Hello, im currently playing around with photoscanning objects in 3D (photogrammetry), so far i always extract displacement and normal maps via the hipoly scan and afterwards apply it to my meshes when rendering in my 3d package. The use is architecture rendering etc. Maybe in future i will try to sell some models on some sites (or my site), or use them for real time useage in games.
Im currently workign with low poly retopologized mesh and applying displacement and normal maps, thats all. But i came across Vector displacement maps, which combine displacement and normal map in 1 maps as far as i understand it.
Should i start using Vector displacement maps inestad of combo displacement+normal in my workflow? (using vray)?
Are displacement+normals map a thing of the past with Vector displacement maps now?
Replies
A vector displacement map is a displacement map that can displace in multiple directions. The data is encoded similarly to how per-pixel normal data is stored in a normal map, however that is where the similarities stop. A VD map represents depth in complex multi-channel form while a normal map records normal direction in the same RGB space.
If you need to use a normal map in addition to a displacement map (standard or vector) depends on what exactly your renderer does with the data. If the mesh is displaced, but the normals are not recalculated, you still need a normal map, this is often the case in game engines.
If the mesh is displaced and the normals are recalculated, you may not need a normal map, this is more common with offline renderers. Even then, normal maps are generally better at representing fine detail. To use a displacement map to represent the same detail as a normal map, you need mesh density that is equal to your texture resolution, aka 1:1 texal to mesh density.
In other words, if im able to get the right result with just displacement map, there isnt any reason to do the same thing instead with vector displacement map?
and yes, if you can do a classical displacement,do it. It will only be one channel of information to displace, not 3, so it will have less of a memory footprint.
You can see what I mean in the third video from the top - http://www.fxguide.com/featured/cryteks-cinebox-an-update/