meanwhile I have completely abandoned sRGB for saving photos (embedding the profile inside the saved pics). Fortunately Pixelfed and most web browsers can deal with non-srgb-pics.
However, I wonder what the best alternative is - AdobeRGB oder Display P3?
My desktop monitor is AdobeRGB (or larger), my laptop Display P3.
I think I won’t be able to see correct colors on my laptop if I save the pics in AdobeRGB, will I? Or the other way round?
Or just use Display P3 when saving on laptop and AdobeRGB when saving on desktop? That’s what I am doing right now. But often I can see that photos that I edited on the laptop do not look exactly the same on the desktop.
I guess if you wanted to try using just one preferred profile you could bring test pairs back in to DT on both devices and compare with snapshots among and between to see how things look to your eye. P3 skews to reds and yellows and adobe towards cyans/greens. I think adobe is still suggested if your destination is print and P3 for screen viewing. So P3 would line up best for phones laptops etc if you are sharing.
This is a question and not a statement from me: I thought sRGB was the preferred profile for universal display including web browsers. Am I mistaken? I personally use Adobe RGB for archival storage and sRGB for JPG. I am happy to be corrected.
A lot of print services and printers only accept SRGB, so if you’re targetting them, I’d use that. If you are doing anything particularly colour critical, use the profile of the printer in question. As for screen, it looks like DCI-P3 is quite commonly supported on TVs and mobile devices, so depending on how you are distributing your images, that’s a good target. Alternatively, if you’re looking for long term future proofing, you could always output directly in rec2020, but I expect you would get issues on displays which don’t know how to handle it.
I’m outputting DCI-P3.
It would seem the major browsers will convert to sRGB by default, if an embedded profile accompanies the image. I’ve demonstrated this for myself with Chrome, others YMMV. With that, you could reasonably encode your images with whatever profile you desire, and browsers will just chunk it down to sRGB for you.
I still use sRGB for export to web destinations, just too lazy to figure out what else to use.
I think most web-browsers are colour managed these day. You’ll probably find misconfiguration from time to time, but you can no-longer assume that wider gamuts are ignored.
Wouldn’t some of those colors be perceptually converted even if an sRGB output profile is used by the browser? For example, when I was testing color management in my image viewer, I used the famous test images: Examples of various wide-gamut images
And even with sRGB as the output, images with wide gamut input profiles had noticeably different colors
Propably you mean Display P3, not DCI-P3? They have different white points and different gamma curves (Display P3 has been derived from DCI-P3 by Apple and is used throughout their devices, mobile and desktop).
This is strange - might be some marketing (mis-)communication on the part of Asus.
This test Asus ProArt P16 review (2025 H7606WP model, AMD Ryzen AI, RTX 5070) (arbitrarily selected) measures a gamma of 2,2 and white point of 6400 K.
So the gamma corresponds to Display P3 (DCI-P3 has 2,6) and the white point is right between Display P3 (6500 K) and DCI-P3 (6300 K).
I also thought so, until literally an hour ago, when I checked Firefox (latest version with its default configuration) on macOS. My iMac is 2020 27" Retina, with a Display P3 screen.
Edit: I was not aware that Discourse converts all image previews to sRGB → you can only see the full gamut (and the difference between Firefox and Safari in this case) when clicking on the above preview and viewing the original, enlarged image file.
Need to remember for viewing PlayRaws etc. …
The “DCI-P3 100%” probably refers to the gamut, and for this it seems to be correct.
The difference in white point is subtle and you easily adapt to it, I agree. However, if you really had the DCI-P3 gamma of 2.6, all (unmanaged) image display would appear completely off i.e. darker and with exaggerated contrast in my understanding.
With my low level of understanding of things, I have thought that maybe the screen is able to make more nuances than what linux/mint/darktable may actually utilize – and that is perhaps the same as you indicate.
But when one also take into account that our vision can in an instant change the perception of one area from being clearly very white to be o so black, and that we also have no control over the screens where others may see our pictures – I think, that until the world has more or less standardized on wider gamut screens, this is not the topic I will spend much time on. (I see the dilemma as to what standard one shall use when storing images now. But I wonder what is really the important point of storing a profile in the picture file …?
I was also suprised to learn about that this summer (via Wide Gamut - Test Page) - it seems, Firefox either has lost or never fully implemented proper wide-gamut support.
