No announcement yet.

Modelling for VR

  • Filter
  • Time
  • Show
Clear All
new posts

    Modelling for VR


    Just a simple question:

    When modelling, let's say a house, is it best to model using one mesh as far as possible? It's completely doable and I have a feeling it could improve performance a lot for VR. (Instead of having let's say a mesh for walls and mesh for ceilings etc.)

    Thanks anyone willing to chip in.

    if you are concerned with the resolution for lightmaps I wouldn't do it that way. Also if you want to have different materials on ceiling and wall surfaces you would end up having several draw calls on that single mesh anyhow.


      Well for most things I normally just do textures in the modelling program (Maya). As I've realized, or have yet to figure out, how to apply material to just one face of an object in Unreal Engine. I think it's possible using UV's, but that's just more work for doing something simple like adding one texture to a face (of a wall), and another to a ceiling face. I'm just trying to figure out what the best work flows are from A-Z for VR. I've been modelling and animating for years, but obviously then for rendering, where performance was never an issue in the final product.

      But thanks for your reply! Not sure how having a single mesh or having different meshes doing the same thing would impact lightmaps? Whether the ceiling and wall are connected or separated it should not make a difference? Or am I missing something.


        I'm coming from the 3Ds max workflow but I'm sure it works similarly in Maya. To get a single mesh to have multiple materials you set a specific material ID to a group of polygons and assign a material with multiple IDs that that mesh in max. Then when you import into UE4 it should come with multiple material channels.

        Having one big mesh mesh would make it a more complicated lightmap. So you would have to turn up the lightmap resolution, resulting in longer light builds and needing more graphics memory comparing to splitting up in several smaller parts.


          Hello guys, i was just passing by and would like to share my experience, i had a model which is contain more than 1500 mesh. I crafted 2 uv, first one for texture and the second one for light map in 3ds max. Whenever import to this giant structure (without compined mesh !) lightmap build perfectly working, but whenever combined this meshes during the import lightmap is not working ! what worked for me: i bring this structure without combining box checked and building the lightmap then merging meshes from developer tools>merge actors.. But i needed this structure as one piece regarding to interaction. And also considering performance with LOD ! in that point after generating LOD, light map completely screwed up ! Now i gave up with light map, now my giant once piece mesh is movable and performance is perfect with LOD.. For your info guys


            Hi Guys

            We do mainly VR dev for Archviz. Our preferred platform for bringing in structures (and you can laugh if you want) is Sketchup. We can bring in a full house with all walls, windows, doors etc as one Mesh. For quick VR walkthroughs it works great. We use one mesh and one movable directional light, with a post process volume. The results are pretty amazing. With baked light, Sketchup is a winner for structures of a building. For decor, you can forget Sketchup of course. We do use a specific modelling method in SketchUp to ensure the UE4 generated lightmaps do not create leaks or inconsistencies. We only use generated lightmaps. But this is of course an Archviz pipe.

            Just some insights.