Providing GPL version of CIE for ColorChecker SG

[Oleksij Rempel]

Hello all,

first of all thank you for creating argyll!

I have ColorChecker Digital SG and ColorMunki Photo - will it be
possible with this device to crate correct CIE and will it be accepted
to argyllcms source?
If yes, what steps should i do to provide it?

[Graeme Gill]

Oleksij Rempel wrote:

Hi,

Post by Oleksij Rempel
I have ColorChecker Digital SG and ColorMunki Photo - will it be
possible with this device to crate correct CIE and will it be accepted
to argyllcms source?

It's possible, although the best way of doing this is to create
a .ti2 file to do the work, similar to the ref/ColorCheckerDC.ti2
provided with ArgyllCMS. See attached for something that
may work for you.

You can then just use chartread -p to read the chart.

Note though that you will have a file that is right for your
particular ColorCheckerSG, and not someting that necessarily
represents the aim points or batch average for all ColorCheckerSG
charts. Other people have posted such reference files.

Note that X-Rite makes available its reference values here:
<http://www.xrite.com/documents/apps/public/digital_colorchecker_sg_l_a_b.txt>,
and there a procedure for converting it to a reference files is given
here <http://www.argyllcms.com/doc/Scenarios.html#PS3>.

Some people have noted that X-Rite's reference values don't seem
to match actual charts very well.

Graeme Gill.

[Oleksij Rempel]

Post by Graeme Gill
Hi,

Post by Oleksij Rempel
I have ColorChecker Digital SG and ColorMunki Photo - will it be
possible with this device to crate correct CIE and will it be accepted
to argyllcms source?

It's possible, although the best way of doing this is to create
a .ti2 file to do the work, similar to the ref/ColorCheckerDC.ti2
provided with ArgyllCMS. See attached for something that
may work for you.
You can then just use chartread -p to read the chart.

Ok, thanks.

Post by Graeme Gill
Note though that you will have a file that is right for your
particular ColorCheckerSG, and not someting that necessarily
represents the aim points or batch average for all ColorCheckerSG
charts. Other people have posted such reference files.

I don't know if ColorCheckerSG == ColorCheckerDigitalSG. At least it has
different branding. See attachment.
IMO, since XRight software do not provide any way for setting Version of
Colorchacker (i don't think there is good autodetection), all users
including Windows, Mac or what ever, use incorrect chart values. I think
Agryll can warn users about it, but still provide some reference for
those who has no spectrometer.

Post by Graeme Gill
<http://www.xrite.com/documents/apps/public/digital_colorchecker_sg_l_a_b.txt>,
and there a procedure for converting it to a reference files is given
here <http://www.argyllcms.com/doc/Scenarios.html#PS3>.
Some people have noted that X-Rite's reference values don't seem
to match actual charts very well.

Same about me.

Yesterday i was plying with it and gathered my first experience. Here
are my notices and questions:
* Each reading of same point by spotread -H provided too different
results. Some times difference is bigger then this:
XYZ: 8.489384 6.420652 10.844807, D50 Lab: 30.449855 22.221768 -21.609957
XYZ: 8.496172 6.425155 10.848450, D50 Lab: 30.460711 22.234237 -21.602624
XYZ: 8.472564 6.406772 10.814047, D50 Lab: 30.416359 22.219140 -21.571466
Probably it makes no sense to use ColorMunki Photo it in high resolution
mode. I would say it is device with 0.1 resolution.

* May be too high noise ration which i got yesterday was related to room
temperature and distance to the room heater

* Results between Xright software and Argyll are almost identical,
except of permanent 0.2 offset for luminescence.
XRight: 30.6, 22.1, -21.8
Argyll: 30.451469 22.147632 -21.731560
Is it critical?

* ColorCheckerDigitalSG has part which announced to have same colours as
original ColorChecker. I measured and compared some fields of
ColorCheckerPassport, ColorCheckerClassic and ColorCheckerDigitalSG.
ColorCheckerPassport and ColorCheckerClassic have almost identical
results (within noise ratio). ColorCheckerDigitalSG has absolutely
different results. For example Violet:
Classic(B4): 30.6, 22.1, -21.8 -- comparable to reported results
Digital(H3): 20.0, 33.3, -30.4 -- different from reported ColorCheckerSG