I was thinking about to create a large model of a terrain in a software like blender and then divide it into pieces and import them into Unreal. My landscape will be a sort of puzzle of many static meshes joined with each other.
My question is if this approach is safe (in terms of permormance for example). My doubt was because my idea is to create a world with many things like plateaus, dungeons etc… and maybe I’m wrong but it seems that a static mesh is more manageable in these terms than the common landscape.
I have another little question, in a mesh it is possible to lock some vertices/edges in the LODs? I mean if there is a way to prevent them from being destroied by the LODs
Landscapes do not support complex shapes, like caves. It should be a heightmap. You can make holes in it tho.
Landscapes have such benefits as dynamic lods and grass, Plus water/rivers and roads sculpting tools. VT can reproject a lot of details on them. (bake)
Yes indeed my doubt was because a landscape doubtless works perfect with foliage, caves made with the “hole” function etc… But at the same time if I want a complex scene like an area with many plateaus, with terrain also on top of them or a pathway between two plateaus to reach the top of them or even a very rocky terrain it seems to work weird also with the hole function. I mean in a scene like this (from Elden Ring):
There are a lot of big “rocks” with terrain on top of them. They are various separated terrains? Or a single terrain cut in many ways with the hole function? Or a puzzle of static mesh planes used as pieces of landscape? Sorry for the messed up question, I tried to explain better possible my doubt but I don’t know how to explain it well
Being a guy doing the same thing (ARPG ala FromSoft), you can certainly use different landscape sections and stick/blend meshes in it to make ‘areas’ just like in Elden Ring.
Landscapes are made of tiles, and you can delete, edit, etc each individually. You can import multiple landscapes, delete tiles and jig-saw them together if you need to keep them distinct for materials, etc.
Making something like Elden Ring w/Unreal landcapes is viable, IMHO.
In the age of Nanite, it’s going to be possible to completely eliminate landscapes if you go mesh-heavy (check valley of ancients demo). It’s not like the future is married to landscapes but otherwise, there really isn’t a way to generate large, expansive environments procedurally the same way with Nanite (yet). As well, Nanite IS coming to landscapes in 5.1(?) so it seems Epic at least sees a future for them.
Yes indeed I saw that in the DEMO they don’t use landscapes so think that I will learn more also about nanite meshes but I think that I will use a mix of nanite and landscapes depending on the specific area.
Anyhow thank you so much for this clarification, you resolved a doubt that I had for a long time
Nanite, as a technology, essentially revolves around clustering triangles and precomputing some type of visibility-matrix so vs regular-culling, you can look-up what you need and render just-about exactly what is needed vs any overlapping, overdraw, etc that can happen with regular meshes.
Nanite as a tech will apply to heightmesh/landscapes at some point (5.1? is my understanding), so at least as far as culling/performance, you will be able to get best of both worlds there. The heightmesh already provides a solid performance improvement over the older landscape tech so I can only see it getting better.
Since the heightmesh is driven by virtual textures, you can use that same texture-cache to seamlessly blend in other meshes.
So, again, I’d be very confident you can do a lot with this stuff.
