On Jul 26, 2006, at 4:16 AM, Joel Guillod wrote:
My guess from my investigations is that there is some issue in the
algorithm used to calculate the imagedata in the rev engine and
which is a function of the screengamma, the image gamma and some
obscure stuff... because extreme color channel values (near 0 or
near 255) are correct and the differences are bigger when in the
middle (near 127). So the gamma correction could be fault.
I'm coming in a little late in this, so I may miss the mark.
I pulled out an old PNG dump script and noticed that most PNG images
have both gAMA and cHRM chunks. Maybe some environments interpret
only one and not both.
IIRC, for painted images, the values refer to the environment that
the painting was done in. If I understand that right, a new
environment has to undo that and then redo the color processing for
the new environment.
I remember something about Windows making some assumptions about
gamma, but I forgot what.
I have a vague memory of long ago checking and seeing that imageData
was _before_ gamma changes. Maybe Rev changed or I slipped in from a
similar neighboring universe. Again, sigh.
I am successful with a test image just now in exporting to the P6
format of PPM, the default export. You might get something other
than P6, so check the PBM/PGM/PPM format. It is easy to parse. It
is for P6, I think, the rgb byte triples in binary right after the
header, that is, right after the 3rd lf in Rev's case. You might
want to compare or use that data. More is online, google for P6 PPM.
I got a compiler error in trying to export a mask.
I hope that gives some hints.
Dar Scott
_______________________________________________
use-revolution mailing list
[email protected]
Please visit this url to subscribe, unsubscribe and manage your subscription
preferences:
http://lists.runrev.com/mailman/listinfo/use-revolution
