Camera "Scanning" -- Use Di-Chro Head to undo Color Mask?

[ColSebastianMoran]

Looking for feedback:

Camera scanning of color negatives. If you use white light, the digital camera image of the negative is very orange, the red channel clips first, and the Gr/Bl channels have condensed histograms. The opposite of ETTR.

So, I started thinking, can I remove the orange mask by using my Di-Chro head to fully offset the orange mask?

Would like feedback on the idea. Does this make sense? Drawbacks? Pitfalls to avoid?

I'll add details below, but that's the essential question: Digital camera scan, Di-Chro head to add the opposite color so the orange mask disappears?

 

[ColSebastianMoran]

In another thread, Huss offered this tutorial for Camera-Scanning of color neg material.

http://www.mfphotography.ca/michael...-guide-to-scanning-film-with-a-digital-camera

He's using a light pad, daylight balanced white light.

 

[ColSebastianMoran]

I set the Di-Cho filters on my Beseler Dual-Mode Duplication for +50 Cyan and +15 Magenta, and I expose with flash, not the incandescent bulb.

This gives me a camera capture in which the orange mask is about neutralized and my histograms are all broad, the WB comes from an area of unexposed film. Notice my WB is 4350K and there is little tint.

Next, I set the blacks and white in my RAW conversion (Lightroom) to expand the histogram both ways staying well clear from any clipping.

 

[ColSebastianMoran]

Next, I move to Photoshop, apply the Color Neg action from Huss's reference (above).

This gives a pretty good color image.

 

[ColSebastianMoran]

The image above is not quite right, but with a little Lightroom tweaking, I have an image that's better than the mini-lab print from 2010. Quite a bit better.

The negative is from a test of a then-new-to-me Yashica GX on Fuji 200 film in 2010.

 

[ColSebastianMoran]

So, anyone have any thoughts on using Cyan/Magenta light from a Di-Chro head to offset the orange mask?

 

[x-ray]

I'm impressed.

 

[Ranchu]

Yes, it looks good, thank you for the report as I was curious about this. I had read that the dichro heads weren't strong enough filtered to counteract the orange mask, but the flash seems to have done it. Good news!

 

[johannielscom]

H, I might have to go scrounge the right parts for a set up like this, for that future moment where my old little Imacon kicks the bucket... 🙄

 

[sevo]

Can't be done literally - it is a mask, not a constant stain, so applying a complimentary constant will not make it vanish. What you can however do is to apply the complimentary colour at the average density of the mask to shift the white point to a more manageable value. But you will still have to undo the actual mask (which presumably is what the mentioned action does).

 

[brbo]

How on earth then did they do optical C-41 prints back then?

 

[ColSebastianMoran]

How on earth then did they do optical C-41 prints back then?

And mini-labs?

I think the mini-lab manufacturers did calibrations for each of the film types. There was enough revenue and they had the resources.

I don't know. How did anyone ever make a C-41 print with good (accurate? believable?) color in a darkroom setting.

 

[ColSebastianMoran]

Can't be done literally - it is a mask, not a constant stain, so applying a complimentary constant will not make it vanish. What you can however do is to apply the complimentary colour at the average density of the mask to shift the white point to a more manageable value. But you will still have to undo the actual mask (which presumably is what the mentioned action does).

@sevo, thanks for this thought. Not sure what to make of it.

Yes, the complimentary cyan/magenta shifts the white point to around 4400K, and just a tiny bit of tint, with flash illumination. That's certainly manageable.

The mentioned action? No, it's not doing anything sophisticated. Here is the heart of the action, auto, find dark and light colors, snap neutral midtones. See the resulting curves adjustment below. I'm amazed that this gives the reasonably good auto result, the first positive image above.

 

[brbo]

Bigger question is how did anyone ever make a C-41 print with good (accurate? believable?) color in a darkroom setting.

Well, I believe that they did it just like you did (for scanning). If "canceling out" orange mask still left some complex colour mask there would be VERY VERY few optical prints made at all.

But since I've never done an optical colour print, I'll wait for sevo's explanation...

 

[ColSebastianMoran]

Doing a little research on sego's tip finds this interesting article.

Yes, it's a mask. Yes, the intensity of the orange at any point depends on the exposure of the magenta and cyan layers at that point. Here's a reference with a clean technical explanation.

http://www.brianpritchard.com/why_colour_negative_is_orange.htm

Quoting from the article, "the amount of mask will be inversely proportional to the amount of dye formed. That is, if no dye is present then maximum mask is present and in the areas of maximum dye formation there will be no mask."

Wow. Sounds troublesome. So rebalanced light works on average but not exactly for other negative densities. How big is the difference? Other routes to address it in a camera-scan file from an RGB sensor? I'm rebalancing based on the unexposed area, the densest mask. Should I rebalance on a midtown grey instead? Mind boggles.

Will not anyone rid me of this meddlesome orange mask!

Or, more seriously, won't anyone create software for properly processing camera-scan files from color negative material?

My conclusions so far:
- These brute force approaches are approximations at best
- Keep the camera-scan RAW file and probably the negatives for important color-neg images
- For less serious hots, a camera-scan gives an image that's pretty good and better than the old mini-lab prints. For these snapshots, keep the JPEG and discard the prints, negatives, and RAW files.

 

[sojournerphoto]

The mask is the challenge. There is a photoshop action that does calculations somewhere in the moving picture hinterlands of the internet that I mean to adapt one day. Hopefullly it can be used with some sensible filtration to maximize the data recorded in each channel too.

 

[ColSebastianMoran]

And, here is someone who purports to have a rigorous technical solution to removing the varying mask. It's re-posting in 2000 from an earlier article at Adobe Forums by Burton.

I haven't tried this and might not.

https://www.dpreview.com/forums/post/499854

 

[Godfrey]

I've not captured any color neg for quite a long time. But...

I use DNG Profile Editor to create customized camera calibration profiles for use in copy scanning both B&W and color negatives. For a color negative, I find the most effect way to remove the orange crossover mask is to use a negative image which has a reasonable amount of unexposed, processed rebate to set up the right profile constants. The crossover mask is consistent through the negative film so images processed with the custom CCP invert to a proper RGB positive which reflects the actual color temperature of the light captured.

Someone asked if anyone had ever produced 'accurate' color prints from negatives. Of course we did: we used reference negatives and a densitometer to calibrate the C41 processing machines. This process also establishes the correct filter pack settings for the enlarger and printing machines; it guarantees color accuracy when done correctly. I'm kind of shocked that film aficionados on this forum would not understand that part of the professional darkroom workflow.

G

 

[brbo]

Someone asked if anyone had ever produced 'accurate' color prints from negatives. Of course we did: we used reference negatives and a densitometer to calibrate the C41 processing machines. This process also establishes the correct filter pack settings for the enlarger and printing machines; it guarantees color accuracy when done correctly. I'm kind of shocked that film aficionados on this forum would not understand that part of the professional darkroom workflow.

And yet you haven't even touched on the question how that workflow (totally) eliminates colour mask. Do you know the answer?

 

[ColSebastianMoran]

I use DNG Profile Editor to create customized camera calibration profiles for use in copy scanning both B&W and color negatives. For a color negative, I find the most effect way to remove the orange crossover mask is to use a negative image which has a reasonable amount of unexposed, processed rebate to set up the right profile constants. The crossover mask is consistent through the negative film so images processed with the custom CCP invert to a proper RGB positive which reflects the actual color temperature of the light captured.

Godfrey, I understand you sample the unexposed area of film. What do you do with that sample other than set White Balance? And, is your CPP applied to the negative image, before inverting? Or to the positive after inverting?

Someone asked if anyone had ever produced 'accurate' color prints from negatives. ... snip ...

My question was never whether. All of us have seen great color prints. More like amazement at how great prints were made optically in wet darkroom, given the complexity of the orange mask.