Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!world!katefans
From: katefans@world.std.com (Chris'n'Vickie of Chicago)
Newsgroups: comp.graphics
Subject: Re: renderman
Message-ID: <1991Jan23.021800.7203@world.std.com>
Date: 23 Jan 91 02:18:00 GMT
Sender: katefans@world.std.com (Chris'n'Vickie of Chicago)
Organization: The World @ Software Tool & Die
Lines: 84

Chris here,


 Minghsun Liu asked:

> I read in the Jan. issue of Byte that several software developers are
> working to incorporate the ability to produce RIB file into their
> modeler products.  Is there a modeler that is capable of produce RIB
> file on the market today?

   Well the ones that I know right off the top of my head are:

VIDI for the Mac
either "Super" or "Swivel" 3-D for the Mac

VGA-Topas for the PC
            This is a new 8-bit version of AT&T's Topas that could serve
        as a very nice front end for PC RenderMan. VGA-Topas can render
        to 8 bits for test rendering, and then output a .RIP file for final
        output.

        Digital Arts is close to having a fully RenderMan compliant modeler
and renderer. Currently the animation part of the program outputs a
"pseudo-renderman" file called a ".rip" file.

        PIXAR states that hierarchy and complex transformations are the
domain of modeling programs, but these motions have to be reduced to simple
"translations" and "rotations" for the RenderMan file. So, how do we pass
these complex motions to the renderer for motion blur calculations?

---------------------------------------------------------------------------

  Eric Haines asked:

> By the way, has anyone other than Pixar developed any non-trivial
> implementation of RenderMan (i.e. trivial == scan-line polygons only, no
> shadows, with all the other features unimplemented)?  Just curious.

    Digital Arts (see above)

  As I understand it, the problem is that the renderer can safely ignore
any rendering specs that it can't handle (ray-tracing, displacement, etc.),
but it _must_ properly render every type of geometry in the specification,
patches, quadrics, NURBS, and all. I guess this would include trimmed NURBS
up to 20th order. Sheesh.

--------------------------------------------------------------------------

  Paul Crowley writes:

>> Has anyone done any work on making things look dirty and worn? 
>> Everything in the computer-graphic world is shiny and new.

and Randy Brown replied:

> There is a bowling alley shot that I believe Tom Porter did for Pixar for
> a poster advertising RenderMan.  The bowling pins looked pretty marked up.
> It's true, though, that even computer-generated imperfections usually look
> too perfect.

    The marks were put on with a texture map of random black marks used
two different ways; first as an ordinary texture (image) map and also as
a displacement map. 
    It might seem kind of dumb, but if you want dirty, beat-up things
simply _scan_ a dirty beat-up thing. Try prying up a well-trod floor tile.
Use that as a texture or bump map, or as a displacement map if you're lucky
enough to have it. Displacement mapping is _wonderful_! I've made an ice-
cream cone that I'm quite proud of....

--------------------------------------------------------------------------

    About the Pixar patents thing - motion blur is one of the greatest
competitive advantages Pixar has over every one else in the commercial
animation world. I don't begrudge Pixar holding back this one thing when
I think how many advances they have introduced to my field. To my knowledge
Pixar has had only two products, the (now defunct) rendering engine (and
software for it), and RenderMan. I don't know how they've managed to remain
in business. You can't run a company like a university. Give'em a break.


                                Chris Williams
                                    katefans@chinet.chi.net

Digital Arts *** TDI Explore *** TOPAS *** Video Toaster *** RenderMan (soon)