Best Practices for Cropping Before Converting Negative

[gabevf]

Hi all,

Got a conversion workflow question.
I tend to get mixed results when converting cropped negatives, vs a whole strip of film.
Is it a good idea to convert a whole strip of negative at a time, and include the negative space between exposures? or perhaps is it better to include some of the border in the cropped version before conversion?

I think this user is experiencing the same thing as me:

Color issue with edge case image

I’ve been using NLP for a while now but this is the first time the algorithm got confused with an image. It’s interpreting the data wrong unless I include a part of the film border. It’s been whitebalanced properly before. First image is when its cropped properly with wrong colors and the second is with some part of the film border included and more correct colors. I can send you the raw file if you need it for debugging. Is it because there is no true black in the image otherwise? …

I find that with individually-cropped negatives, NLP can sometimes get a little heavy-handed in the conversion, and give me extremely saturated colors, weird white balances, blown-out highlights, or crushed darks.

When converting multiple negatives, the conversion tends to give me more accurate and neutral colors/tones. However, the results tends to look under-exposed. I assume that’s because it’s taking the blank space between the exposed parts of the negative into account in the conversion process.

Additionally, I find that individually-converting images from a single shooting session (with similar lighting conditions, camera settings, etc.), or even syncing the scene, might give me different results, so I opt for converting a whole strip at a time so I can get a little more consistency, albeit with the aforementioned under-exposure as a result.

Any thoughts on this? Are there any best practices for the conversion process? Am I doing something wrong, or is this pretty normal?

[Digitizer]

I got weird results when I tried conversions again and again. Once I used the “unconvert” function, results got much better.

What I usually do with a new film type is to create lots of virtual copies and then apply different conversion settings. This gives me side-by side comparisons with relative ease.

So far, my preferred settings are “basic” with saturation set to 3 or lower before conversion, then I set “linear+gamma” and maybe a warmer tone. Btw, I set tolerance to 10%.

I have tried with and without cropping and find that both ways produce nice results. Anyway, I always crop borders off, so I now do it before I launching NLP. I also wb my shots as recommended. My source material are Canon CR2 raw files.

[Digitizer]

Update: images with areas that are very bright can profit from a crop that excludes these areas according to an answer Nate gave in an other post.

[Digitizer]

@Digitizer

Images with areas that are very bright can profit from a crop that excludes these areas according to an answer Nate gave in an other post Terrible Colour

I can confirm that this can help balance images. I also keep my crop tolerance at 15 instead of the default value of 5. This helped to get better images during my tests.