Hello.
I’ve one simple question: What is the best practice (for nanite mesh creation) for this situation where there are small details with different topology?
Weld the small detail to the base mesh? (continuous geometry, and ignore the topology)
or
Have it as separate geo from the base mesh, but attached to the base mesh (1 single FBX and not 2 seperate FBX/imports)?
NOTE: I’ve Manny’s foot in there just for scale reference.
Thank you very much.
Combining geometry into a single watertight/manifold mesh is generally better with Nanite since it will cluster better that way. Only real downside of doing that is disk space from the extra triangles needed to merge the different topo and the time it takes to stitch it together in blender or whatever.
Either way, manifold or not, would also be better to bake out the textures into a single material for the main box part and the filigree if you haven’t already.
The argument we have is that there are benefits that outweigh the weld geo/manifold approach by having this small part as seperate (but attached to the base mesh, 1 FBX), even if there is a very small nanite overdraw of this part.
* And this is the reason for this post, hoping for some sort of an official (an Epic’s staff answer) assistance and/or guidance for handling situations like this one, but I doubt any of the staff will even bother with things like this:
a- More effective render culling of this small corner part since is its own nanite cluster.
b- In 5.8 this small part can be voxelized if it remain seperate.
c- Base mesh topology remain untouched and retains it clustering effectiveness.
Also, what do you mean exactly by?:
“would also be better to bake out the textures into a single material”
Thank you for your reply.
