Color separation negatives - CMYK no digital?

[gnd2]

When creating color separation negatives from transparency film, CMY is straight forward with RGB filters, CMYK is simple in the digital realm, but I got to wondering, what about CMYK without going digital? Theoretically should be possible, but is there a practical means of accomplishing this? If so, how?

[Gary Beasley]

If I remember right it is made from the CMY images stacked so just the areas common to all three is used. Also what is called undercolor subtraction can be implemented to reduce the color loading where black is printed for a cleaner look, especially when printing on a newpaper press where the blacks can look muddy with purely CMYK.

[Drew Wiley]

Depends on your output medium. Some CMY colorants will overlap to a create deep visual black; some won't, and require an additional K separation. But it was routinely done for decades before the term digital meant anything other than a hair or wart on your finger. Usually, a light yellow filter and lith film were involved; but again, depends on the specifics. Those with experience in the offset printing industry would know best.

Pigment and ink-based color printing techniques often require a supplemental black, including color carbon and carbro, as well as modern inkjet printing (which often demands more than one kind of black - hardly progress in that respect!) On the other hand, transparent dye based printing media, like dye transfer and chromogenic prints, generally do not.

What exactly do you have in mind?

[gnd2]

Mostly academic at this point, but I've been researching color carbon, but was starting to wonder why some processes like C-prints and chromes didn't seem to need a dedicated black while other processes did.

I came across a video on color carbon which used black. Then I started reading a Kodak doc on dye transfer to try and understand the calibration process, but of course it only covered CMY which got me wondering.

I was starting to think a plain BW negative from the original for black, but now that you mention it Gary, I suppose stacked CMY would be more accurate. Maybe a positive from the stacked CMY as a mask to reduce the CMY deposits where the black will be?

Of course, if there are pigments that will produce a nice black and save the extra effort that sounds ideal. I'm guessing it's not quite that simple :)

Started looking for color separation filters, but the #29 deep red and #61 deep green seem to be getting difficult to find. Lee has a 47b gell for only about $16, but only #25 and #58 to go with it :(

[Drew Wiley]

Typical matched sets would be either 25 red, 58 green, and 47 blue, or the technically more precise deeper 29 red, 61 green, and 47B blue. But yes, some of these are getting harder and harder to find, although just about anything could be custom coated at higher price if necessary. Lee poly pseudo-gels are optically inferior, if you're thinking about enlarging images through them. Balancing a set of separation negatives involves a LOT of work and film quantity up front to establish your standardized protocol, and requires a b&w transmission densitometer, along with full-range matching curve plotting. But it's also all output related; and therefore, don't expect dye transfer examples to be ideal for pigment prints. In fact, there are all kinds of variations to both. You really have to customize it to your own specific process pigment or dye set. It's a wild West out there! Some regard that as fun, some don't.

You could also see if someone like Jim Browning could do the color separations for you using a drum scanner and image-setter. That would simplify the personal battle.

A true black pigment is hard to achieve. Lampblacks trend purple - totally evident when toned down a little with white. Mineral blacks trend greenish. Mediocre black reproduction is the Achilles heel of current Inkjet print technology. If you ever get the opportunity, look at some skilled dye transfer prints with really deep shades in them - the gradation of black just goes on and on way down there, while highlight repro is generally far less satisfactory. With inkjet, it's just the opposite.

[Sam L]

Spinel black pigment is available from Kremer:
https://shop.kremerpigments.com/us/shop/pigments/47400-spinel-black.html

[gnd2]

Maybe it wasn't clear when I said "pigments that will produce a nice black", but I wasn't referring to a true black pigment. I meant finding CMY pigments that would combine to produce acceptable black, so that the black separation could be avoided altogether.

I was referred to a book that stated the black separation could be produced from the original through a yellow or salmon filter. Not really any more detail than that though, and I haven't been able to find any details on how to do the undercolor subtraction. I have another book on the way from amazon, maybe it will have something, but I'm not getting my hopes up.

