Hey Nate,
I’m really not sure if this even is possible based on the current code source.
however coming from darkroom printing, I find that Paper reacts much more predictable than NegLabPro does. It seems like the program is a bit to sensitive to minor tone deviations in the Negative. I often find myself having to adjust the White balance by quite a lot on two negatives coming from the same strip of film, located next to each other, shot in same time, same setting, lighting etc.
Before You ask, I have tried the Match settings, and match Scene Buttons, and found the results to be rarely satisfactory.
Don’t get me wrong, NegLabPro is great and I enjoy using it.
Just maybe an option to select multiple Negs and let the engine calculate the median inversion to all of them?
Do Others have similar Issues to mine?
Kindly Yuval
I see similar things, but I’d not call them an issue - but a consequence of how NLP works.
espiacally in Portait sequences i find it difficult and therefore an Issue to match the Skintones so it does not look weird.
However I will have to check if 2.3 is maybe more consistent.
I think the idea of an average or median inversion is interesting. Not being familiar with the code of course 
Yes, I’m working on that for v3, which I’m calling “roll analysis.” To get a good result, it is a bit more complicated than averaging, but that’s the basic idea.
There are a lot more variables to control when it comes to scanning. What scanner or camera model did you use? What was the light source and the CRI of the lightsource? Was everything properly masked off and even, or are tiny differences in brightness causing color issues? Did you use the same exposure values during digital capture across the entire roll, or did you leave it on “auto-exposure”, which could create differences in exposure between similar negatives?
Because there are so many factors, Negative Lab Pro uses an evaluative model (rather than a fixed model) for inversion. This works for a broader range of negative inversions, but there are cases where it breaks down, and it’s for those cases that I’m building “roll analysis.”
But to some extent, there will still also be more variables in a digital setup than a darkroom, and thus there will be some trial and error involved in getting the workflow dialed in.
-Nate