Re: metamerism and metameric color - more
Re: metamerism and metameric color - more
- Subject: Re: metamerism and metameric color - more
- From: Peter MacLeod <email@hidden>
- Date: Thu, 21 Jun 2001 11:09:00 -0700
At 10:34 AM +0200 6/21/01, Rudy Vonk wrote:
Peter MacLeod wrote:
> According to Wyszecki and Stiles (1982) that is exactly the definition..
> P. 184:
> "Definition of Metamerism
> Metameric color stimuli are color stimuli with the same tristimulus
> values but different spectral radiant power distributions.
I appreciate the post and realize that I am getting a little pedantic,
but my personal impression is that the definition you quote is
semantically "hollow". In addition, it may even be incorrect because it
excludes the very realistic possibility that the metameric objects,
under the illuminant with which they look the same (or have identical
tristimulus values), actually *do* have identical reflectance power
distributions (i.e. the difference in SPD under another illuminant is
limited to wavelengths absent from the first illuminant).
There are other definitions, and I personally find Fred Bunting's not
only simpler and more practical, but also more accurate:
> A relationship
> between two color samples
> whereby they appear to match in color
> when viewed under certain conditions
> (usually illuminant), but not under others.
--
Well not to beat a dead horse :)...but the fact that metamers under a
given illuminant may not match under different illuminants
(which changes the SPD of both samples, obviously) is an important
one, but that is not metamerism, it's "no longer" metamerism
(which is essentially what Fred's definition says,) i.e. the condition
for metamers:
integral(SPD_1 * cone_l) = integral(SPD_2 * cone_l)
integral(SPD_1 * cone_m) = integral(SPD_2 * cone_m)
integral(SPD_1 * cone_s) = integral(SPD_2 * cone_s)
for a pair of SPDs, for each of the three cone responsivity
curves, might not hold. Fred's definition is a nice explanation
of metamerism, but it doesn't tell one anything about _why_
the samples are metameric or not. With the colorimetric definition
one can do actual computations, and predict whether samples
are metameric, or how far apart they are, under different
illuminants, etc. Abstract explanations are useful pedagogical
tools, but one eventually needs to understand the precise
definition if one is to do any real work, right?
--Peter