Re: [CinePaint] strange handling of SIGTERM

On Tue, Aug 18, 2009 at 2:42 AM, Frank Peters<[email protected]> wrote:

> If you want to evaluate your workflow and troubleshoot any color
> management issues, the best approach is to use some color profile
> test images.  A search should uncover a lot of these test images,
> but here is a link for a small set:
>
> http://developer.apple.com/technotes/tn2006/downloads/tn2115_Images.zip
>
> The corresponding description is here:
>
> http://developer.apple.com/technotes/tn2006/tn2115.html

Not sure about "any issues", given a bewildering number of things to
consider, but that's a cute little trick -- thank you for pointing it
out.

> These test images contain embedded profiles which can be extracted
> using ImageMagick or GraphicsMagick:
>
> convert test_file.tif test_file.icm
>
> The extracted profile can then be used in Cinepaint along with
> suitable work space (e.g. Adobe RGB 1998) and display profiles
> (e.g. sRGB) to see if color management is functioning properly.

It's good to know about IM's way of extracting a profile, but it
turned out, I could test Cinepaint without extracting a profile: it
reads the embedded one (although I think I read somewhere that it
didn't).

> Another tool to evaluate color management is tifficc (or jpegicc)
> from the LittleCMS package.  It will allow you to accomplish
> profile transformations manually (i.e. outside of any program)
> to compare with transformations that should occur inside the
> program.

That's also good to know, although IM would do it just as well. For
the record, and for the benefit of future visitors, here's what I
observed:

* these test images tell me the profile is not used when viewed with
unmanaged software, such as gthumb
* the word "not" disappears when viewed in Cinepaint, Krita, &c.
* it disappears as well after profile conversion:

tifficc -o/usr/local/share/color/icc/sRGB.icc RGB/TestRGB16.tif test.tif
convert RGB/TestRGB8.jpg -profile /usr/local/share/color/icc/sRGB.icc test.jpg

(IM complained about an "unknown field" in the test tiffs, but did the
job all right)

Gimp offers to convert the embedded profile, and the images do look as
if they were "color-managed", if conversion is confirmed.

I must say that the RGB tests looked more convincing than Gray tests,
where the "not" only disappeared (almost) when the images were viewed
exactly at the right angle; if looked at from just above the screen
axis, it was very conspicuous, especially on the mac, whose display is
extremely sensitive to the angle of viewing. But that's just a nuance.

So, the embedded profile gets read and applied everywhere it is
supposed to be, in particular, in Cinepaint. In fact, I could not make
it not apply the embedded profile -- it seems to ignore all settings
in the "Default behavior" group. The only way I could mess up the test
image in Cinepaint is by assigning any other profile to it.

At any rate, this resolves one of the uncertainties; onwards to the
remaining 200 or so.

Thanks,

--Gene

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july