University of Wollongong Research Online

Faculty of Engineering and Information Faculty of Informatics - Papers (Archive) Sciences

October 2003

MPEG-21: goals and achievements

I. Burnett University of Wollongong, [email protected]

R. Van de Walle Ghent University, Belgium

K. Hill Rightscom, London, UK

J. Bormans Interuniversity MicroElectronics Center, Belgium

F. Pereira Instituto Superior Tecnico, Portugal

Follow this and additional works at: https://ro.uow.edu.au/infopapers

Part of the Physical Sciences and Mathematics Commons

Recommended Citation Burnett, I.; Van de Walle, R.; Hill, K.; Bormans, J.; and Pereira, F.: MPEG-21: goals and achievements 2003. https://ro.uow.edu.au/infopapers/46

Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library: [email protected] MPEG-21: goals and achievements

Abstract MPEG-21 is an open standards-based framework for multimedia delivery and consumption. It aims to enable the use of multimedia resources across a wide range of networks and devices. We discuss MPEG-21's parts, achievements, ongoing activities, and opportunities for new technologies.

Disciplines Physical Sciences and Mathematics

Publication Details This article was originally published as: Burnett, I, Van de Walle, R, Hill, K et al, MPEG-21 Goals and Achievements, IEEE Multimedia, Oct-Dec 2003, 10(4), 60-70. Copyright IEEE 2003.

This journal article is available at Research Online: https://ro.uow.edu.au/infopapers/46 Feature Article

delivery and consumption by all the players in the delivery and consumption chain. MPEG-21: MPEG-21 context MPEG-21 is the newest of a series of standards being produced by the Moving Picture Experts Group (see http://mpeg.telecomitalialab.com), Goals and more formally known as ISO/IEC JTC 1/SC 29/WG11. MPEG has a long history of producing multimedia standards: MPEG-1 (1993), MPEG-2 (1996), MPEG-4 (1999), and MPEG-7 (2001). Achievements MPEG-1 and MPEG-2 resulted in many suc- cessful commercial products and services, such as video-CD, DVD, digital television, digital audio broadcasting (DAB), and MP3 (MPEG-1/-2 audio Ian Burnett layer 3). University of Wollongong, Australia More recently, MPEG-41 standardized audio– visual coding solutions that address the needs of Rik Van de Walle communication, interactive and broadcasting Ghent University, Belgium services, and many combinations of these mod- els. At the heart of MPEG-4 technology is an Keith Hill object-based representation approach in which a Rightscom, London scene is modeled as a composition of both nat- ural and synthetic objects with which a user may Jan Bormans interact. Through this new coding approach, Interuniversity MicroElectronics Center, Belgium MPEG-4 creates new opportunities for users to play with, create, reuse, access, and consume Fernando Pereira audio–visual content. MPEG-4 is currently used Instituto Superior Técnico, Portugal on the Internet (such as Quicktime 6) and is being adopted in several other communications contexts. he appetite for content consumption Building on MPEG-4’s object-based vision, the MPEG-21 is an open and easier access to information con- MPEG-72 standard concentrates on the descrip- standards-based tinues to increase rapidly. Access tion of multimedia content. It enables quick and framework for devices are also evolving so that we efficient searching, identification, processing, multimedia delivery haveT a wide choice of terminal and network and filtering of multimedia material. This lets and consumption. It capabilities. The result is that people, both in users (both humans and machines) use and inter- aims to enable the their personal and professional lives, are increas- act with non-text-based multimedia content use of multimedia ingly creators as well as consumers of digital directly and efficiently. resources across a media. They thus demand solutions that deliver Together, the MPEG-1, -2, -4, and -7 standards wide range of accessible and advanced multimedia creation and provide a complete, powerful, and successful set networks and consumption on many platforms. These new of tools for multimedia representation. Other devices. In this content providers and the traditional media multimedia-related standards (and proprietary article, we discuss sources share a core set of concerns: management solutions) exist, such as JPEG and JPEG 2000. MPEG-21’s parts, of content, repurposing content based on user However, widespread deployment of multimedia achievements, preferences and device capabilities, protection of applications requires more than this loose col- ongoing activities, rights, protection from unauthorized access/ lection of standards. The problem is that overall and opportunities modification, protection of privacy of providers multimedia consumption and commerce remain for new and consumers, and so on. The MPEG-21 nontransparent and aren’t happening on a large technologies. Multimedia Framework initiative aims to meet scale. In late 1999 and early 2000, MPEG exam- these challenges by enabling the use of multime- ined several questions: Do all existing dia resources across a wide range of networks and multimedia-related standard specifications fit devices. In this article, we discuss MPEG-21, an together? Does anybody know how they fit open standards-based framework for multimedia together? Do we have specifications for all the

