Copyright © 2010 W3C® (MIT, ERCIM, Keio), All Rights Reserved. W3C liability, trademark and document use rules apply.
This document defines the Ontology for Media Resource 1.0, a core vocabulary to describe media resources on the Web. It is defined based on a core set of properties which covers basic metadata to describe media resources. Further it defines syntactic and semantic level mappings between elements from existing formats. The ontology is supposed to foster the interoperability among various kinds of metadata formats currently used to describe media resources on the Web.
This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the W3C technical reports index at http://www.w3.org/TR/.
This is the a public Working draft of the Ontology for Media Resource 1.0 specification. It has been produced by the Media Annotations Working Group, which is part of the W3C Video on the Web Activity. The Working Group expects to advance this specification to Recommendation Status.
Please send comments about this document to public-media-annotation@w3.org mailing list (public archive).
Publication as a Working Draft does not imply endorsement by the W3C Membership. This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress.
This document was produced by a group operating under the 5 February 2004 W3C Patent Policy. W3C maintains a public list of any patent disclosures made in connection with the deliverables of the group; that page also includes instructions for disclosing a patent. An individual who has actual knowledge of a patent which the individual believes contains Essential Claim(s) must disclose the information in accordance with section 6 of the W3C Patent Policy.
1 Introduction
1.1 Purpose of this specification
1.2 Formats in scope
1.3 Formats out of scope
2 Terminology
3 Property value type
definitions
3.1 URI
3.2 String
3.3 Integer
3.4 Float
3.5 Date
4 Property definition
4.1 Core property
definitions
4.1.1 Description of the approach
followed for the property definitions
4.1.2 Core properties
4.2 Property mapping table
4.2.1 Rationale regarding
the mapping table
4.2.1.1 Semantic
Level Mappings
4.2.1.2 Syntactic
Level Mappings
4.2.1.3 Mapping
expression
4.2.2 The mapping table
A Candidate Additional
Elements
B References (Normative)
C References (Non-Normative)
D Acknowledgements (Non-Normative)
This section is informative.
This document introduces the Ontology for Media Resource 1.0, a core vocabulary to describe mapping between different metadata formats of media resources, mostly targeted to media resources on the Web. It is defined based on a core set of properties which covers basic metadata to describe media resources. For example creator is a common property that is supported in several existing metadata formats, and is therefore part of the core vocabulary. This vocabulary describes a mapping between different schemas containing this property. Ideally, the mappings should be semantics-preserving, but this is not achieved with the first version of the Ontology (Media Ontology 1.0), because of the difference in nature of the properties in the mapped vocabularies. Their extension is not exactly overlapping and their values might differ in syntax too. For example the propertydc:creator from Dublin Core and the propertyexif:Artist defined in EXIF are both mapped to the property Creator of our Ontology, but the extension of the property in the exif vocabulary (the set of values that the property can refer to) is more specific than the one of Dublin Core. Mapping back and forth with our ontology as reference will hence induce a certain loss of semantics. This is inevitable if we want to achieve a certain amount of interoperability.
The Ontology defines mappings between a set of vocabularies and a set of core properties in our own namespace. Although some of these properties can seem to be redundant with the Dublin Core set, we defined our own namespace for several reasons:
Dublin Core is one of the vocabularies that we take into account in the mappings.
The Dublin Core set does not cover all of our needs, we would hence still have to create properties in our own namespace.
More importantly, the Dublin Core properties have been created with a set of restrictions, although loose, and we might want to apply other restrictions to our properties. we have to have "our hands on" the set of properties to be able to control or constrain their behavior, and cannot be dependant on an external source of authority for the definition of our core mapping. For a practical use of the Media Ontology in an API, we define type restrictions for our properties that go beyond generic Dublin Core specification.
The core set of properties and mappings, i.e. our ontology, provides the basic information needed by targeted applications (see Use Cases and Requirements for Ontology and API for Media Object 1.0) for supporting the interoperability among the various kinds of metadata formats related to media resources, and particularly media resources on the Web. In addition, the ontology will be accompanied by an API that provides uniform access to all elements defined by the ontology. Although the set of properties is now limited, it already constitutes a proof of concept.
The initial version of this document contains only a limited set of properties with their corresponding mappings to a subset of the vocabularies listed in section 1.2 Formats in scope. The mapping was done to a subset of these vocabularies as a starting point, the subset chosen was voted by the Working Group as the vocabularies at the core of the gorup participant's expertise. This set of vocabularies can be later expanded to more or all of the formats in scope of the Ontology. This document is nevertheless being published at this stage of mappings, with the aspiration to gather wide feedback on the general direction of the Working Group. In particular we would like to encourage feedback on section 4 Property definition.
This specification defines an ontology for cross-community data integration of information related to media resources, with a particular focus on media resources on the Web. The purpose of the ontology is to help circumventing the current proliferation of video metadata formats by providing full or partial mappings towards existing formats.
