The Australian Broadcasting Corporation has completed a proof-of-concept trial of CSIRO’s Continuous Media Web (CMWeb) application for integrating rich media content with searchable metadata.
Rob Garnsey, head of systems at ABC’s new media and digital services, said the technology has potential for the wider Internet and within the corporate network.
“This technology is about extending the functionality of the Internet to rich media such that management of digital assets is easier,” Garnsey said. “Being able to keep track of what’s in a digital file provides a framework for identifying content in catalogues which can be used for retrieval and research.”
Garnsey said the trial went well and the ABC’s expectations were met.
“We are in a relatively early stage with this technology, and although we have no firm plans, we will to talking to CSIRO about adopting it further,” he said. “The ABC has terabytes of rich media data that is waiting to be digitised so the potential for this technology here is definitely there.”
CSIRO CMWeb senior software engineer Conrad Parker said the proof-of-concept went “as hoped”.
“The proof-of-concept involved translating one news bulletin’s worth of video from the NewsML [markup language] format to Continuos Media Markup Language (CMML) which took about two days,” Parker said. “Because its based on existing Web infrastructure it worked better than expected.”
CMWeb uses the Annodex media format which encapsulates the rich media file and CMML allowing indexing and searching of specific moments within the file.
Parker said CMWeb has been in development for two years under Silvia Pfeiffer and after being announced in June, the ABC is the first to trial the technology.
“We are very excited about it,” he said. “Media indexing does for media what HTML does for text therefore enterprises can manage their media assets in a way that’s integrated into the organisation such as within an Intranet.”
Instead of watching the CIO’s presentation in terms of time, the point of interest can be sent as a URL which is embedded in the media file, according to Parker.
“For enterprises it is efficient to use rationalised media sources to deliver content across the Internet,” he said. “These can be company news bulletins or recorded meetings.”
Regarding the data repository requirements for CMWeb, Parker said this is very flexible.
“Media companies tend to have different databases in much the same way text was handled 15 years ago,” he said. “Annodex can be put in front of any database that supports a scripting language like Perl.”
Parker described the process of setting up CMWeb as “similar to setting up a Web site”.
“For the trail we installed Annodex on Red Hat 9.0 with Apache 2 as the Web server,” he said. “CMWeb can also be integrated with search engine facilities that can index a repository of Annodex media. This allows for searching of the specific sections of rich media files.”