Reloaded !

“I'm sitting in a former naval barracks in Alameda, California, watching the digital assembly of a human face. Bones, teeth, glistening eyes. Layer upon layer. Finally the hair and skin, the creases and tiny scars that make us who we are. The face blinks and breathes. Then it snarls, and my skin crawls.
Agent Smith is back, and he's pissed.

”People get really preoccupied with, 'Are you going to top yourselves this time? Are you really gonna come up with a zinger?'“ Gaeta tells me. ”My job has nothing to do with making zingers. The point is not to knock you over with a visual trick. The point is to be able to construct events that are so complex, in terms of what human bodies need to do, that the total 'effect' is impossible choreography. 'My God! It looks real, but it just can't be.'“

If the dojo fight in The Matrix was a kung fu sonata, the Burly Brawl is a symphony. Neo tears the sign from the ground and wields it as a kendo sword, vaulting pole, and battering ram. A woman walking by can't believe what she's seeing; suddenly her body is hijacked, she drops her grocery bag, and another Smith charges into the fray. Whole battalions of Smiths arrive, mount assaults, attack in waves, scatter, regroup, and head back for more. (At ESC, one massive pile-on was dubbed the ”Did someone drop a quarter?“ shot.) In the thick of it, Neo is dancing, chucking black-tied bodies skyward, pivoting around the signpost, and using shoulders as stepping-stones over the raging river of whup-ass.

Creating the Burly Brawl, however, is a taller order than inventing Bullet Time. To portray Neo in hand-to-hand combat with more than 100 Agent Smiths in the old way would have required Escher-like tangles of crisscrossing still-camera rigs and years of compositing. What Gaeta needed was a virtual camera that could fly through the 3-D scene - as free from the laws of space and time as Neo is from the physical laws of the Matrix.

”The concept of Bullet Time had to graduate to the true technology it suggested,“ he says. ”For Reloaded, we had to finish the job so that we could get relentless, uninterrupted, and editable chunks of Neo in the zone.“

While the topography of the human face is the hardest to simulate digitally, it turns out to be one of the easiest to map photogrammetrically. It has fewer shadows and occlusions than, say, the city of Paris. The language of the face communicates maximum information through the subtlest inflections. The interfaces of our souls are designed to be read in a heartbeat.

Then Reeves and Weaving each sat down on a stage in front of five Sony HDW-900 video cameras. The massive datastreams from these cameras - one gigabyte a second - were treated like holy water; even the cameras' color-correction software was disabled to prevent any loss of data. Instead of recording to tape, which requires compression, the cameras were modified to send uncompressed data to a bank of high-end PCs that stored it on a huge disk array. ”The scene in that room was surreal,“ Gaeta recalls.

”