Managing Digital Video Content

manjula patel

Managing Digital Video Content

Manjula Patel reviews the two-day workshop on current and emerging standards for managing digital video content held in Atlanta, Georgia, 15-16 August 2001.

“Managing Digital Video Content” [1], a two-day workshop on current and emerging standards for managing digital video content took place on 15-16th August, 2001, in Atlanta, Georgia. The workshop was sponsored by ViDe, the video development initiative [2], the Southeastern Universities Research Association, SURA [3], Internet2 [4] and the Coalition for Networked Information, CNI [5]. Approximately 180 delegates attended, the majority from the States, peppered by one or two from Europe and Australia. Note that most of the presentations are available from the workshop web site [1].

Standards-Based Content Management

The theme for the first day was “standards-based content management”. The workshop kicked off with an overview of ViDe digital video initiatives [2] by Jill Gemmill, current chair of the video development group. This was followed by an overview of SURA and Internet2 digital video initiatives by Mary Fran Yafchak, IT Program coordinator for SURA [3] and co-chair of the Internet2 [4] digital video work group.

The keynote address for the day was entitled “Globally sharing information assets” and was given by Clifford Lynch, Executive Director of the CNI [5]. The major themes running through the keynote were the lack of experience in handling video assets in terms of metadata, rights management, delivery of video and infrastructure components. Lynch highlighted the fact that video assets are a lot more complex than the conventional textual materials that we are accustomed to dealing with, and that technology is only just reaching the point where it is becoming feasible to do things with video. Although tools for creating (bad) video are now widely available, the real issue in managing digital video is that of stewardship. Capturing of events such as workshops, conferences and seminars is fairly common nowadays, however it is often done with a blatant disregard fir IPR and rights issues –too many events are recorded in an ad-hoc manner. Lynch argued that libraries are well-placed to exercise stewardship over digital video assets given their experience in this area. However, it will require a significant change in mindset, for example they would have to be aware of activities on campus and take an active role in capturing them.

The next presentation, “Putting the pieces together”, was given by Grace Agnew, Assistant Director for systems and technical services, Georgia Technology Library and chair of the ViDe working group on digital video access working group [6]. People tend not to use video as a “normal” information object, so that there is a real danger of “ghetto-ising” video assets. Agnew called for the immersion of moving images into mainstream information. Also, many image archives are small, often home-based without a web presence at all, hence data interchange is a critical aspect. In addition, there is a tremendous amount of analogue material which needs to be taken into account.

XrML [7], ODRL [8], MPEG-7 [9] and MPEG-21 [10] are likely to become important for rights metadata, while SOAP [11], OpenURLs [12] and open video bucket architecture [13] could feasibly provide “smart transport” for such information.

Jean Hudgins, Georgia Tech Library and Dan Kniesner, Oregon Health and Sciences University Library –both members of the ViDe digital video access working group then provided an overview if the ViDe Dublin Core application profile for digital video[6].

This is based on qualified Dublin Core[14] with extensions for video.

Lunch was followed by a session on digital video accessibility by Chris Hodge, University of Tennessee. There are a whole host of issues which need to be addressed in this area. Although research is being undertaken in various areas such as speech recognition and automatic captioning, there is currently no system available that would be adequate on the scale required.

Mairead Martin, Advanced Internet Technologies, University of Tennessee and Doug Pearson, Indiana University then spoke on the relative merits of XrML [7] and ODRL [8] in the context of rights metadata. Also, multimedia in an object-oriented MPEG-4 space implies that permissions would be required for each object, raising issues with regard to the granularity of descriptions. We should also bear in mind that Digital Rights Management (DRM) is not just about enforcing restrictions, but also about managing interactivity, reuse and reproduction of digital assets.

The rest of the afternoon was taken up by three concurrent breakout sessions: “Applying DC to digital video assets”; “Applying rights metadata to digital video assets” and “Implementing OAi using the ViDe DC database”. It was refreshing not to have to choose which ones to attend and which to miss out on, since all three sessions were run three times to provide all the opportunity to attend all three –although this was somewhat hard on the people involved in running the sessions!

Emerging Technologies and Asset Management Implementations

The second day began with Michael Nelson, University of North Carolina, talking about the Open Archives Initiative [15] –OAi : Past, Present and Future. It appears that distributed searching which was in vogue a few years ago, is now falling out of favour. In a typical internet environment, difficulties are encountered when the number of nodes reaches 100 or more. The OAi protocol is concerned with harvesting metadata (not full-text). Content editing, input and deletion of records are all outside the scope of the protocol and must be handled in the back-end systems.

The second keynote address was given by Jane Hunter, Distributed Systems Technology Centre (DSTC), Australia. She spoke about “MPEG-7: Transforming digital video asset description”. A comprehensive coverage was provided of the MPEG-7 ISO/IEC standard [9] being developed by the Moving Picture Experts Group (MPEG) [16]. The standard is due out in September 2001. MPEG-1, MPEG-2 and MPEG-4 are mainly concerned with compression, whereas MPEG-7 focuses on metadata. The richness and complexity of MPEG-7 made it apparent, to me at least, that there is scope for an MPEG-7-Simple template. Hunter also gave a brief overview of MPEG-21 [10] which provides for a multimedia framework and covers content management and usage as well as IP management and protection, amongst many of its other facets.

Following on from Jane Hunter’s coverage of MPEG-7, Anna Benitez, University of Columbia, described several interesting applications which involve making use of MPEG-7 implementations. The applications are research projects being undertaken by the Digital Video and Multimedia group [17].

During the rest of the afternoon, attention turned to other implementations for managing digital video. Jim DeRoest, Research Channel and University of Washington, described the high-end UW MediaWeb Project, followed by Jon Dunn, who spoke about digital music and audio projects at Indiana University [18]. Kathy Christensen of CNN News archives, then detailed issues relating to digital asset management at CNN, where they are investigating face and voice recognition, for example to identify or search for reports by a particular reporter. The development of footage.net [19], which claims to be the largest union catalog of footage on the web, was then described by its founder, John Tariot. He explained that neglect of standards had resulted in difficulties with performing cross-searching, fielded searching and sorting, for example on dates.

The workshop closed with a vendor panel comprising representatives from Virage[20], Blue Angel Technologies [21] and Ascential Software [22], all of whom had been demonstrating their systems throughout the workshop in an adjoining room.

Conclusions

The workshop proved very successful, all credit due to its organisers. The two day programme was packed full of interesting and informative presentations and discussions, which generated a definite “buzz” throughout its duration. The major theme to watch out for in the future is that of digital rights management (DRM) –an area touched upon by many of the speakers and one which needs a lot more attention.