here is a first screen of my alternative terrain (texturing) demo (early stage at this moment). The mapping is randomly generated an no heightfield is given.
I'm using tile mapping based on Wang. he implementation is done by a util-class for vitual tilemap generation and a gridmesh for tilemapping. If the code is cleaned, it will posted here later. (this week i hope)
Future work:
The virtualmap generator are enough abstract, to implement different versions of generators. A Random generator is done today and seen in the demo.
Future versions are imagebased or hightmap -generators, to exactly texture the grid.
many work todo ^^..
edit:
The grid is to big today. It needs split of in a quadtree or octree for speedup. For testing only ^^
It will be incompatible to terrain PAGE, since its cannot take up enough texture data. A clear interface for a terrain would be good. (for spezial data methods etc.)
like getHeightAt(), setHeightAt(...)
Great stuff! I'm looking forward to that overhang texturing, how are you going to do it?
Ok, this is a bit tricky.
I need a little bit more grid for the overhang. For the grid i use subdivision on there faces. The data is given by a map for shifting the vertex of the heightfield in heigher regions.
We (Galun und me http://www.world-of-mystery.de/client/) also using shader for terrain texturing. We are using a splatting shader but i'm not very satisfied with the result. It looks not really nice. The texture repetition is to clear.
better textures(not programmer art) + more differences in the splatting(more textures and more complex splat-alpha) really help removing the visual repetition…as does objects like vegetation etc
More different textures means that more texture units have to be used. I don't kown if multiple renderpasses is an option here. My experience is that they lead to a noticeable slowdown.
I really have no clue what you mean with "complex splat-alpha" and in what way it would help.
Is this meant in combination with more textures?
Right objects like vegetation will hide this effect.