home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.amiga.graphics
- Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!rpi!masscomp!mark
- From: mark@westford.ccur.com (Mark Thompson)
- Subject: Re: arenderman update!
- Message-ID: <1992Aug26.221226.21108@westford.ccur.com>
- Sender: usenet@westford.ccur.com (UNIX news)
- Organization: Radiant Image Productions, Merrimack NH.
- References: <Bt0nyu.6En@chinet.chi.il.us> <1992Aug17.204956.10418@westford.ccur.com> <Bt82vA.8q0@chinet.chi.il.us>
- Date: Wed, 26 Aug 1992 22:12:26 GMT
- Lines: 52
-
- It took a while to reply, but I've been super busy....
-
- In article <Bt82vA.8q0@chinet.chi.il.us> katefans@chinet.chi.il.us (Chris'n'Vickie of Chicago) writes:
- >>speed is definitely not one of RenderMan's strengths.
- > Not true. Digital Arts 3.8 is the most fully RenderMan complient renderer
- >yet, and frankly it runs rings around everything I have found on the Amiga.
-
- I did briefly take a look at Digital Arts at Siggraph but was unaware that
- they were now fully RenderMan compliant. It looked quite good.
-
- >>Doing things like breaking all primitives up into stochastically
- >>super-sampled sub-pixel micro polygons is extremely compute intensive
- > Not really. In real terms the slowest part of the renderer is still
- >determining visibility.
-
- While the RenderMan specification does not define any specific method for
- determining visibility, the REYES render typically uses a Zbuffer which
- is computationally trivial. However......
-
- >Breaking the "primitives" into micro-polys is
- >actually quite quick, although it _is_ hard-disk intensive.
-
- While dicing may be quick, those micro-polys must be stochatically super
- sampled with a Z interpolation for each sample point. So while the method
- of visibility determination is simple, the shear number of calculations
- forced by super sampling makes quite a burden. Ofcourse this means you
- get nice clean surface intersections :-)
-
- > From what I've seen, the REYES algorthym is faster than any raytracer.
-
- True, but typically much slower than the more standard scanline/Zbuffer
- rendering methods.
-
- > On the system mentioned above, I get broadcast quality animation at a very
- >reasonable 5 minutes a frame.
-
- Not bad, not bad at all. One can only hope that an Amiga implementation
- fairs as well. I would really like to see that happen.
-
- >Of course this presupposes that the
- >operator is not afraid to get his fingers dirty in the RenderMan .RIB
- >file. This leaves our Mac friends out in the cold.
-
- :-) :-) :-) :-) :-)
-
- %~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~%
- % ` ' Mark Thompson CONCURRENT COMPUTER %
- % --==* RADIANT *==-- mark@westford.ccur.com Principal Graphics %
- % ' Image ` ...!uunet!masscomp!mark Hardware Architect %
- % Productions (508)392-2480 (603)424-1829 & General Nuisance %
- % %
- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-