60 1070-986X/03/$17.00 © 2003 IEEE Published by the IEEE Computer Society necessary technical elements for multimedia Interoperability is the driving force behind all transactions? Which standard activities are most multimedia standards. It’s a necessary require- relevant? Who is making the “glue” that will let ment for any application that requires guaranteed these standards fit together? communication between two or more parties. Many high-quality multimedia standards are From a more philosophical view, interoperability being created to meet the needs of different com- expresses the user’s dream of easily exchanging munities. However, the result is an array of stan- any type of information without technical barri- dards offering parallel or competitive solutions. ers. To achieve this goal, we must standardize both In contrast to this a transparent multimedia the content structure and a minimum set of com- world requires the creation of a single, big picture munication processes. The key to effective stan- view of multimedia. dardization is to create a minimum standard that Currently, multimedia technology provides normatively defines a minimal (but complete) set the different players in the value and delivery of tools that will guarantee interoperability. Such chain with an excess of information and services. minimum specifications provide space and a foun- However, no complete solutions exist that let dif- dation for competitive, proprietary, and alternate ferent communities—content, financial, com- developments (which wouldn’t be normative) of munication, and computer and consumer tools that don’t need standardization to obtain electronics and their customers (each with their interoperability. This enables the incorporation of own models, rules, procedures, interests, and technical advances—lengthening the lifetime of content formats)—interact efficiently using this the standard—as well as stimulating technical and complex infrastructure. Developing a common product competition. A standard also has impor- multimedia framework would facilitate coopera- tant economic implications because it allows the tion between sectors and support a more efficient sharing of investment costs and the acceleration implementation and integration of the different of application deployment. A further advantage is needs of each community, resulting in an that an open standard reduces consumer reliance enhanced user experience.3 on sole proprietary solutions; this is essential if we One problem, however, is that in the desire to want truly transparent multimedia usage and con- achieve interoperability, the framework may vio- sumption to be a reality. late the requirement to protect the value of the MPEG has been proactive in identifying cur- content and the rights of the rights holders. From rent multimedia initiatives and inviting collabo- the alternate view, digital rights management ration in the context of MPEG-21. Examples of (DRM) systems—which protect those rights—can envisaged collaborations are the Open eBook prevent interoperability if they use nonstan- Forum, International Telecommunications dardized protection mechanisms. Realizing an Union, Open Mobile Alliance, and so on. The open multimedia infrastructure thus requires an aim of this process is to maximize interoperabil- integrated and suitably normative DRM system. ity, minimize the overlap between concurrent Thus, MPEG-21 seeks to create the big picture activities, and share common technologies. of multimedia standards. It aims to guarantee To ensure that the standardization of a big interoperability by focusing on how the elements picture encompassed both technological and user of a multimedia application infrastructure should demands, MPEG established a formal develop- relate, integrate, and interact. Finally, where ment process as follows: open standards for elements are missing, MPEG is creating new MPEG-21 parts to fill the gaps. 1. Define a framework supporting the MPEG-21

vision statement. October–December 2003 MPEG-21 objectives A key assumption of MPEG-21 is that every 2. Identify the multimedia framework’s critical human is potentially an element of a network components. involving billions of content providers, value adders, packagers, service providers, consumers, 3. Understand how the framework’s compo- and resellers. Thus, besides client-server-based nents relate and identify where standards applications, peer-to-peer networking and the technology gaps exist. resulting flexibility of user roles has been an underlying part of MPEG-21 thinking since the 4. Involve relevant (and complementary) stan- early days of the standardization process. dardization bodies in this effort.

61 Corporate software for download Student records with (staff with licenses only) rights of access

Downloadable computer base training Video tour of (rights of use: UoW students only) UoW research activities

Campus tours with multilanguage soundtracks http://www.uow.edu.au

Web site link Digital item Distance delivery content - 1. Prospective on a fee per subject basis Online lecture notes student (students enrolled in subject only) 2. Student 3. Graduate 4. Staff 5. Visitor Methods for Interactive Digital Item interaction campus maps

Figure 1. Example 5. Assess each of the unavailable technologies. with other users. All users must also express and Digital Item. If they fall under MPEG expertise, then manage their interests in Digital Items. MPEG develops the corresponding standards. In practice, a Digital Item is a combination of If not, engage other bodies to achieve their resources, metadata, and structure. The resources development. are the individual assets or (distributed) content. The metadata describes (distributed) data about 6. Integrate the relevant available and devel- or pertaining to the Digital Item as a whole or also oped technologies. to the individual resources in the Digital Item. Finally, the structure relates to the relationships At the time of writing, this process has been among the parts of the Digital Item—both underway for some two and a half years. To date, resources and metadata. An example of a Digital several parts of the MPEG-21 framework are fully Item might be a presentation of a university, standardized and several others are in an including photos, videos, animation graphics, advanced state of development. textual information, news related to the universi- ty’s research activities, e-learning material, navi- Basic concepts and elements gational information driven by user preferences, The basic concepts in MPEG-21 relate to the and so on (see Figure 1). The Digital Item is thus what and who within the multimedia framework. the fundamental unit for distribution and trans- The what is a Digital Item that’s a structured digi- action within the MPEG-21 framework. tal object with a standard representation, identifi- cation, and metadata within the MPEG-21 Parts of MPEG-21 framework. The who is a user who interacts in the MPEG-21 is organized into several indepen- MPEG-21 environment or uses a Digital Item, dent parts primarily to allow various slices of the including individuals, consumers, communities, technology to be useful as stand-alones. This organizations, corporations, consortia, govern- maximizes their usage and lets users implement ments and other standards bodies, and initiatives them outside MPEG-21 as a whole in conjunc- around the world. The user roles include creators, tion with proprietary technologies. For example, consumers, rights holders, content providers, dis- we use MPEG-2 video together with MPEG-2 sys- tributors, and so on—there’s no technical distinc- tems but not with MPEG-2 audio in the context tion between providers and consumers.4 All parties of the US digital TV system. However, although that must interact within MPEG-21 are catego- we may use the various parts independently, rized equally as users. They assume specific rights they were developed to give optimal results when

IEEE MultiMedia and responsibilities according to their interaction used together.

62 The MPEG-21 parts already developed or cur- These technologies link information that rently under development are identifies and describes content directly to the content itself. 1. Vision, technologies, and strategy: describes the multimedia framework and its architectural 12. Test bed for MPEG-21 resource delivery: provides elements with the functional requirements a software-based test bed for delivering scal- for their specification.5 able media and testing/evaluating this scalable media delivery in streaming environments. 2. Digital Item Declaration (DID): provides a uni- form and flexible abstraction and interopera- MPEG-21 achievements ble schema for declaring Digital Items. Here we provide an overview on currently available MPEG-21 technologies. For those under 3. Digital Item Identification (DII): defines the development (parts 4 and 8 through 12), we con- framework for identifying any entity regard- sider only those parts that, at the time of writing, less of its nature, type, or granularity. are at a high level of maturity.

4. Intellectual property management and protection Vision, technologies, and strategy (IPMP): provides the means to reliably man- The MPEG-21 technical report describes the age and protect content across networks and multimedia framework and its architectural ele- devices. ments with the functional requirements for their specification.5 5. Rights Expression Language (REL): specifies a This part is available as a free downloadable machine-readable language that can declare document from the International Organization rights and permissions using the terms as for Standardization Web site at http://www. defined in the Rights Data Dictionary. iso.ch/iso/en/aboutiso/introduction/index.html.

