koobas.hobune.stream
Is it more performant to have multiple objects in a parent tree or to use armatures on a mesh?
So I have a very simple model (kind of like a minecraft human model) where it is made of cubes and rectangular prisms. Each part of the model is separated and is its own object, but there will only be rotation animations for each part, i.e. I do not need any mesh deformation. So, one option is to have each part inside a tree of parent/child objects that I can rotate around their respective pivot points. (For example, the root will be the body object, and the neck, wings and feet will be a child of that object.) The other option is to join all the objects into one mesh, then use an armature and assign the vertex groups for each bone. Which one is more peformant? Since I am making a mobile game, I want it to run as smoothly as possible.
The reason I am unsure is because the armature and vertex groups seem like they would take extra computing, but I have also heard that one mesh is better than a few meshes, and one object better than a few objects.
I recommend doing what is easiest for you to do. Because the performance gain will be very small, even for mobile devices.
