- From: Hugh Sasse <hgs@dmu.ac.uk>
- Date: Thu, 22 Apr 1999 18:23:44 +0100 (BST)
- To: Robert Praetorius <RPraetorius@AspenRes.Com>
- cc: www-amaya@w3.org
On Thu, 22 Apr 1999, Robert Praetorius wrote: > > Something to consider, which is non-trivial, and obviously > > not based on code I have not seen :-) -- > > > > [detailed descriptions of various black-image-or-text-on- > > black-background and white-on-white problems deleted] > > Amen. I have wished for some kind of browser intelligence in > this area since I first used Mosaic n years ago. I'd hate to hafta be > the one to figure out a simple, reliable way to do it, though:-) Thank you. Well, I was thinking along the lines of: HSV space is based on a circle, with hues (H) around the outside and greys in the middle, and it extends out of the plane to give the intensity (V) from black to white. The Saturation (S) is the distance from the middle out, so (h,s,v) gives the co-ordinates of any colour (like (theta, r, z)). So my idea was that the points in the space, which represent an individual colour would repel eachother like charged particles, and so those close together would move furthest in a time step. If the "charge" is proportional to the number of pixels then that may help... This is all very well, but it is computationally expensive, colours which fly past others will make images look odd, pixels that are exactly the same colour when they should not be, still will retain that problem. A linear transformation to fill out the space as best as possible (like a monochrome contrast stretch) would help with some, but not all of these problems, but as I say, it needs someone to he hit by a bolt of inspiration, really. > > -------p--a--s--s--i--o--n-----n--e--e--d--s-----a-----f--a--c--e------- > "oncology recapitulates philately" --Mark Maxson Robert M. Praetorius > "balance, not symmetry" --Mark Stanley home: rmp@MA.UltraNet.Com > (attribution by Stigler) work: RPraetorius@AspenRes.Com > Hugh hgs@dmu.ac.uk
Received on Thursday, 22 April 1999 13:23:54 UTC