- This topic has 10 replies, 2 voices, and was last updated 6 years, 1 month ago by
.
-
Posts
-
Hi,
I have an Iiyama XU2492HSU and a Spyder5. I have calibrated/profiled it several times with Displaycal with different settings and now I am wondering which settings would be best for me.
Above all, I have noticed that different color-managed applications appear to interpret the same monitor profile in slightly different ways. E.g. Photoshop in general appears to lighten up deep shadows stornger than other programs such as Affinity or GIMP, and in general it also shows a slightly cooler image.
I have tested this by opening the ECI test image in different color-managed programs.
The differences appear to be very slight, however sometimes a very dark color may appear as black or the other way round.
I have tried the following settings in Displaycal: with and without calibration (gamma 2.2, srgb, profiling only etc), profile types curves+matrix and xyz lut+matrix.
How can I know which profile/settings are best for me? Which profile/application does actually display my photos correctly?
Thanks in advance
b.
Hi,
both Affinity Photo and Photoshop use relative colorimetric with black point compensation for screen rendering (the latter is generally desirable so there is no black clipping on displays with nonzero black level). The difference in near black rendering comes from different black point compensation implementations. BPC (= a form of perceptual mapping of only the lightness axis) when applied on-the-fly is a kind of “dynamic” rendering not set in stone, meaning there are vendor differences between implementations. Affinity Photo’s BPC is closer to that of the open source littleCMS (e.g. GIMP).
A way to eliminate those rendering differences is to use the exact same rendering intent without BPC, i.e. convert to profile, select the current display profile, select rendering intent “Perceptual”, disable BPC, convert. Note that this is only guaranteed to work for DisplayCAL-generated cLUT profiles, because their perceptual table is employing its own baked-in BPC.Thanks for the answer.
If I disable BPC, the image looks the same in PS and Affinity, as far as the shadows are concerned, i.e. dark colors are darker in PS than with BPC. The entire image (neutral grey) ist still more green in PS and more magenta in Affinity (it is barely noticeable tough).
But I think I a more confused now than before. The exact steps that you suggested only appear to be possible in PS, but the image does not change after conversion. However if I convert the image to my screen’s profile in Affinity, the darkest colors are lighter, i.e. very similar to the image in PS with BPC.
I still don’t know which redering/profile is more correct.
The entire image (neutral grey) ist still more green in PS
Try disabling OpenGL rendering in PS (settings -> performance -> use graphics processor, uncheck, restart PS).
But I think I a more confused now than before. The exact steps that you suggested only appear to be possible in PS, but the image does not change after conversion.
The DisplayCAL BPC is closer to Adobe BPC, so there is likely little visual change (in fact, in most cases it should look the same).
I still don’t know which redering/profile is more correct.
All the different implementations are equally valid. If you want consistent results, don’t use dynamic BPC rendering (although unfortunately in Affinity and PS, the only way to circumvent it is by converting to the display profile manually, as outlined above, because there is no way in these applications to set a default rendering intent for display. GIMP on the other hand seems to allow this).
Disabling the graphics processor did fix the color issue, but doesn’t that mean that my PS will be slower?
I am still quite confused. I tried it again with different profiles but converting to screen profile in PS only seems to work with the cureves-type profile.
My favorite Raw software is RawTherapee, which displays the image exatly the same way as PS. However I am thinking about ending my Adobe subscription. The trouble is though that other image editors do not display the image like RT. RT only darkens the shadows if I activate softproof (no printer profile selected) with a certain kind of screen profile.
Disabling the graphics processor did fix the color issue, but doesn’t that mean that my PS will be slower?
Depends on what you’re doing. I doubt you’ll notice it.
I am still quite confused. I tried it again with different profiles but converting to screen profile in PS only seems to work with the cureves-type profile.
Not sure what you mean. You can’t just use arbitrary profiles to convert to, otherwise there will be another implicit conversion to the display profile. It has to be the currently assigned display profile. Alternatively you can also set up softproofing with the display profile in the same way (only in Photoshop, won’t work in Affinity for some reason).
The trouble is though that other image editors do not display the image like RT.
When the same intent is used (perceptual, without dynamic BPC), the result should be exactly the same (bar rounding differences).
Well, I think I have come up with something. I have opened the test image and other photos in Opera, which has, as far as I know, the best color management as far as browers are concerned (it does recognize screen profiles, doesn’t it? I mean it has no fake color management). And the test image looks exactly like in Affinity. Consequently, PS must be wrong. However, as you wrote, this can be fixed by deactivating BPC in PS. At the moment I am using an XYZ Lut-profile, calibration is sRGB, I have used more than 1000 fields for profiling.
I hardly ever print my photos, and if I do I do it with dm/cewe. So I guess what the browser shows me must be “right” for me. The only question is whether other people’s browers/screens show my photos like mine. But I guess I can test that since we have more than one computer.
RT ist still a problem though. Nothing happens if I deactivate BPC. The image only looks similar to Opera if I activate soft-proofing without printer profile.
There is still confusion, but maybe less.About the “arbitrary profiles”: they are not arbitrary, I am just experimenting all the time and so I have created several profiles for my actual screen – but with different settings. So when converting, I am using the active screen profile.
You know, you are actually understanding color management but I am just trying this and that in order to make the image look the same. I think I actually understand very litte 😉
I have opened the test image and other photos in Opera, which has, as far as I know, the best color management as far as browers are concerned
Hmm. Isn’t Opera based on WebKit (Chrome)?
You know, you are actually understanding color management but I am just trying this and that in order to make the image look the same. I think I actually understand very litte ?
It’s a learning curve for sure. You’ll gain more confidence over time 🙂
