RE: Colorimeters and third-party developer support
RE: Colorimeters and third-party developer support
- Subject: RE: Colorimeters and third-party developer support
- From: Roger Breton <email@hidden>
- Date: Tue, 23 Aug 2011 15:47:34 -0400
Bob,
Graeme Gill wrote :
> I think my summary hit the mark somehwat better than the "we see 20
> degrees", while not delving into all the minute detail of cone and rod
> distribution.
To which Bob replied :
> What I find amazing is that we are still using the 1931 2-degree standard
observer that was based on research done in the 1920s
> apparently with just a dozen men all living near London! One can't blame
the original researchers for their choice of subjects, but for
> the CIE to use this data (after all sorts of jiggery-pokery to make it
more usable for their aims) to represent all human color vision
> seems incomprehensible to me - a biologist. Has anyone repeated these
tests with a decent-sized sample of people - men and
> women of different ages from different countries?
The color science literature abounds with studies that have revisited the
original 1931 2 degree Observer, in the explicit view of proving it wrong or
improving on it. Yet, despite it's known shortcomings, like the y curve,
who's modeled after the 1924 V-lambda photopic curve, who is a few
nanometers off in the shorter end of the spectrum, and a host of other
aspects I'm not aware of, is still preferred because, on "average", they
yield good predictions. I specifically once asked Mark Fairchild the
question and his reply was "yes, it works, on average". So, who am I to
argue with all the research he's conducted over the years, with the CIE and
at his lab, in Rochester?
Keep in mind, too, that the apparatus to recreate the original color
matching experiment is not for the faint of heart. Today, it might be
relatively to set a bipartite field with tuneable lasers ($$$$) but,
remember that, at the time of Wright and Guild, they had to create their
primaries from very sharp physics. Which made their conclusions all the more
worthwhile and withstood the test of time.
> Reading all this stuff, it doesn't surprise me in the least that Roger
finds that, since nothing ever matches, color accuracy is not the
> name of the game. Sorry Roger if that is not quite what you said. ;
No, you are quite right. I think the pursuit of absolute accuracy is not the
panacea because of the limitations of current instruments or surround
factors Graeme alluded to or other unknown processes that are not taken into
account in the overall evaluation of color matching (it's an endless field
of scientific study -- "fascinating", as Mr Spock would put it). Which makes
me grateful that colorimetry gets me close to my goal, still. When I have
all the factors of adaptation I know of right, the relative size of the
stimulus I'm comparing right, a dimly lit room at 64 to 32 Lux, I know I am
not far from where I should be. I mean, I maybe a perfectionist, though,
and, admittedly, many people I work with would just be happy with what they
get out of off-the-shelf solutions but not me, I look for that extra,
because I can see the difference. And it bugs me. That's why I keep making
the case for effective ICC profile editing tools -- hope X-Rite NEVER
retires ProfileEditor (don't you dare??!). This is what I tell my clients,
there is only so much the "best" combination of
tools/hardware/media/lighting will do today. The rest is up to a human.
Color is a judgement after all?
But I'm surprised by how much "garbage" is routinely judged acceptable in
this industry. No one wants to take the extra time to fundamentally
understand what's going on with color? Very few actually care. I'm glad many
of these people actually hang out on this List ;-)
Thank's for your kind patience and bearing with me.
Best / Roger
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden