Videoconferencing is one of those nice ideas that has never much caught on because it's so pricey. However, with the terrorist scare during the Gulf unpleasantness, lots of executive types suddenly didn't feel too much like flying, evidently placing a fairly high price on their lives since they turned to videoconferencing in droves. The reason for videoconferencing's high cost is limited bandwidth. Transmitting the equivalent of two television shows in real time so both can react to the other takes a lot of bandwidth, and bandwidth is still expensive. The best way to reduce the bandwidth requirement is with video and audio compression techniques, but to keep everything operating in real time (as opposed to virtual time?) the compression must be done in hardware and even then it has to have a fast and tight algorithm to be of any use. All this increases the price of the video coders/decoders, which aren't cheap to start with.
I'm no expert on what's involved with videoconferencing, but it seems to me that the standard solutions are using the brute force method. Unfortunately, the brute isn't large enough. However, if you think about the technologies that we have available today, it wouldn't be all that hard to whip up a quick and dirty videoconferencing system that wouldn't need special hardware or software. Consider this. Timbuktu allows two Macs to share screens over a modem link, even in color. Therefore Farallon (and probably the people who do Carbon Copy/Mac) have a pretty good idea how to send commands rather than data across the link. Assuming similar systems on both ends, a command can produce a large amount of data that would take much longer to transmit. Add to the Timbuktu technology (along with a little video wizardry thrown in for realism) the animation capabilities of interFACE, or whatever Bright Star calls their talking head product these days. Most of the time when you're talking to people they're just sitting in about the same position, so interFACE could simulate the other person's head quite well. The one part of my system that doesn't exist as far as I know is a technology to record sound, compress it tightly, transmit it, and expand on the other end, all while synchronizing with the simulated head. This system wouldn't be as complete as a true videoconferencing system, but it would be a workable substitute that wouldn't require several hundred thousand dollars in hardware and connect charges. It probably won't happen, but it's not for the lack of innovation consulting on my part.
Back in the real world, a company called PictureTel has shown a prototype videoconferencing system using a IIci or a high-end PC clone. From the sounds of it, PictureTel's system will do more than just allow two people to see each other and talk at the same time. You should be able to include graphics and other computer-generated information directly, since the entire system is running on a Mac or PC. PictureTel will also be making the coders/decoders for Macs and PCs, presumably for NuBus and the PC buses (ISA, EISA, and MCA). No hurry to get your order in though, PictureTel won't be going to market with this stuff until 1992 or 1993.
An interesting solution to the bandwidth problem comes from Digital Access Corp., which markets a line of modems called Fracdial. The Fracdial modems allow users to choose how much bandwidth they want, up to the 1.5 megabips speed of a T1 link. The modems accomplish this feat by using multiple digital lines and synchronizing them so the applications on either end think they have a single high-speed line. Sometime this fall the Fracdial modems will get a bridge that will allow videoconferencing for about $2500, which is apparently a lot less than standard videoconferencing bridges. The other feature Digital Access has planned for the Fracdial modems is a direct interface to Unix workstations, which should increase the product's popularity significantly.
Related articles:
Communications Week -- 24-Jun-91, #357, pg. 19, 22
