In 2011, L.A. Noire's complex face capturing technology looked like the next big thing for video games, or maybe digital entertainment in general. And perhaps it still will be, but it's amazing to see what difference a year makes. It took Team Bondi a rig of 32 cameras and years of work to capture the face of Detective Phelps and the rest of L.A. Noire's cast. Now, look at what a filmmaker and programmer have done with a Kinect and a DSLR.
Working with Carnegie Melon University's Golan Levin, filmmaker Jonathan Minard and artist/programmer James George put together a little project called "virtual cinematograhy." A high resolution image of Levin's face is captured by an SLR and grafted onto a Kinect depth model. Instant 3D model.
Levin is captured with camera and Kinect answering Reddit AMA questions. Because there's only one Kinect, the model isn't perfect--motion sometimes obscures parts of Levin's face, causing blank spots to appear. Additional cameras would solve that problem, and still at a much lower cost than the elaborate setup used to create L.A. Noire.
This is the kind of technology that could completely reshape virtual filmmaking someday. Motion capture won't just be used to create lifelike animation which is then attached to artist-created 3D models. Motion capture will insert a real actor into a virtual world, complete with completely real movements and facial expressions.