Am Samstag, 23. November 2013, 19:10:13 schrieb Marie-Noëlle Augendre: > 2013/11/23 Tobias Ellinghaus <[email protected]> > > > It uses what your system reports as the profile for that monitor. That's > > correct behaviour of darktable. Currently we are just trying to help you > > debugging your system so that the correct profiles are set. > > So, dt doesn't seem on its most behaviour on my computer, as both the > system (I can check it via the GUI) and Gimp (checked using the Filter as > you've told me yesterday) find the corresponding ICC profile for each > screen. > > > Ok, another thing I would like you to test is how the image looks in > > darkroom > > when you set the display profile to sRGB, compared to GIMP. > > I get the closest rendering with the output profile set to the actual > screen profile, the one that has been defined in the system. > If I set it to sRGB, the picture becomes a bit more saturated than in GIMP
I will refer to this later in this mail. [...] > > ps aux | grep colord > > colord 596 0.0 0.0 339184 1628 ? Ssl 09:29 0:00 > /usr/libexec/colord > marie-n+ 25918 0.0 0.0 112664 936 pts/0 S+ 18:50 0:00 grep > --color=auto colord This is good news and might be the key to get a working system for you! Did you compile darktable yourself or do you use a precompiled package? > > Last but not least: > > > > Could you please run these commands and tell us what the output is: > > xprop -display :0.0 -len 32 -root _ICC_PROFILE > > _ICC_PROFILE(CARDINAL) = 0, 0, 84, 56, 97, 114, 103, 108, 2, 32, 0, 0, 109, > 110, 116, 114, 82, 71, 66, 32, 88, 89, 90, 32, 7, 221, 0, 8, 0, 19, 0, 19 > > > xprop -display :0.0 -len 32 -root _ICC_PROFILE_1 > > _ICC_PROFILE_1: not found. These two commands tell us that the laptop profile got correctly loaded by your system while the profile for the big screen didn't make it to the appropriate xatom. This isn't necessarily surprising since the _ICC_PROFILE_1 xatom is less known -- many applications only check and set the _ICC_PROFILE. It however explains that darktable can't find the correct profile. I am not entirely sure why it's using the laptop profile instead, but that doesn't matter too much right now. > > hexdump <your laptop profile filename> | head > > 0000000 0000 3854 7261 6c67 2002 0000 6e6d 7274 > 0000010 4752 2042 5958 205a dd07 0800 1300 1300 > 0000020 0800 2a00 6361 7073 6e2a 7869 0000 0000 > 0000030 0000 4ca3 0000 5432 0000 0000 0000 0000 > 0000040 0000 0100 0000 d6f6 0100 0000 0000 2dd3 > 0000050 7261 6c67 ce25 faca c38d 689d 5eb3 772a > 0000060 16e4 eae4 0000 0000 0000 0000 0000 0000 > 0000070 0000 0000 0000 0000 0000 0000 0000 0000 > 0000080 0000 1500 6564 6373 0000 8001 0000 8700 > 0000090 7063 7472 0000 0802 0000 4b00 6d64 646e Converting the beginning "00 00 38 54" from hex to decimal numbers we get "0 0 56 84" which is also the start of the xatom from above (every pair of numbers is switched due to byte order). So it's indeed the laptop profile that is loaded into your xserver. > > hexdump <your bigscreen profile filename> | head > > 0000000 0000 7855 7261 6c67 2002 0000 6e6d 7274 > 0000010 4752 2042 5958 205a dd07 0800 1300 1200 > 0000020 1f00 1200 6361 7073 6e2a 7869 0000 0000 > 0000030 0000 1e6d 0000 f956 0000 0000 0000 0000 > 0000040 0000 0100 0000 d6f6 0100 0000 0000 2dd3 > 0000050 7261 6c67 a0e8 4d71 4fc1 bcfd 0c32 e581 > 0000060 9582 8db2 0000 0000 0000 0000 0000 0000 > 0000070 0000 0000 0000 0000 0000 0000 0000 0000 > 0000080 0000 1500 6564 6373 0000 8001 0000 8100 > 0000090 7063 7472 0000 0402 0000 4b00 6d64 646e This one differs in the 3rd byte already so it's not the one we see. This confirms our observations from yesterday. > I don't whether this can be an issue, but in looking for the actual place > these profiles were stored, I found out that the laptop one is in > /var/lib/colord/icc/ > and the big screen in /home/marie-noelle/.local/share/icc/ The second is the default for user supplied profiles, but this shouldn't matter. > > Please be assured that I really want to get this sorted out. > > For me, the problem is temporarily fixed as I can modify the output profile > for each picture; but I'll do my best to provide the information you'll > need to reach a more complete solution. Now that we know that the laptop screen has the correct profile assigned while the big screen has none at all we can try to make some sense of your observations: - darktable only loads the laptop profile since that is all it can find. - GIMP loads the correct profile, because for the laptop screen it can get the correct profile from the xserver, however for the big screen it can't. It still manages to provide the correct result because you manually assigned the profile for the big screen in your GIMP preferences (as can be seen in that screenshot). What gimp now does is using the manually supplied profile as the fallback when no xatom profile can be found, which leads to the correct result in your case. Lucky circumstances. - I don't know about the other applications, maybe they are querying colord, maybe they are lucky as GIMP. So, what can we do next? I guess the simplest would be to check why darktable doesn't get the correct profile from colord (and I am confident that colord has the correct profiles set. At least on my system that's quite reliable). What is the output of ldd `darktable` | grep colord ? > Thanks, > Marie-Noëlle I guess we are slowly getting there. :) Tobias
signature.asc
Description: This is a digitally signed message part.
------------------------------------------------------------------------------ Shape the Mobile Experience: Free Subscription Software experts and developers: Be at the forefront of tech innovation. Intel(R) Software Adrenaline delivers strategic insight and game-changing conversations that shape the rapidly evolving mobile landscape. Sign up now. http://pubads.g.doubleclick.net/gampad/clk?id=63431311&iu=/4140/ostg.clktrk
_______________________________________________ Darktable-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/darktable-users
