How does nanite work in ue5 ???

I’m sure the nitty gritty is a secret. But i’m still curious about the “basics” or what else you guys can talk about?
From the descriptions out there, it basically LOD’s so that each polygon is at least a pixel small. And keeps it that way regardless of distance from cam.
Makes a ton of sense. And I can totally see 3d being able to do that.

What I don’t understand is how the “lod’ing” of the triangles doesn’t have its own massive performance hit?
For example, if you run an optimize function in 3d software (like 3dsmax) it takes a while to calculate. Even for realtively lowres meshes.

How does the ue5 engine do this in realtime without smashing the framerate?

Just curious thanks.

Check Digital Foundry channel in YouTube and answers are there analyzed by experts.