pyle@lll-lcc.UUCP (03/08/87)
I am looking for any and all people who have implemented a ray tracing program. Sharing experiences and ideas and asking questions has been very helpful in increasing my understanding. If anyone has implemented a ray tracer with distributed ray tracing, I would especially appreciate hearing from you. Response by E-Mail would be fine. Thanks, Ernie Pyle
shankar@srcsip.UUCP (Subash Shankar) (09/04/88)
For those of us who are ignorant, can somebody give a brief overview of what ray tracers are, and how tracing algorithms generally work. Thanks.
kyriazis@rpics (George Kyriazis) (09/05/88)
In article <8085@srcsip.UUCP> shankar@ely.UUCP (Subash Shankar) writes: >For those of us who are ignorant, can somebody give a brief overview of >what ray tracers are, and how tracing algorithms generally work. >Thanks. For a introduction to ray-tracers a very good text is David Rogers "Procedureal elements to Computer Graphics". Nevertheless, I am making a small summary here: Ray tracers are programs that generate very high quality pictures, but (usually) take a fair amount of time. The principle under which they work is very close to the way we see. When we are looking, light rays leave light sources, hit several objects, and finally they reach our eye. That is very inefficient to simulate, since most of the rays leaving the light source(s) do not reach us. The technique that is followed is the following: Instead of following rays from the light sources, we follow the rays these rays in the reverse direction, ie. beginning from our eye. Assuming that we have a screen infront of us with N pixels, we only have to follow N rays now. Of course these rays bounce off objects, and eventually reach a light source or get colored from the background. When a ray finally terminates, it is evaluated and it's color is returned to the parent ray (when a reflection is taking place the reflected and maybe the refracted ray are called children rays and the incident one is called parent ray). Finally the tree of rays going off from every pixel is evaluated and the pixel is colored with that color. Usually a ray tracer spends most of its time doing ray-object intersections, therefore a fast intersection routine will make a fast ray tracer. Things are not so simple though, and people are continuously trying to find ways to make them faster and faster. Hope I answered the question for those people. George Kyriazis kyriazis@turing.cs.rpi.edu ------------------------------
chu2_ltd@uhura.cc.rochester.edu (Jim Huang) (04/17/89)
I've noticed many requests for ray tracers recently. Many of these, including the mtv and qrt flavors, can be anonymously ftp'd from drizzle.cs.oregon.edu [128.223.4.1]. +-------------------------------------------------------------------------+\ | Jim Huang University of Rochester | \ | P.O. Box 27918 | | | Rochester, New York 14627 | | | (716) 274-3706 ARPA: chu2_ltd@128.151.224.17 | | +-------------------------------------------------------------------------\ | \__________________________________________________________________________\|