New Profiles: - This ones are little better than the last ones.
IMG_0173_Faithful.icc (25.4 KB) IMG_0173_FineDetail.icc (25.4 KB) IMG_0173_Landscape.icc (25.4 KB) IMG_0173_Neutral.icc (25.4 KB) IMG_0173_Portrait.icc (25.4 KB) IMG_0173_Standard.icc (25.4 KB)
How they were done (for my self future reference):
This point of this guide is to have “better” colors as the standard matrix right from the start, emulating the colors of the camera software, which I obviously find nicer.
Download:
it8.cht ( Amazon Drive )
https://www.sobek.su/Images/IT8.png (image of a IT8 Target with good resolution)
it8.ods ( Target It8.ods (68.5 KB) )
source_data.txt ( source_data.TXT (15.8 KB) )
Sources used:
https://encrypted.pcode.nl/blog/2010/06/28/darktable-camera-color-profiling/
http://www.hutchcolor.com/Spreadsheets_to_go/Formulae.xlsx
Open the IT8 target on the screen (for the reference I used a 4K 27”) and take a photo with RAW from a certain distance so that the pixels of the screen are not noticeable (a litle bit out of focus may help). Make sure the target doesn’t fulfil the whole picture so that there is not so much distortion.
Like this:
Import the RAW on DPP
Choose Picture Style – Faithful and Gama adjustment as non linear. Make sure everything else is off.(like Sharpness). Use camera white balance.
Choose Work Color Space and Profile as Wide Gamut RGB.
Choose Convert and Save and save a 16-Tiff with embed ICC. Add _DPP_Faithful or something else for better organization.
Import the RAW to Darktable, apply this settings:
Crop the image to same size of DPP output (im my case to 6000x4000px), use the margins settings to align the crop:
Its OK if the difference is just 1 pixel (like 6000x3999)
Crop the image to exterior square and apply the same crop on Darktable to the DPP .Tiff (copy and paste crop histogram)
For both images select the color picker with Lab values and make sure the white square on the bottom left has 96 L for both the Tiff and the Raw:
Because the crops are not 100% the same one have to add two live samples for each image:
Export the DPP TIFF to PFM.
“You need to select “Lab” as output color profile. This color space is [probably] not visible in the combobox by default. You first need to enable it by setting allow_lab_output to TRUE in $HOME/.config/darktable/darktablerc . Alternatively, you can start darktable with:
darktable --conf allow_lab_output=true
Then select “PFM (float)” as output format and press “export” to generate the source image file. “ (https://www.darktable.org/usermanual/en/darktable_chart_making_input_images.html)
Export the Darktable TIFF with the following settings:
TIFF 16 bit
rec2020
Absolute colorimetric
One has to do the rest on Linux:
Open $darktable-chart
Import the last exported DPP .Tiff and the it8.cht for the chart menu.
Adjust:
Go to to process Menu and export the RAW data as .csv (one may have to select the patches with gray ramp to be able to click on the export buttom). Save the file with a meaningful name.
Open the .cvs on libreoffice and paste the values on the downloaded LibreofficeCalc it8.ods file in the green table. The values are converted from LAB to XYZ on the BCD Colums (the conversion formula comes from here Spreadsheets / http://www.hutchcolor.com/Spreadsheets_to_go/Formulae.xlsx )
Make sure the A to L tables are correctly copied (like A1 to A1 and not A1 to GS0).
Now copy the source_data.txt and rename it to something useful like DPP_Faithful.txt
Copy the new XYZ values from the formula file it8.ods from colums A to D and from lines 5 to 292 to the DPP_Faithful.txt . Save the file.
Note: if one have more than one reference its better to copy the lastdarktable.tiff several times with meaningful names like lastdarktable_Faithful.tiff, lastdarktable_Neutral.tiff, etc.
Now with ArgyllCMS installed navigate to the folder of the Darktable .TIFF (the last one exported) and put it on the same folder the it8.cht and the DPP_Faithful.txt
Open terminal and write:
$scanin -v -p -a -dipn lastdarktable_Faithful.tiff it8.cht DPP_Faithful.txt
If everything goes right two files were created. The diag.tif should confirm that the reading was right. The boxes should be smaller like the ones on the $darktable-chart. If not increase the BOX_SHRINK value in the it8.cht file.
Now create the icc file (edit the obvious fields…):
$colprof -v -A "Canon" -M "Canon EOS M100 DIGITAL" -D "Canon M100 DPP Faithful" -C "Copyright Note" -q l -a m -u lastdarktable_Faithful
Im my files I had an avg error ranging from 7 to 12.
Copy the icc files to the right Darktable folder (on windows C:\Users\Username\AppData\Local\darktable\color\in ).
Check on darktable if the colors are “right” comparing the DPP and the standard matrix. If the image has a lower saturation, something went wrong.
