I would like to know if it’s possible to implement some basic features of modeling software in Unreal Engine while taking advantage of its own capabilities. The specific tasks are as follows:
- Adding vertices, constructing edges and faces, and using these dynamically generated meshes in the scene.
- Dynamically modifying the UVs of a mesh.
- Performing skeletal binding and generating skeletal meshes (if point 1 can be achieved, skeletal meshes can also be generated).
- Dynamically generated meshes being used on a large scale in the scene, with the potential for performance optimization such as batching support.
- Attaching game-generated or imported textures and other data to meshes at runtime (this can be achieved by creating a powerful material).
- Supporting Lumen and Nanite (optional).
Are there any possible solutions to achieve these features?
Replenishment：
I have a basic understanding of Unreal Engine’s official PMC (Procedural Mesh Component) and the third-party RMC (Realtime Mesh Component), and I have looked at their APIs. However, both seem to have their limitations:
-
PMC has significant performance issues when there are a large number of meshes in the scene. I have tried using batching and backface culling, but there is still considerable performance overhead.
-
RMC, while having significant performance advantages, lacks sufficient documentation, and I am unsure if it can meet my requirements. Additionally, the latest version doesn’t support Lumen (though I haven’t verified if PMC supports it).
So I am wondering if there is a way to call more low-level interfaces to achieve this, or if anyone familiar with RMC can clarify whether it can meet these needs?