Quote: "Also note that checking if the object is hidden takes longer to perform than actually hiding it, same with checking if it is in screen."
I don't know about the checking if it's hidden versus hiding it. I never did any benchmarks.
If the objects aren't in screen and are showing, you're just wasting processing time displaying objects that are out of view. The trade off is the check if they're in screen, versus always being displayed if they are not in view. For a plain or two you can leave them visible even behind the camera and it won't matter much. A tree, the polygons will probably take more processing power than the in screen check. Run some tests. Find out what works best.
If you are displaying things like a question mark or an indicator of an action, you could use a 2d image and paste it at the item's screen location when the camera is within range. For this, you'd most likely want an in screen check or your 2d image would show up on screen even if the object was behind the camera.
The news letter also mentions stacking objects. Kevin Picone had an excellent example from some years back of storing all of the usable objects in arrays and managing their display with that. You could really manage what you are trying to accomplish and squeeze out some good performance by implementing some of these techniques.
Enjoy your day.
