[argyllcms] Re: Using two cameras to create a (pseudo) profile
- From: Bernhard Bablok <bablokb@xxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Sat, 21 Jul 2012 12:58:50 +0200
Am Sat, 14 Jul 2012 14:57:52 +1000
schrieb Graeme Gill <graeme@xxxxxxxxxxxxx>:
> Bernhard Bablok wrote:
> > Hi,
> >
> > is the following procedure possible with the use of argyll-tools?
> >
> > * print a target (e.g. with targen)
> > * create a cht-file for the target
> > * take an image with camera A
> > * create the reference-file from the image from camera A
> > * take a second image with camera B
> > * create a profile using the reference-file created with camera A
> >
>
>
> > I know this is not really the catholic way, but I want to make the
> > colors of camera B look like the colors of camera A. BTW: "camera A"
> > could also be a film-camera+film-scanner (even using BW-film).
>
> Hi,
> I'm not sure why you want to deviate from a normal
> device space transformation. Simply profile camera A and profile
> camera B, and then create an RGB to RGB space transform from
> camera B to camera A device space. The transform can be on the fly
> using a CMM/cctiff or application, or using a device link (collink).
>
> ie. to be explicit, you want to emulate this:
>
> Scene CIE -> camera A processing -> RGB of camera A
> -------------- = camera A B2A ------------------
>
> with this:
>
> Scene CIE -> camera B processing -> RGB of camera B
> -------------- = camera B B2A ------------------
>
> So overall you want this:
>
> Scene CIE -> camera B processing -> RGB of camera B
> -------------- = camera B B2A ------------------
> -> inverse camera B processing -> Scene CIE
> -------------- = camera B A2B ------------------
> -> camera A processing -> RGB of camera A
> -------------- = camera A B2A ------------------
>
> and camera B A2B + camera A B2A is a normal device link.
> (This ignores any issues with gamut, white point etc.)
>
> Graeme Gill.
>
Hi,
would that also work if camera A is a black-and-white film (+scanner)?
IMHO, normal profiling for A will always fail, since there are no
color-values in the measurement.
I'm not through with testing, but taking
the script from Iliah I might be able to make the results of camera B to
mimic the results of the black-and-white film.
Bernhard
Other related posts:
