Best method for complex collisions?

Hello fellow devs, quick question on collisions…

I have a bunch of cave interior static mesh’. they are around 1200 polys each.

If I create custom UCX objects the combined UCX mesh’ come to around 850 polys but of course the collisions are a lot less accurate.
If however I optimise the mesh I also get it down to around 800 polys.

So… given the poly count of the static mesh objects and the fact that there are only ever 3 or 4 shown onscreen at one time, which would be the best option in terms of performance vs results?

  1. Use the UCX files despite being less accurate?
  2. Use the optimised mesh with display set to none? (this is more accurate but is a whole other mesh added to the blueprint)?
  3. just use complex collisions as simple on the original mesh (400 faces really isn’t that much and is less of an overhead than using a second hidden mesh)?

I don’t need to simulate physics on these static mesh objects but they are sat inside a blueprint.

Thanks for any help you can offer.


Just use complex collisions as simple. If you have super detailed mesh then you can always just use lower lod level for collisions. You can have hundreds of thousands vertices for static collisions even on mobile.

Indeed using lower LOD levels is the best way to achieve higher precision collision.

Hi guys, this is indeed the way i ended up going. it turned out the 4 meshes with around 1200 faces each was really nothing as far as performance goes.

Thanks for the replies though!