Am 05.03.07, 14:41 -0300 schrieb gga: > > Kai-Uwe Behrmann wrote:
> > HDR is only possibly with matrix profiles, best without gamma and outside > > the above described way. > Yes, that's one reason for me to wanting to avoid doing lcms transforms > and needing the data raw and not wanting lcms to do any processing. > > Basically, I am hoping lcms' cmsHPROFILE (_lcms_iccprofile_struct ) is > keeping its data as float and allowing me to get to that data directly. > Or is lcms all 16 bits as soon as data is read (in which case I guess > lcms will be pretty much useless to me, and I'll need to it all with CTL)? I am afraid the profile itself contains only 16-bit values in the LUT part of its tables at a extent. The matrixes can be of a signed 15/16-bit type, thus reaching from -32767 to 32767 with 16-bit precision after the point. Still, matrix profiles with gamma 1.0 seems to be a good option. As a hint Greg Ward recommended to use XYZ with a illuminant of type E. Possibly you need some intermediate step to transform from a CLUT colour description or chain as described to a matrix based RGB. The matrix RGB description could remain final or go to XYZ. HDR data should allread be in gamma 1 matrix colour space. For such data a a conversion through the limitted range of curves or LUT may result in clipping. At least a splitting in matrix based HDR and possibly CLUT based LDR transforms should occure. regards Kai-Uwe Behrmann -- development for color management www.behrmann.name + www.oyranos.org + www.cinepaint.org ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
