Stories
Slash Boxes
Comments
NOTE: use Perl; is on undef hiatus. You can read content, but you can't post it. More info will be forthcoming forthcomingly.

All the Perl that's Practical to Extract and Report

use Perl Log In

Log In

[ Create a new account ]

chaoticset (2105)

chaoticset
  (email not shown publicly)
http://chaoticset.perlmonk.org/
AOL IM: chaoticset23 (Add Buddy, Send Message)
Yahoo! ID: illuminatus_foil (Add User, Send Message)

JAPH. (That's right -- I'm not Really Inexperienced any more.)

I'm not just here, I'm here [perlmonks.org], and here [javajunkies.org] too, I ramble randomly in my philosophical blog [blogspot.com] and my other blog [blogspot.com]. Soon I'll come in a convenient six-pack.

Journal of chaoticset (2105)

Thursday June 05, 2003
02:44 PM

Fake Plants

[ #12639 ]
Turns out there's a LOT of work being done in the computing world to produce realistic-looking, not-rendered-in-years fake plants. Me, I'm a lug, I figured this was something that had been worked out by now.

Turns out it isn't, and there's a lot of stuff detailing different approaches, but they seem to be evenly divided into Stuff That Works And Is Slow and Stuff That's Not Great But Is Fast.

I suspect LOD is the way to handle this, but if that's the case then this is going to have to be phenomenally clever. The LOD will have to interleave with whatever geometric solution I'm using, so that each plant doesn't have to be rendered individually.

(Hum...if I had a set of self-modifying formulas, ones where the state could be stored and the rest of the tree restored from them, then the LOD wouldn't matter, it could be generated by dividing (roughly) the info about the tree by the LOD for the current representation. This would result in a finite number of possible trees for each species, would reduce the amount of detail per tree to a (possibly) manageable amount, and still have the kind of effect desired. Hopefully.)

I wonder if some sort of grouping algorithm could be produced, so that after certain LODs are passed a cluster of trees becomes a stand, and a larger group becomes a forest, etc. This would also allow for better imaging on smaller plants, because you could still focus on anything larger than, say, single blades of grass, but from even twenty feet they'd become indeterminate bushy things.

The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More | Login | Reply
Loading... please wait.
  • I wonder if some sort of grouping algorithm could be produced, so that after certain LODs are passed a cluster of trees becomes a stand, and a larger group becomes a forest, etc. This would also allow for better imaging on smaller plants, because you could still focus on anything larger than, say, single blades of grass, but from even twenty feet they'd become indeterminate bushy things.

    The biggest problem with LOD algorithms I've seen is in handling the transitions. If you're a FPS gamer I'm sure you'r