I did a print of a ColorChecker from data from BabelColor.
Created the tiff file from the Lab values manually and printed it using
the profile I want to check.
The average delta L* is -1.78, not very I believe.
Here is all the data.
ID Name L* a* b* L* a* b* dE* dL* da*
db* dCh
1 A1 38.2220 14.3334 15.3457 35.9142 13.7526
15.0621 1.99 -1.95 -0.74 -0.28 0.65
7 B1 62.4629 34.8743 57.6940 60.6768 34.5032
56.8047 1.56 -1.54 -0.37 -0.89 0.96
13 C1 29.6514 16.2705 -52.9038 26.7594 16.5299
-51.0011 2.43 -2.19 0.26 1.90 1.92
19 D1 96.1463 -0.6548 3.4579 94.0714 -0.3537 3.6570
1.33 -1.24 0.45 0.20 0.36
2 A2 66.6732 17.0922 17.3976 64.5268 17.5595
17.3022 1.79 -1.75 0.54 -0.10 0.48
8 B2 40.7353 8.8011 -42.7412 38.2207 9.4617
-42.0075 2.34 -2.20 0.66 0.73 0.99
14 C2 56.0738 -39.1555 33.3647 54.0247 -39.8266
32.1141 2.06 -1.94 -0.67 -1.25 1.42
20 D2 81.3125 -0.1517 0.5805 79.3528 0.3442 0.5721
1.54 -1.35 0.74 -0.01 0.50
3 A3 51.1623 -5.4645 -21.9859 48.4936 -4.8027
-22.1608 2.73 -2.67 0.80 -0.17 0.68
9 B3 52.2329 47.2192 15.9864 50.6864 46.4749
16.5942 1.62 -1.54 -0.75 0.61 0.96
15 C3 42.7964 55.9794 27.7372 41.7218 55.5474
28.6662 1.13 -0.98 -0.43 0.93 1.02
21 D3 66.5474 0.0535 0.5786 64.6257 0.4010 0.3188
1.67 -1.57 0.52 -0.26 0.43
4 A4 43.5512 -13.3035 21.2860 40.9930 -13.8798
20.2642 2.46 -2.32 -0.66 -1.02 1.17
10 B4 30.6517 20.5312 -21.2601 28.6027 19.3222
-21.3416 1.75 -1.58 -1.29 -0.08 1.21
16 C4 82.0183 3.5113 79.6733 80.2312 2.7142 79.4778
1.31 -1.22 -0.80 -0.20 0.82
22 D4 51.3854 -0.0921 -0.3136 48.9312 -0.1841
-0.5532 2.47 -2.45 -0.14 -0.24 0.26
5 A5 56.4900 8.6953 -25.0619 54.0171 8.9641
-24.6407 2.38 -2.33 0.30 0.42 0.50
11 B5 72.8860 -22.9759 57.8776 70.5982 -23.8539
56.0386 1.95 -1.73 -0.88 -1.84 2.04
17 C5 51.8542 48.0150 -13.5492 50.1620 46.9324
-13.3578 1.72 -1.69 -1.08 0.19 1.10
23 D5 35.4914 -0.0037 0.4777 32.6954 -0.1854 0.2604
2.30 -2.27 -0.27 -0.22 0.28
6 A6 71.6661 -32.6250 0.5290 69.6704 -32.9291
-0.1201 1.60 -1.53 -0.31 -0.65 0.72
12 B6 72.6079 20.3094 68.0319 70.9563 19.8156
65.3073 1.43 -1.25 -0.49 -2.72 2.77
18 C6 51.2280 -29.5649 -27.5832 49.5486 -28.9127
-26.9776 1.71 -1.68 0.66 0.61 0.89
24 D6 20.4286 0.4309 0.0691 17.9123 1.4260 -1.4175
2.62 -1.73 1.49 -1.49 1.79
On 3/19/2022 10:43 AM, graxx@xxxxxxxxxxxx wrote:
So, Yves, what chain of conversions will you be going through :
- From IT8 “RGB” to printer “RGB”?
First, do make sure the output of whatever Input RGB to Wolf IT8 is “good”?
