I’ve got architectural models which I’ve made a lot of changes to in Houdini which is why I don’t use datasmith. I’m targeting windows 10 VR so I’m using static lighting in ue5. The problem I have is that I’ve got lots of non connected geometry in a single static mesh from the architecture file and this means I need to use huge light maps, for them like 2k is still showing blue and giving low res shadows. An example of this would be a group of columns and beams that are all over the building, in an ideal workflow you would just have one mesh for the column and instance it around but with the architecture files all the columns would be one static mesh and all have a single light map which would need to be huge. Replacing all the columns manually would take forever, I’ve also thought it’s possible to use Houdini engine to instance the meshes but both of these would be extremely exhausting. Is there a better workflow around this or do you just have to deal with huge light maps and long build times and are there any long term consequences of this?
Thanks.
I think making the building modular would work better. Definitely worth testing.
It’s better for static lighting and memory usage.
With instancing, though, I think each instance still needs its own light map. I don’t understand how it can work otherwise.
The problem is it can be modular. Basically you’ve got the architecture modelling software like Revit that you output a model from and maybe do some topology and uv cleanup but the whole model was made in an architectural software and is a completely custom design.
Maybe I’m misunderstanding but basically with the column example I’ve got a 50+ columns in a single static mesh so they’re all in a single light map uv. I need to set the light map resolution to something like 4k for that static mesh for it to look decent. Is it not more efficient to have a single static mesh of one column with a light map size of like 512 instances around?
That’s exactly what I’m saying.
But: much smaller lightmap, yes, but I -think- each column needs its own lightmap, because that’s the layout of the shadows is stored. For sure, though, UE will work better with modular assets.
I see thankyou. Do you think there is a better way to go about this or do I need to just isolate a single column and copy and place it manually. Like a way to split a mesh in individual meshes by connectivity? There is also the problem that some assets are bigger than others but I wouldn’t want the uv size to change so maybe I would need to have some sort of world based material? I think if I have to choose between manual placement or longer build times I will just choose longer builds for now.
Yes, it’s all pros and cons. In the end, all you can do is try different methods, and see what happens.
I think that with modular pieces, you’ll get better looking lighting, but will not save much on memory.
As for placing the columns etc, you can do it with a blueprint.