3D model and world size technique

[tikvoid]

I am currently using glTF and Blender and I have always had this question regarding not only glTF but FBX, DAE and OBJ as well.
I would like to make good quality 3D models but at the same time lightweight, I remember that at the time of PS2, at the launch of Shadow of the Colossus, it was a big new: huge map, small but clearly visible character, camera that saw it perfectly and giant bosses.

The game was running at about 18 ~ 27 fps (?) But it was still new because it at least ran without visible problems or with a very large and stressful lag.

My final question is: How was this done and how could I do it in Blender, with glTF preferably, how can I adjust the size of the 3D model and the “world” so that it is lightweight and clearly visible? Is this done in Blender or the OpenGL / ES application? Texture streaming? Async loading? Some tips?

[GClements]

Level of detail, billboards, skybox. For Blender, you can use “baking” to calculate the lighting for high-detail models and incorporate it into the texture for low-detail models. For billboards/skybox, you’ll probably need to do this inside the engine using render-to-texture.