It seemed to me after reading this that whenever possible I should work in the widest color space possible, as even if I cannot see a difference today I may be able to see a difference looking back on my art in 10 years on a different/better screen. So, I went over to Krita and picked the biggest color space to set as my new default.
Unfortunately, this seems to alter my art already completed to have different colors when I reopen old files. I’m concerned that this means even if I paint new pictures in the new color space, I may not understand substantially how the colors on my screen will differ from the colors on the screens of others. (I know there is always variation, but the shift in colorspace changing them so much on my own screen is making me worried that it will be severe on the screens of others.)
Can someone help me understand why my old images change their colors so much when the colorspace is changed from Krita’s default sRGB to the ACES profile available?
If I have a beefy computer, are there reasons not to use the larger colorspace if I want to post my art online to share it?
Does it matter if I am not printing my art that I care about this at all?
Thank you.
Trying to really understand digital painting as a medium
I think using the most popular settings is fine. The quality and settings of other people’s displays are uncontrollable, and if you care about other people, you’re better off posting with the most commonly used standards. On the internet, sRGB 8bit is more than enough.
Screens may be better 10 years from now, but you’ll definitely be better 10 years from now too. Just grasp the moment and don’t think too far ahead.
I think the color space is a matter of suitability, not a matter of advanced replacing backward. If there is no real need, there is no need to delve into ACES or the like.
I think what I don’t understand is how to gauge suitability. What makes one colorspace better than another for a given application?
What is the point of being in one colorspace versus another, particularly if it’s likely that moving to a larger colorspace will make your art appear differently to other users?
What could be maybe invented in 10 years doesn’t really matter because you can not predict how your colors will look then. When you use a large color space now, one your current display can’t handle, you can obviously not see the out of gamut colors correctly and not only do you have to guess how it could look on a screen not invented yet, the colors on your current screen are wrong too. Worst case is that the colors are bad on every display.
Color spacees also can influence how some blending modes work snd influence your workflow. You should focus on what is needed now snd works best for most. A lot of displays in use nowadays aren’t even able to handle 100% sRGB so when your does, you’re already ahead of most people.
The default color profile active for new documents is sRGB-elle-V2-srgbtrc with 8bit depth (if not changed). It’s an sRGB non-linear color space, a Creative Commons variant of sRGB that fixes some things for modern displays. You can read more about it in the description in Krita’s color space browser. This is the go-to profile when you want to work for the web and since it is based on what is basically the baseline standard for images, will also be good enough for most other things. Maybe you want to try the linear space at some point (it changes mixing of colors) which is okay as long as you remember to select the correct profile when exporting.
Personally, unless I don’t have an absolute requirement to work in a different color space, I don’t so I rarely do. Especially since my Monitor screen tablet is only calibrated for sRGB anyway and Krita’s internal conversion (you can set this up in Settings → Configure Krita → Color management → Display) can only do so much.
The “ideal colorspace” is a matter of opinion, I’m afraid.
In any case, when working in 8 bit/channel it’s a bad idea to pick the largest gamut just because you can. The precision is just very limited, and something like Rec. 2020 is too large to prevent color banding, it’s designed for at least 10 bits per channel. Also any other colorspace will suffer conversion loss if the end result is displayed as 8bpp sRGB, no matter if you, some hoster, or the user’s browser converts it.
Plus you really can’t control what various sites do on forced conversions of uploaded images, be it social media, imgur silently serving a WebP version unless explictly using the direct link, any kind of forum etc.
That’s why people tell you “if in doubt, stick with sRGB”.
Thank you to everyone that replied! I’m not sure who to mark as Solution since I feel that several people contributed to the knowledge/understanding I was really looking for.