Tracking MPEG-4 At Streaming Media East
Non-Natural Codecs (BIFs, Facial, 3D, Structured Audio, AFX)
The afternoon panel on Non-Natural Codecs in MPEG-4 had to be the most technical session of the conference, where we were able to hear from the researchers developing the next generation of software applications for creating MPEG-4 content. This session focused on the features of the standard that could take it far beyond the streaming media content we know today by implementing an object-based coding mechanism. With less emphasis on products, there was a great deal more technical knowledge passed along on how these features actually get implemented and their applications for future business areas.
Below is a description of the panelist’s presentations, but for a more general background on BIFs and object-coding, please check out my tutorial on "MPEG for All."
• Eric Petajan, Chief Scientist and Founder of face2face animation, inc. discussed the facial and body objects that we might start seeing in the gaming market. He offered a very impressive demo of an audio+facial animation streamed at 14 kbps over a wireless modem. Facial animation only consumed 2 kbps of that data rate, while body animations that we'll start seeing in version 2 of their application will be in the 4-8 kbps range.
• Gabriel Taubin, Research Staff Member at IBM T.J. Watson Research Center discussed different methods for 3-D mesh encoding in detail including topological coding and progressive forest splitting. The latter transmits coarse meshes before finer meshes in a stream and would therefore be beneficial for scalable applications over various network connections.
• Brane Zivkovic, President of Soundball, Inc. covered the technical details of structured audio, synthetic coding and text-to-speech applications built into MPEG-4. The essence of these components lies in being able to create multi-level audio tracks, which play back richer sounding audio across many different devices. Research efforts in this arena continue at MIT and Berkeley, with engines for content creation and delivery still under development at Soundball.
Alexandros Eleftheriadis, Associate Professor of EE at Columbia University moderated a discussion of applications between the audience and the panelists. Clearly, many of these components are still in the research phases, but these are the companies to watch for cutting-edge MPEG-4 applications in the coming year.
Page 4: Applications >>