[Digikam-users] Re: Color Management Jeopardy

[Martin (KDE)]

Am Mittwoch, 16. März 2011 schrieb David Vincent-Jones:
> With so many persons using so few (relatively) camera models I am
> really surprised that there is not a wealth of camera profile data
> sitting out there for common use.

There is no such thing like ONE camera profile. First: it is not the 
camera you have to profile but the combination light-lens-camera. Next 
you may need different kind of profiles (matrix, lut ...). Every user 
has different requirements in the quality of the profile. There are 
different calibration pattern out there (IT8, ...)

What do you need a profile for, if you know only one of this 
parameter? I for myself build several profiles under different lights 
and situations. It is not that easy to build a really good one. Just a 
slightly bit of green in the background and your profile has to little 
green. You need a reproducible environment. I only use self-made 
profiles for special situations. For all others the default 
(darktable) profiles are quite good.

For raw development I currently use darktable. This software uses some 
general profiles (where available) and get really good results. But 
the main point is not the profile, but the base curve. Only if the 
colour (after white balance of course) is wrong there is a need of 
profiling (or colour correction).

> 
> Canon, unfortunately, now bury their information in a welter of
> propitiatory files, it almost appears that this manufacturer is
> trying to force the world into its software. Shame!

This is in no way different than Nikon (in some parts better in others 
worse). I am not forced in any parts to use canon software for raw 
developing. At least for linux there are lots of software out there 
(digikam, ufraw, rawstudio, darktable, photivo - just to name the open 
source ones). And some of them are realy good, even better than the 
canon software. I think the same is true for Nikon.

> 
> Maybe it needs the open-source community to assemble and
> disseminate reliable information.

As I already said, just a different lens with another coating and you 
get a slightly blue shift (or may be yellow). Most likely this is not 
visible but the same may be true for profiles as well.

Don't take me wrong: I like the idea of profiling and I use it. And 
thanks to Graeme we have all tools we need to do it as open source. 
And it may be worth the effort. But you have to take the effort by 
yourself and profile your workflow.

Martin

> 
> David
> 
> On Wed, 2011-03-16 at 12:47 -0400, Elle Stone wrote:
> > On 3/15/11, Graham Dicker <graham.dicker at antecor.com> wrote:
> > > First question that I have is how do I deal with the
> > > raw images out of my Pentax K200D? There is no embedded profile
> > > in the raw files and I cannot find one for this camera
> > > anywhere on the internet. On the CD that came with the camera
> > > there are icc files for sRGB and AdobeRGB but I don't know if
> > > these have any bearing on the processing of PEF/DNG files.
> > 
> > Hi Graham,
> > 
> > I haven't finished exploring the digiKam raw converter. But there
> > are three choices for the camera profile: "None", "Embedded",
> > and "Custom".
> > 
> > "None":
> > I assume that "None" means use the default camera matrix in the
> > dcraw code. That's the only logical conclusion, and the output
> > using "None" is identical with UFRaw output, after adjusting for
> > exposure. The only thing is, if you use any "Workspace" color
> > space other than sRGB, digiKam doesn't seem to handle the Tone
> > Response Curve ("gamma") correctly. (TRCs aren't always a true
> > gamma curve; sRGB doesn't have a true gamma TRC; digiKam seems
> > hard-coded to use the default dcraw TRC, which may or may not be
> > exactly the same as sRGB - I'm too lazy to track down the answer
> > - but the default dcraw TRC is most definitely not a true gamma
> > curve.)
> > 
> > The digiKam raw converter uses libraw, which gets its camera
> > matrices from the dcraw c-code. dcraw c-code does have a camera
> > matrix for the
> > 
> > K200D:
> >  { "PENTAX K200D", 0, 0,
> >  
> > 	{ 9186,-2678,-907,-8693,16517,2260,-1129,1094,8524 } }
> > 
> > The matrix above is what is used to convert your interpolated raw
> > file from the "raw" color to a working space.
> > 
> > "Embedded":
> > "Embedded" is an option from the dcraw code that only applies to
> > a few cameras. Quoting from the dcraw documentation
> > (http://www.cybercom.net/~dcoffin/dcraw/dcraw.1.html):
> > 
> > "+M or -M Use (or don't use) any color matrix from the camera
> > metadata. The default is +M if -w is set, -M otherwise. This
> > option only affects Olympus, Leaf, and Phase One cameras."
> > 
> > 
> > "Custom":
> > "Custom" requires that you have your own custom camera profile,
> > for which you'd need a target, the least expensive of which is
> > made by Wolf Faust
> > (http://www.colorreference.de/targets/index.html - you'd need to
> > purchase the C1 target).
> > 
> > So I'm not sure how to respond to your question regarding the
> > K200D - is the problem that you aren't happy with the colors you
> > are getting? Or you are not sure about the digiKam settings to
> > use? Or something else altogether? (Feel free to send a private
> > email if the problem isn't really digiKam-related.)
> > 
> > Martin is correct when he says that the manufacturer's camera
> > profile, if you could even find one, isn't likely to work very
> > well when using any software other than the manufacturer's own
> > proprietary raw rendering software.
> > 
> > Many people (including myself) do choose to create their own
> > custom camera profile because they are not happy with the colors
> > coming from the default dcraw camera matrix. My own custom
> > camera profiles are simple, general purpose matrix profiles,
> > just like the ones that dcraw uses, only the colors are nicer. A
> > lot of people who own Nikon and Canon cameras feel that the
> > default dng matrices produce anemic colors for these cameras.
> > 
> > Elle
> > _______________________________________________
> > Digikam-users mailing list
> > Digikam-users at kde.org
> > https://mail.kde.org/mailman/listinfo/digikam-users
> 
> _______________________________________________
> Digikam-users mailing list
> Digikam-users at kde.org
> https://mail.kde.org/mailman/listinfo/digikam-users