Hi everyone,

I've been working on a small personal project for a long time now and a while back I stumbled across Geometry Clipmaps for terrain rendering. I've (very slowly) been getting my head around how it all works on the rendering side of things. However I'm having some issues deciding how to produce the data from which to render my terrains. I've had a lot of different opinions on the most suitable approach and so I figured I'd come to a good source of facts and opinions!

I've mostly been focused on a large set of "chunked" files, all of these stored at the different resolutions that I might need. So if there are 9 levels display, then I'd have the entire terrain stored in 9 different sets of chunks, each in its own resolution.

Another idea I had was to just access one large file, but I'm not sure how suitable this would be considering the scale of the game world. For each of the higher levels (lower resolutions) I'd need to do checks at intervals, rather than every point. Is the searching time required by this going to be a big overhead that makes it impossible?

Also someone suggested just using a custom file format system, each file containing either a specific areas data for all levels or maybe all the data for one specific level?


As is probably quite obvious, I'm getting myself a bit confused a muddled with all the ideas. So please, any ideas (especially any that you guys have got working yourselves) are eagerly awaited!

Thanks!

The first solution you mention there is roughly what I was originally considering.. I guess I really am going to need to learn a lot more about loading on the fly and the torodial array stuff that was mentioned.