- This topic has 22 replies, 3 voices, and was last updated 2 years, 3 months ago by
.
-
Posts
-
Calibrated is OK (under profile idealization required by macOS), uncalibrated is not ok.
Calibrated profile stores a 2.2 power law gamma so as explained above when showing an sRGB image in must lift dark greys from RGB number in image to RGB number in display (and colrospace mapping). It works as intened and you see a sRGB image as it should (under profile idealization required by macOS, your display response has a lower gamma value near black… on an accurate/detailed profile it should lift less)
Uncalibrated profile stores the fake TRC that apple sets to all profiles created by EDID detection, close to sRGB, so color managed apps when showing an sRGB image make little to none TRC correction. BUT that profile is FAKE, since report shows that display DOES NOT track default profile TRC. You were seeing sRGB images darker that it should in near black greys. This is NOT becasue factory calibration, but because factory calibration and default profile do not match. If you set all calibration settings in displaycal to native/as measured it will create a custom profile with no calibration = TRC will be accurate (under macOS limitations). It will show images as intened (regarding TRC, not whitepoint) in photoshop and darker greys will show brighter than with default profile…because they should look that way.
Outer non color managed displays are not controlable, by definition, although you can use softproof on a “modeled sample remote display” profile and mark “preserve RGB numbers” to see it non color managed in such display as a way to predict how they will see them. For example non color managed AMOLED android phones. You can create such profiles from synhetic profile editor in displaycal folder or by measuring such phones with remote (web) measurement in DisplayCAL (make sure to turn off autobrightness and keep a finder over the screen if you do that)
I understand the science behind it and thanks for thorough explanation. However, one thing still sticks out and that is that the photo printed on profiled printer looks more like uncalibrated screen than calibrated one. The near black greys are darker and the colors are more saturated.
I meant: paper has very “low contrast”. If you are matching shadows, white paper is much lower “brightness” than your screen. If you match white paper with screen, black ink reflects “a lot” of light. That’s why when you softpoof contrast drops.
And since you have an spectrophotometer… you can measure it, actually it is stored in your custom printer profiles.Hence you were making all your comparisons with printed copy wrong: not using softproof or not matching whites. You can match whites (assuming proper light source for printed copy, good SPD etc) by lower monitor brightness till it matches light spurce or rising the output of light spurce to match monitor. As a hint lux (light spurce for printed copy) = pi * cd/m2 (monitor)
Also due to TRC profile idealization required by macOS (infinite fake contrast TRC, “Blackpoint compensation”: for 0 input is 0 output, instead of actual black color output relative to white), color management engine is lifting a little extra when you softproof. The higher the contrats in display is, the less this matters.
You can see this in “expected TRC/gamma value” vs “measured TRC/gamma value” in DisplayCAL reports.However, one thing still sticks out and that is that the photo printed on profiled printer looks more like uncalibrated screen than calibrated one. The near black greys are darker and the colors are more saturated.
Vincent is generally right, but I add here some words:
- Print profiles made by ColorMunki devices are not well, Munki doesn’t see violet and even part of blues, it has significant color error in dark, it is not well on textured materials due to single direction patch lighting, official profiling tool is weak.
- We actually don’t have reference D50 simulating light, so better to check images near window when weather is sunny. Practically it is better to make synthetic mixed light source or measure this light (.sp file) for Argyll colprof.exe profiler. With i1Pro you may also include OBA influence in the measurement.
- We usually set displays much brighter than prints that are seen under room lights. Prepress works with dark displays. My display has 90 cdm brithness. Our contrast perception depends on ambient brightness. Thus, DisplayCAL has correction for room brightness in lux, but it is actual for high lux in the room, it won’t correct your prints.
- Don’t compare print to screen holding it directly near display, that’s synthetic conditions. But, turn you head and check the print aside from display, our eyes need around a minute to adapt to different light and different image holder. There also can be parasite lighting on print coming from display.
- Check for your print output workflow correctness. More reliable way is to turn off profile correction in printer driver and use “colors corrected by Photoshop”. You have two different color transformation ways for photo printing: relative colorimetric + black point compensation and perceptual. The first one will output all in-gamut colors correctly and clid out-of-gamut colors, the second one deforms all colors to avoid clipping, so it keeps color gradients and avoid clipping. Both methods recalculate all colors to paper white point, when it is seen under print profile light source (D50 or some else, i.e. it is usually not D65 of display, but typical papers have cool tint).
- Yes, papers have much less technical contrast, but our vision filters it well. Gloss papers may have up to 280:1 contrast, while matt ones are limited to ~50:1. But our perception of display screen differs from paper perception. I can’t clearly explain it, but it is a fact. At least display surfaces have specific ambient light reflection and liquid crystals show some different colors and contrast in different directions even on IPS screen. That is why black emulation and paper emulation is Photoshop works with significant error. Of course, you have to compare print with soft proofed image, i.e. not with sRGB ets. original.
- Display white consists of red, green and blue, so when you compare two displays or display with some other image holder, you should equalize brightness levels.
Thanks Vincent and Aleksei for detailed explanations. I understand now everything so much better.
- We usually set displays much brighter than prints that are seen under room lights. Prepress works with dark displays. My display has 90 cdm brithness. Our contrast perception depends on ambient brightness. Thus, DisplayCAL has correction for room brightness in lux, but it is actual for high lux in the room, it won’t correct your prints.
Dark displays are not mandatory. ISO 3664 + ISO 12646 with TCL130 in P2 “appraisal” aims for 500lux D50 to match a monitor D50 160cd/m2 (+- your visual adjustment to match). There was a thread on LuLa
Yuji LED VTC D50 pass babelcolor ISO 3664 test without issue, although with a “C” regarding OBA/UV content. They are very good.
Keep in mind that actual brightness is not set in stone, you just need to get a match, by modifiying monitor output or by modifying light source for printed copies.
A hint is lux = Pi * cd/m2
