I want to make a FPS game with destructible buildings. The first thing that comes to mind is to presegment the buildings into multiple gameobjects. So if something hits that segment, it will disappear and instantiate some debris. However, I’ve been hearing that the more gameObjects in the scene the more lag. If I build a entire city with about 50-100 peices per building, that will probably equate to a ton of drawcalls. I could batch render objects with shared materials but I need specular and normal maps on my buildings with batch rendering does not support.
From what I read, minecraft clones do not have each cube be one gameObject. Instead they have them be part of one enormous mesh. So I’m wondering if I should do something similar with my buildings. Have each building just be one gameObject and then write some sort of script that modifies the mesh to create holes when it gets hit in certain areas.Would this be better performance wise than having the building be broken into dozens of gameObjects?
thanks
Roger0