6. Rights Data Dictionary (RDD): specifies a dic- Digital Item Declaration tionary of key terms required to describe The Digital Item Declaration part is the sec- users’ rights. ond part of MPEG-21 (ISO/IEC 21000-2).6 It spec- ifies a uniform and flexible abstraction and 7. Digital Item Adaptation (DIA): defines descrip- interoperable schema for declaring the structure tion tools for usage environment and content and makeup of Digital Items. Through the Digital format features that might influence the Item Declaration Language (DIDL), we can transparent access to the multimedia con- declare a Digital Item by specifying its resources, tent—notably terminals, networks, users, and metadata, and their interrelationships. the natural environment where users and ter- The ISO/IEC 21000-2 describes this DID tech- minals are located. nology in four main sections:

8. Reference software: includes software that ❚ Model. The DID model describes a set of implements the tools specified in the other abstract terms and concepts for defining MPEG-21 parts. Digital Items. Within this model, a Digital Item is the digital representation of a work (for 9. File format: defines a file format for storing example, a digital music album, an e-book, or

and distributing Digital Items. a piece of software including setup and con- October–December 2003 figuration information). As such, the Digital 10. Digital Item Processing (DIP): defines mecha- Item is the thing that’s acted upon (managed, nisms for standardized and interoperable pro- described, exchanged, collected, and so on) cessing of the information in Digital Items. within the model.

11. Evaluation methods for persistent association ❚ Representation. The DIDL is based on the terms technologies: documents best practices in eval- and concepts defined in the DID model. It uating persistent association technologies contains the normative description of the syn- using a common methodology (rather than tax and semantics of each of the DIDL ele- standardizing the technologies themselves). ments, as represented in XML.

63 ❚ Schema. The complete normative XML schema that can affect an item’s configuration. for DIDL comprises the entire grammar of the DID representation in XML. ❚ A selection describes a specific decision that affects one or more conditions somewhere ❚ Some detailed examples. Illustrative examples of within an item. DIDL documents are provided to aid in under- standing the use of the specification and its ❚ An annotation describes a set of information potential applications. about another identified element of the model without altering or adding to that element. Major concepts in the context of the DID include the following:6 ❚ An assertion defines a full or partially configured state of a choice by asserting true, false, or unde- ❚ A container is a structure that groups items cided values for some number of predicates and/or containers. We can use these groupings associated with the selections for that choice. of items and/or containers to form logical pack- ages (for transport or exchange) or logical ❚ A resource is an individually identifiable asset shelves (for organization). Descriptors allow for such as a video or audio clip, an image, or a the labeling of containers with information textual asset. A resource may also potentially that’s appropriate for the purpose of the group- be a physical object. All resources must be ing (for example, delivery instructions for a locatable via an unambiguous address. package or category information for a shelf). ❚ A fragment unambiguously designates a spe- ❚ An item is a grouping of subitems and/or com- cific point or range within a resource. ponents bound to relevant descriptors. Items Fragments may be resource-type specific. may contain choices, which can be customized or configured. Items may also be conditional (on ❚ A statement is a literal textual value that con- predicates asserted by selections defined in the tains information, but not an asset. Examples choices). We consider an item that contains no of likely statements include descriptive, con- subitems an entity—a logically indivisible work. trol, revision tracking, or identifying infor- An item that contains subitems is a compilation. mation (such as an identifier as described in MPEG-21 part 37). ❚ A component binds a resource to a set of descrip- tors. These descriptors are information related ❚ A predicate is an unambiguously identifiable dec- to all or part of the specific resource instance. laration that can be true, false, or undecided. Such descriptors will typically contain control or structural information about the resource Figure 2 shows some of the most important ele- (such as bit rate, character set, start points, or ments within the DID model and how they’re encryption information) but not information related. Figure 3 (on p. 66) shows a simple exam- describing the content within. ple of a DID, based on the Digital Item in Figure 1. Here, the DID is a container consisting of one glob- ❚ An anchor binds descriptors to a fragment, al descriptor and one item. The latter consists of which corresponds to a specific location or two descriptors, one choice, and two subitems. The part of a resource. choice lets the user choose between a video pre- sentation and a textual presentation. When the ❚ A descriptor associates information with the user chooses the video presentation (and only enclosing element. This information may be a then, due to the condition in the first subitem), the component (such as a thumbnail of an image user will receive an MP4 file (research.mp4). When or a text component) or a textual statement. the user chooses the textual presentation (and only then, due to the condition in the second subitem), ❚ A condition describes the enclosing element as the user will receive a text file (lecturenotes.txt). being optional and links it to the selection(s) that affect its inclusion. Digital Item Identification The third part of MPEG-21 (ISO/IEC 21000-3)7

IEEE MultiMedia ❚ A choice describes a set of related selections specifies

64 ❚ how to uniquely identify Digital Items (and Container parts thereof), Item Item Item Descriptor Component Descriptor ❚ how to uniquely identify intellectual proper- ty related to the Digital Items (and parts Descriptor Component thereof), Component Resource Descriptor Descriptor ❚ how to uniquely identify description schemes, Resource Resource ❚ the relationship between Digital Items (and parts thereof) and existing identification sys- Component Item tems (such as the International Standard Component Descriptor Recording Code, International Standard Book Descriptor Number, International Standard Serial Number, Resource digital object identifier, and so on), and Resource

❚ the relationship between Digital Items (and parts thereof) and relevant description schemes (such as the International Standard Figure 2. Some Digital Item Declaration model elements and Musical Work Code, International Standard their relationships. Text Code, and so on).

