home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Light ROM 4
/
Light ROM 4 - Disc 1.iso
/
text
/
maillist
/
1994
/
aug94.doc
/
000179_owner-lightwave-l _Thu Aug 4 17:03:40 1994.msg
< prev
next >
Wrap
Internet Message Format
|
1995-03-23
|
3KB
Return-Path: <owner-lightwave-l>
Received: by netcom7.netcom.com (8.6.8.1/Netcom) id PAA12450; Thu, 4 Aug 1994 15:59:24 -0700
Received: from pixar.com by netcom7.netcom.com (8.6.8.1/Netcom) id PAA12414; Thu, 4 Aug 1994 15:59:07 -0700
Received: from dino.pixar.com by pixar.com with SMTP id AA07021 (5.65c/IDA-1.4.4 for <lightwave-l@netcom.com>); Thu, 4 Aug 1994 15:55:49 -0700
Received: by dino.pixar.com (/\==/\ Smail3.1.25.1 #25.14) id <m0qWBh3-00033uC@dino.pixar.com>; Thu, 4 Aug 94 15:55 PDT
Message-Id: <m0qWBh3-00033uC@dino.pixar.com>
Date: Thu, 4 Aug 94 15:55 PDT
From: bjorke@pixar.com (Kevin Bjorke)
To: lightwave-l@netcom.com
Subject: Re: LW to RenderMan...
Sender: owner-lightwave-l@netcom.com
Precedence: list
Reply-To: lightwave-l@netcom.com
John Foust writes:
> The problem is, Renderman doesn't like zillions of polygons. It
> certainly won't render as fast as it does with simpler
> spline-and-patch based objects.
RMan is actually perfectly happy with polygons, especially in recent
years. However, you will probably benefit from tuning your output RIB
so that they will render efficiently. Many of the default RMan settings
are tuned for patch models, but the simpler case of polygons (guaranteed
planarity and the implicit acceptability of gouraud-shaded results, for
examples) should often be tuned differently.
In addition, many polygonal models (especially ones with polys that aren't
really planar) can often benfit from having quadrilaterals encoded as
bilinear patches and subsequently enmeshed as single entities.
> If you want to access
> Renderman-specific features that aren't in LightWave, you'll resort
> to hand-editing your scenes like every other power-user of
> Renderman. I find it amazing that so many users of Renderman are
> tweaking their scenes and shaders by hand.
All that means is that those users have inferior methods of outputing
well-formed RIB streams. Just making a RIPointsPolygons() call and nesting
it inside AttributeBegin/AttributeEnd is RIB, but not good RIB.
> Still, we plan to add "RIB entity file" export to InterChange, and
> theoretically, we can add import someday, too.
In fact, RIB was never initially meant as a model format -- though Showplace
can read it, as can a few other scattered programs, such as NatPix RMaster
or a modeler I wrote long,long ago called "MMaker" for the SGI (it started out
just as a RIB-file checker, and eventually grew into a modeler). The main
problem is generality -- if you import RIB, you better be able to deal, in
some way, with arbitrary-basis-matrix patches, NURBs, Booleans and so forth.
--------------------------------------------------------------------------
Kevin Bjorke | If you draw a picture, and then draw more and
Animation Scientist | more pictures, and put them on TV, then dino-
Hi Tech Toons | saurs will turn into birds! - Rebecca