Well I'm glad I'm not the only person that actually sees what the next step in gaming needs to be about. Unfortunately, you'll go into the same type of depression I did, when you realize this isn't doable in any serious way on next-gen console hardware if you still plan on having that hardware run a video game.
Personally I feel in order to make things such as soft body physics and accurate material creation, the first step is really going back to square 1 and getting rid of rasterization just so we can get the materials right and do away with so much hackery and texturing. Put the focus back on geometry and high polygon counts. We need to run all of this in a ray-traced engine, which itself requires more power than we can realistically currently afford yet, then they have to handle the physics behind an increased polygon count. A much, much larger polygon count than what we are currently used to, because when you start deforming geometry, it needs to be made of lots and lots and lots of bendable and breakable geometry subdivisions. If you want to bend that car door, well it needs to be made out of lots of subdivided triangles. You can't do this with a detailed 1080p brick texture plastered on a flat plane to create a brick wall and some bump mapping or tesselation, which is currently what we're accustomed to. If you want to blow up that wall, you need to actually have it made of 3D bricks, which crack in tiny little pieces themselves. Same goes for soft body cloth dynamics. You want to bend a table cloth? The polygon count needs to go from 1-20 polygons to thousands, 10s of thousands or more. Or ....the holy grail.....fluid dynamics. We need both ray-tracing with IOR capability and an insane polygon count as well as a capable enough box of doing serious computation to run a complex fluid sim.
Nvidia's PhysX demo got me excited...
Still waiting for a game that actually uses PhysX to manipulate water like this and still calls itself a video game and not a demo.
Log in to comment