RE: [Algorithms] Ray picking in editors - best done with collision volumes, or visual volumes?

[Rowan W. <ro...@ir...>]

my 2 cents:

- picking in an editor must be per poly. if you just use BV's only, =
things get to be a pain when you are dealing with densely packed objects =
and/or manipulating the scene from a distant viewpoint

- as you say, not all objects have collision data. so its best to just =
use the triangle render data. you can quickly whip up some simple =
raycast routines specifically for the editor (eg linear iteration =
through all polys in an object)... they dont need to be uber fast coz =
you arent stressing them in realtime use

rowan

> -----Original Message-----
> From: Megan Fox [mailto:sha...@gm...]
> Sent: Friday, 26 November 2004 2:29 PM
> To: gda...@li...
> Subject: [Algorithms] Ray picking in editors - best done with=20
> collision
> volumes, or visual volumes?
>=20
>=20
> It seems like either could work, but both appear to have (fairly
> major) drawbacks, so I'm curious what others have done here.
>=20
> On the one hand, I could use my visual bounding volumes for the
> picking.  This has the advantage that anything visible can be selected
> and worked with, but my concern is that my visual bounding volumes are
> generally not that accruate - they're just a best-fit AABB per mesh,
> that's it.  It seems like I'll have tons of cases where it becomes
> difficult to select one object due to another piece of intersecting
> cruft (like a particle fountain or light object or whatever).
>=20
> On the other hand, I could go with ray picking via my collision
> routines.  This has the advantage of fairly accurate bounds for all
> the objects, but carries the significant disadvantage that just not
> everything is going to BE collidable.  Now I suppose I could get in
> the habit of always defining a collision volume for an entity "just to
> be thorough", even if said volume was just a trigger-type sphere, but
> this strikes me as fairly wasteful.  I could also get in the habit of
> keeping collision data around for these visual-only entities but only
> using it when editing, except then I'm maybe just wasting effort when
> I should just go with the pre-existing visual AABBs.
>=20
>=20
> What have people used in their engines, and which ends up being "good
> enough"?  This doesn't have to be pixel-perfect, my artist is just
> getting tired of typing entity names into the console manually every
> time he wants to toy with their position or a value ;)
>=20
>=20
> Thanks,
> -Megan Fox
>=20
>=20
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from=20
> real users.
> Discover which products truly live up to the hype. Start reading now.=20
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> GDAlgorithms-list mailing list
> GDA...@li...
> https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list
> Archives:
> http://sourceforge.net/mailarchive/forum.php?forum_id=3D6188
>=20