|
|
"Norbert Kern" <nor### [at] t-onlinede> wrote:
> =?ISO-8859-1?Q?J=F6rg_=27Yadgar=27_Bleimann?= <yaz### [at] gmxde> wrote:
> > High!
> >
> > *Sigh* ...would still take several new computer generations before I can
> > render a flight over the blooming springtime steppes of Bactria in
> > bearable time... er, in realtime of course! I just hope my grand-nephews
> > or -nieces would be inclined to raytracing then...
>
>
> I can imagine of several tricks to simulate a good vegetation in realtime.
> By example you can store millions of predefined coordinates and calculate
> only the visible ones...
>
> Norbert Kern
Another way would be, as Christopher Hormann mentioned, to create an
acceleration structure. In simple terms, instead of checking all plants
every time you need to check a given plant's neighbors, you check a subset,
ie the nearest plants within a given radius.
A simple techinque you may be able to try is to divide up your plants into
cubic regions, and only check the plants that reside inside the cube
containing the original plant, plus a certain number of surrounding cubes.
Beutiful work, by the way! I can't wait until you release it :)
Regards,
George
Post a reply to this message
|
|