Presentation of the intermediary results of the CELTIC project Mediamap on the EBU Production Technology Seminar by Maarten Verwaest.
(Geneva - 2010, February 3rd)
4. Side-Effects Difficult to exchange, to manage and to access media Often stored as unstructured data or in proprietary systems Different aspects of a single item are often managed by different systems Manual annotation is expensive and yet not scalable(!) Content dispersion Excessive Numbers of Copies and File Transfers Various formats - Transcoding bottleneck Cumbersome supply chain integration Newsfeeds Field Production (EFP, Outside Broadcast,...) Co-production 4
5. 5 Mediamapis a Eureka Celticprojectfinanced by the DGE (France) - IWT & IRSIB (Belgium) Q4 2008 – Q2 2011 Total effort: 59 man years http://www.mediamapproject.org
6. Challenges – From re-use to real-time co-production Production applications - deliver content in stead of bare essence Production tools must preserve available metadata; Value-added services inject high-level metadata; Metadata - From plain text to machine-readable descriptions Each format or genre requires a full-featured ontology (e.g. IPTC/NewsML-G2) Workflow - From “disclosure by annotation” to “information integration” trace the item throughout its lifecycle, i.e. the script, raw footage and edited clips; Semantic technology (RDF, SPaRQL) - express, store and query the relationships A virtual media asset management system implements a collaborative production environment and reduces content dispersion to a minimum. 6
7. 7 Digital Silos Collaborative Model Corporate Systems Media Asset Management systems enable re-use and collaboration. Custom development is expensive and results in moderate stability. A user-friendly online platform will enable real-time co-production. Independant developers will create reusable apps, thus resulting in increased stability and lower cost Digitisation has enabled increased functionality Corporate Network Central Media Asset Mgnt Access via Internet ApplicationPlatform Infrastructure
16. Outstanding Issues Each broadcaster has a data model > There is no formal or de facto standard vocabulary MXF is a wrapper for essence > There is no standardised syntax for metadata Production Tools only produce essence > They don’t support the editorial context MAM systems are designed to support archive processes > Production processes require a fundamentally different logic Creative processes usually happen off-line 16
17. Mediamap – Beyond the State of the Art Research and Development The Unified Semantic Entity (USE) is a powerful wrapper of audiovisual content. It ensures preservation the semantic context; The Open Semantic Bus (OSB) extends of ISO/OAIS. It implements cross-system references to make information persistent. Prototypes Advanced Product Engineering methods; Develop and integrate a smart camcorder; Media Asset Mgntservices designed to support production processes. Validation - Use Cases “Birthday of the Expo 58 in Brussels” Congo 60 Final Demo on Collaborative Production - TBD 17
22. Open Semantic Bus (WIP) 22 Based on FP6 Memories ‘AXIS’ Acquisition, eXchange, Indexing and Search of Audiovisual Material Based on ISO 14721:2003 (OAIS) reference model RDF and OWL based definitions for data and semantics Mediamap extensions NewsML-G2 translation Integration with script editing, raw material acquisition and craft editing
28. Semantic technology is a necessary condition to deal with distance – in space, in time or introduced by different systems and cultures. However, semantic awareness requires more than RDF and OWL. We still need a adequate vocabulary and a proper data model. The most important source of descriptive metadata is pre-production or script editing and we still need a standard vocabulary (an “ontology”) and a syntax (a “wrapper”) to support (non-news) editorial systems – A-S-A-P! In the advent of Web 3.0 (exchange of information),we need to think beyond interoperability. Production tools must understand and preserve the context in which the content is being created, so that it will be exploitable for other people (elsewhere, later, different). 28