The following table lists the formats that were selected by the working group as in-scope, along with the identifiers which are used as prefixes to identify them in this specification.
Note:
This specification is based on a review of existing formats and the properties they provide. This review does not aim to be complete, and this specification does not aim to cover all properties defined in these formats. The choice of properties is motivated by their wide usage.
Identifier | Format | Example | Reference |
cl11 | CableLabs 1.1 | cl11:Writer_Display | Cablelabs 1.1 |
cl20 | CableLabs 2.0 | cl20:Producer | Cablelabs 2.0 |
dig35 | DIG35 | dig35:ipr_name/ipr_person@description='Image Creator' | DIG35 |
dc | Dublin Core | dc:creator | Dublin Core |
ebucore | EBUCore | ebuc:creator | EBUCore |
pmeta | EBU P-Meta | pmeta:Contribution | EBU P-META |
exif | EXIF 2.2 | exif:Artist | EXIF |
frbr | FRBR | frbr:Person | FRBR |
id3 | ID3 | id3:TCOM | ID3 |
iptc | IPTC | iptc:Creator | IPTC |
it | iTunes | it:©ART | iTunes |
lom21 | LOM 2.1 | lom21:LifeCycle/Contribute/Entity | LOM |
ma | Core properties of MA WG | ma:creator | 4 Property definition |
media | Media RDF | media:Recording | Media RDF |
mrss | Media RSS | mrss:credit@role='author' | Media RSS |
mets | METS | mets:agency | METS |
mpeg7 | MPEG-7 | mpeg7:CreationInformation/Creation/Creator/Agent | MPEG-7 |
nmix | NISO MIX | nmix:ImageCreation/ImageProducer | MIX |
qt | Quicktime | qt:©dir | QuickTime |
media | SearchMonkey Media | media:type | MediaMonkey |
dms | DMS-1 | dms:Participant/Person | DMS-1 |
tva | TV-Anytime | tva:CredistsList/CredistItem | TV-Anytime |
txf | TXFeed | txf:author | TXFeed |
vra40 | VRA Core 4.0 | vra40:agent | VRA |
xmp | XMP | xmpDM:composer | XMP |
yt | YouTube Data API Protocol | yt:author | YouTube Data API Protocol |
The following formats have been decided to be out of scope for this specification.
MPEG-21: It is not media description in narrower sense.
This section is normative.
Any Resource (as defined by [RFC 3986]) related to a media content. Note that [RFC 3986] points out that a resource may be retrievable or not. Hence, this term encompasses the abstract notion of a movie (e.g. Notting Hill) as well as the binary encoding of this movie (e.g. the MPEG-4 encoding of Notting Hill on my DVD), or any intermediate levels of abstraction (e.g. the director's cut or the plane version of Notting Hill). Although some ontologies (FRBR, BBC) define concepts for different such levels of abstraction, our ontology does not commit to any classification of media resources.
A property is an element from an existing metadata format for describing media resources, or an element from the core vocabulary defined in this Working Group. For example, the Dublin Corecreator element and the Media Ontology creator element are properties. A property links a Media Resource with a value: dc:creator links a given resource with the value of its creator (Dublin Core specifies: "Examples of a Creator include a person, an organization, or a service."). The properties can have structured and/or unstructured values. The set of properties selected to be part of the Media Ontology Core vocabulary is listed in section 4 Property definition.
The notion of Mapping refers to the description of relations between elements of metadata schemas; in our case the mapping concerns a subset of the "in scope" Vocabularies, and the properties of the core vocabulary of the Media Ontology. These Mappings are presented in section 4.2 Property mapping table.
Property value types are the types of values used in a property. For example, the property dc:creator can have either string or URI as value types. Property value types are defined in sec. 3 Property value type definitions. They are relying mostly on XML Schema data types [XML Schema 2].
Note:
Currently the data types are defined in terms of XML Schema, part 2. The Working Group intends to check and potentially modify these definitions, to ensure compatibility with the return types defined in API for Media Resources 1.0.
URI "Uniform Resource Identifier" is defined in [RFC 3986]. In this specification the term URI is used since it is well known. However the term is used as meaning IRIs "Internationalized Resource Identifiers (IRIs)" [RFC 3987], that is URIs which may contain non-escaped characters other than ASCII. The data type is anyURI .
A String
value is represented using the XML Schema string
data type.
A Integer
value is represented using the XML Schema integer
data type.
A Float
value is represented using the XML Schema float data
type.
A Date
value is represented using the XML Schema dateTime
data type.
This section is normative.
This list of core properties has been defined by making an initial set of mapping propositions from the list of vocabularies in scope. From this original mapping table, it has been checked which of the properties were supported by most of the vocabularies, and which ones were judged useful by WG resolution. A list of properties by a cross validation between the member's opinion and the popularity of the properties across the vocabularies has been selected in order to get some level of objectiveness.
