So let me preface this with saying I’m still very much learning how to create art. Most the tutorials I’ve seen seem to create meshes that are one single connected mesh by building parts out piece by piece by using edge loops and extruding faces. Is it considered bad practice to have details on a mesh not actually connected to the mesh? For instance I have this building :
I created the windows and doors and accessories individually and kinda pasted them to the building. In the game you can’t go inside and nothing on the building is animated. Being able to separate pieces for reuse allows a faster workflow for me and decreases the chances that I’ll have messed up topology. It gets imported into UE4 as a single static mesh. Is this method ok or are there hidden downsides to doing things this way?
As Maximum Dev said, that’s fine. I wouldn’t rely on it entirely or overdo it, like a very simple mesh made from 15 or so separate elements. I was working on a project where i was UV mapping for other artists, since I was the only one who knew how, and I was given a go-kart / box car style vehicle (stylized and simple) which was made from roughly 50 elements: Each wheel, Wheel segment, pipe segment and so on. That’s what I would call overboard. I modified it to about 3 elements (wheels and connecting parts x2 and the main body)
From that screenshot and your description, it looks perfectly fine.
Thanks for the advice guys! Yeah i’m not trying to have too many separate pieces, probably just doors and windows and thing’s I’d reuse. My next question is whether or not I should remove the hidden faces. On every window and door there is a side that the user will never see that’s just taking up space in the UV’s and increasing my poly count. I read some place that closed meshes are better in the engine for lighting? Does anyone know anything about this? Should I remove those hidden faces or leave them?
You can remove them safely if they’re just never visible. Like the backside of the door where the seam would never be visible to the camera/player. It really depends. I would say the possibility of having an issue with a hard border are less then having issues with shadow bleeding from entirely dark backfaces that were not removed.
As mentioned above remove the invisible polygons. But in your case, you can keep the other side of the windows and texture them as well, because you can then easily rotate it 180 degrees then and make some variation to the texture look.
They don’t need to be welded together–though I will say that due to how the lighting system works it will look better if it was all welded together.
It takes too much time to do that though, and most of the time the lighting will look OK if they aren’t welded but sometimes people do have issues with lighting because the meshes are intersecting.
Best practice is no holes in the facade. If the backface is inside the building not a problem to remove it. If the backface is exposed to world space then the ray trace from the lighting source will not be blocked and shadows will not be created as if the building was not there in the first place.
I couldn’t find a definitive answer for this and YouTube has different opinions on the matter, so I thought to post my question here to not create an already existing post.
First of all, I’m not talking about big meshes like buildings for example, that they need to be seperated in parts for better performance.
But, I’m talking for things like a model I’m currently working on: It’s a sci-fi container that has a few separate geometries, like handles or big screws, etc…
It’s better to merge them all to the main model? or keep them separated as it is now?
Watching at some youtube videos, one sais separate geometries and the other sais unified geometry has better performance result on Unreal Engine.
Looking at ‘quicksell bridge’ assets, all of them are merged geometries.
