Game Engines and HeightMaps...

Dave Jr

I've just watched this tilable texture video:

https://www.youtube.com/watch?v=VEGAkV_LOfo

in which the user makes a tillable texture within zbrush and captures the texture via its grabber. However he also grabs the heightmap data and uses this in marmoset2 to displace the geometry in addition to using a normal map.

What I wanted to ask was; do engines commonly use heightmap data in this way or would those tilablle meshes be geometry instead?

Cheers

Quack!

Height maps are going to be used much more for this generation. From higher use of parallax displacement to tessellation, they are all driven by proper heightmaps.

With that said, verts are cheaper then expensive shaders in most cases. So if your mesh tile is simple enough, using geometry may save resources over using a shader based affect. It is situational.

Dave Jr

Quack! wrote: »

Height maps are going to be used much more for this generation. From higher use of parallax displacement to tessellation, they are all driven by proper heightmaps.

With that said, verts are cheaper then expensive shaders in most cases. So if your mesh tile is simple enough, using geometry may save resources over using a shader based affect. It is situational.

So in relation this this, geometry such as rubble, rocks, cliffs etc would all be heightmap data or geometry?? I.e use the heightmap in zbrush to generate a low-res base with which youd then use your normal map on, or a flat plane and rely on the heightmap data?

Quack!

Again it is situational and depends on your scene budget. Plenty of games will use a combination of all of these things.

Heightmaps for terrain are often a start, and then the artist will take over from there, morphing the shapes into something they need. In general the heightmap will contain large scale shapes for the terrain, then extra detail will be layered on top, in the form of multi-layer textures or actual geometry.