Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!cs.utexas.edu!uunet!mcsun!ukc!edcastle!cs.ed.ac.uk!jsp
From: jsp@cs.ed.ac.uk (John Spackman)
Newsgroups: comp.graphics
Subject: Re: Ray Tracing Optimization Idea
Keywords: Exploit coherence
Message-ID: <8173@skye.cs.ed.ac.uk>
Date: 25 Mar 91 10:47:36 GMT
References: <1991Mar21.183340.19319@dartvax.dartmouth.edu>
Sender: nnews@cs.ed.ac.uk
Reply-To: jsp@lfcs.ed.ac.uk (John Spackman)
Organization: Laboratory for the Foundations of Computer Science, Edinburgh U
Lines: 24

npw@eleazar.dartmouth.edu (Nicholas Wilt) writes:
>I had a simple, interesting idea to make tracing primary rays more
>efficient.  Cache the last object intersected; intersect with it
>first and if hit, determine whether it is still the nearest object as
>follows:
>	1. Fire a ray from the intersection back at the eye
>	2. If no object is intersected, the cached object is still nearest
>	3. If any objects are intersected, find the farthest intersection
>	   that's not beyond the eye.  This is the new nearest intersection.
This scheme is very similiar to `shadowing caching' whereby each light
source is assigned a record of that object (if any) found to cast a full
shadow for the previous shadow ray traced to that light.   This object
is queried first for the next ray to expolit coherence.

Your method of exploiting view ray coherency could trace rays FROM the
eye rather than TO the eye.  If the cached object were hit once more,
the distance to this intersection allows many objects to be rejected even if
their bounding volume is struck but at a GREATER distance, since this
is a lower bound on the distance to any intersection with the contents.

|: JANET: jsp@uk.ac.ed.lfcs :: ARPA: jsp%lfcs.ed.ac.uk@nsfnet-relay.ac.uk  :|
|: John Spackman, Computer Science, Edinburgh University, Room 2417 JCMB,  :|
|: The King's Buildings, Mayfield Road, Edinburgh EH9 3JZ. Tel 031 650 5125:|