Narrowband Sat Correction Artifacts [Chroma Denoise]

[vburun]

When scanning in a narrowband RGB negative with highly saturated colors, the narrowband trichrome correction toggle under the process tab results in a very clamped down color with wormy artifacts (especially with reds):

Alternatively, turning off the narrowband trichrome correction, turning 'Dye Separation' to 1.50 and 'Separation Damping' to 1.00 results in a much cleaner, pleasing result:

[thetalkingdrum]

Interesting, could you share the source files? I'd like to take closer a look at the worm stuff.

In general I would probably try to solve this with the crosstalk matrix, finding a balance you like and save it as custom matrix you can use on all your scans. That is kind of the right tool. But if dye separation together with damping works for you that might also be fine.

[thetalkingdrum]

Also, is this tri chrome or single shot?

[vburun]

@thetalkingdrum
Here are the source files: https://we.tl/t-UnLCZfdYtcYXThkO
This is a tri-chrome scan calibrated and made with the Camera Scanning tab.

[vburun]

Okay I figured out what was causing it

I tried to solve this with the crosstalk matrix but it didn't have an effect on the artifact I'm seeing. It seemed to have something to do with the way that Narrowband correction handles oversaturated colors. Setting Dye Separation and Separation Damping to max just pulled those colors down in a different way. Changing the input ICC to sRGB also highlighted the same artifacts a little, but less so than Narrowband.

Turning down Chroma Denoise took away those artifacts. I used Chroma Denoise to counter some of the digital chroma noise that came through in the final image, but it was set too high. Narrowband simply amplified those artifacts.

That being said I still do prefer the Dye Separation hack for oversaturated colors compared to Narrowband

[thetalkingdrum]

Thanks for sharing the files! It made me find another tri chrome bug (unrelated to your issue).

Yeah, I saw the chroma de-noise artifacts. I will look into it and see if that is something I can trigger on other scans as well. The digital chroma noise you are seeing is probably just film grain though.. There is very little noise on these modern senors when shooting at low iso.

Nice that the Dye Separation / Damping is able to solve the saturation problem. That was part of the original idea behind the tools. The narrowband button is just an icc profile, so it will not be a perfect fit for every source (or taste), but I think it does an acceptable job here. You can increase dye separation slightly if you find it too muted. The red roses here are extremely saturated without any kind of damping, so some kind of fix is needed. I was able to get results I thought looked ok with eyeballing the crosstalk mixer as well.