tags:

views:

450

answers:

2

I'm finishing up on a 3D planet with ROAM (continuous level of detail).

My goal now is to have good quality render using textures.

I'm trying to find a way I can use a tiling system (small good textures combined), but in a way I can take advantage of my CLOD mesh.

Current algorithms (from what I've found) using this tiling systems produce a huge texture and then direcly apply it. That is not what I want... the planet is very big, and I want more power than simply increasing the texture size.

Is there any known algorithm/opengl feature for this kind of stuff?

I don't know much about shaders, but Is it possible to create one that paints a objects alone... I mean, not giving the texcoords, but putting the right color for every pixel (not vertex) of the mesh?

PS: My world is built using perlin noise... so I can get the height in any world point (height map with infinite resolution)

+1  A: 

You have used 3D Perlin noise for the terrain, why not generate the texture as well? Generally, programs like Terragen, Vistapro and the like use altitude to randomly select a range of color from the palette, modify that color based on slope, and perhaps add detail from smaller textures based on both slope and altitude. In your case, distance could also modify detail. For that matter, 2d perlin noise would work well for detail texture.

Have you modified the heightmap at all? Something like an ocean would be hard to achieve with pure 3d Perlin noise, but flattening everything below a certain altitude and applying a nice algorithmic ocean texture (properly tuned 2d Perlin noise with transparency below a certain level) would look good.

R Ubben
Can you give more info on your second sentence? Any sample or link on the algorithmic ocean using perlin noise? mainly how it would handle lights...
mrlinx
Well, look at this slide. (http://www.noisemachine.com/talk1/9.html) Perlin is demonstrating cloth, but leaving out the operations he introduces to separate the fibers would make nice water, especially if you were careful with the specularity. What he did on the marble vase at the beginning of that presentation would work very well for a more alien ocean. I think slide 22 is the best for water - just a sum of the absoulte value of a series of calls to the noise function.
R Ubben
A: 

This article:
http://johanneskopf.de/publications/blue_noise/
does something similar for 2D images. Check out the video.

shoosh