I already have it done. I just made two tile-maps into half of an isometric grid and put them on each other to make a whole, I removed blend and added erase tiles on contact. I'm using a large invisible object to erase on contact, this prevents visual problems. I was trying to avoid many objects since the area that needed to be a solid color is large. To prevent visual problems I would probably need to use 8 or even 16 objects, at that point I might as well just render large tiled objects for the ground instead of using tiled background and remove the black/white blend layer.
I thought tile-maps didn't take up much CPU/ memory vs objects?
Is it better performance wise to have 8/16 of the same objects, or 2 tile-maps that repeat the same single tile to make the black area outside of where the diamond is supposed to be?
I'm not sure how each is processed, seems to me tile-maps uses less of everything somehow.
Is there a huge difference or should I just tile objects for the ground and forget the black/white mask?