Filter for scanning color negs

I have a good slide copy attachment and want to digitize some color negs. I've read much about how to do this, but almost nobody seems to compensate the orange mask with filters, choosing to do all the work in software. Why would one not filter the original scan/copy and put less burden on the post processing? Finally, can anybody suggest the optimum filter to complement the orange mask?

The orange mask is not just a background color, it is a mask to enhance the dynamic range of the film. It is composed of the unused color developer. In short it is not "orange" per se, rather "Not CMY". Where there is color, there is less masking. It is hard to imagine a filter which would pass CMY and not orange.

 

In theory, the orange mask is removed by isolating and optimizing each of the primary colors in Levels (Photoshop). This is a rather laborious process, and small errors make color balancing difficult. A better solution is to use the back end of scanning software, like Silverfast HDR, to do the conversion.

Ah! That's what I didn't understand. I thought it was a uniform mask, but apparently not. Thanks!

put less burden on the post processing

 

In any case, my scanning software has never complained about being overworked when I'm scanning CN film

:D

Anything that makes the job easier would be welcome. However the orange mask differs significantly between emulsions, even within the same brand. For example compare Fuji Reala and NPH400, or roll film and 35 mm.

 

The only shortcut I've found, as mentioned above, is to use scanning software for the conversion. HDR in Silverfast refers to an unprocessed negative scan. SilverfastHDR will process digital camera copies of negative color film in the same way it processes a raw scan made using Silverfast.

 

There are a few caveats. First, make sure the white balance of the camera is set using an empty film holder, so it is not influenced by the orange mask. Secondly, crop the negative "scans" closely to eliminate everything but the image area. "Clear" borders overwhelm the conversion process. That's also true when using a film scanner.

Thanks. As you might guess, I've never digitized CN film. Shot lots of it and even wet printed some, but that was quite a while ago. I almost wonder if I'd be better off just scanning an old 4x6 print of the same neg.

That would be okay for sharing with friends and family. However if you "scan" the negatives with a camera, you can make high quality prints as large as you like, or create an impressive slide show on an HD or UHD television.

I thought it was a uniform mask, but apparently not.

 

Conrad, it actually IS a mask, but when used it has the effect of a uniform filter so its effects CAN be countered by the use of a filter. (This is exactly what we do with the filter packs in conventional printing.) As to why people don't use such filters for "scanning," I dunno. I'd guess that the effect is relatively small with respect to the range of the "scanner," so is maybe seen as not worth the trouble? Again, I dunno for sure.

 

The basis of the mask is that, according to my understanding, a couple of the standard film dyes have unwanted absorptions in a different color range. So the very clever method to deal with this problem was to use dye couplers (one of the precursors to actual dye) in the film that started out with the same "wrong absorption," or color. The couplers that are turned into dye lose their original color, but the resulting dye has the same "wrong absorption." So the final result is that there is appears to be an overall blanket of constant filter over the entire film image. And it can probably be countered, as you suspected, with a complementary filter.

 

To estimate what the corrective filter would be, you could look at a film data sheet for the"characteristic curves." The lowest part of the density curves show what's on the "clear" film. A Kodak sheet for Portra shows D-min (minimum density) of roughly 0.20, 0.60, and 0.80 for "R" "G," and "B" (R correlates to cyan dye, G to magenta dye, and B to yellow dye). To neutralize the orange mask color, you would want to have all be equal, more or less. So, based on densitometer readings, you would want to use something like a 60cc cyan filter plus a 20cc magenta filter. Note: to put a little perspective on this, the 60cc filter attenuates to roughly a factor of 4 times, equivalent to an 2 f-stops. So the quality gains are probably not a very big deal unless you are "scanning" at a very high ISO speed. Or if your camera has very tiny pixels.

 

I carried this a little farther, and without looking closely at spectral data, it looks like this MIGHT be fairly close to a Wratten 80B or 80C color conversion filter. (Note: the Wratten 80A converts from tungsten to daylight, roughly, and these are slightly weaker than that.)

 

If you were thinking of spending money to get such a filter, it might be worth checking more carefully (the Status M densitometer readings look only at narrow spectral zones). If you have a Kodak filter handbook (pub B-3) you could take the spectral data, then sum densities with Kodak's "Spectral-Dye-Density Curves," right next to the characteristic curves (you'd need to estimate values from the "Minimum Density" curve). Then see how flat the resulting curve is.

 

Of course, if anyone goes this far, they should also be "scanning" with a daylight-balanced light source, such as a flash. Or whatever is the camera's native sensitivity.

Thanks Bill! I keep thinking about this and what it may come down to is just filtering so that the camera histogram is more or less even between the channels, thus taking advantage of the full dynamic range of all three channels. I do have a decent densitometer, but haven't used it much in this digital age. I also have a good selection of filters and probably wouldn't have to buy anything. Worth a shot when I get things set up.

Thanks, Bill! I keep thinking about this...

 

Yep, you were right from the start. The only question is whether you'll be able to see the improvement. But if you already have the filters (I had a suspicion you would), it costs almost nothing to try. From the looks of things it can buy you about two stops of sensitivity on the weakest channel.

 

To use the histogram you can probably just set up on blank film. And be sure to set the camera white balance to the most likely "native" sensor response, probably daylight or 5500K. Anything shot in non-daylight color will have offset color in the film, but the mask itself will remain a constant.

I'm trying to shoot more RAW, so white balance isn't really a thing. @digitaldog clued me in that the histogram isn't an accurate measure of what's happening with RAW either, so I'll just see what I get when I edit the files, and go back and correct if necessary.

I've been working on my process for color negs with a light table and a macro lens. I have a Kaiser Slimlite Pro, an X-Pro2 with the 60mm macro, a tripod with the center column flipped and a Digitaliza film holder. This isn't perfect, but what I've figured out for getting a good base to work from is:

 

Import to Capture One. Use the ICC Generic profile for my camera and Linear Response curve. Use the White Balance picker on blank film outside the frame. Crop within the frame to clean up the histogram.

 

Go to the Levels tool and pull the lower left slider to the right side of the usable part of the histogram and the lower right slider to the left side. This inverts the negative. Make finer adjustments to put the white point and black point where I want them.

 

Now I can freely use the crop and rotate tools to get the frame I want (I like to include some film border).

 

The process for B&W is the same, but I turn on "Enable Black & White" and don't worry about the white balance.

 

If you use Lightroom, you can do something similar except you don't have a linear option and instead of levels you flip the diagonal line in the curves and adjust the endpoints to the desired black and white points.

I'm trying to shoot more RAW, so white balance isn't really a thing.

True, but misleading. You can assign any white balance to a raw file, but you have to know what it is without regard to the orange mask. You could write it down on a sticky note, or measure it with the camera and record it in EXIF data and apply it automatically when opening the RAW file.