On ProPhotoRGB
On ProPhotoRGB
- Subject: On ProPhotoRGB
- From: Marco Ugolini <email@hidden>
- Date: Sat, 25 Feb 2006 20:10:43 -0800
Hi to all.
While viewing ProPhotoRGB with the ColorThink grapher, I noticed something
that has not yet been discussed in this forum, to my knowledge.
Please go to my web page to view a few images first, before I go on:
<http://home.mindspring.com/~marcoug>
(Warning: I will leave the images up for viewing until the end of this
coming week. I will take them down after that.)
Now, this is what I have observed:
1. First image: If you plot ProPhotoRGB with ColorThink 2.2 using a 2D view
in Yxy, you can see that a considerable portion of the space of that profile
falls outside the spectrum locus (the horseshoe-shaped container that
represents humanly-visible colors according to the experimental CIE data on
the standard human observer).
2. Second image: Here I show the result of plotting ProPhotoRGB in 2D using
Luv (a more uniform environment, in which color distances correspond more
closely to the way they are perceived by human observers): the portion of
the color space outside the spectrum locus appears even more strikingly
large. I am not sure what to make of this, but it would not seem to me a
desirable feature that the ProPhotoRGB working space plots colors that fall
outside the spectrum locus, i.e., outside the confines of human perception.
3. The third image shows how ColorThink 2.2 plots another large-gamut space,
DonRGB, in a 2D view in Yxy.
4. The fourth image shows how DonRGB, plots in 2D in Luv. In both cases,
DonRGB appears to behave well, straying out of the locus only by a tiny area
in the lower left, nothing that is even closely as dramatic as was the case
with ProPhotoRGB.
5. The fifth image is a 2D view of ProPhotoRGB, this time using XRite's
ColorShop. This shows a "cobweb view" with deviations. What strikes me here
is to see that there are at least 3 areas exhibiting what appear to be
obvious problems. I circled and numbered these three areas. Area 1 shows a
strangely and very obviously "pinched" end in one of the six hue lines. Area
2 shows a *crossover* in the cobweb lines that unite similar points in two
adjacent hue lines: can't be good, can it? Area 3 shows an *overlap* in
which things look like they are "falling all over themselves" (pardon my
unscientific description): I would love to be told that this is something of
no consequence, though it certainly does not seem that way.
6. The sixth image shows a "cobweb view" too, this time of DonRGB. No
pinched points, no crossovers, and no overlaps here. This looks like a nice,
well-behaved profile, so far as this view of it goes.
Could anyone please explain:
a) What are the meaning and consequences of ProPhotoRGB straying into areas
outside the spectrum locus? Dan Margulis has said many times on his forum
that ProPhotoRGB deals in "imaginary colors", as a way of dismissing its
usefulness. Whether or not he is trying to score points in his own favor and
actually knows what he is saying, does he nonetheless have a point, possibly
in spite of himself?
b) What do the uneven, pinched and overlapping steps in ProPhotoRGB
indicate? Nothing to worry about? And why not?
Thank you to those who will reply.
--------------
Marco Ugolini
Mill Valley, CA
_______________________________________________
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