Now here’s a little headache I just ran into. Hopefully I’m just having a fundamental misunderstanding of masks or the DxO order of operations that can be cleared up.
I’m working on this photo of the (pre-renovation) Notre Dame, and there’s this one tiny window whose light bothers me, in the middle-right on the frame. I put a small Auto Mask over it, reduce highlights and saturations, and think that’s that.
What happens is that this local adjustment changes the colour tone of the entire image. This is most clearly visible around the altar, which switches from yellow to red and back again depending on whether I enable or disable this tiny little mask.
What’s even more confusing is that the effect is reversed in the preview window versus the exported image. The preview window tends to red with the mask enabled; the exported image however tends to red when it’s disabled.
Can anyone tell me what’s going on here?
Could the use of DxO Smart Lighting and ClearView be the cause of the global effects? Are they applied after local adjustments? I also know the colour fidelity of the preview window is generally not to be relied on, but a complete reversal like this just seems weird.
There are numerous possibilities. First, is the image preview magnification at 75% or greater? And do you see any difference when using the loupe tool to examine the area where you see the colors change? If the image in the loupe looks different than what you see in the preview at 75% or greater magnification, then noise reduction might be having an effect.
Does what you see in the preview match your exported image if you turn on Soft Proofing? It should - and it might be different than what you see if Soft Proofing is disabled.
If you disable your global ClearView Plus adjustment, does the problem go away? Yes, ClearView Plus can introduce color shifts and might be affected by local adjustments.
Likewise, try disabling Smart Lighting and see what happens. That is even more affected by local adjustments, but shouldn’t be causing much of a color shift if any.
Thanks @Egregius for your ideas! When looking into these, I realised a systematic investigation is almost impossible because what the preview window shows seemingly depends on how I got there, not on the actual settings.
The following video shows me testing the effect of CA corrections. Taking the steps of 1) switching off CA and 2) enabling the local adjustment, shows no influence of CA (and after that, also not of the local adjustment). Taking these same steps in reverse, using undo, does show an influence of both.
I tried to figure this out a few times throughout the day, but there is no clear answer.
First of all, I didn’t touch the disabled AutoMask for the small window (your original post), which is also not enabled in either video.
When using the history as described, I did not see that color shift on the stairs under the altar, but on the lower half of the priest’s vestment while the LAs were off.
What I also noticed is that your Master file appears a bit redder than your Virtual Copy, to which you also applied the unsharp mask.
And since the VC shows a slightly different perspective, I tried to correct this (on a copy) and interestingly, the overall color impression has shifted slightly towards reddish, although less than on the Master.
… …
Hmmm – you’ve really cranked up the color and sharpness and I suspect that such strong settings show up differences. Changing the perspective (and cropping) results in a different set of luminance and color values on which the calculation for Smart Lighting, ClearView Plus and Chromatic Aberration is based.
The brightness and color shift I noticed on the priest’s robe occurs once, but does not reoccur with further activation/deactivation of CA. Perhaps it is a temporary miscalculation, but no idea.
Thanks for looking into it, @Wolfgang. I did indeed crank up a couple of sliders in an attempt to counteract the unfortunately somewhat out-of-focus original. The issue does remain however when I switch off all those corrections. The biggest headache still comes from the fact that toggling the local adjustment seems to have an effect on previous steps in the history (when undoing), whereas applying those same settings directly does not have the same effect. This is what the videos show: the local adjustment change at the top of the history is enabling/disabling that small local adjustment, but its effect is only visible when undoing it. I think I’ll open a support ticket and see if they have any ideas.
the „somewhat out-of-focus” you tried to counter seems to be due to the critical long exposure time combined with the wide open lens and heavy perspective corrections in post-processing. And I think that’s what you can get out of the shot.
As a side effect every correction, no matter how small, becomes visible, which means living with it.
Don’t know if you have FilmPack available, what I used here. DSC_2637.NEF.dop (83,9 KB) → VC 2
Thanks again, @Wolfgang! Yes, by “out of focus” I just meant “blurry”, regardless of the actual reason. I do have FilmPack installed as well, so I may copy some of your adjustments in the DOP. I appreciate you taking the time to make a suggestion.
In fact, it would be nice if this forum had a “RAW/DOP exchange” section, where users can request such suggestions. Especially for more challenging shots, it’s always interesting to see what others make of it. But that’s a thought for another day.
I will post in this thread if/when I get a reply from support about this peculiar behaviour.
My scrutiny shows that the out of focus is more than likely motion blur. You used a speed of 1/10 sec which, on a less than sturdy tripod, is not really fast enough.
You have also over-exposed the lower row of windows, as can be seen in this screenshot from FastRawViewer…
