I made some modular assets in Maya and exported them to Unreal.
The assets have issues with what seems to be bleeding and some dark spots, as highlighted in the pictures I will attach.
The light maps were automatically generated inside of Unreal, as I seriously suck at unraveling and making my own in Maya.
The devs seem to think that the auto generation works well, not sure why I am having problems.
I understand that there should be padding between each face, so why didn’t the auto generation create padding?
If someone says I should create my own lightmap, can I ask how you would group the faces? For instance, the seams that you on both the floor and the walls (They form an x) end up being so tiny in width that I cannot even put them on a grid line in the UV editor.
Also, how would I even know which faces to group? And how would I even group them? I have lots of experience in Maya except for when it comes to UVs, unfortunately.
I recommend creating your own UVs. You should set your grid based on the desired resolution of your lightmap. With these being modular assets you should be able to get a uniform texel density pretty easily - this will help keep shadows uniform in resolution if they cast across multiple modules.
Unwrap faces in such a way that minimises seams. The worst of those dark lines appear when the edge of a UV cell intersects a lightmap texel (Which is why it’s important to snap to grid where possible)
Have a read of this thread and the linked reddit thread.: Baking lightmaps Correctly without increasing lightmap resoltuion - Rendering - Unreal Engine Forums
First off, don’t construct simple flat objects that way, it provides next to no benefit and instead makes a performance hit due to increased draw calls. If at all possible try to make surfaces out of a single object, it will have little impact on memory and will look better and perform better. The only downside is having to manage more objects.