Through the website, anyone can sign up for MPEGIF's News,
Discussion, and Technology mailing lists.


Join the forum
Membership in MPEGIF will put you in touch with your future clients, partners,
suppliers and competitors. It will put your company’s name on the list of the top
companies at the leading edge of MPEG-4. For the price of a one-page
advertisement in a trade magazine, you may join the forum and might meet your
next customer there.

Help Drive Success
MPEGIF has a unique and broad spectrum of members, coming from all
industry segments - all individuals who are focused on MPEG-4 and have
decisive roles in their companies. The time is now to communicate your value
proposition to other members.

To join MPEGIF, or to find out what activities will benefit your company, visit:
http://www.mpegif.org.


whitepaper_mpeg-if_understanding_mpeg4

More Related Content

Similar to whitepaper_mpeg-if_understanding_mpeg4 (20)

More from Aniruddh Tyagi (20)

Recently uploaded (20)

whitepaper_mpeg-if_understanding_mpeg4