[argyllcms] Re: what am I missing

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Thu, 07 Feb 2013 12:57:22 +1100

prathana shrestha wrote:

> I am trying to calibrate a scanner using Kodak Q60 target.  But I'm not able 
> to obtain
> the LAB values corresponding to the scanner RGB values, as given in the 
> 'profcheck'
> file.

Hi,
        I'm not sure what you mean by that. Typically the Lab/PCS/CIE values 
corresponding
to the test chart values come from the test chart vendor, or are something
that you have measured using a spectrometer. For instance, the Kodak
Q60 target references files are at <ftp://ftp.kodak.com/gastds/Q60DATA/>

scanin creates a .ti3 file that has the corresponding PCS + device values
in a single file, created from the CIE reference file + the RGB patch
value from the scan. This is the file that profcheck also uses.

> Please help, what am I doing wrong?Here is the procedure I am following: - the
> correction matrix is generated using: 'colprof.exe -v -am -qu' (I expect the 
> matrix
> value represents XYZ in D50)- the matrix is modified by using: Bradford matrix
> (D50-to-D65)

You can also check that you have got your math right by using absolute intent
with icclu, since that also undoes the D50 white point transform.

But what about the RGB channel curves ? You have to apply them
before the matrix to do a conversion from scanner RGB to XYZ.

> and XYZ-to-sRGB conversion matrix. (to apply the correction matrix to
> scanned RGB values and convert them to sRGB in D65)                Bradford 
> matrix
> =[0.9555766 -0.0230393  0.0631636;
>  -0.0282895  1.0099416  0.0210077;
>   0.0122982 -0.0204830  1.3299098]

Hmm - that's not the Bradford matrix, but did you mean that this is the
Bradford space D50 to D65 matrix ?

> XYZ-to-sRGB conversion matrix=
> [ 2.3706743 -0.9000405 -0.4706338;
>  -0.5138850  1.4253036  0.0885814;
>   0.0052982 -0.0146949  1.0093968];

Hmm - that's not the D65 XYZ to sRGB matrix. According to the copy of the 
standard
I have, it should be:

  3,2406 -1,5372 -0,4986
 -0,9689  1,8758  0,0415
  0,0557 -0,2040  1,0570


 - the modified matrix is applied to the scanner RGB values to obtain corrected 
sRGB values.

What about the sRGB per channel non-linearity ?

Graeme Gill.


Other related posts: