Return-Path: <cthulhu!kenb@yang.cpac.washington.edu>
Received: from aerospace.aero.org by antares.aero.org (4.1/AMS-1.0)
	id AA07000 for  /u/strauss/bin/mail_handler.pl strauss; Mon, 28 Dec 92 13:18:13 PST
Received: from yang.cpac.washington.edu by aerospace.aero.org with SMTP (5.65c/6.0.GT)
	id AA16612 for strauss@antares.aero.org; Mon, 28 Dec 1992 13:18:06 -0800
Posted-Date: Mon, 28 Dec 92 13:12:41 PST
Received: from cthulhu.UUCP by yang.cpac.washington.edu
	(4.1/internet->uucp gateway 2.4.2) id AA02231; Mon, 28 Dec 92 13:19:38 PST
Received: by  cthulhu  (NeXT-1.0 (From Sendmail 5.52)/NeXT-2.0)
	id AA01594; Mon, 28 Dec 92 13:12:41 PST
Date: Mon, 28 Dec 92 13:12:41 PST
From: cthulhu!kenb@yang.cpac.washington.edu (Ken Birdwell)
Message-Id: <9212282112.AA01594@ cthulhu >
Received: by NeXT Mailer (1.62)
To: rman@aero.org
Subject: Re: prman and alternatives

> Also, prman does not seem to do ray tracing in 

> the sense of providing reflections and shadows automatically, and 

> reflection maps don't seem to work so well for close-quarters 

> convoluted reflecting.

RenderMan, as it exists on the NeXT and at Pixar, is not a ray tracer  
but just a fancy Z buffer system.  The fact that you can coax prman  
to output photo-realistic images is a testament to its flexibility,  
but its designers never intended mirrored balls to be its forte.   
RenderMan's main strength (IMHO) is its ability to render motion blur  
and depth of field effects with relatively minimal cost, as well as  
its ability to do displacement maps.  With the addition of the  
RenderMan shading language, all of these features become easily  
accessible.

If you're just rendering stills and need multiple reflections (but  
for some reason don't need any radiosity illumination) you should  
think about writing or  adapting an existing ray tracer to the  
RenderMan spec.  I wrote one from scratch in about 3 months, but its  
inability to do displacement maps and it's poor performance on Spline  
surfaces left a lot to be desired.   There are now a few new  
techniques for handling the hundreds of thousands of triangles needed  
to do a good ray traced RenderMan (such as the ZZ-buffer method), as  
well as new techniques for cooked (radiosity) images, so I'm sure  
you'll have better luck than I did.

> I'm  not planning on making any movies!

Too bad, that's what RenderMan does best!


