IBM Corp in the US has announced a prototype system that it says enables a video picture to be composed while the data is still compressed. This, it reckons, dramatically cuts the required processing power. It says that the system’s computational requirement is 1m operations per second, compared with about 50m for systems using the conventional method of decompressing data before it is displayed. The prototype videoconferencing system is claimed to enable unlimited numbers of users to communicate without slow motion or mismatches between the sound and picture. It is called the Multimedia Multiparty Teleconferencing system, and the company says that when mass-produced the product could sell for around the price of an everyday personal computer add-in board: it has applied for a patent for the system’s enabling ideas. Other features include the ability to save and re-play conferences, and to integrate private conversations into group conferences. In addition to straightforward videoconferencing, the company says that the system is suitable for such applications as electronic blackboards, and computer-aided design. While a prototype system is on trial running under AIX Unix on a PS/2, the company says that adapting the system for OS/2, Windows, or RS/6000 should be relatively simple. Similarly, at the moment the product is limited to local network usage but IBM believes that long-distance applications should be practical.
