Re: Particle systems

[Jan Ekholm <chakie@infa.abo.fi>]

On Tue, 7 Dec 1999, Keith Lucas wrote:

>I'm just concerned about the amount of effort in culling a huge list of
>particles... even if they can be done by regarding them pretty much as points.
>Also I was thinking about cheating and using some of the particle emissions
>several times: instead of 16 fires all throwing off particles, you could have 4
>drawn 4 times, which quarters the amount of the physics to do, even if not the
>drawing... or am I being completely paranoid about this?

Maybe hierarchical bounding boxes for the particles would be nice in this
case? You encapsluate the bounding boxes of all particales for the entire
explosion, smoke, whatever into one big bounding box and then use that for
high-level culling. If the box is culled you can of course ignore all
particles inside it, but if it is visible you have to go deeper into the
hierarchy of bounding boxes, in your case the individual particles.

But then, I'm still struggling with implementing something similar into my
scene-graph, so I might be far off here... :-)

I think you need som kind of scene-graph that manages your data on a
higher level than the raw display-lists, vertex arrays etc.

>What would you do about particles that can themselves emit particles? I'm
>thinking of things like sparks that could break up when they land into smaller
>sparks or maybe the debris from an explosion which trails smoke particles behind
>it.

Create new bounding boxes for the 'subparticles'.

---------------------+------------------------------------------------------
 Jan 'Chakie' Ekholm |    CS at Åbo Akademi University, Turku, Finland
    Linux Inside     | I'm the blue screen of death, no-one hears you scream