I find photolab has bad color for skin and for other things too, maybe only for Nikon cameras (D850 here).
DxO colors are not their strong point. Too much agressive and unatural. White balance give sometime weird result (no way to find good balance - when nxstudio is ok - often need to go cooler than needed and then remove blue to get better colors). It is way better with neutral color rendering but still often need lot of work and is often unable too acheive desire result. Not their strong point.
1 Like
2058 and 2059 were taken with the same equipment, just seconds apart.
Why is the pic 2059 so reddish, but the pic 2058 is not?
This is going to depend on the colour temperature that you use, both in-camera and in post-processing.
I never rely on automatic white balance, preferring to use 5600°K and 0 magenta/green.
Then I adjust the temperature, either manually or, if I can find something that I know was a neutral grey/white, using the pipette.
But colour rendering is such a subjective thing, it would be difficult to always please everyone.
I have tried several different apps and they all seem to vary their rendering of colour temperature, so I don’t see how you can say one is “correct”
I appreciate everyone’s efforts to sort this issue. Seems like it should have been sorted ages ago.
1 Like
Why white balance in camera would change RAW (when white balance is adjusted manually in photolab) ?
Is photolab changing its way of creating manually adjusted white balance related to white balance metadata ?
Indeed each software has its own way to manage/create colors.
But some colors are imprinted on the subconscious from birth. This is particularly true for skin colorS. And this color doesn’t admit of any major deviations when rendering wants to be natural. And here I find photolab not very accurate (with my cameras of course, I don’t know for other ones).
But I know, tastes and colors are …
…hard for me to say. Haven’t seen such terrible skin rendering as shown in the OP yet. I’ve seen issues with (saturated) reds in exported images, but not in fair skin tones and well-exposed captures.
As for white balance, I got the impression that something has changed in the last few months. I used to be able to WB in Lightroom DNG files exported by PhotoLab of photos taken with UniWB settings. This does not work unconditionally any more and I’ve filed a ticket with support.dxo.com. We’ll (hopefully) see.
which camera model ?
for your race… I have no clue about right skin colors of other races, sorry
EOS 5D and/or 5D Mk III.
Things look like the values for Temp and Tint have shifted either in DPL or LrC with the result that the available range is now too narrow to compensate the green cast.
This is what I get with output from PhotoLab versions 2 etc., customised with “No Corrections” and exported as DNG with all corrections (AC) and optical corrections only (OC) respectively and where available. Differences in output are already noticeable without setting WB:
The cast goes away when I use the “Legacy” WCS instead of the “DxO Wide Gamut” WCS as can be seen in the white balanced photos below:
While using the legacy WCS can be used as a workaround, I don’t consider it to be a solution though. The cast is much more obvious on screen than here (second images in 2nd and 3rd rows use “Wide” WCS and have a magenta cast that cannot be compensated easily.
I was pretty sure something like this would be noticed 
1 Like
I’ll use photolab for other races, so 
it is not about using PL - it is about that I can’t say if some hue of “black” or “yellow” is right or wrong for “black” or “yellow” skin … which is totally natural and nothing racist
Me neither.
I was kidding, but I maybe I should not try in english.
I occasionally have issues with reds and oranges especially from my Q2. If I do, I go to CO. Similarly, I go to PL if I have a “troublesome” image in CO. I don’t know enough to be able to dissect why this happens because it doesn’t happen frequently enough. Of course, I could stick with either one and get good results.
Yep, made the same experience. And I have little hope here. Support is already failing because of far more mundane problems.
This document is enlightening. Recommended. Thanks Chakonari.
It is, but problem remains.
Only PL can’t process this image the right way.
Even daktable produce a usable result.
@Pat91 @anon78744791
And what would happen if there (in the shooted scene) is a “pure” green or red or blue color (say exactly the same as sensor filters colors) far more luminous than other colors ?
Should it be clipped ?
Because as I understand this is what darktable method recommend (correct me if I understood wrong).
This topic is very instructive and interesting.
PS : I understand white balance plays a role in this, but result could lead to this situation.
IMHO, that wouldn’t change anything. The reason for which a pixel is more luminous than others is not relevant.
What did you do exactly to change the camera model in NEF?
Just exiftool CameraModel? Can you provide us with the command line?
Your observation looks like one of the key points, sorry to have missed it.
The NEF you have provided looks OK in PL7.4/Win11.
1 Like