Path: utzoo!attcan!uunet!husc6!mailrus!ncar!ames!elroy!gryphon!keithd From: keithd@gryphon.COM (Keith Doyle) Newsgroups: comp.graphics Subject: Re: Palette Optimization Message-ID: <9871@gryphon.COM> Date: 21 Dec 88 17:09:33 GMT References: <12745@cup.portal.com> <571@epicb.UUCP> Reply-To: keithd@gryphon.COM (Keith Doyle) Organization: Trailing Edge Technology, Redondo Beach, CA Lines: 33 In article <571@epicb.UUCP> david@epicb.UUCP (David P. Cook) writes: >To do this, you must go to a more complex method, commonly called >Color Cube Compression (devised by Paul Heckbert). >The >original article for this method may be found in SigGraph proceedings. I'm still chasing a copy of this article, though I have been able to piece together much of it from various postings and other references. The one question I have about the color-cube compression is that it appears the weighting of the color values is not addressed in these or the related algorithms. Example: let's say a given image is using 4096 colors, R 0-15, G 0-15, B 0-15. If 90% of these colors are used by only one pixel, yet another image using the same colors has a similar characteristic but the 10% colors used more often are completely different colors. The basic algorithm would seem to select the exact same palette based on the colors used, without taking which colors were used how much into any consideration. While I can see that when you are tallying the number of data points within a given sub-cube you could treat a color used by 300 pixels as 300 data points instead of 1, this could result in a single cube being selected for subdivision containing the most data points, yet all these points represented by a single color and unable to be subdivided further. I suppose then, the algorithm could mark such a sub-cube as unable to be further subdivided, and then ending the algorithm if all the cubes have been so marked if it occurs before reaching the desired number of cubes by subdivision. Is this a correct approach? The references I've been able to find so far leave out minor details like this. Keith Doyle keithd@gryphon.COM gryphon!keithd gryphon!keithd@elroy.jpl.nasa.gov