Note, however, that the DII standard doesn’t of this standard. However, it does describe the specify language’s syntax and semantics. REL can declare rights and permissions using ❚ new identification systems for the content the terms as defined in the RDD. It provides mech- elements for which identification and anisms to support the use of digital resources in description schemes already exist and are in publishing, distributing, and consuming of digi- use (for example, it doesn’t attempt to replace tal movies, digital music, electronic books, broad- the ISRC, as defined in ISO 3901,8 for sound casting, interactive games, computer software, and recordings) and other creations in digital form, in a way that pro- tects digital content and honors the rights, condi- ❚ normative description schemes for describing tions, and fees specified for digital contents. It also content. specifies access and use controls for digital content in cases where financial exchange isn’t part of the The DID and DII parts are closely related; we terms of use and supports exchange of sensitive or can associate identifiers covered by DII with private digital content. Digital Items, containers, components, and/or REL also provides a flexible, interoperable fragments thereof by including them in a state- mechanism to ensure that personal data is ment in the DID (see Figure 4 on p. 67). Examples processed in accordance with individual rights. of likely statements include descriptive, control, REL supports guaranteed end-to-end interoper- revision tracking, and/or identifying information. ability, consistency, and reliability between dif- ferent systems and services. As such, it offers

Rights Expression Language richness and extensibility in declaring rights, October–December 2003 The fifth part of MPEG-21 (ISO/IEC 21000-5)9 conditions, and obligations; ease and persistence explains the basic concepts of a machine- in identifying and associating these with digital interpretable language for expressing the rights contents; and flexibility in supporting multiple of users that act on Digital Items, components, usage/business models. fragments, and containers. It doesn’t provide specifications for security in trusted systems, pro- Rights Data Dictionary pose specific applications, or describe the details This sixth part of MPEG-21 (ISO/IEC 21000-6)10 of the accounting systems required. Nor does it specifies a RDD for use within the MPEG-21 address the agreements, coordination, or institu- framework. The RDD forms the basis of all expres- tional challenges in building an implementation sions of rights and permissions as defined by REL.

65 IEEE MultiMedia Figure 3. Example Digital Item declaration.

66 The RDD comprises a set of clear, consistent, Item Descriptor structured, integrated, and uniquely identified terms to support REL. The RDD specifies the structure and core terms and specifies how fur- ther terms may be defined under the governance Item Descriptor Item of a registration authority. Use of the RDD will facilitate the accurate Component Descriptor exchange and processing of information between interested parties involved in the administration Descriptor of rights in, and use of, Digital Items. Item Descriptor The RDD also supports the mapping and Item Descriptor transformation of metadata from the terminolo- Descriptor gy of one namespace (or authority) into that of Component Descriptor another namespace (or authority) in an auto- Statement mated or partially automated way, with the min- imum ambiguity or loss of semantic integrity. Descriptor uri:ABCDEF The RDD is a prescriptive dictionary, in the sense that it defines a single meaning for a term represented by a particular RDD name (or head- Statement word), but it’s also inclusive in that it recognizes the prescription of other headwords and defini- uri:ABCDEF myID:1234567 tions by other authorities and incorporates them through mappings. The RDD also supports the circumstance that the same name may have dif- ferent meanings under different authorities. The RDD has audit provisions so that we can track but the foundations. The standard is the starting Figure 4. Relationship additions, amendments, and deletions to terms point for a dynamic dictionary and not the fin- between the MPEG-21 and their attributes. ished article. Digital Item Declara- The RDD recognizes legal definitions as, and tion (gold boxes) and only as, terms from other authorities that we can MPEG-21 ongoing activities Digital Item Identifica- map into the RDD. Therefore, terms that are Besides the MPEG-21 parts previously dis- tion (pink boxes) parts. directly authorized by the RDD neither define cussed, work on other MPEG-21 parts has begun. nor prescribe intellectual property rights or other legal entities. Digital Item Adaptation The RDD standard and the RDD are distinct This seventh part of MPEG-21 (ISO/IEC things. The standard provides, first of all, a 21000-7) specifies tools for the adaptation of methodology for making the dictionary. The stan- Digital Items.11 dard then contains the initial group of terms—the One of the goals of MPEG-21 is to achieve RDD standardized terms—defined according to interoperable transparent access to (distributed) this method. Finally, the standard also provides advanced multimedia content by shielding users for a registration authority that will manage the from network and terminal installation, manage- ongoing work of adding to the dictionary. ment, and implementation issues. To achieve this For its part, the dictionary also has the char- goal requires the adaptation of Digital Items (see

acteristics of an ontology. It provides a set of Figure 5, next page). As shown in this conceptu- October–December 2003 well-defined terms for use in rights expressions al architecture, a Digital Item may be subject to a for Digital Items. In recognition of the great resource adaptation engine, a description adapta- diversity and complexity associated with multi- tion engine, or a DID adaptation engine, which media content, it also represents as many differ- produces the adapted Digital Item. For the adap- ent definitions of terms as its users require. tation, it’s essential to have available not only the Further, it shows their relationships in a struc- description of the content but also a description tured way to support the mapping and transfor- of its format and of the usage environment so mation of terms between different schemas and that content adaptation may be performed to pro- systems. For these reasons, the standardized vide the user the best content experience for the terms in the dictionary aren’t the definitive list, content requested. While MPEG-72 dealt with the

67 ing of the Digital Item, downloading Digital Item and presenting individual media Declaration resources, and performing tasks adaptation Adapted Digital (methods) that an author requires engine Digital Item Declaration Item Declaration for a particular Digital Item. While Digital Item processing is still in the Identifiers Description Identifiers adaptation early stages, Digital Item methods Rights expressions engine Rights expressions have been defined in more detail. The concept of Digital Item Descriptors Descriptors Resource methods is that, on receipt of a Digital Item adaptation Digital Item Digital Item, users will have a list of Adaptation tools engine Adaptation tools processes (methods) available that they can apply to the Digital Item. Resources Resources These methods provide a mecha- nism for a user to specify a preferred set of procedures by which the Figure 5. Digital Item content description problem, the description of Digital Item should be handled. A simple illustra- Adaptation content format and usage environments wasn’t tion of this functionality would be a music album architecture. addressed, although it’s now the target of the Digital Item with an add track method. The MPEG-21 DIA specification. method would let a user add a new track to the Recently, multimedia adaptation has become Digital Item in the preferred format (that is, in the a reality in the wireless application protocol and correct place in the hierarchy and with the cor- 3rd Generation Partnership Project mobile com- rect descriptors). MPEG intends for the methods munication arenas. However, we’ll need to even- to be run on the Digital Item method engine tually focus on problems of adaptation for a (DIME), which supports standard base operations broader range of content and conditions. Since (Digital Item base operations, which are analo- MPEG-21 Digital Items are a package of varied gous to a programming language’s standard multimedia resources, this leads to the problem library of functions). Users could then create of adapting single multimedia resources and methods for their Digital Items from these base maximizing the user experience of complete operations by using a standardized syntax (Digital resource packages in a variety of contexts.12 Item Method Language). Figure 6 shows how the DIME will be used as the core of an MPEG-21 pro- Digital Item Processing cessing node. The figure indicates the interfaces One of the important aspects related to the between the DIME and various parts of MPEG-21, Digital Item concept is that the DID is a static the user, and the platform’s facilities. declaration of only the information (structure, resources, and metadata) in the item; in the dec- Outlook laration itself there’s no implied processing of Based on the MPEG requirements14 and the that information. This is important because it enhanced requirements resulting from new usage contrasts with, for example, HTML-based Web cases,4 MPEG-21 continues to standardize rele- pages where the presentation of the information vant new technologies, and for more investiga- is intermingled with the information itself. tive areas, generate technical reports. The While the separation of the information from the following issues are under consideration: usage of the information in Digital Items has sig- nificant advantages (different users can receive ❚ Definition of an interoperable framework for different presentations of the same information), IPMP15 that improves and adds to already it also means that on receipt of the static decla- available IPMP tools defined within the con- ration, a user has nothing that indicates how the text of the MPEG-4 standard. information should be processed. Thus, DIP is the area of MPEG-21 that encom- ❚ Evaluation of methods for persistent associa- passes all the aspects of processing a static Digital tion technologies (MPEG-21 part 11).16 Item from a user (and application) perspective.13 It includes such processing as downloading the ❚ Definition of MPEG-21 reference software in

IEEE MultiMedia Digital Item, requesting IPMP and rights process- general and a software test bed for MPEG-21

68 resource delivery in particular (MPEG-21 parts ble today, no current end-to-end solutions let dif- 8 and 12).17 ferent user communities interact in an interop- erable and efficient way. ❚ Definition of an MPEG-21 file format.14 MPEG-21 guarantees such interoperability by focusing on how the elements of a multimedia ❚ Specification of MPEG-21 event reporting application infrastructure should relate, integrate, mechanisms that will monitor and commu- and interact. Key MPEG-21 parts have already nicate among peers and users the events relat- reached the status of international standard (or ing to Digital Items and/or the programs and will reach it during 2003) while new technologies devices that operate on them at any given continue to be considered for inclusion in new time.18 This will let users, for example, moni- MPEG-21 parts. The strength of MPEG-21 is that tor and record the use of Digital Items, it’s a standard grown from clear multimedia resources, and their descriptions. industry requirements for interoperability. As such, the requirements, existing standards, and ❚ Exploration of requirements and technology current work are firmly based on usage cases and for highly scalable audio and video coding. scenarios provided by that industry. MM In this context, MPEG is considering how these developments can be optimally Acknowledgments aligned with MPEG-21 in general and with We’d like to thank all the MPEG-21 members MPEG-21 DIA in particular.19 for their interesting and fruitful discussions, which have substantially enriched our technical Opportunities for technologies knowledge. Because we wrote parts of this article One of the key aspects of MPEG-21 is that it’s on the basis of the official MPEG documents, the a standard framework and not a complete, imple- MPEG community is in a way a coauthor. mentable solution. Hence, the multimedia and signal processing communities have many References opportunities to use novel techniques and solu- 1. F. Pereira and T. Ebrahimi, ed., The MPEG-4 Book, tions within the framework. Part 7 provides an Prentice Hall, 2002. example of this: DIA.11 This part of the standard 2. B.S. Manjunath, P. Salembier, and T. Sikora, eds., gives the community a rich set of metadata that Introduction to MPEG-7: Multimedia Content describes the context of delivery of resources (for Description Language, John Wiley & Sons, 2002. example, terminal, network, natural environ- 3. MPEG Requirements Group, MPEG-21 Overview, ment, and personal preference information). We ISO/MPEG N5231, Oct. 2002. can use these as input to signal processing algo- 4. MPEG Requirements Group, MPEG-21 Use Case rithms that adapt the media itself or higher level Scenarios, ISO/MPEG N4991, July 2002. software that chooses appropriate media streams 5. Multimedia Framework—Part 1: Vision, Technologies based on content and context metadata.20 In all and Strategy, ISO/IEC 21000-1:2002, 2002, http:// these spaces, MPEG has left the algorithms and www.iso.ch/iso/en/ittf/PubliclyAvailableStandards. the innovation open to researchers and develop- 6. Information Technology—Multimedia Framework ers while providing the standardized infrastruc- (MPEG-21)—Part 2: Digital Item Declaration, ture, since they don’t impact interoperability. ISO/IEC 21000-2:2003, Mar. 2003. Hence, while a number of parts of MPEG-21 7. Information Technology—Multimedia Framework are in place, and others are following close (MPEG-21)—Part 3: Digital Item Identification, ISO/IEC 21000-3:2003, Mar. 2003. behind in the standardization timetable, October–December 2003 researchers and developers should find the frame- 8. Information and Documentation—International Stan- work a useful place to start when building new dard Recording Code (ISRC), ISO 3901:2001, 2001. and innovative research tools. The advantage of 9. MPEG MDS Group, MPEG-21 Multimedia creating compatible solutions with the MPEG-21 Framework, Part 5: Rights Expression Language (Final framework is the broad market base that instant- Draft Int’l Std.), ISO/MPEG N5939, July 2003. ly becomes available from interoperability. 10. MPEG MDS Group, MPEG-21 Multimedia Framework, Part 6: Rights Data Dictionary (Final Conclusion Draft Int’l Std.), ISO/MPEG N5842, July 2003. Although electronic content creation, distrib- 11. MPEG MDS Group, MPEG-21 Multimedia ution, consumption, and trade are already possi- Framework, Part 7: Digital Item Adaptation (Final

