Hollywood is not high tech. Hollywood is not even industrial. Hollywood is feudal.

Power in the film and television businesses has always lain in the ability to control people, in building a web of allegiances so strong and deep that a king (a movie studio head or TV network programming boss), can order his lords (producers) to field an army of skilled warriors (directors, writers, and actors). Superagents like Michael Ovitz (the Black Knight) put together packages of mercenaries for hire by the highest bidder. Motion Picture Association of America president Jack Valenti plays the Pope. There is even a guild system (the Directors Guild, Writers Guild, and Screen Actors Guild, for example). Everyone else, including the audience, is just a peasant.

Like all feudal societies, Hollywood relies on the building of palaces and other acts of conspicuous consumption to show who is strong and who isn't. Power is held closely and formal corporate structures exist mainly to satisfy the banks. Business is often done in public over baronial lunches at the Ivy or royal dinners at Spago. Valet parking is provided. Honors are handed out to nearly everyone. Billy Crystal plays the MC.

Suddenly it makes perfect sense why Japanese companies like Sony and Matsushita were attracted by American movie studios; they were already comfortable with the system, a thriving Samurai society with no threat of a renaissance in sight.

Yeah, but why am I, a computer guy, writing about this? Because I sense that the entertainment business is poised for a new type of technological revolution. And no, it's not the Internet. That revolution already fizzled as a delivery channel and will be forgotten for another five to 10 years, when suddenly it will once again smack us all upside our heads. I'm talking about a revolution not in how movies are delivered, but how they are made. And any revolution that is going to be successful has to fit with the power structure of the society.

In the entertainment business, technologists have always, though reluctantly, given way to the showmen. Edison may have invented the movies, but his idea of entertainment was putting four or five different subjects on an eight-minute reel. He stubbornly fought against Edwin S. Porter, who wanted to make films that would tell stories. When D.W. Griffith's "The Birth of a Nation" finally defined the modern film in 1914, movies were already more than 20 years old. His camera may have been the same as Edison's, but Griffith's use of editing, the closeup, the moving shot, the fade, the cut, and the dissolve were what changed the world. "The Birth of a Nation" was 12 reels long, cost $100,000 to make, and grossed $15 million. It was a triumph of technique over technology.

What Griffith did was turn Edison's engineering exercise into an art form. This happened again when talking pictures appeared in the late 1920s, and when Hollywood first opposed and then absorbed the television production business of the 1950s. And it's about to happen again as Hollywood embraces MPEG-7, shifting the technical emphasis from what can be presented using the new tools, to how it can be presented.

The Motion Pictures Expert Group standards (MPEG-1, -2, -4, and now -7) have long defined how audio and video can be compressed and decompressed. But there is a lot more to entertainment than compression. There is how a scene is presented. This idea of a multimedia description language — a kind of PostScript for movies — began with MPEG-4. The idea then was to make it easier to compress video for low-bandwidth connections if we could separate each scene into components. This has long been done in animation, where characters are photographed against a background scene. In the context of MPEG-4, then, it ought to be possible to compress the foreground and background separately, saving lots of bandwidth in the process since the background can be held in cache memory and rendered over and over again. It's a good idea, but this capability of MPEG-4 has hardly ever been used. That's why MPEG-7 takes this idea even further.

MPEG-7 is called a Multimedia Content Description Interface, and compression is only one of its functions. MPEG-7 takes every aspect of a scene to a lower level of abstraction. For visual material, a lower abstraction level would be a description of shape, size, texture, color, movement, trajectory, and position. Audio material can be reduced to key, mood, tempo, tempo changes, and position in sound space. The highest level abstraction would give semantic information like, "This is a scene with a barking brown dog on the left and a blue ball that falls down on the right, with the sound of passing cars in the background." Taking this simple scene description and rendering it with all the information from the lower abstraction layers (this is what a ball looks like, acts like, sounds like) ought to produce something like a movie scene. The greater the amount of information available at each abstraction layer, the more realistic the movie, all within some very realistic bandwidth constraints, which was the idea in the first place. MPEG-7 replaces compression with description for visual media in the same way the PostScript page description language replaced bitmaps with bezier curves. Beyond a certain level, with MPEG-7 it ought to be hard to distinguish the rendered scene from reality. At least that's the idea.

Enter once again the artists and businessmen to apply this idea to a new function. The ultimate use of computers in film production would be to completely eliminate the studio, sound stages, sets, even the actors, reducing the making of movies to rooms of computers and rich people taking power lunches. The actors would be replaced by synthespians, who will exist only as libraries of gestures and expressions in a computer bank somewhere.

Ironically, the greatest problem with synthespians isn't technical but legal. "Replacing actors is not impossible," says "Titanic" director Jim Cameron, "but it is a really strange legal area. The film company may own the character, but the star owns his likeness. What, for example, does Harrison Ford think about an 'Indiana Jones' video game? Sure we can replace actors, but this business is not based on actors, it's based on movie stars, and those stars are as important off the set as on in the business of celebrity. We might well be able to replace the star on the set, making the production process more efficient and the stars forever young, but we'd still need them to do publicity."

But what about taking, say, the 70-odd episodes of the original "Star Trek" series, extracting all the gestures and expressions of a 35-year-old Captain Kirk, and then using the computer to produce all-new episodes with the original cast? "Sure, we could do it," said Carl Rosendahl, producer of Dreamworks' "Antz," the computer-animated feature. "But it would be expensive, and we'd rather be applying computer technology to doing new things, things that couldn't be done any other way. Though with William Shatner's limited dynamic range, I'm sure that analyzing four or five old episodes would be more than enough."