Hi! On Fri, Feb 5, 2021 at 1:00 AM Elle Stone <ellestone@xxxxxxxxxxxxxxxxxxxx> wrote: > A misconception I keep seeing on various forums needs to be corrected: > > GIMP-2.10 does *NOT* produce correct editing results in color spaces > other than sRGB. Neither does GIMP-2.99. > > Editing in AdobeRGB, ProPhotoRGB, Rec2020, etc WILL produce *wrong* > results for many operations, and unless you are thoroughly conversant > with the underlying code, or else have a way to compare results with a > properly color-managed editing application, you don't have any way to > know what's right and what's wrong. It's best to stick with editing only > in GIMP's built-in sRGB color spaces. > > The same is true if you are using GIMP-2.99: Some things that don't work > in GIMP-2.10, do work in GIMP-2.99. Other things that actually do work > in GIMP-2.10, don't work in GIMP-2.99. > > About two years ago major changes were made in babl and additional > changes were made in GIMP-2.99, messing up stuff that still works in > GIMP-2.10. For awhile progress was being made in GIMP-2.99 on extending > the arena of "what actually works", some of which progress is from bug > reports I filed and in some cases helped to fix - it seems nobody else > was testing the new code to see what actually did work. > > I was able to write code that fixed some of the bugs I reported for > GIMP-2.99 color management. But once I reached the point where further > coding requirements exceeded my coding ability, progress simply stopped, > with everyone else saying "some day" proper color management for GIMP > would be a priority. I began to feel like the best way to make sure a > bug would never get fixed, was to have the dreaded "Concepts: Color > Science" tag attached to it. > > Since autumm of 2013 I've been participating in GIMP development, mostly > in the area of color management (editing in color spaces other than > sRGB) and color science (making sure GIMP code produces correct results > for things like layer blend modes, Curves and Levels, AutoStretch, > Luminance, and so on; and adding code for things like LCh color pickers > and blend modes). > > Participating in GIMP development used to be challenging and enjoyable. > But over the last couple of years my interest in and patience with the > slow pace of progress regarding GIMP color management have dwindled to > the point of disappearing altogether. > > If someone else feels like helping with GIMP color management and color > science, here's a list of still-open bug reports that I reported after > the migration to gitlab, most of which have to do with color > management/color science: > > > https://gitlab.gnome.org/GNOME/gimp/-/issues?scope=all&utf8=%E2%9C%93&state=opened&author_username=ellestone > > Here are bugs that I opened before the migration to gitlab: > > https://gitlab.gnome.org/GNOME/gimp/-/issues?scope=all&utf8=%E2%9C%93&state=opened&author_username=bugzilla-migration&search=ellestone > > The most important color management bugs still open from before the > migration to gitlab are these: > > * Replace hard-coded sRGB parameters to allow editing in other RGB > working spaces (opened 6 years ago): > https://gitlab.gnome.org/GNOME/gimp/-/issues/594 - in some ways this bug > is obsolete as current GIMP color management issues are less about > actual hard-coded values and more about a lack of any way to convey the > required "not sRGB" color space information to various sections of code > that need this information. > > * Decomposed to LAB images have the wrong ICC profile assigned (opened 4 > years ago): https://gitlab.gnome.org/GNOME/gimp/-/issues/883 > > * Address various limitations of LCMS soft proofing (opened 4 years > ago): https://gitlab.gnome.org/GNOME/gimp/-/issues/976 > the > and hoping very much that GIMP will find new developers that them a lot > of energy and some interest and expertise in color management and color > science. > * Support for high bit depth RGB (and LCH?) color palettes for painting > (opened 2 years ago): https://gitlab.gnome.org/GNOME/gimp/-/issues/1328 > > Similar searchs in https://gitlab.gnome.org/GNOME/gegl/ and > https://gitlab.gnome.org/GNOME/babl/ will turn up a few additional color > management issues. > > Best of luck to all, > So it feels like you are saying goodbye. If so I hope it's only a temporary one and you'd be back if we get more active on the color management side. We will, we definitely will, because it's a major part of what GIMP 3 is supposed to be about anyway. I personally have a high theoretical interest on this topic, but not a practical one unfortunately (because we work in sRGB), which is why it is both high in priority for me (because we need it for GIMP 3) and low (compared to what we actually do day to day with GIMP). I have always been hoping that Mitch and Ell would be back soon as they are much better suited than me to decide on these topics anyway IMO. These are actually very intimidating topics. I do understand them (or so I believe), when scratching the head very hard, but I feel much better when I am not alone to validate your patches. Adding to this the existence of my main priorities, it's just easy to put these patches on the side temporarily "for later". 😕 > It's been stated over and over again by a particular GIMP dev that GIMP > development works best if people do what they are interested in doing. > The question is which current GIMP devs other than myself have an > *interest* in dealing bugs tagged with "Concept: Color Science"? I am guessing I might be this dev. I hope this doesn't mean I am part of why you lose interest in contributing. I do believe dev work best when everybody does what they like best. But I am also aware it can become frustrating when you are alone doing it, when it feels it's not priority to others and when you have unreviewed patches waiting forever. Sorry for this! It's really not my intent, it's just hard for me too. 😢 Anyway yes it's a real problem when we lack developers. 2020 has been a particularly dreadful year. Well we got some nice new developers, but 2 of the other main many-year-long developers got on a pause (one with no news at all, so we fear it might be more than a pause, though hope remains!). So yeah very bad year. 😱 Anyway whatever you decide, I wish you the best continuation in anything. I do hope you will still stay around, at least just a bit (like do not hesitate to comment if you need, even if just once in a while). And I would name your gitlab nickname if I need, so do not feel forced to answer, but we would be thrilled if you continue doing so from time to time. In any case you will always be considered part of GIMP. 🤗 Jehan Elle Stone > > -- > https://ninedegreesbelow.com > Color management and free/libre photography > _______________________________________________ > gimp-developer-list mailing list > List address: gimp-developer-list@xxxxxxxxx > List membership: > https://mail.gnome.org/mailman/listinfo/gimp-developer-list > List archives: https://mail.gnome.org/archives/gimp-developer-list > -- ZeMarmot open animation film http://film.zemarmot.net Liberapay: https://liberapay.com/ZeMarmot/ Patreon: https://patreon.com/zemarmot Tipeee: https://www.tipeee.com/zemarmot _______________________________________________ gimp-developer-list mailing list List address: gimp-developer-list@xxxxxxxxx List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list List archives: https://mail.gnome.org/archives/gimp-developer-list
