- This topic has 2 replies, 3 voices, and was last updated 11 months, 2 weeks ago by
.
-
Posts
-
This is the one setting I’m lost on. The default calibration setting is gamma 2.2. I’ve watched several videos when using displaycal tha sayt sRGB is the best as it allows the full gamut of the monitor. Half the internet says gamma 2.2 and the other half sRGB. Then the replies get so deep into the weeds of technical caveats it’s a rabbit hole and I’m lost. Whole paragraphs on clipping, tone, curve etc, etc. Yes this is all very important, but most summations don’t make it clear which to use. I did search the forums here for something concise, but to no avail.
I have a IPS monitor I just got and plan on editing my photos in Photoshop/Camera Raw with Adobe RGB color space. I would like accurate shadow detail along with (naturally) accurate colors.
I’ve watched several videos when using displaycal tha sayt sRGB is the best as it allows the full gamut of the monitor.
Nonsense.
Half the internet says gamma 2.2 and the other half sRGB. Then the replies get so deep into the weeds of technical caveats it’s a rabbit hole and I’m lost. Whole paragraphs on clipping, tone, curve etc, etc. Yes this is all very important, but most summations don’t make it clear which to use. I did search the forums here for something concise, but to no avail.
I have a IPS monitor I just got and plan on editing my photos in Photoshop/Camera Raw with Adobe RGB color space. I would like accurate shadow detail along with (naturally) accurate colors.
Color managed enviroment : does not matter as long as display icc profile’ TRC stores ACTUAL “gamma” value of disply post calibration.
Non color managed: some video players, games, DOES matter and with a very limited contrast display sRGB TRC is usually best to avoid.So usually aim to 2.2 to avoid extremely lifted greys non color managed on a display trying to simulate sRGB with low contrast window.
Keep in mind that AdobeRGB and srGB colorspaces that assume ideal infinite contrast with RGB 0 => 0nit
They key point is to visualize a 2D graph RGB INPUT vs GAMMA value. 2.2 is constant. sRGB has that shape you can see online.
If your display has limited contrast it won’t be able to track 100% any of these curves. the graph will “bend” to lower gamma values near black.If your IPS 1000:1 display is perfectly calibrated to AdobeRGB 2.2, it won’t be really 2.2 near the blacks. It can’t (1000:1).
If display icc profile stores perfect gamma 2.2 from white to black and you render an sRGB image it will lift very dark greys, it MUST do it because color managed will compensate the differences between sRGB and 2.2.
If display icc profile stores actual gamma of display, including non zero brightness at RGB 0, some “SIMPLE” color managed engines like Apple’s and most browsers may clip blacks because they are out of gamut (Photoshop won’t).So you’ll need to find an equilibrium position depending on what you do when using limited contrast displays.
As a general rule for newbies and if you had a >800:1 display after calibration… try to aim to gamma 2.2 and store “post calibration” monitor behavior in a “single curve matrix profile” with black point compensation.
THis means that after calibration, the resulting ICC profile will store near perfect 2.2 gamam with fake infinite contrast, so sRGB images on Photoshop will have an unwanted but very subtle lift near black RGB 0If if bothers you, do the same with no black point compensation in DisplayCAL. Display ICC profile will store actual gamma, which cannot be 2.2 near black. This way Photoshop will simulate the best it can these “ideal, infinite contrast colorspaces” like sRGB and AdobeRGB by applying its own “black point compnesation”, rendering fake grays to do not clip near black (because your limited contrast display CANNOT SHOW THEM, no mmatetr what you do).
The handicap is that “simplified” color management on Firefox or macOS desktop may crush very dark greys to black… because they are out or gamut. The color engine lacks of “black point compensation” to “fake” dark greys by lifting them so you can percieve certain amount of brightnes separation between them (not crush) at the expense of accuracy.To amplify and add to the points Vincent is making:
sRGB was never intended to be used as a display space, it was intended, like Rec.709, to be an encoding model that would produce effective results when viewed on a standardized CRT. The key features of these spaces are:
HDTV primaries which offer good balance of saturation with minimal coding inefficiency for 256 levels per channel (it’s important to not waste scarce bit depth on rare colors)
D65 white— subjectively appears brighter and whiter on low intensity displays compared D50 without going blue.
Overall tonal response close to L* for best coding efficiency (per primaries above).
Effective contrast range appropriate for CRTs with black level just above veiling glare in dim to dark viewing environments (CR 10,000 to 50,000).
A spliced OETF with a linear slope near black to ensure finite contrast window, avoid arithmetic hazard of infinite slope in camera internal calculations, and subdue dark noise typical of early CCD sensors.
sRGB was for PCs, assuming a dim environment (office)
Rec709 was for HDTV production assuming dark production environment.
The intended display was a CRT (with power law response, gamma 2.3 to 2.5) with a variable peak white and black level set for comfortable viewing in a range of environments, and incurring a subjective contrast enhancement power response delta of up to approx 1.1— in other words, television broadcast mastering targets gamma 2.2 overall response and includes an end-to-end assumption that the display (receiver) alignment will have a slightly higher gamma to improve subjective contrast, where the final alignment is up to the subjective preferences of the display user.
In this light, given a controlled viewing environment and calibrated display using fully ICC color-managed presentation for web and video will lead to slightly less pop because sRGB and 709 content will not receive the subjective contrast expansion with local color mgmt. But other factors can overwhelm this effect.
The video / web end-to-end assumption for sRGB and Rec 709 content do not apply to photography and pre-press in general because typically such content is captured into spaces intended to be handled by an ICC pipeline all the way to presentation, by which the finer points of display calibration are hidden behind the color mgmt module which will take display calibration into account.
Today, sRGB profiles and the range of decode options defined by Rec BT.1886 offer standardized ways to convert web and legacy video assets into more modern content encoding systems which target standards that do not include HDTV’s end-to-end assumption. So for example, a video grader will apply a profile with 1886 reference EOTF to legacy HDTV content when importing into resolve then mix it with other content to target UHD or DCI or whatever.
It is all of these considerations taken in the aggregate that lead to default monitor calibration based on gamma 2.2 to 2.4, which provides excellent results for most applications, and where if you have more rigorous alignment needs you will choose them in context of the specifics of your workflow.
