[argyllcms] Lost in the display calibration jungle
- From: Anders Torger <torger@xxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Tue, 12 Oct 2010 20:59:32 +0200
I'm a hobbyist photograph and a newbie concerning CMS. I've played
around with Argyll and i1 Display 2 and read alot on the subject. I'm
mainly interested in display calibration for now.
The first thing I had to realize is to loooower the expectations.
Instruments are not accurate and they drift, screens drift too and has
poor adjustments controls, and what the correction profiles can do is
rather limited and will introduce some artifacts (banding etc). When
perfect is impossible, one needs to know which type of errors that are
acceptable and which ones are a problem. It's there I'm lost.
My application is to postprocess photos. When I adjust the colors to be
just right, I want to feel confident that what I see is close enough to
what is seen on a "perfect" system. I spend quite some time to get the
right colors, and if my screen is too much off target, that time is
wasted, simple as that.
I have a wide gamut display (Dell 2408WFP), and I like the concept
although it introduces some CMS problems, so my next display will
probably be wide gamut too. I publish most things on the web, and
browser icc profile support is coming strong and wide gamut displays
become more widespread so I see that it is likely that I will use
AdobeRGB on many/most published web photos quite soon into the future.
I have looked at various colorimeters, but to a newbie like me it seems
like
- They are not very accurate, some (perhaps all) are so inaccurate they
cannot really be trusted out of the box, but need calibration.
- There may be even larger accuracy problems for wide gamut
calibration.
Then there are spectrometers. Price goes up significantly, and it hurts
shelling out the money for a hobbyist like me. But I need something that
works and I can trust, I rather spend 1000 dollars on something that
fulfills my expectations than 100 dollars on something that really does
not do the task I need done.
But then some say that colorimeters are really better at calibrating
displays because they are faster and they measure dark colors more
accurately. Assuming that the colorimeter has been calibrated with a
spectrometer first. So then you need a spectrometer *and* a colorimeter.
However, perhaps there's some overkill here to what I want to achieve?
I'm thinking that even if spectrometers can't measure the darkest colors
very good, it probably won't matter much for display calibration at
least, since when I adjust a photo I don't put another tint on dark
colors than on midtones and light, I look at the photo as a whole. It
seems to me that it is much more important to have correct midtones and
bright colors than dark. But then again I don't know how dark "dark" is,
perhaps spectrometers get problems already in the lower midtones? That
it takes longer time to measure is not a big problem. Measurement takes
time even with the fast colorimeters. If it takes 5 or 40 minutes does
not really matter to me, as long as the job is properly done.
(I'm thinking that the "dark tone" problem could be handled like
avoiding too much correction down there, since correction may be quite
off and just introduce artifacts. Getting the right balance between
correction artifacts and accuracy is also an interest, but that's the
programmer/engineer part of me talking, first-hand I want to satisfy the
photograph in me.)
What I hope to get from a spectrometer is
- Accuracy out of the box
- I do accept some deltaE, but it should be reasonably right.
- What kind of accuracy can be expected out of the box?
- No need to recalibrate the instrument regurarly with a reference
calibrator (that kind of service is not really available around here)
- The need to recalibrate the instrument with another instrument
regularly like seems to be required for colorimeters I see as a
real pain. I really would like to avoid that. If required I would
probably need to buy both instruments myself.
- 'Good enough' performance for my need, I can feel confident that I
will not need to redo all my color adjustment work if I would see my
photos on a "perfect" system later on.
For obvious reasons (=price) Colormunki Photo/Design seems attractive.
I've read that it has problems with drifting (probably heat related),
but I'm not sure if it is so much that it will be a real problem or if
it is just theoretical.
What do you more experienced CMS users think? Will a consumer
spectrometer like Colormunki Photo be ok, or is disappointment likely?
Pro grade equipment the only things that work? Or is a calibrated
colorimeter so much better than a (entry level) spectrometer that it is
unwise to spend the money on a spectrometer (the calibration of
colorimeter problems is there though)?
/Anders
Other related posts:
