- This topic has 5 replies, 2 voices, and was last updated 5 years, 2 months ago by
.
-
Posts
-
I ran DisplayCAL using an older GretagMacbeth i1 Pro (rev A) spectrophotometer, with default settings except for turning on Black Level Drift Compensation, as advised. (Results attached.) I installed the profile and am using the D.Cal Profile Loader, not Windows color management.
When I open a gray-scale step tablet image in GIMP (2.10.8, with Color Management activated), the low tones are brighter than I would expect/like to see. The RGB 10 10 10 step is displayed on screen as 21 21 20, the 20 step as 29 30 29, etc. (Screen values read using Just Color Picker.) I do astrophotography with lots of dark sky in pictures and detail rising up only slightly from the dark. When an image pixel value of 10 is displayed as 20, I darken it, because it’s brighter than I want…and then when the resulting image is displayed on an unmanaged monitor, I’ve got, like, an image value of 5, which is way too dark.
Is this likely a consequence of the i1 Pro’s limited sensitivity at the dark end? If I calibrated/profiled using an i1 Display Pro would the 10 step map closer to 10 10 10 in display space? Do you have any sample calibrations for a Dell P2715Q monitor done with an i1 Display Pro? (Or for a Dell U2415?)
Thanks for looking into this–and for all the time you’ve spent on the software & support!
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Hi,
in a color managed environment, how bight any given set of (source) RGB values appear on screen is a function of the source profile, not the display profile. Assuming your display is calibrated to gamma 2.2 (the default), and your source profile is sRGB, this means the RGB values sent to the display (after conversion from sRGB to the display profile) are likely around the range you measured with Just Color Picker.
So, in short: You cannot rely on the RGB values (alone) to assess luminance, as it’s the profile(s) that give them meaning.
By “brighter” I mean in comparison to 0 0 0 black. Color managed (as it stands), the 10 step is a distinct shade of gray, whereas I am used to 10 being almost indistinguishable from black. (And yes, display gamma is 2.2, source profile is sRGB, and the monitor covers about 98% sRGB space–not wide gamut.)
Does the lower black accuracy of i1 Pro affect display profiles on the low end?
Color managed (as it stands), the 10 step is a distinct shade of gray
This what it should be. R=G=B 10 in sRGB has a relative lightness (L*) of around 3.
Does the lower black accuracy of i1 Pro affect display profiles on the low end?
Not necessarily. The i1 Pro is rated to around 0.1 cd/m2, which should be fine for (e.g.) typical 1000:1 contrast, IPS panel, computer displays.
I redid the calibration using an sRGB tone curve and am now a happy camper.
Most of the photo processing software I use is not fully color managed, it assumes the display profile is sRGB, and things work better for me if the hardware adjustment comes close to sRGB rather than a gamma curve. I now see substantial agreement between the non-CMS software and Gimp/Firefox/Affinity Photo, with the CMS apps representing color saturation more accurately. Though I’m probably seeing (in Gimp) nearly the same step tablet luminances as before, it helps (me) that non-CMS apps aren’t making it seem like the tones are too light.
I had been using i1 Match 3 before DisplayCAL and thought it was calibrating to sRGB, but from the CMS/non-CMS differences I first observed in Affinity Photo, i1 Match is obviously calibrating to something like a gamma curve. I thought my displays were “calibrated,” but the #10 step on the tablet was being displayed with a too-low luminance by apps that weren’t mapping the image profile to the display profile.
Users whose software is not 100% fully-color-managed might be helped by getting a nudge towards calibrating with an sRGB tone curve. Is there a precision difference between using sRGB vs. Gamma 2.2? Does that become less of an issue as video LUT depth increases from 8 to 10 to 12 bits?
Is there a precision difference between using sRGB vs. Gamma 2.2? Does that become less of an issue as video LUT depth increases from 8 to 10 to 12 bits?
When the videoLUT bit depth is limited to 8 bits, you’ll have less likelihood of calibration artifacts like banding if the target response is close to the native response.