Otherwise, you’ll be carrying inaccuracies down to the printer profile and blaming errors on the printer profile. Work “modularily”.
/ Roger
*From:*argyllcms-bounce@xxxxxxxxxxxxx <argyllcms-bounce@xxxxxxxxxxxxx> *On Behalf Of *Yves Gauvreau
*Sent:* Friday, March 18, 2022 5:38 PM
*To:* argyllcms@xxxxxxxxxxxxx
*Subject:* [argyllcms] Re: Prints to dark, how to fix this?
Tomorrow I'll do the following test, I'll use the Lab values of the Wolf IT8 target, and print them using the profile I made for this paper that give me problem. From there we'll have data to work with and we'll see if I have imaginary problems or not.
Yves
On 3/18/2022 4:25 PM, graxx@xxxxxxxxxxxx wrote:
Yves,
If you are convinced “it is the profile”, why not print a small
chart with a series of known color patches? That way, you could
compare what the print gives you relative to what you know the
colors to be (Lab Input vs Lab Output).
You don’t say how much light is being reflected to your eyes from
the print? Compared to how your monitor is calibrated, Luminance
wise? If you want to put all the chances on your side, the amount
of light reflected towards your eyes from the print should not be
too far off the amount of light emitted towards your eyes by the
monitor. You could use a simple photographic light meter to
compare the amount of light incident on your face from both the
monitor and the print, being illuminated by your LEDs.
Keep in mind that, having a high CRI is not a guarantee of
anything. Ultimately, your eyes tell the story. I assume the paper
has no optical brighteners.
In the “printing world”, the ISO-3664 standard provides a simple
“viewing conditions” description: if your monitor is calibrated,
say, to 160 cd/m2 and your viewing booth has an illuminance of 500
Lux, at the print level, then any observer would be considered in
a standardized viewing condition and could make judgement of print
vs monitor. Of course, to make this kind of comparison, it is
expected to be carried at a D50 level. If your monitor is
calibrated to D65 and your LEDs don’t provide the same
chromaticities, then you’re one strike down. And even if both give
you the same D50 chromaticities, it’s possible that you may not
still observe a satisfactory match. Colorimetry is a bitch.
How does the match look, a picture of the Munsell ColorChecker
chart viewed on your monitor vs a physical chart viewed under your
LED?
Any other suggestions?
/ Roger
*From:*argyllcms-bounce@xxxxxxxxxxxxx
<argyllcms-bounce@xxxxxxxxxxxxx>
<mailto:argyllcms-bounce@xxxxxxxxxxxxx> *On Behalf Of *Yves Gauvreau
*Sent:* Friday, March 18, 2022 4:09 PM
*To:* argyllcms@xxxxxxxxxxxxx
*Subject:* [argyllcms] Re: Prints to dark, how to fix this?
On 3/18/2022 12:01 PM, Alan Goldhammer (agoldhammer) wrote:
This video from Andrew Rodney is useful:
https://www.youtube.com/watch?v=iS6sjZmxjY4
What happens if you print a standardized image from your printer?
I forgot to try this.
Are the colors and brightness level correct?
I use Photoshop to soft proof the image before printing and I make
a copy without soft proofing active and I try to match them as
best as I can. Whatever I do here, the print is much darker, it's
unacceptable to me when I have to tweak my image before printing,
that's guess work.
What is the light source you are viewing the "too dark images"
under? Perhaps your viewing light source is not bright
enough. Remember, prints are reflective; the monitor is
transmissive. Personally, I think your screen brightness is
very low. Are you using it in a dark room?
I use a high CRI > 95 5000K 800 Lumens Led bulb to view my prints
and about 2 feet above the print. My screen is a BenQ SW270C that
I hardware calibrated with my I1 Pro 2 and on top of that I made
another profile with DisplayCal and I get pretty good results,
nothing above 1.0 DE2000 and I view my screen in about 30 lux of
light measured at 2 feet from my screen.
I'm practically certain it's the profile but I don't know how to
test this hypothesis and if I'm right, I don't know either how to
correct it so I don't have to tweak my image all the time, I'm not
convinced at all that this is the way to go.
Yves