69 Committee Draft), ISO/MPEG N5845, July 2003. 12. F. Pereira and I. Burnett, “Universal Multimedia Keith Hill is a founding partner Experiences for Tomorrow,” IEEE Signal Processing of Rightscom, London, a strategy (Special Issue on Universal Multimedia Access), vol. consultancy for digital content. 20, no. 2, Mar. 2003, pp. 63-73. He has contributed to the major 13. MPEG MDS Group, MPEG-21 Multimedia standards activities related to Framework, Part 10: Digital Item Processing (Working management and protection of Draft), ISO/MPEG N5855, July 2003. intellectual property and identification systems for 14. MPEG Requirements Group, MPEG-21 rights holders. He was a primary contributor to the Requirements, ISO/MPEG N5873, July 2003. launch of MPEG-21 in 1999 and continues to play a 15. MPEG Requirements Group, MPEG-21 Architecture, leading role in defining standards that enable the secure Scenarios and IPMP Requirements, ISO/MPEG delivery of content within a multimedia framework. He N5874, July 2003. has a BMus (honors) from the University of Sheffield. 16. MPEG Requirements Group, MPEG-21 Multimedia Framework, Part 11: Evaluation of Persistent Association Tools (Working Draft), ISO/MPEG Jan Bormans heads the Interuni- N5875, July 2003. versity Microelectronic Center’s 17. MPEG Requirements Group, MPEG-21 Multimedia Multimedia Image Compression Framework, Part 12: Resource Delivery Test Bed Systems group in Leuven, Bel- (Working Draft), ISO/MPEG N5640, July 2003. gium. His main research interests 18. MPEG Requirements Group, Current Vision on Event are the efficient design and Reporting in MPEG-21, ISO/MPEG N5871, July 2003. implementation of embedded systems for advanced 19. MPEG Requirements Group, Scalable Video Coding multimedia applications. He’s the Belgian head of del- Requirements, ISO/MPEG N5880, July 2003. egation for ISO/IEC’s MPEG and SC29 standardization 20. P. van Beek, et al., “Metadata Driven Multimedia committees. H’s also the MPEG-21 requirements editor Access,” IEEE Signal Processing (Special Issue on and chairman of the MPEG liaison group. He earned Universal Multimedia Access), vol. 20, no. 2, Mar. his PhD in applied sciences at the Vrije Universiteit 2003, pp. 40-52. Brussel (VUB), Belgium.

Fernando Pereira is an associate Ian Burnett is an associate pro- professor of telecommunications fessor at the University of at the Instituto Superior Técnico, Wollongong, Australia. His cur- Lisbon, Portugal. His current rent research interests are in mul- research interests are in video timedia processing and delivery, analysis, processing, coding and speech and audio coding, 3D description, and multimedia interactive services. He has audio, and audio separation. He received his BSc, been participating in the work of ISO/MPEG for many MEng, and PhD in electrical and electronic engineer- years, notably as the head of the Portuguese delegation, ing from the University of Bath, UK. chairman of the MPEG Requirements group, and by chairing many ad hoc groups related to the MPEG-4 and MPEG-7 standards. He received his MSc and PhD Rik Van de Walle is a professor of in electrical and computer engineering from Instituto multimedia systems and applica- Superior Técnico. tions and head of the Multimedia Lab at Ghent University, Belgium. Readers may contact Fernando Pereira at Instituto His current research interests Superior Técnico, Instituto de Telecomunicações, Av. include multimedia content deliv- Rovisco Pais, 1049-001 Lisboa, Portugal; [email protected]. ery, presentation and archiving, coding and description of multimedia data, content adaptation, and interactive For further information on this or any other computing (mobile) multimedia applications. He received his MSc topic, please visit our Digital Library at http://computer. and PhD in engineering from Ghent University. org/publications/dlib. IEEE MultiMedia

70