The following information is available for each property:
Rough description of purpose
Mappings to existing formats
Editorial note | |
This core list of properties is neither closed, nor pretend to be exhaustive and that the group still looks at rationale for including properties to be considered in the final list of properties. The candidate additional properties taken into account by the WG are summarized in A Candidate Additional Elements. In addition this table will be elaborated for further information. |
Name | DataType | Description |
Descriptive Properties (Core Set) | ||
Identification | ||
ma:identifier | { identifier:URI, type:String } | The URI identifying a resource, which can be either an abstract concept (Hamlet) or a "Representation" (an instance or file accessible at a given URL). See use case 4.4 of the Annotating Media Fragments Use Case document. |
ma:title | { title:String, type:String } | The title of the document, or the name given to the resource. |
ma:language | String([BCP 47]) | The language used in the resource. Recommended best practice is to use a controlled vocabulary such as [BCP 47]. |
ma:locator | URI | The URI at which the resource can be accessed (e.g. a URL, or a DVB URI). |
Creation | ||
ma:contributor | { identifier:URI|String, role:String } | A pair identifying the contributor and the nature of the contribution. E.g.actor, cameraman, dirctor,singer, author, artist (Note: subject see addition of contributor type). |
ma:creator | { identifier:URI|String, role:String } | The authors of the resource (listed in order of precedence, if significant). |
ma:createDate | { date:Date, type:String } | The date and time the resource was originally created. (for commercial purpose there might be an annotation of publication date). |
ma:location | { [name:String], [longitude:Float, latitude:Float, altitude:Float], [geoLocation:String] } | A location where the resource has been shot/recorded. |
Content description | ||
ma:description | String | A textual description of the content of the resource. |
ma:keyword | String | A descriptive phrase or keyword that specify the topic of the content of the resource. |
ma:genre | String | The genre of the resource. Recommended best practice is to use a controlled vocabulary such as the EBU vocabulary. |
ma:rating | { identifier:URI, value:Float, max:Float, min:Float, context:String } | A pair identifying the rating person or organization and the rating (real value). The scale of the rating should be provided. |
Relational | ||
ma:relation | { identifier:URI, relation:String } | A pair identifying the resource and the nature of the realtionship. E.g. transcript, original_work. |
ma:collection | URI|String | A name of the collection from which the resource originates or to which it belongs. |
Rights | ||
ma:copyright | { copyright:String, identifier:URI } | The copyright statement. Identification of the copyrights holder (Digital Right Management is out of scope for MAWG, see the [PLING] for more information on this topic). |
ma:policy | { policy:String, type:String, identifier:URI } | Description of the policy of the media resource or reference. see the [PLING] for more information on this topic. |
Distribution | ||
ma:publisher | URI|String | The publisher of a resource. Examples of a Publisher include a person, an organization, or a service. |
ma:targetAudience | { identifier:URI, classification:String } | A pair identifying the issuer of the classification (parental guide and issuing agency, targeted geographical region) and the value given in this classification. |
Fragments | ||
ma:fragments | { role:String, identifier:URI } | A list of pairs of fragment role and fragment identifier. Fragment types can be spatial, temporal or tracks (e.g. chapters, favourite scenes, subtitles). |
ma:namedFragments | { label:String, identifier:URI } | The list of named fragments annotated for this resource (pairs of label and fragment identifier). |
Technical Properties | ||
ma:frameSize | { width:Float, height:Float } | The frame size. For example: w:720, h: 480. It is required to use a pixel unit. |
ma:compression | String | The compression type used, e.g. H264. Note: it is possible to use extended MIME type as value for this property, see [RFC 4281]. |
ma:duration | Float | The actual duration of the resource. It is required to use a second unit. |
ma:format | String | MIME type of the resource (wrapper, bucket media types). |
ma:samplingrate | Float | Audio sampling rate. It is required to use a sample/second unit. |
ma:framerate | Float | Video frame rate. It is required to use a frame/second unit. |
ma:bitrate | Float | Average bit rate. It is required to use a kbps unit. |
ma:numTracks | Integer | Number of tracks. |
As a first step to build the Media ontology, a set of commonly supported properties by the aforementioned vocabularies (metatdata schemas) has been listed. This list, henceforth referred to as "the Media Ontology", is the basis for vocabularies matching. Its namespace is "ma", for Media Annotation. We provide a first set of mapping propositions between the Media Ontology and a subset of vocabularies taken into account in this Working Group. These mappings specify both the semantic and some elements of the syntactic correspondances between the Media Ontology and the considered vocabularies' properties.
