DxO PL6 - playing with camera profiles

DxO PL6 v6.1.1 (b86)

Canon R5 raw = 012A1001.CR3 from DPReview

DxO PL6 “No Correction” preset applied for a start

DxO PL6 WB measured off “C4” patch ( most neutral a/b from Lab data and yet bright enough for S/N ) … DxO WB sliders were Temperature: 6924 + Tint : 0 ( not that sliders nominal values matter give how screwed the whole WB process is, just a note what they were )

CC24 data from BabelColor PatchTool = CC_Avg30_spectrum_CGATS.txt
(there are some differences between CC24 color pigments before and after 2014 - we assume DPReview still used pre 2014 version based on the text under the chart - it looks like CC24 from 2010 redesign , already X-Rite - not Gretag MacBeth ) = https://www.xrite.com/service-support/new_color_specifications_for_colorchecker_sg_and_classic_charts

BabelColor PatchTool v5.5 ( https://babelcolor.com )

Adobe DCP profile is a dual illuminant profile extracted from ACR code ( by removing all external DCP profiles from the system, converting CR3 to DNG forcing DCP profile to be embedded there and extracting it afterwards with dcptool, then inserting a linear curve ) - so we have a matrix profile ( 2 x CM + 2 FM + LTC )

DxO PL6 [Working Color Space] = Version : Classic (Legacy)

DxO PL6 [Color Rendering] = DCP Profile + … + Intensity : 100 + PSC : 0

DxO PL6 [Export to Disk] = Action : Export as TIFF + Quality : 16bit + ICC profile : ProPhotoRGB ( CC24 fits in sRGB gamut with one F3 “cyan” patch going outside sRGB but still well within AdobeRGB gamut ) + PSC : OFF

NOTE:: granted in a ideal situation we will use our own target individually measured multiple times to average (and discard clear outliers) with our own spectrophotometer (either with a stable LED light /like the recent iPro3/ or properly heated Tungsten light /like what we have in pre iPro3 devices/ to address tungsten vapor deposit drift ) + our own source of light (again measured) like Solux 5000K halogen (no longer sold as of 2023 - you still can get Eiko 4700K halogen though ) + light box covered w/ all black matte material from all 6 sides around + will be using a single illuminant profile for that illumination light specifically, but alas

TOP/LEFT triangles are from the REFERENCE data & BOTTOM / RIGHT triangles are from DxO conversion results

RIP Solux, you were good !


R5 matrix DCP profile in DCPTool notation

Canon R5 Color & Forward Matrix + Linear Curve.xml (2.5 KB)

so a pure matrix profile does a decent job to reproduce at least CC24 colors in DxO PL6… now shall we test what a mess WideGamut from DxO is ?!

off the bat doing nothing but switching to DxO Wide Gamut makes all that

you don’t need the numbers here

a normal person 'd ask - how is the world replacing internal working space with something of a wider gamut can do something like this with colors that were well fitting in whatever legacy working space DxO was using … beats me - so can we fix some brightness to save a day ?!

Did DxO provide use with some guidance before tramping on stuff that worked for DCP camera profiles ?