Ah, yes, Qt uses a quite particular layout. When they say "RGBX" they
mean not the
memory organization but an uint32 containing the pixel. Then, since x86
is little
endian, channels got swapped. I was tricked with that as well.
Just create a QImage with rgb 1, 2,3 and examine the memory to see what
I'm
talking about.
Glad to see you solved the issue
Regards
Marti
El 28/09/2012 13:34, Andreas Rettig escribió:
Ok, I get exactly the same result like that from Photoshop when I
change my code to this:
rgbInTemp[j] = col.green(); //before it was blue
rgbInTemp[j + 1] = col.blue(); // before it was green
rgbInTemp[j + 2] = col.red();
So something with the colors seems to be wrong...
Am 28.09.2012 um 11:23 schrieb Marti Maria <marti.ma...@littlecms.com
<mailto:marti.ma...@littlecms.com>>:
even with that parameter set, the colors are different. Absolute
colorimetric was just one Intent I tried, but it was the same
problem with all the other intents.
There is always much more red in the lcms-version.
Here is a comparison with INTENT_RELATIVE_COLORIMETRIC.
https://www.dropbox.com/s/nsgfp2c9cab6g6w/pslcms.jpg
Something is different in your setup. Is your ampel.icc identical to
the embedded profile? Is "sRGB Profile.icc" same as the canonical
"sRGB Color Space Profile.icm" used by photoshop? Are you using ACE
as color engine in photoshop?
You could try jpgicc utility to make sure the embedded profile is
being honored. The profiles involved in your transform are simple
matrix-shaper, which are very simple. I don't discard a bug, but I
cannot reproduce it.
Regards
Marti
------------------------------------------------------------------------------
Got visibility?
Most devs has no idea what their production app looks like.
Find out how fast your code is with AppDynamics Lite.
http://ad.doubleclick.net/clk;262219671;13503038;y?
http://info.appdynamics.com/FreeJavaPerformanceDownload.html
_______________________________________________
Lcms-user mailing list
Lcms-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user
