Procedural textures, is this the future?

[R_Yell]

Hope not! But some people are promoting them as a reliable way of producing game art:

http://www.bit-tech.net/gaming/2006/11/ ... Gam/1.html

I guess this is could make sense in consoles. Not too much RAM available compared to PC, lots of processing power to generate those textures even on the fly, and all hardware is the same so no compatibility issues.

[Warby]

i played the roboblitz demo yesterday that uses that specific texturing technic and while i must say that robooblitz certainly features them ost hideouse textures i have ever witnessed in a 3d game they claim the game is a total of 50 mb large (in this new article they even claim the textures are only 3 mb ) how comes my steam had to download 300 mb for the dmeo ? which is only a fraction of the entire game

and the ram aspect doesn't really fly either because as far as i understand this they get GENERATED on load up ( there was a really long progress bar starting the game i assume that was it ) and after that will fill the texture memory just like any texture does !

and how they try to suck up to crysis there ... embarrassing

for the lose !

/edit don't get me wrong i love everything auto generated and small but this is just not it !

[e-freak]

the xbox360 port is 50 mb

[zaphod]

I would think the real breakthrough would be when they can render out the procedural textures straight from the video card without having to generate them and store them as bitmaps.

[Sindwiller]

Yeah, such stuff isn't new. There are much programs who utilize such algorithms to combine multiple layers of imagery and renders to one texture, or even create them 100% from scratch. 99& of all demo's are using those algorithms with a very high benefit for executable size. I also know a guy who's developing a game using mostly such textures ("Artists are too slow and too inefficient."). But the big question is: Where the fuck is the artistic part there?! :roll:

[e-freak]

sindwiller if you would have red the article you would have seen that they redeveloped/new developed the algorithms. you can't really compare that to demos.

[Zacker]

The large size of regular bitmap textures is indeed a problem, especially if you want them to be high res. Procedural generation of them is a good idea, but I still believe that it's need to be more mature before we can use it widely.

[Defrag]

Procedural methods don't really offer enough flexibility right now -- you tend to find that most procedural textures are generated using similar methods (perlin noise [see the ubiquitous render->clouds in PS...] etc). There's not enough control or finesse available to make game textures as a replacement for hand-made textures imo. I mean, there's enough there to do some stuff and certain materials can be done pretty well, but how do you account for wear and tear, paint splashes, ripped paper, moss, water damage, scuffs and all that kind of thing? I'd say that procedural textures can complement existing methods, but they're a long way away from replacing them.

[Pericolos0]

The problem with procedural textures is that you can almost always see that they are procedural textures. Anyways megatexturing is the future, as memory ammounts andbandwidths keep increasing.

EDIT:

the article discussion is amazingly stupid btw:

http://forums.bit-tech.net/showthread.php?t=123874

One could only imagine what Gears of War would look like with procedural textures!

This type of texturing is exactly what the industry needs! Not bloat like J Carmaks approach.

And if anyone has seen the textures used in Roboblitz, they look U3 engine quality!

hahaha

The images produced are worthy of Crysis