EDIT - I’ve worked around this - code posted below
I’ve just started with the occlusion culling, and after reading existing threads I still have a couple of questions.
First - Is there anywhere that explains the OC logic? - if so, I can just figure this out myself.
I have a basic canyon like scene. I want optimal performance, obviously.
All I want to do is specify what I assume are view areas - A space where once inside, all objects in that space are visible. Anything completely outside is not. I think this is far simpler than the OC algorithm is suited for. But would also be much more useful in this case. I could write a script to do this in Start, but I’d rather avoid it.
I’m having two issues at the moment.
First, I’m getting a ‘pop’ at certain locations, where objects become ‘viewable’ even though they should be viewable much sooner. Cell size, I imagine.
Second, things outside view areas aren’t reliably being culled, whereas I would prefer they would always be. I gather this again has to do with the each cell determining what is viewable.
So my questions
- Should I just do this with triggers?
- What is the advantage of specifying view areas versus one large view area in an outdoor scene. It doesn’t appear to matter, excepting making things more complex.
- What is the in game overhead (aside from the massive render time) with having lots of view-cells? What about the same number of view cells in multiple view areas?
- What happens if some of the static objects are mesh combined (like trees) at runtime? Nothing good, I figure. I guess I’ll try to merge the mesh combine and export as OBJ scripts.[/b]