I need to make a large building 30x50 meters.
What would be the best way to make a fairly big building with this modular mesh, in terms of performance? Is it better to just make the building as one big mesh?
Hey @Longrid!
So the answer to this question is a bit hard to get with this much information. Is the building going to have insides? Or is it one large outer piece?
TYPICALLY though, I would say the overall answer is going to be yes. If the insides will be loaded separately, say when the front door is opened, I’d have one for the outside, and one for the innards. Even then, you can use level streaming to make each floor a separate Static Mesh to lower draw.
All that aside, you can use smaller meshes to craft the larger one, then highlight them all (select the centermost LAST because that will be the anchor point) and in the details window use “Convert to static mesh”. 
That will make all the smaller pieces merge into one larger piece.
1 Like
Yes, this is a large street building measuring 35/50 meters.
Did I understand correctly that I can make it entirely from modular meshes and combine it? What will happen on the optimization side?
When I have built the entire level, will the uv map merged mesh work correctly? Light map, lods?
Yes!
It should require fewer draw calls. Be careful not to merge everything in the project, though, as if any part is on-screen it will render the entire thing.
Do not merge the entire level. It will be very bad for optimization. Think of it as a cookie! Merging chocolate chips, flour, egg. Then baking it you have a cookie.
Don’t merge the entire bake sale. Merge the level blockout- the floors, the ceilings, the walls- but don’t merge the table and chairs inside with it. You want that stuff to be de-rendered when unseen.
1 Like
It turns out that the entire object is drawn on the screen. Objects like houses or something like that are worth connecting. Because there will be a lot of draw calls from separate parts. And there will be fewer draw calls from one mesh.
Large objects should be made with one mesh only when it will not appear on the screen often.
For example, a long street wall. This should be done in modular parts. Because by combining it into one mesh, it will be fully rendered when it hits the screen. And that’s wrong. The player has no way of seeing one end if he is on the other end.
Did I understand you correctly?
Many thanks for your answers
So the confusion I think you’re hitting is file size vs object dimensions. A big, huge building that just has a few windows tends to take less resources to draw than a detailed coffee cup. It all depends on the Level of Detail (LoD). So “How many polygons is this thing” is more important than “How big is this thing.” So things inside the building should only load in when necessary, but typically the outside walls can all be one piece. So with that in mind- the more complicated (size) something is, the longer it takes and more resource intensive it is. But the dimensions (measurable in length/width/height) have very little to do with it.
1 Like