Claims are being made that Sub D is not optimized for use in games.
A user on Reddit recently addressed Infinity Ward’s tech demo for the upcoming Call of Duty: Ghosts, in which the developer stated that it was using Sub D technology for rendering high-polygon models.
According to the user, models created with Sub D use an aggressively high level of polygons which can’t be handled by current gen hardware in real-time. Even if they could be rendered, it would be a waste of resources. Considering the Call of Duty: Ghosts will be coming to the Xbox One and PS4, maybe the Sub D technology is only for next generation platforms (and PC)?
Of all people, an employee from Rare responded to the post stating that, “You certainly can use sub division surfaces in games. At Rare we implemented an engine using sub div surfaces for Perfect Dark Zero – back when we were creating the Gamecube engine, before it got X-Boxed. The reason we eventually went for View Independent Progressive Meshes, was because sub div surfaces destroyed your vertex cache.
“You had to generate new vertices and it destroyed any kind of coherency on your index cache (VIPMs also suffer the last unless platform supports invalid zero area triangles). I assume modern hardware has come a LONG way in this regard. It certain does NOT increase the number of polygons infinitesimally – instead it allows you to generate as many as you need (and can process) at the required level of detail.”
Infinity Ward has been the subject of controversy concerning the “new” engine for Ghosts, when it was revealed that it was only a heavily modified version of the engine that had been used in other Call of Duty games till date. The whole “realistic fish AI that swims away from you” bit didn’t help matters either. But it’s hard to deny that Sub D isn’t a new technology and thus, not worth such a big deal that Infinity Ward has made it out to be.