NetNews Usenet Archive 1992 #18/NN_1992_18.iso/spool/comp/graphics/8756

Internet Message Format | 1992-08-12 | 1.9 KB

Path: sparky!uunet!utcsri!torn!news.ccs.queensu.ca!qucdn!polowinj
Organization: Queen's University at Kingston
Date: Wednesday, 12 Aug 1992 17:19:06 EDT
From: Joel Polowin <[e-mail]>
Message-ID: <[e-mail]>
Newsgroups: comp.graphics
Subject: Re: Is there a mathematical relationship RGB -> wavelength?
References:  <[e-mail]>
Lines: 27

 > For curiosity's sake, I was wondering if it is possible to take a set
 > of RGB color fractions and come up with an approximate wavelength of
 > the observed mixture.  I suppose this could be done by just
 > interpolating between some known values in a table, but I would like
 > to know if there is a way to actually calculate the wavelength.  Of
 > course, I'd also be interested in a formula to compute the RGB
 > fractions for a given wavelength.

If you're interested, I've written a (relatively) short function in C
which takes arguments of RGB values, computes a "brightness" value and
the intensities of the two spectral colours (ROYGBV) present in the
"colour" being described, and returns this information along with a
description in words.  For example, given RGB of 230, 208, 179, it
determines that this has a "brightness" of 80%, a grey level of 70%,
orange 15%, yellow 5%: "lt, sl yelsh org".  (The routine needs some
tweaking yet; on my monitor, this colour looks more like a yellowish
orangey grey.  But the spectral-colour separation isn't bad.)  If
you have wavelength info for some kind of arbitrary standard Orange
and Yellow, you could probably then come up with a reasonable
calculation of a wavelength for something which is 15% orange and 5%
yellow.  What exactly you'd do to get the wavelengths of colours which
combine red and violet I don't know off-hand.  The whole scheme is kind
of arbitrary, but it seems to work fairly well.

Joel
 [e-mail], [e-mail],
   [e-mail]