home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.sgi
- Path: sparky!uunet!usc!sdd.hp.com!saimiri.primate.wisc.edu!ames!eos!prevost
- From: prevost@eos.arc.nasa.gov (Michael Prevost)
- Subject: Re: what happens to big color values?
- Message-ID: <1992Dec30.055731.11018@eos.arc.nasa.gov>
- Organization: NASA Ames Research Center
- References: <C002M8.3D8.1@cs.cmu.edu> <1hqi3lINNi51@fido.asd.sgi.com>
- Distribution: comp
- Date: Wed, 30 Dec 1992 05:57:31 GMT
- Lines: 36
-
- kurt@cashew.asd.sgi.com (Kurt Akeley) writes:
-
- >--
-
- >In article <C002M8.3D8.1@cs.cmu.edu>, gleicher@CS.CMU.EDU (Michael Gleicher) writes:
- >|>
- >|> When lighting calculations are taking place, what happens when color values
- >|> are > 1?
- >|>
- >|> I'm computing the lighting "manually" in a program as well as displaying it on
- >|> the screen. For points whose RGB value is computed > 1, I'm not seeing the
- >|> expected results. For instance, what I compute as an RGB value of
- >|> (1.5,1.4,.9), I'd expect to appear nearly white (since I expected the values >
- >|> 1 to be truncated), but it appears yellow.
- >|> The things which could be happening are:
- >|> 1) truncation (this is what I expected, but doesn't seem to be the
- >|> case)
- >|> 2) normalization (eg: make the vector be unit magnitude)
- >|> 3) divide by the largest element
- >|> 4) something else
- >|>
- >|> If someone could fill me in, I'd be most grateful.
-
- >The color components should be independently clamped (not truncated)
- >to 1.0, resulting in a color near white. What hardware are you running on?
-
- There was also for a short period a bug in the color routine that didn't
- clamp the values correctly. This was fixed on the next release. It was
- IRIX 3.???. It was fixed by 4.0.
-
-
- --
- Michael Prevost
- Sterling Software
- moffett Field Ca.
- prevost@eos.arc.nasa.gov
-