Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!caen!uwm.edu!psuvax1!hsdndev!dartvax!ascutney!quinn
From: quinn@ascutney (Jerry Quinn)
Newsgroups: comp.graphics
Subject: Re: Ray Tracing Optimization Idea
Keywords: Exploit coherence
Message-ID: <1991Mar28.152931.17790@dartvax.dartmouth.edu>
Date: 28 Mar 91 15:29:31 GMT
References: <2448@umriscc.isc.umr.edu> <1991Mar26.183524.1547@pixar.com> <3754@ssc-bee.ssc-vax.UUCP>
Sender: news@dartvax.dartmouth.edu (The News Manager)
Organization: Dartmouth College, Hanover, NH
Lines: 39

In article <3754@ssc-bee.ssc-vax.UUCP> coy@ssc-vax.UUCP (Stephen B Coy) writes:

>Just a thought.  Since objects that are closer to a light tend to
>cast larger shadows it would seem beneficial to shoot shadow rays
>from the light to the surface rather than the "normal" way.  This
>way the shadow cache should tend to have slightly more coherence.
>The extra cost should be zero but I'm afraid that the benefits might
>also be as slight.  Comments?
>

If there are a lot of small objects between the objects being shadowed
and the one close to the light, you could get some benefit, since the
cache would fall off the small objects often and cause retraces,
whereas the object close to the light is still valid, but I think the
same thing can exist in the other direction.  Simply change the
position of the object to shadow and the light source.  The correct
answer requires more knowledge than we're likely to want to wait for.


                            0 (light)
			  
			  ------  (object blocking light)

                          --
                           --
			       --  (Objects that get cached)
                             --	
			
                         --------- (object to shadow)



>
>Stephen Coy
>coy@ssc-vax.UUCP


Jerry Quinn
quinn@sunapee.dartmouth.edu