Please do correct me if I’m wrong but I’m guessing that my level design might not be suitable to take advantage of precomputed visibility yet.
The level is fine, I think the approach, and even what is currently up on my site is a little misleading. You can use multiple Precomputed Visibility Volumes and should only keep these in playable areas. The visibility cells are only used when the camera is within the cell, so those areas that are player will never get to shouldn’t need any visibility cells.
So if we look at my mock-up of your area. On the right I have a single PCV volume and left is using multiple PCVs. By reducing the number of cells that need to be generated that data is never generated for those cells that we’ll never access and in turn keeps your PCV data much lower so that you don’t have to worry about as much being loaded in.
That’s the general premise there for approaching this. However, with your example and where you show the camera, the level would never use precomputed visibility since the camera is never within a cell. So my suggestion here is an old trick of having it visible in your level when you do the light bake. Disable Cast Shadow and Actor Hidden in Game. This allows you to still precompute the needed cells, keep the camera within the cells, and reduce the overhead of having visibility cells generated for these areas you’ll never need them in.
