[argyllcms] Re: Camera calibration: LUT only as good as matrix?
- From: Torsten Bronger <bronger@xxxxxxxxxxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Wed, 03 Jul 2013 22:10:53 +0200
Hallöchen!
Iliah Borg writes:
> [...]
>
>> I do my RAW converting with Darktable, which uses the Adobe
>> matrix for my camera. This profile is also used by dcraw (and
>> probably many other tools), and it is very close to my own
>> calibration.
>
> A shot of IT8 on a leather case is not a calibration. [...]
You mean that it is too inaccurate to be used as a calibration?
>> So, the calibration not only should be independend of the way you
>> got the TIFF, it even must be. (Unless the RAW converter is
>> broken.)
>
> Calibration is very dependent on how you are getting the shot. In
> particular, what light was it.
Calibration as I mean it is determining the position of the
(imaginary) primary colours of the camera in XYZ colour space. This
is valid in linear approximation, and independent of the shot.
> Here is a simple experiment. [...]
>
> Sorting by the error value one can see that in both cases deep
> blue-green patch B7 is the worst offender but with the filter on
> the lens the error is very tolerable while without it it is
> extremely significant. [...]
Well, I cannot explain these observations without experimenting with
such matters myself. But one can optimise to one particular
photograph because every photo covers a different part of the
possible colour space. In other words, the error remains the same,
you just move it into those colours that don't occur in your
picture. But for practical reasons, I am not interested in a
profile which is valid for *all* my pictures with that camera.
>> And correcting exposure and WB are mere linear operations, which
>> are hardly error-prone.
>
> What is linearity? We can say sensor is linear if it closely
> follows matrix profile.
There certainly is non-linearity in the sensor, but I was unable to
show it for my camera -- the line was perfectly straight. And then,
you needn't only be able to measure it, you must also see it with
the eyes. I personally doubt that any non-linearity is noticable
with modern cameras.
> Even lenses are not linear in a certain sense, [...]
What do you mean with this?
Tschö,
Torsten.
--
Torsten Bronger Jabber ID: torsten.bronger@xxxxxxxxxxxxxxxxxxxxx
or http://bronger-jmp.appspot.com
Other related posts:
