I am unfamiliar with what Unity’s limits might be and I’m curious to know what might be the best way to attack a racing-type game where the track is long an winding and in several places the player can see a long way into the distance so serving up road chunks as required would not seem to be a viable option.
Is it normal to bring in the full track as one huge model or are there other (better) ways to keep performance high?
Thanks
Depends. You can use a large model. You’re mainly concerned with what is in view of your camera. So having a massively large model, in theory, could perform the same as a small model, given you are rendering the same amount in view at one time. The main issue would be if you have enough memory to load up this massive model. If you’re trying to create an infinite driving experience, I’d probably programatically stitch together multiple repeatable sections.
If you look at any driving/racing game, they don’t draw infinitely long distances. You ALWAYS have to kink the road with a turn, a billboard, a hill or some such thing to cull the view. Wipeout, Midnight Club, Need for Speed, Forza, Gran Turisimo, GTA all have turns.
As for “one big mesh”, I wouldn’t suggest that unless you have a low amount of detail. It would be best to break up the mesh into chunks based upon visibility/distance.
Thanks ppl.
Is there any way to know numerically how far the actual view distance is at any given point? Does Unity have a setting or can report what the view distance is atm or is that
Well you could raycast to get the distance, but the camera has a max render distance set anyway