Re: [Gimp-developer] 10-bit display color depth support (30-bit, deep color)

[Magnus Larsson <k magnus larsson bahnhof se>]

Den 2021-12-05 kl. 19:06, skrev Liam R E Quin:
> On Sun, 2021-12-05 at 12:58 +0100, Magnus Larsson wrote:
> > I am using EIZO Color Edge CG279X monitor and the quoted message was
> > from Nov 2020.
> Have you tried a more recent gimp 2.99, and made sure the image is in
> 32-bit floating point linear precision/mode?
I am building GIMP_2_99_8-110-g44f6ee36fe. That is the latest git commit 
from https://gitlab.gnome.org/GNOME/gimp.git 
<https://gitlab.gnome.org/GNOME/gimp.git>.

The test files from EIZO are:

 * 8 bit non-linear integer for the 8 bit gradient and
 * 16 bit non linear integer for the 16 bit gradient.

This format ought to suffice for a test comparing the two files linked 
from the EIZO site. I also prefer to execute the test as instructed by 
EIZO with their original files.

I see no difference if the files are converted to 32 bit floating point 
linear light in GIMP | Image | encoding.

What is the final precision of the preview generated in GIMP main window 
by GTK3?

It seems the GIMP main preview window is stuck to 8 bit per color 
channel, or my whole system is stuck on 8-bit per color channel.