Roll Analysis Produces Inconsistent Colors Across Varied Lighting Conditions?

Image Editing Help
1

NLP v3.0.2
Camera: Fuji GFX 100S
Lens: Pentax 120 F4
Tethering: Fujifilm Tether App
Light: Raleno 192 Video Light

F5.6
SS 1/10th
ISO 100
EV +0.7
Manual Focus

Hi everyone,
I’ve recently been experimenting with the Roll Analysis feature in Negative Lab Pro, and while it seems promising, I’ve run into some color consistency issues.

What I’ve noticed is that the first few images on the roll often look great after conversion, but as I move further down the strip, the colors start to get weird—funky tones, unnatural tints, etc. After some trial and error, I suspect this might be because the lighting conditions varied significantly across the roll.

For example:

  • First third of the roll: shot on a sunny day
  • Second third: cloudy/misty light
  • Final third: nighttime or artificial lighting

If I run Roll Analysis across the whole roll, it seems like NLP assumes a uniform lighting scenario, which doesn’t match reality. This makes me wonder:

  • Does Roll Analysis work best only when the lighting is consistent across the roll?
  • Is the algorithm applying some kind of shared baseline that breaks when the scenes vary too much?
  • Should I be avoiding Roll Analysis in mixed-light situations and just convert each frame individually?

For now, I feel like I get better results doing individual conversions, even if it’s more time-consuming. Curious to hear if others have experienced this too, and if there are any best practices for handling mixed-light rolls.

Thanks in advance!

2

I found that roll analysis can help with camera scanned images which were taken with consistent lighting and the very same exposure settings. If one of these things change (like lighting in your examples) roll analysis can still make a difference - but not necessarily in the way we want.

My lesson learned is to try RA with series of captures taken under similar conditions, specially if individual images don’t show all colours, e.g. no reds. In these cases, I also often try with RA and without - in order to find the starting point closer to where I want it to be.

3

Yes, based on my findings as well, this is good advice.