Senor Barborito wrote:

I need clarification wrote:

So, where is the wow factor going to come from in TEH FUTURE OF VIDEO GAMES?

I'll throw in some guesses since I'm bored - they're all original to me but I'm sure twenty other people at least had these ideas earlier (always the case) . . .

Senor Barborito wrote:

Back when I was . . . 19, I think, I made the prediction that eventually all objects in games will be parametric (I know Carmack has had some ideas along these lines at various points but I don't know if it's been anymore than idle fantasy on his part). IOW there will be a standard algorithmic representation of a base (nearly platonic) 'form' for a vase, and one fills in values (or lets the computer fill in randomized values, more likely) for height, width, perturbations in shape (10% chance of there being a handle, etc.) and texture is similarly constructed from such random values.

A nice idea but I'm not so sure it was realistic. It certainly would answer the growing crisis of creating enough detail to meet the capacity of video cards. One good open library of platonic forms and nobody would have to bother making 90% of all art in video games again other than to tweak values. Converts the entire content crunch into a massive distributed effort that is progressively solved with successive video games.

Getting all your values to match a consistent 'theme', however (Ming dynasty), would be a real load of fun.

Senor Barborito wrote:

Second idea is a more recent one (six months back), essentially an urge to go back to very low-detail worlds such as we saw in Quake 1, and actually do them right. Dynamics, fracture values, sheering - you name it. On the rendering side throw in a very rough-realtime radiosity. I once read of an interesting proposal to do this via creation of reflective cubemaps for each visible surface, and then rendering the cubemap to the surface as an additive lightmap. The end result is that whatever was visible from the surface in question would cast a very slight amount of light upon it. An interesting idea but I never saw so much as a good working techdemo of it.

Needless to say all of this idea is way beyond my level of competency - but of these ideas this one especially interests me. A sort of cry to go back and fix all the things we missed the first pass through in physics and in rendering. Call it 'low-fi gaming', maybe make it a retro thing like the Tenebrae project.

Senor Barborito wrote:

Third idea is easily implementable by anybody interested (mech and 3D RTS games only), I thought of it a few weeks back. I'm sure everybody remembers the indie game jam where a bunch of ex-LGS and Game Developer Magazine folks wrote an engine centered around rendering as many individual sprites as possible. Why don't Mech games do something similar for cities? Wouldn't even have to be 80,000 sprites - just use 10,000 to create the illusion of ten thousand of people fleeing in the opposite direction of your mech(s) whenever you storm a city in the game. Throw in some looping/overlayed 'panicked crowd' sounds and voila - a really heightened illusion that you are actually storming a fucking city. If you really want to go all out with the immersion here, set it up with an LOD system so that as objects approach they eventually flesh out to Quake1-character polycounts.

Senor Barborito wrote:

Fourth idea will be out shortly, I'm sure - bump+spec mapping for a space engine. Essentially the same tech (scaling multi-million polygon characters down to a few thousands + bumpmaps) used for Doom III. This would really make a lot of space games look a lot better, and with the general lack of anything in the way of complex geometry to render but ships and space stations, they're perfectly situated for it. When in the Hell Freelancer didn't do this is beyond me, but I suspect it had a lot to do with Microsoft telling them to finish the game sometime this decade.

--SB