Re: CIECAM97 & the idiosyncracies of perception
Re: CIECAM97 & the idiosyncracies of perception
- Subject: Re: CIECAM97 & the idiosyncracies of perception
- From: Nathan Moroney <email@hidden>
- Date: Fri, 28 Sep 2001 17:45:04 -0700
- Organization: HP Labs, PTD, CCR
Quick follow-up to some of the CIECAM97s posts:
1. Full credit for CIECAM97s must go to CIE TC1-34 (including
Drs. Hunt, Fairchild, Luo, Nayatani and others). I am grateful
to be working with them on TC8-01. I am the chair of this TC
which is attempting to revise CIECAM97s for color management
applications. For the gory details go to:
http://www.colour.org/tc8-01/
2. One of the most immediate and significant improvements that
CIECAM97s provides relative to CIELAB is improved blue
constancy. CIELAB and to a lesser extent CIELUV tend to shift
to purple for a constant blue gradient. CIECAM97s is a
considerable improvement. For a visual example, see:
http://web.hpl.hp.com/personal/Nathan_Moroney/
3. Other advantages include a better chromatic adaptation
transform, additional perceptual attributes (as previously
mentioned) and the possibility for image-specific appearance
based adjustments (something I showed at Seybold SF).
4. Yes, CIECAM97s is complicated but one of the objectives of
TC8-01 is to simplify, if possible, without sacrificing
performance. There will be a trade-off between complexity
and performance. The s at the end of CIECAM97s is supposed
to be for 'simple'.
5. Yes, someday CIECAM97s (or CIECAM2x) may make its way into
the larger color management world. It will take time because
of the trade-off mentioned in item 5 but I believe CIELAB and
CIELUV are starting to show their age (see items 3 & 4).
CIECAM97s is still very much a work in progress and I would
not expect users and developers to gain much by wading
through the equations at this time.
6. CIECAM97s is not a magic bullet. There are still gamut mapping,
rendering intents, user preferences, device characterization,
color variability and other issues to be addressed. Having
said that, even the current state of CIECAM97s is useful
enough to be used right now in some HP printers. It is new
enough that there is limited documentation or technical
marketing material but it sneaks into some product reviews.
For example the 2250TN:
http://www.zdnet.co.uk/reviews/rstories/0,3040,e7106659,00.html
7. Im going to pass on the D65 vs D50 for D50 light booths. I'll
weasel out by saying there is ongoing research on the topic of
adapted white points and light booth technology/variability.
Thanks for the comments & questions,
Nathan Moroney