This is what I am shooting for. .
I was able to achieve this result without issue in earlier versions of RT which included the initial release of 5.9. I can’t say for sure when the problem started, but it’s definitely there on 5.9-206
Hi Mike,
Sorry: I really do not grasp what’s your question 
Grey patch #5 is supposed to have a Lab value of 49.7 0.03 -0.22
Your top image shows RGB values, hence of no interest here.
Your bottom image shows Lab 49.6 0 -0, which sounds natural,
because you asked RT to whitebalance on that patch!
(Use the exposure module to tweak the L value to 49.7.
Then what happens to the a and b values?)
Do you mean that old RT versions showed 49.7 0.03 -0.22 after
white balancing on that patch?
Have fun!
Claes in Lund, Sweden
No, but I was able to tweak the temp and tint to achieve 0.03 -0.20. Cannot get to those values now in the latest version 5.9-206
Aha! So that is what you meant.
Could you send me your raw file?
Click my avatar and send it as a message.
Nikon D5100 Camera neutral
Nikon D5100 Camera Neutral.dcp (271.3 KB)
Brilliant (?) idea:
Untick Observer 10 deg instead of Observer 2 deg
in the White Balance module.
More info here: https://rawpedia.rawtherapee.com/White_Balance
Have fun!
Claes in Lund, Sweden
I tried that Observer 2 deg still not getting it. I’m sure that someone smarter than me could do it.
Foyer_camera neutral_white.NEF (14.2 MB)
Nikon D5100 Camera Neutral.dcp (271.3 KB)
Using camera profile tone curve and look table, RCD-Bilinear. As basic as it gets.
I went all the way back to RawTherapee_dev_5.9-7 which I know worked okay. Easy Peasy!
Too bad the newer versions are more difficult for me to work with regarding WB.
@Lawrence37, could please link me to the download for RawTherapee5-6643merge. I believe it was this version that allowed me to open into PS from RT.
Also, is there a place where I can see all of the available versions which are still available after 5.9-7 for download.
Thanks,
Mike
All builds are available here as long as they are not more than 90 days old: Actions · Beep6581/RawTherapee · GitHub
There’s also select pre-dev builds which never expire: Release Pre-dev Builds · Beep6581/RawTherapee · GitHub
1 Like
Just a few points you might want to consider before chasing this one:
There have been changes that influence colour/representation/info shown, most notably:
These, but especially 6643/6640 have an impact on colours and initial representation. Observer 10 vs 2 was already pointed out, but even the “old” version 2 renders differently when using an old vs a new RT version that is set to observer 2.
About chasing the exact numbers shown in the Sekonic chart: This isn’t worth the time/effort and is actually impossible to do…
The examples shown by you are all done while in fit to screen mode. Zoom in to 100% and you will see that the numbers shown will differ:
Fit to screen view:
Same one but now @ 100%:
Placement of the spot itself: I moved the spot slightly upwards, all else is the same:
There’s also the age/condition of the target and/or possible dirt on the lens/sensor to take into account. The latter, for example, being a definite issue with the RAW you provided.
The example you show with the correct numbers isn’t correct. The a* channel should be 0.03 and not 0.3 as shown. Another reason why going for the exact numbers isn’t viable: RT’s precision shown is only 1 decimal point deep and you would have to start rounding up/down.
All of these will show different numbers. Which to use to get the correct result?
My advise: Be precise, but don’t overdo it (talking from experience 
).
The difference between 49,7/0,03/-0,22 as mentioned by the target reference sheet and, say, 49,6/0.0/-0.2 as shown in the first @100% example aren’t noticeable, especially when looking at them on a monitor.
Also make sure that you are working zoomed in at full resolution (100%) when doing this.
Anyway, hope this ramble gives you some peace of mind, although it might not have answered the why does RT version X.Y behave differently compared to version X.Z question.
EDIT: Clarified something.
1 Like
Thanks Jacques,
I understand what you are saying and your point is well taken. For my photo editing needs there really is no need to obsess over slight variations in the target values.
Thanks for all your support,
Mike