Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!apple!brutus.cs.uiuc.edu!wuarchive!gem.mps.ohio-state.edu!ginosko!usc!venera.isi.edu!raveling
From: raveling@isi.edu (Paul Raveling)
Newsgroups: comp.graphics
Subject: Re: How to map 24-bit RGB to best EGA/VGA palette?
Keywords: RGB EGA VGA color
Message-ID: <9559@venera.isi.edu>
Date: 8 Sep 89 18:51:04 GMT
References: <593@celit.com> <3129@cbnewsm.ATT.COM> <7743@cbmvax.UUCP> <13319@well.UUCP> <586@celit.com> <4862@eos.UUCP> <13381@well.UUCP> <126@vsserv.scri.fsu.edu>
Sender: news@venera.isi.edu
Reply-To: raveling@isi.edu (Paul Raveling)
Distribution: na
Organization: Information Sciences Institute, Univ. of So. California
Lines: 41


> It's not clear that this is HSV but it was in response to someone who
> asked about it (I don't have the paper.  Can you elaborate Paul?).

	It wasn't HSV, but I don't recall details any more, don't
	have the code, and can't find the paper.  (ISI's library has
	shrunk a bit while it's occupying temporary quarters -- it's
	permanent home floor is being remodeled.)

> This makes sense because most of the quantizations that you'll ever do
> involve colors which are going to be close in any color system.  Colors
> which are close in RGB are also going to be close in HSV, HLS, YIQ, CMY,
> CIEXYZ, whatever.  The relative distances will only change slightly and
> relatively rarely will they cross over.

	In this hack for visualizing images I've added a switch
	to show them in either RGB or HSV.  It appears that the
	patterns in RGB space would be much easier for quantization
	to work on than those in HSV.

	Colors tend to form relatively well-defined shapes in
	RGB space, somewhat reminiscent of clouds.  In in HSV
	they spread out into relatively more planar forms that
	look more like curtains without clearly defined "top"
	boundaries.  HSV seems to be better for showing all chroma
	info that's present, but in most cases there's a lack of
	clearly defined boundaries -- colors are more diffuse,
	grading gradually from high-density areas to low density
	without well-defined edges or surfaces.

	I'll probably try some other color spaces as time goes on,
	but RGB still looks most promising for quantization algorithms.
	It makes me think mother nature (aka "the laws of physics")
	operates naturally in either RGB or something with 1-1 mapping
	to it.  Trying to account for human perception is where it
	gets more confusing.


----------------
Paul Raveling
Raveling@isi.edu