Hi Yves,
I initially indeed used identity luts created by 3DLUT Creator.
Since this is a hassle (and the luts are sometimes slightly off) I reverted
to creating them myself.
This really is very easy, the .cube file is well documented (the required
lut dimension is in the header).
By default, the data are in r,g,b order in the domain 0..1.
The reasons to create abstract profiles which are utilized during profile
creation are these:
1. Allow for user preferences of colours, and/or
2 .Fix colour ranges which were off, for instance by perceptual /
saturation mapping, and/or
3. Allow for a better usage of the printer gamut.
Due to the mapping, usually from a default space like sRGB / Adobe98 / ...
to a printer colour space a great deal of the gamut is 'lost'.
With precision edits, taking care not to touch memory colours, these
'voids' may be partly restored creating a more pleasing, popping, output.
The edits are rather dependent on the printer type (its gamut shape, that
is).
All this obviously is not about accuracy...
Greetings,
Auke
Yves Gauvreau wrote:
Auke,
I'm curious again on how you get your identity reference cube file, do you
use one created by 3DLut Creator without any color adjustment to the image
or something else?
Maybe I missed your previous post on this, sorry if I did, what's your
reasons to "embed" an abstract profile?
Thanks,
Yves
On 10/21/2021 10:01 AM, Auke Nauta (auke.nauta) wrote:
Hi Yves,
The .cube files generated from 3DLUT Creator contain the name of the ICC
profile used by the image to create the colour adjustments.
I use this ICC profile to convert from the rgb values in the .cube file to
CIELab.
This is done for both the reference .cube (identity) and adjusted .cube.
The Lab values are written to standard .ti3 files.
I just used a simple header for them. They only need the Lab data to
function with 'refine'.
---
CTI3
DESCRIPTOR "Argyll Calibration Target chart information 3"
ORIGINATOR "Argyll target"
CREATED "Wed Oct 20 14:47:45 2021"
DEVICE_CLASS "OUTPUT"
TARGET_INSTRUMENT "GretagMacbeth Spectrolino"
NUMBER_OF_FIELDS 4
BEGIN_DATA_FORMAT
SAMPLE_ID LAB_L LAB_A LAB_B
END_DATA_FORMAT
NUMBER_OF_SETS 82991
BEGIN_DATA
1 4.86944 0.40069 -3.57315
2 5.67259 -0.06737 -3.04615
3 6.52494 -0.50601 -2.66309
...
END_DATA
---
Greetings,
Auke
Yves Gauvreau wrote:
Auke,
curious, to know how you do "b. Create pseudo .ti3 measurement files from
the .cube files (identity and adjust)." below.
