[argyllcms] Re: Noticeable differences between calibrated white points
- From: Sam Berry <samkberry@xxxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Wed, 27 Oct 2010 09:37:23 +0100
On 26 October 2010 21:05, János, Tóth F. <janos666@xxxxxxxxxx> wrote:
>
>
> And what do you think about the limiting factor?
> Is it the sampled wavelength range ( the capability of the sensor or the
> physical UV filter in all ColorMunkis and some i1pros - I don't know
> yours...), the sampling resolution, the H3 "agressive coating" layer on the
> panel's surface, the unusual spectrum of the WCG-CCFLs and LEDs (I think you
> have RGB LEDs for a H-IPS display) combined with the limitations of the
> standard observers...? Is it the panel technology itself?
>
There was a study done in 1997(ish) by somebody Shaw who tested the
different observers against the original and newer data. The spread of data
between observers was around 4dE, all as measured by lab spectroradiometers.
So before you get into instrument errors, ambient light, monitor
non-linearities and varying observation diameters, you've already got plenty
noticeable errors. I suspect that the 1931 observer disparities and the CM
errors cover the white point error pretty much entirely.
>
>
> To be honest, I currently think that this Munki was an absolute waste of
> money if it can't really calibrate my display.
> May be the factory calibration drift (after the temperature change, backlit
> luminance setup in the OSD, etc...) was in the same range where I am now.
> And ok, I also calibrated the tonal response but it's not far from my target
> and ColorMunki is not perfect either (mostly if we talk about the shadow
> details...).
> If I need to set up the white balance with my eyes, I could keep with the
> cheap i1LT (which I sold).
> Because I need to question the accuracy of the gamut measurements as well
> (if it can't find the WP, they should be drifted as well, I think...)!
>
I find the i1 pretty good at measuring the gamut. Certainly your CM is a
lot, lot better than the i1LT on the displays you have. Giving it up just
because it can't match a white point exactly (which may not ever be possible
due to failures of observer metamerism ie 2degree vs 10 degree) seems like
throwing the baby out with the bathwater. Remember that even most
professional displays have a non-uniformity of upto 3dE across the screen.
Upto 10dE is not that unusual on lesser screens. Expecting a perfect match
is completely impossible. If you don't like the match, do it by eye. You'll
probably find you can't do it much better than the CM. Matching one thing
throws another out.
>
>
> So, I can't trust in the "affordable but relatively expensive" (- for a
> home user) instrument.
> I can't use my other wide-gamut display as a reference for manual WP setup.
> What should I do?
>
Relax your expectations I think is probably your best option. Unfortunately.
If it helps, I find my i1 matches my S-PVA to my standard gamut screens very
well. I expect your CM will too. Again, the screen will look warm and green
on first glance, but the colours should look right when looking in detail.
>
> - Sell this display and find one with native sRGB gamut (and may be a
> different panel technology?)
> -> Can anyone recommend me a 10+ bit display with 1920x1200 or 2560x1440
> resolution, good responsiveness and relatively "viewing angle free" panel
> which has nearly 100% sRGB gamut coverage without over-saturation?
> I think you can't. That's why I bought my current display. :) They don't
> make any good displays with sRGB gamut nowadays.
> On the other hand, this H-IPS is a real winner in any other ways: It's
> fast, it's contrast ratio is high enough.
>
Unfortunately there is no easy answer. I would suggest you try my trick of
viewing a small monochrome image before you give up. They may well match,
bluey blacks or not. If you don't want to accept the contrast loss that
sorting the grey balance out entails, then you can't expect perfect colour
matching either. Shadow colour casts do strange, unpredictable things to
colours.
Buy a Dell Ultrasharp 2209WA for relatively small amount if you want an
actually useful standard gamut reference screen.
>
>
> - Should I get a small and cheap CRT display to use it as a reference for
> the manual WP setup? (And use the chromacity coordinaltes from the
> manufacturer's ICM file for gamut emulation?)
>
You could, but I wouldn't bother. Soon the world will expect files to look
right on wide-gamut screens, and then people with standard gamut screens
will get wide gamut screens to calibrate too :)
Ignore the factory ICM if you do!
If you have kitted your room out with a good D50 (not grafilite!)
ambient/viewing light, measure that and use it as a reference. If you
haven't kitted out your room with good D50 light, and are using something
else, then being obsessive about monitor WP is even less worthwhile.
>
> - Should I forget about all of these things, set back the factory settings
> on the display, delete any LUTs and profiles (or live them as they are now),
> sell the Munki, and live like the usual "blind" users do? :D
>
If it's an NEC PA241, then that might work! Or just accept what the CM tells
you, and also accept it's never going to look the same as a different panel
technology, ever.
:)
Sam Berry
www.satsumatree.co.uk
Other related posts:
