How can I best optimize this volcano level I have modeled out in Blender for Unity & VR?
polycounter lvl 2
Hello there! I'm new to the forum but was hoping I could maybe get some advice / insight on a project I'm currently tackling. 
I'm working on a relatively simplistic stylized volcanic island for a VR game where you fly around with a jetpack (visually and conceptually similar to something you'd find from Nintendo: Mario, Animal Crossing, Breath of the Wild, etc.). I've spent the past couple of months working on and ironing out the design of this volcano based off concept art as well as in-game tests. Currently, it lives as a large sculpted mesh inside Blender since that was the easiest and most fluid way for me to make fast iterations. Currently we've been using the full size mesh for testing.
Now, I'm needing to start breaking it down into parts, polishing, and optimizing things. My two challenges to figure out are general optimization and texturing. My current strategy is breaking the mesh down into visually logical chunks, which I'll show below. I am familiar with topics such as LODs, static batching, occlusion culling, and keeping polycount / memory down. I'm struggling to figure out what the best strategy is for something big like this. I'm used to working on pieces of this scale / magnitude.
Am I on the right track?
The volcano:
Ortho & Isolated
Current Rough Breakdown Into Pieces
Current plan is to polish each part (polycount, appearance, etc) and in the end give them LODs, etc. I'm not sure how I'm going to be texturing this yet, but in the end it will be something relatively simple likely without the need for super detailed textures.
Thanks for reading this far and checking things out! I realize this is awfully general, but I'd love to hear your thoughts and if I'm going about this the right way. ♥
Replies
I am pretty sure 1 million triangles is quite alot for a VR scene, you would be able to reduce it alot by baking the sculpt onto lowpoly meshes. I am not sure if you have planned to do that. :)
Generally I think going for a modular approach is good because like you mentioned the engine can use occlusion culling to reduce the draw calls alot. ( I think you have activate it manually in your camera settings)