Perhaps using sRGB as the input color profile does make some sense. Darktable/rawspeed would I think normally demosaic the raw file into its internal color space (RGB). Using sRGB as the input color profile means that no remapping occurs after demosaicing, and so we have the image in the intended linear color space. Or maybe I am talking rubbish so I will shut up now.
Jim On Tue, Jan 8, 2019 at 11:40 AM Jim Robinson <[email protected]> wrote: > I have now checked the vignetting data with some real images and they do > a good job. Whilst sRGB is indeed the color space set in the camera and > does appear in the metadata of the raw file I don't think that it can be > relevant to the raw data. I see from the dcraw man page that olympus raw > files have an embedded color matrix that the -M switch (used in > calibrate.py) ignores. > > *+M* or *-M*Use (or don't use) any color matrix from the camera metadata. > The default is *+M* if *-w* is set or the photo is in DNG format, *-M* > otherwise. > Besides DNG, this option only affects Olympus, Leaf, and Phase One cameras. > Maybe darktable/rawspeed uses the the embedded matrix. This could perhaps > explain why there is very little difference between use of rec2020 and > enhanced color matrix as input profiles. > > Anyway I am happier now, but maybe this is just a problem/solution with > .ORF files. > > regards, > > Jim > > >>> >>> >>>
_______________________________________________ Lensfun-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/lensfun-users