The presented mappings are "one way" so far, i.e. the semantics is of a relationship between one Media Ontology property and one or more property/ies of considered vocabularies. For example, ma:copyright is mapped to both xmpDM:copyright and dc:rights (as part of the XMP standard [XMP]); the same property is mapped to exif:Copyright (see [EXIF]). Unfortunately, no semantic relationship can be thereof inferred between the properties in XMP and in EXIF. The mappings that have been taken into account have different semantics: the properties of the different vocabularies can be:
Exact matches: the semantics of the two properties are equivalent in most of the possible contexts. For example, ma:title matches exactly vra:title.
More specific: the property of the vocabulary taken into account has a semantic that takes into account only a subset of the possibilities expressed by the property defined in this Working Group. For example in DIG35, ipr_names@description and ipr_person@description are more specific than the property ma:publisher to which it is mapped.
More generic: the inverse of the above, the property of the vocabulary taken into account has a semantic that is broader than the property defined in this Working Group. For example, the DIG35location is more general than the ma:location.
Related: the two properties are related in a way that is relevant for some use cases, but this relation has no defined semantics. For example, in Media RSS, media:credit is related to ma:creator.
This list of relations between vocabularies (or informal mappings) and the "Core Media Properties list" is published as a table; its purpose is to get feedback from the communities that are currently using the different vocabularies: the people or companies actually using the different vocabularies could proof-read our interpretation of the vocabularies and comment on the proposed mappings.
Syntactic level mappings declare the correspondence between two semantically equivalent properties but with a different syntactic expression. Its most evident case is the date formatting, but some others may appear.
Editorial note | |
Currently the mapping table for some of the formats does not contain information about syntactic mapping, but this information will be added in the following version. |
The mapping expression corresponds to the concrete implementation or representation of the mappings defined in the previous paragraph, both at a semantic level and at syntactic one.
SKOS (acronym for Simple Knowledge Organization System) is currently a Recommendation of the W3C Semantic Web activity that defines a vocabulary for representing Knowledge Organization Systems (i.e. vocabularies) and relationships amongst them. In SKOS the mapping properties that we take into account in the mapping table are expressed as: skos:exactMatch, skos:narrowMatch, skos:broadMatch and skos:relatedMatch. Some more fine grained definition of the properties has still to be done: we need to define their formal properties (if they are symmetric, etc) to enhance more efficient concrete mappings. The mappings have to be as precise as possible to be efficiently used in the related API.
The following mappings are established from the Media Ontology's core properties to various multimedia metadata formats. This list of formats is not closed, nor pretends to be exhaustive. The group still looks at rationale for including and excluding formats to be considered in the final mapping table;
Below table shows the considered candidates for extending the core properties. In addtion, concerning to addtional information of candidate additional elements, see the wiki page of the MA WG.
Name | Description |
Descriptive Properties (Core Set) | |
Accessibility metadata | This is concerned to annotative material for accessibility. E.g. link to transcript, what 'axes' the media can be adapted -- such as closed captions, audio description of video, video contrast. |
Fictional character | Specify fictional character impersonated by a contributor with role "actor". This is supported in the metadata set of the Timed Text WG. |
Technical Properties | |
Media types | Media types for which certain technical properties are relevant. If necessary, this could evolve into media specific profiles. |
This document is the work of the W3C Media Annotations Working Group.
Members of the Working Group are (at the time of writing, and by alphabetical order): Werner Bailer (JOANNEUM RESEARCH), Tobias Bürger (University of Innsbruck), Eric Carlson (Apple, Inc.), Pierre-Antoine Champin ((public) Invited expert), Ashish Chawla ((public) Invited expert), Jaime Delgado (Universitat Politècnica de Catalunya), Jean-Pierre EVAIN ((public) Invited expert), Philip Jägenstedt (Opera Software), Ralf Klamma ((public) Invited expert), WonSuk Lee (Electronics and Telecommunications Research Institute (ETRI)), Véronique Malaisé (Vrije Universiteit), Erik Mannens (IBBT), Hui Miao (Samsung Electronics Co., Ltd.), Thierry Michel (W3C/ERCIM), Frank Nack (University of Amsterdam), Soohong Daniel Park (Samsung Electronics Co., Ltd.), Silvia Pfeiffer (W3C Invited Experts), Chris Poppe (IBBT), Víctor Rodríguez (Universitat Politècnica de Catalunya), Felix Sasaki (Potsdam University of Applied Sciences), David Singer (Apple, Inc.), Florian Stegmaier ((public) Invited expert), John Strassner ((public) Invited expert), Joakim Söderberg (ERICSSON), Thai Wey Then (Apple, Inc.), Ruben Tous (Universitat Politècnica de Catalunya), Raphaël Troncy (CWI), Vassilis Tzouvaras (K-Space), Davy Van Deursen (IBBT).
The people who have contributed to discussions on public-media-annotation@w3.org are also gratefully acknowledged.