Hallöchen!
Post by Novák LeventePost by Torsten Bronger[...]
Do you mean http://darktable.org/redmine/issues/9038 with
"vignetting correction"? I toy with the idea of implementing
9038 in LensFun.
[...]
Yes, for the vignetting (flatfield) part. But there is also the
dark frame removal feature which removes thermal noise. This can
not be done in lensfun as it depends on the sensor chip not the
lens. Also, even for the vignetting part if someone uses a
telescope without a regular photo lens attached it does not really
make sense to use lenfsun.
I've pondered this thoroughly and came to the conclusion not to take
"Lens" in "LensFun" too literally. After all, even TCA and
vignetting correction depend on the sensor if you are pedantic.
Moreover, LensFun always has the information which sensor was used.
Currently, it only takes the cropfactor info out of it -- but why
not more. Besides, at least three of the things you correct with
flatfields (vignetting, coloured corners, sensor dust) heavily
depend on the lens model and settings.
The API has to be extended by just one field: the timestamp of the
picture.
I don't say that I do that, but only because I don't know whether I
will have the time and skills. I'm halfway through with flatfield
correction in a standalone Python program. This would be needed to
be translated to C++ then (which is not my primary language).
Post by Novák Levente[...]
Can't the source code of RawTherapee be used for these? It does
contain these features.
I haven't looked at the code but I doubt this is worth the effort.
Typically, you have to change much when porting code from one
program to another, and the "core" of this feature is a simple
pixel-wise multiplication, possibly with an interpolation.
Tschö,
Torsten.
--
Torsten Bronger Jabber ID: ***@jabber.rwth-aachen.de
or http://bronger-jmp.appspot.com