Nikonscan 4.0.3 C41 inversions and RA4 paper colour signature question

[warden]

But thanks for chipping in, and I agree with you - this was useful if anything to revisit the humble, built-in color inversion performed by Vuescan, which does a really really good job. Kudos to Mr Hamrick!

I’ve been enjoying this thread so thanks for making it and sharing your tests!

 

[Steven Lee]

err... that's not how it works. You should have detected and reported the 'heavy color casts' before I revealed the labels. Your judgement is now influenced by any implicit bias linked to you now knowing the labels

Heh. Sorry I missed the part where I was supposed to guess the tool, I was simply selecting the most pleasing/natural image. Looking back, my pre-reveal picks were Vuescan and ColorPerfect, consistent with my post-reveal conclusions.

 

[radialMelt]

I've been experimenting more with ColorPerfect since you revealed the results, albireo. To anyone else who also wants to try it out but is scared off by the UI, check out these couple of videos:





They recommend foregoing any of the color correction stuff, simply setting white/black points to maximize data, and then moving to PS or LR. It's working well for me so far, putting out very flexible positives, even on some very challenging (poorly processed) c-41 negs of mine. Comparing these next to files I inverted with SmartConvert revealed that I have been far too haphazard with clipping when using it.

 

[albireo]

I’ve been enjoying this thread so thanks for making it and sharing your tests!

Thank you Warden, glad to hear.

 

[albireo]

I've been experimenting more with ColorPerfect since you revealed the results, albireo. To anyone else who also wants to try it out but is scared off by the UI, check out these couple of videos:





They recommend foregoing any of the color correction stuff, simply setting white/black points to maximize data, and then moving to PS or LR. It's working well for me so far, putting out very flexible positives, even on some very challenging (poorly processed) c-41 negs of mine. Comparing these next to files I inverted with SmartConvert revealed that I have been far too haphazard with clipping when using it.

Nice. I think this guy's tutorials are better than those made by Colorperfect's author actually. Nice find.

 

[radialMelt]

Well I decided to purchase ColorPerfect and put it to task re-inverting some older, challenging neg scans of mine. While it definitely isn't a silver bullet, it seems that the workflow outlined in the videos above - paying close attention to clipping and color space - is giving me significantly more flexible and clean files to further color correct in LR. This is probably obvious to most. In my case, it has highlighted some serious errors I have been making in my color inversions up to this point. And yeah, all of my SmartConvert stuff is completely clipped in the highlights. Crazy I didn't notice it before.

 

[Tom Kershaw]

This is probably obvious to most. In my case, it has highlighted some serious errors I have been making in my color inversions up to this point. And yeah, all of my SmartConvert stuff is completely clipped in the highlights. Crazy I didn't notice it before.

- In my experience colour inversions are something that one needs iterative experience with over time to begin to understand what is going on.

 

[radialMelt]

- In my experience colour inversions are something that one needs iterative experience with over time to begin to understand what is going on.

Very true! I've been doing it for a few years now and constantly learning new things. Moving forward paying much closer attention to clipping and color spaces seems to be a rather big discovery for me.

I don't want to derail the thread entirely, but on the topic of scanning... I'm noticing some of my scans have unrecoverable, clipped blacks even before being processed. For example, loading into ColorPerfect and adjusting the black point (aka "BP Tails"), a certain amount of pixels are unrecoverable/clipped. My thinking is this could be two things:

- severely underexposed negs. In my case I'm scanning Vision3 250D that was shot at ISO 400. I don't think it is that underexposed. Processing errors are a real possibility as well.
- scanning error? too much gain blasting out the areas of little density?

At the office right now but I can provide an example linear .tiff when I get home should anyone be curious. All input is welcome!

 

[koraks]

I'm noticing some of my scans have unrecoverable, clipped blacks even before being processed.

Scan as a positive and invert manually. You'll learn some more about what's in your negative. You can then always proceed to whatever convenient way of color balancing, but at least you'll understand what the possibilities are. Spoiler: blacks never look "clipped" on properly scanned color negative film (not even if it's underexposed). The shadows tend to slope off quite gently into black. If you see a hard cut, it's virtually guaranteed to be a scanning error.

 

[radialMelt]

Scan as a positive and invert manually. You'll learn some more about what's in your negative. You can then always proceed to whatever convenient way of color balancing, but at least you'll understand what the possibilities are. Spoiler: blacks never look "clipped" on properly scanned color negative film (not even if it's underexposed). The shadows tend to slope off quite gently into black. If you see a hard cut, it's virtually guaranteed to be a scanning error.

Sorry, to be clear, I am scanning as a positive and inverting later (currently with ColorPerfect).

I have noticed that in ColorPerfect some negatives have fully recoverable blacks (no clipping reported), whereas with others (particularly this roll of 250D) ColorPerfect is reporting some amount unrecoverable, clipped blacks.

And this is all just based on the histogram/reported data, not a qualitative analysis of the image itself.

 

[albireo]

Find your own thread, you two! Get off my lawn!

@radialMelt glad you decided to experiment with Colorperfect.

One thing I commented on and that it's worth exploring further is the color profiles in the tool: some seem to be spot on (many Kodak mainstream film ones IME) but others are way off, almost as if they're based on old data.

Anyhow, let's make a Colorperfect thread with examples/lessons learnt!

 

[koraks]

Alright, then I'd invert manually and see what you got. Or not invert at all and just inspect the low density areas and see how they look. This will help to rule out some oddity with this particular app.

 

[Tom Kershaw]

Kodak Ultramax 400. Top: CP, Bottom: NS. In this instance I prefer the CP interpretation with the more naturalistic and slightly less saturated colour palette.

NikonScan / ColorPerfect comparison 4

 

[radialMelt]

Thanks for sharing! I would say the CP interpretation is a bit green for my tastes, but I think that speaks to a bigger picture message - you can pretty much get to any result you want with these tools. It comes down to preferred workflow, I think.

I have re-inverted a few tricky rolls of mine using CP and am really liking the results. The workflow is a bit clunky, but the data/information it gives regarding clipping and colorspace has been a game changer for me. I'm also finding the WB to be more consistent from frame to frame than the other solutions I've used.

Thanks again for the great discussion!

 

[Steven Lee]

And yeah, all of my SmartConvert stuff is completely clipped in the highlights. Crazy I didn't notice it before.

Thank you for pointing this out. I too have re-visited my SmartConvert results and indeed, highlight clipping is a problem. I am puzzled by the author not seeing it and asking for samples. It clips highlights every time with 100% reliability. All he needs to do is to compare the output to a manually inverted image.

By the way, this may be intentional. The clipping is perfectly consistent with Noritsu/Frontier results, at least that's what all labs I've tried gave me. Mimicking lab scanners, optimizing for speed+volume and not having manual white point control is the stated design goal of SmartConvert. I doubt the lomo-crowd educated by youtube even cares.

 

[radialMelt]

Mimicking lab scanners, optimizing for speed+volume and not having manual white point control is the stated design goal of SmartConvert.

I love the controls and speed of operation, and UI in general. But sadly the output and lack of data (e.g., histogram) are deal breakers for me. Kinda wish I could get a refund...

 

[Tom Kershaw]

The clipping is perfectly consistent with Noritsu/Frontier results, at least that's what all labs I've tried gave me. Mimicking lab scanners, optimizing for speed+volume and not having manual white point control is the stated design goal of SmartConvert.

I've always scanned my film on the Nikon Coolscan, so don't have any Noritsu or Frontier results to compare. However, I've often thought that the issue with film processing as a service is in the scanning stage in terms of available equipment and technology.