By the end of this year the World Wide Web, already swamped with textual information, will be flooded with audiovisual files as bandwidth ramps up.

And since all of those huge digital and audio files will need to be managed, demand for CSIRO's recently released Annodex Continuous Media Markup Language (CMML) is certain to grow, says Project Leader Dr Siliva Pfeiffer.

A world first, CMML allows users to search, access, navigate and query time-continuous media the way HTML does for text. Dr Pfeiffer says after CSIRO released the Firefox plugin on Slashdot in February there were 4000 downloads in a single day, and uptake since then has continued to be huge.

CSIRO chose Firefox as a development platform over Internet Explorer (IE) because the Firefox's plugin architecture was easier to work with and the browser was supported across several platforms. The initial Windows version has now been followed by OS/10- and Linux-capable plugins.

"We only just brought out these plug-ins in January, so that was the most important step really, for people to be able to use the new data; now the next important step is that people create web sites that have that sort of information inside it so that they are hosting audio and video that can be called with a format that we have developed.

"We've had quite a few requests from people who want to start setting up these websites," she says.

The plug-in is a file format that allows the inclusion of Metadata in enriched audio and video files, and subdivides those files into a sequence of separate clips. An Annodex search delivers actual video content. Unlike tools like Google's beta video search, which search a program's closed captioning text then deliver an excerpt from the text and a single still image from the program, an Annodex search provides the user with a detailed summary of the video content, an interactive list of video clips, and hyperlinks to additional material.

Anodex has been developed within CSIRO for the last five years and is being issued as an open international standard through the IETF, the standards body that provided TCP/IP.

"We've been doing research into multimedia for a while - probably about 10 years or more - and the big challenge in multimedia recently has been the automatic extraction of content information from audio video files," Dr Pfeiffer says.

"The problem that we all see is that everything is going digital in audio and video, and we'll all be swamped by lots and lots of information with digital audio and video files as many people will put their files online. Lots of people have large collections of audio and video files and we can't really manage them - you can't really get to information inside them, address points of interest or interesting events inside videos."

CSIRO has been addressing the issue along two lines. The first is audio video content analysis using signal processing to allow automatic extraction of information from audio and video files. The second line of research has been into the build up of infrastructure within which users can manage information, including the ability to point into audio and video files, recombine clips out of audio and video files, hyperlink clips to other information, and search within files.

CSIRO has licensed the software under the Mozilla Public Licence, the GNU General Public Licence and the GNU Lesser General Public Licence. All source code for Annodex is available from the CSIRO'S Web site.

"The whole research field in multimedia has been moving towards this for a while but there is no standard that pulls things together they way we have. There's a few different things that are happening, a few other standards that are happening in the field, but we are the only ones really to be fitting it into the whole infrastructure of the World Wide Web."