I use PL5 for initial processing (DeepPrime, sharpening, some basic adjustments), then SEP for better control of B&W appearance.
I’m not sure what you meant by “it’s already converted”, but perhaps it’s because I was a little loose with my language when I described 850nm as “B&W straight out of the camera”. It’s not really B&W. It still has to be demosaicked - it’s just that at those high wavelengths, the red, blue, and green Bayer elements have pretty much the same response (if you are curious, I have written something up about how the different Bayer elements work at near infrared wavelengths). The result is that the image “looks” B&W.
I suspect that you and are talking past each other a little bit in terms of what behavior I’m expecting to see, so I reinstalled my old Google-era NIK package to confirm that I’m remembering the past behavior right. See this screenshot. This is the behavior that I am trying to achieve in SEP3. Notice how the selection from the monument control point picked up the sky, but that part of the sky selection was negated by the 2nd control point placed just to the right of the monument.
. Adding 1 more monument control point and a few other large “neutral” control points selects just the monument.
This is exactly the behavior that I am used to from SEP, but cannot figure out to achieve in the latest version. It also happens to be the behavior you described in your first reply to me earlier today. To me, this behavior was the most powerful aspect of NIK control points and probably the biggest reason I became a SEP user. The fact that I can’t seem to get this behavior anymore frustrates me, particularly as the change seems to have been a conscious decision on the part of DxO.
BTW, thank you for all the time you have spent trying to help with this.