[argyllcms] Re: targen question
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Fri, 20 Dec 2019 18:02:39 +1100
Yves Gauvreau wrote:
Hi,
here is bit of a ti1 file generated by targen. I've tried to convert line 1
XYZ value to
its RGB counterpart and vice-versa using the Lindbloom calculator here
Any idea why?
yes, because you are examining the target generate guide values. These are
exist just to assist chartread in recognizing the patches. They aren't
guaranteed to be accurate values of any particular colorspace. They
may correspond to the target -c profile XYZ values (and do for
the -n option). Note that they are absolute values.
If you want accurate values run the .ti3 through fakeread with
your chosen ICC profile.
Your suggestion of using this command for generating o sequence of 30
Neutrals from L* 1
to L* 30 doesn't quite get me what I want, yes I get 30 gray patches but from
RGB 0 to 1.
They correspond to the 30 neutral values with a L* value from 0 to 100.
targen -d3 -c sRGB.icm -e0 -B0 -f0 -n30 outfile
If you want 1 to 30, generate 0 to 100 and edit the result:
targen -d3 -c sRGB.icm -e0 -B0 -f0 -U -n101 outfile
[I notice that there is a limitation in the current targen which rounds
values within 2% of 0% or 100% down/up that causes the L* 1 and L* 99
values to be rounded. I'll change this in the next release to have
a smaller tollerance for the rounding.]
For instance, patch 10 is RGB 10.76989 10.76989 10.76989
xicclu -s100 -ff -ia -pl sRGB.icm
10.769890 10.769890 10.769890 [RGB] -> MatrixFwd -> 9.999983 -0.535200
-4.349175 [Lab]
(L*a*b* being D50 relative, while absolute white point of sRGB is D65, i.e.
L*a*b* 100.000000 -2.387816 -19.404026).
Or lookup the relative value:
xicclu -s100 -ff -ir -pl sRGB.icm
10.769890 10.769890 10.769890 [RGB] -> MatrixFwd -> 9.999983 0.000000 0.000000
[Lab]
Cheers,
Graeme Gill.
Other related posts: