[Darktable-users] strange blue coloring of Canon raw files (EOS 700D)
Hi, blue colors in the raw files of my Canon 700D are turned into much darker blue or purple in darktable when applying the usual corrections (sharpen, orientation, base curve). The jpg files produced by the camera show a blue which is much closer to the real color. Is this a known problem? What Is the best way to correct this? I have the impression that this does not occur in rawtherapee. To get an idea I've uploaded the raw file with the darktable xmp file, the jpg and the jpg created by darktable: https://www.dropbox.com/sh/mgcoja286dbrub9/AACK2wqO9pSpBtcoAWyc6BHJa?dl=0 Thanks in advance! Regards, Gerd -- ___ Darktable-users mailing list Darktable-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-users
Re: [Darktable-users] strange blue coloring of Canon raw files (EOS 700D)
On 07.01.2016 09:36, Christian Mandel wrote: > Am 07.01.2016 um 00:02 schrieb Robert William Hutton: >> On 07/01/16 09:21, gersch...@gmail.com wrote: >>> blue colors in the raw files of my Canon 700D are turned into much darker >>> blue or purple >>> in darktable when applying the usual corrections (sharpen, orientation, >>> base curve). The >>> jpg files produced by the camera show a blue which is much closer to the >>> real color. >>> Is this a known problem? What Is the best way to correct this? I have the >>> impression that >>> this does not occur in rawtherapee. >> >> Yeah, this problem has been present for quite some time. I see it in shots >> taken with my Canon 60D >> but not in those taken with my 5d3. It's cased by the input color profile; >> try changing the profile >> from "enhanced color matrix" to "standard color matrix" or even "rec2020". >> >> The ultimate solution is to profile your camera using a color target: >> >> https://encrypted.pcode.nl/blog/2010/06/28/darktable-camera-color-profiling/ >> >> Darktable used to accept these custom profiles for inclusion in the main >> project, but this is >> currently not happening pending a rework on the way the input end of the >> pipe works (though I don't >> remember the details). None of this stops you using your own custom profile, >> though. > Thanks for your information. I've tested different input color profiles but I still have the impression the blue is too saturated. Is there someone with a Canon camera who would share their optimised color profile or darktable style? Thanks also for the hint to try RawStudio profiles. Regards, Gerd Regards -- ___ Darktable-users mailing list Darktable-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-users
Re: [Darktable-users] strange blue coloring of Canon raw files (EOS 700D)
> blue colors in the raw files of my Canon 700D are turned into much darker > blue or purple > in darktable when applying the usual corrections (sharpen, orientation, > base curve). The > jpg files produced by the camera show a blue which is much closer to the > real color. > Is this a known problem? What Is the best way to correct this? I have the > impression that > this does not occur in rawtherapee. Yeah, this problem has been present for quite some time. I see it in shots taken with my Canon 60D but not in those taken with my 5d3. It's cased by the input color profile; try changing the profile from "enhanced color matrix" to "standard color matrix" or even "rec2020". The ultimate solution is to profile your camera using a color target: https://encrypted.pcode.nl/blog/2010/06/28/darktable-camera-color-profiling/ Darktable used to accept these custom profiles for inclusion in the main project, but this is currently not happening pending a rework on the way the input end of the pipe works (though I don't remember the details). None of this stops you using your own custom profile, though. >>> >> >> Thanks for your information. I've tested different input color profiles but >> I still have >> the impression the blue is too saturated. >> Is there someone with a Canon camera who would share their optimised color >> profile or >> darktable style? >Have you already tried setting "gamut clipping" in "input color >profile" to "linear Rec2020 RGB"? >https://www.darktable.org/usermanual/ch03s04s03.html.php#d0e7524 >> Thanks also for the hint to try RawStudio profiles. >> Regards, Gerd >Roman Hi, yes, I tried Rec2020 RGB with gamut clipping but now all colors look rather pale. I think that the colors in the standard color matrix look good except the blue being too dark. Regards, Gerd -- Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now! http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140 ___ Darktable-users mailing list Darktable-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-users
Re: [Darktable-users] strange blue coloring of Canon raw files (EOS 700D)
On 09.01.2016 09:21, Roman Lebedev wrote: > On Sat, Jan 9, 2016 at 1:49 AM, gersch...@gmail.com > wrote: >>>>>>> blue colors in the raw files of my Canon 700D are turned into much >>>>>>> darker blue or purple >>>>>>> in darktable when applying the usual corrections (sharpen, orientation, >>>>>>> base curve). The >>>>>>> jpg files produced by the camera show a blue which is much closer to >>>>>>> the real color. >>>>>>> Is this a known problem? What Is the best way to correct this? I have >>>>>>> the impression >> that >>>>>>> this does not occur in rawtherapee. >>>>>> >>>>>> Yeah, this problem has been present for quite some time. I see it in >>>>>> shots taken with >> my Canon 60D >>>>>> but not in those taken with my 5d3. It's cased by the input color >>>>>> profile; try >> changing the profile >>>>>> from "enhanced color matrix" to "standard color matrix" or even >>>>>> "rec2020". >>>>>> >>>>>> The ultimate solution is to profile your camera using a color target: >>>>>> >>>>>> https://encrypted.pcode.nl/blog/2010/06/28/darktable-camera-color-profiling/ >>>>>> >>>>>> Darktable used to accept these custom profiles for inclusion in the main >>>>>> project, but >> this is >>>>>> currently not happening pending a rework on the way the input end of the >>>>>> pipe works >> (though I don't >>>>>> remember the details). None of this stops you using your own custom >>>>>> profile, though. >>>>> >>>> >>>> Thanks for your information. I've tested different input color profiles >>>> but I still have >>>> the impression the blue is too saturated. >>>> Is there someone with a Canon camera who would share their optimised color >>>> profile or >>>> darktable style? >>> Have you already tried setting "gamut clipping" in "input color >>> profile" to "linear Rec2020 RGB"? >> >>> https://www.darktable.org/usermanual/ch03s04s03.html.php#d0e7524 >> >>>> Thanks also for the hint to try RawStudio profiles. >> >>>> Regards, Gerd >>> Roman >> >> Hi, >> yes, I tried Rec2020 RGB with gamut clipping but now all colors look rather >> pale. I think >> that the colors in the standard color matrix look good except the blue being >> too dark. > Just to be sure: i meant: in "input color profile" module, set "gamut > clipping" to "linear Rec2020 RGB". > Not: in "input color profile" module, set "profile" to "linear Rec2020 RGB". > >> Regards, Gerd > Roman. Thanks Roman for the clarification. This setting improves the blue color tone a bit but it still looks different in color and darker. I've done a few more test pictures with lots of different blue objects. In many cases the raw files looks good in darktable using the standard input color profile. Only in some cases the blue becomes dark or purple. I have to test if this occurst at special situations, e.g. bright sunlight. Regards, Gerd -- Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now! http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140 ___ Darktable-users mailing list Darktable-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-users
Re: [Darktable-users] strange blue coloring of Canon raw files (EOS 700D)
On 09.01.2016 09:21, Roman Lebedev wrote: > On Sat, Jan 9, 2016 at 1:49 AM, gersch...@gmail.com > wrote: >>>>>>> blue colors in the raw files of my Canon 700D are turned into much >>>>>>> darker blue or purple >>>>>>> in darktable when applying the usual corrections (sharpen, orientation, >>>>>>> base curve). The >>>>>>> jpg files produced by the camera show a blue which is much closer to >>>>>>> the real color. >>>>>>> Is this a known problem? What Is the best way to correct this? I have >>>>>>> the impression >> that >>>>>>> this does not occur in rawtherapee. >>>>>> >>>>>> Yeah, this problem has been present for quite some time. I see it in >>>>>> shots taken with >> my Canon 60D >>>>>> but not in those taken with my 5d3. It's cased by the input color >>>>>> profile; try >> changing the profile >>>>>> from "enhanced color matrix" to "standard color matrix" or even >>>>>> "rec2020". >>>>>> >>>>>> The ultimate solution is to profile your camera using a color target: >>>>>> >>>>>> https://encrypted.pcode.nl/blog/2010/06/28/darktable-camera-color-profiling/ >>>>>> >>>>>> Darktable used to accept these custom profiles for inclusion in the main >>>>>> project, but >> this is >>>>>> currently not happening pending a rework on the way the input end of the >>>>>> pipe works >> (though I don't >>>>>> remember the details). None of this stops you using your own custom >>>>>> profile, though. >>>>> >>>> >>>> Thanks for your information. I've tested different input color profiles >>>> but I still have >>>> the impression the blue is too saturated. >>>> Is there someone with a Canon camera who would share their optimised color >>>> profile or >>>> darktable style? >>> Have you already tried setting "gamut clipping" in "input color >>> profile" to "linear Rec2020 RGB"? >> >>> https://www.darktable.org/usermanual/ch03s04s03.html.php#d0e7524 >> >>>> Thanks also for the hint to try RawStudio profiles. >> >>>> Regards, Gerd >>> Roman >> >> Hi, >> yes, I tried Rec2020 RGB with gamut clipping but now all colors look rather >> pale. I think >> that the colors in the standard color matrix look good except the blue being >> too dark. > Just to be sure: i meant: in "input color profile" module, set "gamut > clipping" to "linear Rec2020 RGB". > Not: in "input color profile" module, set "profile" to "linear Rec2020 RGB". > >> Regards, Gerd > Roman. Hi, I've actually found a photo where the blue color situation of Canon raw files can be seen very well: Have a look at https://www.dropbox.com/s/8ikzw34pcvpstqf/IMG_7993.CR2?dl=0 https://www.dropbox.com/s/ykhmyu12c9o89sa/IMG_7993.JPG?dl=0 The real color of the shirt is very close to the blue in the JPG but it appears to be purple in the raw image if the standard base curve is applied. The setting suggested by Roman improves the color a bit but not enough. Regards, Gerd -- Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now! http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140 ___ Darktable-users mailing list Darktable-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-users
[Darktable-users] compiling tools/basecurve
Hi, I'm trying to compile tools/basecurve using the manual given at https://github.com/darktable-org/darktable/tree/master/tools/basecurve and run cmake -DCMAKE_INSTALL_PREFIX=$YOUR_INSTALL_PATH -DCMAKE_BUILD_TYPE=Release .. I get the following error message: -- checking for module 'exiv2' -- package 'exiv2' not found CMake Error at /usr/share/cmake-2.8/Modules/FindPackageHandleStandardArgs.cmake:108 (message): Could NOT find EXIV2 (missing: EXIV2_LIBRARY EXIV2_INCLUDE_DIR) I've intalled dt via apt-get, so I didn't compile it myself. exiv2 is also installed via apt-get. Any ideas what could be the problem? Thanks and regards, Gerd -- Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now! http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140 ___ Darktable-users mailing list Darktable-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-users
Re: [Darktable-users] strange blue coloring of Canon raw files (EOS 700D)
Hi, Ok, I managed to compile and run the basecurve calculation program. The basecurve I get looks very similar to default canon curve. The colors look a bit closer to what I expect. Still a certain blue tone turns to purple while other blue tones stay blue. I wonder if there is another solution, e.g. by changing color zones. Thanks for your help so far! Gerd On 16.01.2016 14:46, johannes hanika wrote: > hi, > > that example image is totally not dark saturated enough blue to > trigger the Lab conversion turns purple colourspace problem (which is > solved by the gamut clipping, as roman mentioned). i suppose you could > try to measure a basecurve from a CR2/jpg pair for that camera [0], > that affects colour saturation a lot, too. > > cheers, > jo > > [0] https://github.com/darktable-org/darktable/tree/master/tools/basecurve > > On Sat, Jan 16, 2016 at 10:24 AM, gersch...@gmail.com > wrote: >> On 09.01.2016 09:21, Roman Lebedev wrote: >>> On Sat, Jan 9, 2016 at 1:49 AM, gersch...@gmail.com >>> wrote: >>>>>>>>> blue colors in the raw files of my Canon 700D are turned into much >>>>>>>>> darker blue or purple >>>>>>>>> in darktable when applying the usual corrections (sharpen, >>>>>>>>> orientation, base curve). The >>>>>>>>> jpg files produced by the camera show a blue which is much closer to >>>>>>>>> the real color. >>>>>>>>> Is this a known problem? What Is the best way to correct this? I have >>>>>>>>> the impression >>>> that >>>>>>>>> this does not occur in rawtherapee. >>>>>>>> >>>>>>>> Yeah, this problem has been present for quite some time. I see it in >>>>>>>> shots taken with >>>> my Canon 60D >>>>>>>> but not in those taken with my 5d3. It's cased by the input color >>>>>>>> profile; try >>>> changing the profile >>>>>>>> from "enhanced color matrix" to "standard color matrix" or even >>>>>>>> "rec2020". >>>>>>>> >>>>>>>> The ultimate solution is to profile your camera using a color target: >>>>>>>> >>>>>>>> https://encrypted.pcode.nl/blog/2010/06/28/darktable-camera-color-profiling/ >>>>>>>> >>>>>>>> Darktable used to accept these custom profiles for inclusion in the >>>>>>>> main project, but >>>> this is >>>>>>>> currently not happening pending a rework on the way the input end of >>>>>>>> the pipe works >>>> (though I don't >>>>>>>> remember the details). None of this stops you using your own custom >>>>>>>> profile, though. >>>>>>> >>>>>> >>>>>> Thanks for your information. I've tested different input color profiles >>>>>> but I still have >>>>>> the impression the blue is too saturated. >>>>>> Is there someone with a Canon camera who would share their optimised >>>>>> color profile or >>>>>> darktable style? >>>>> Have you already tried setting "gamut clipping" in "input color >>>>> profile" to "linear Rec2020 RGB"? >>>> >>>>> https://www.darktable.org/usermanual/ch03s04s03.html.php#d0e7524 >>>> >>>>>> Thanks also for the hint to try RawStudio profiles. >>>> >>>>>> Regards, Gerd >>>>> Roman >>>> >>>> Hi, >>>> yes, I tried Rec2020 RGB with gamut clipping but now all colors look >>>> rather pale. I think >>>> that the colors in the standard color matrix look good except the blue >>>> being too dark. >>> Just to be sure: i meant: in "input color profile" module, set "gamut >>> clipping" to "linear Rec2020 RGB". >>> Not: in "input color profile" module, set "profile" to "linear Rec2020 >>> RGB". >>> >>>> Regards, Gerd >>> Roman. >> >> Hi, >> >> I've actually found a photo where the blue color situation of Canon raw >> files can be seen >> very well: >> Have a look at >> htt