[Drew Wiley]

The basic problem is that most pigments sets are too opaque to allow strictly CMY workflow if you need a deep pure black. It CAN be done in tricolor carbon - I've seen it - but even with pigments, not all are by any means created equal in permanence, especially in the case of relatively transparent watercolor pigments. Every painter knows this. The silver bullet would be nano-pigments so small that they are for all practical purposes transparent. But so far, that holy grail has been impossible to achieve technically except in transoxides (look at the surface of Mars for an example of those kinds of hues). I once had an inside track with industrial manufacturers on that same hunt, and learned a lot of the who and why of the difficulty.

Some commercially available micro-ground pigments come close, particle-size wise, but aren't entirely ideal either, color-wise. It would be almost impossible to grind something that small yourself, especially in batch to batch consistency. I just haven't personally had the time to run the tests myself, but do have a pretty good idea where I'd personally start if this were my game. And it all just depends if you need a true process set, or just something OK usable. It's a more involved topic than one might initially think. But there's also a lot of interest in it at the moment.

[neil poulsen]

. . . CMYK is simple in the digital realm, . . .

I'm not so sure about this, if one is relying on an Epson driver. These drivers require an RGB input, which it then converts to CMYK output. So, one might create a CMYK file in Photoshop. But when it gets sent the the printer (and its Epson driver), it's converted to an RGB file, and then again converted back to CMYK.

Back in the mid to late 2000's, I used a ColorBurst RIP that replaced the Epson driver. With the RIP, I was able to create a true CMYK printer out of my Espon 4000. There was an intervening CMYK to CMYK profile to preserve accurate color that converted from whatever CMYK space I was using in Photoshop to the printer's CMYK color space. But, RGB never entered the picture; it was a straight CMYK to CMYK workflow.

This was relevant for the following reason. Graphic artists would often work in the CMYK color space. Though not a graphic artist myself, I was working on a logo. Using the RIP in the way described above, I was able to realize a purity in color that I could never have achieved, had I been working in RGB.

[Pieter]

I'm not so sure about this, if one is relying on an Epson driver. These drivers require an RGB input, which it then converts to CMYK output. So, one might create a CMYK file in Photoshop. But when it gets sent the the printer (and its Epson driver), it's converted to an RGB file, and then again converted back to CMYK.

Back in the mid to late 2000's, I used a ColorBurst RIP that replaced the Epson driver. With the RIP, I was able to create a true CMYK printer out of my Espon 4000. There was an intervening CMYK to CMYK profile to preserve accurate color that converted from whatever CMYK space I was using in Photoshop to the printer's CMYK color space. But, RGB never entered the picture; it was a straight CMYK to CMYK workflow.

This was relevant for the following reason. Graphic artists would often work in the CMYK color space. Though not a graphic artist myself, I was working on a logo. Using the RIP in the way described above, I was able to realize a purity in color that I could never have achieved, had I been working in RGB.

The problem is the CMYK workspace is only a simulation. Monitors are RGB.

[Drew Wiley]

Inkjet printing is an entirely different ballgame, Neil. It does not involve process colors at all, but very complex programmable combinations of pigments, dyes, and lakes with the priority being both the abiiity to pass through those tiny printer nozzles, along with realistic programmability via analytic geometry color mapping software. One of the things I find disappointing about such inks is their lack of transparency, as well as their so-so uneven rendition of blacks.

Monitors, just one more hoop to jump through, interpolated just like Pieter explained above. Real pigments and dyes behave somewhat differently, depending on how they reflect back light.

Otherwise, graphics pre-press programs have been used in the past to facilitate post-scan modernized commercial versions of carbon printing, such as the Polaroid Permanent process, Ultrastable, and Evercolor. But all of these were also inevitably linked to the same halftone workflow, and therefore rendered an analogous halftone look, and not real transparent color either. All also proved to be expensive commercial bellyflops before too long. By way of contrast, look at competent dye transfer prints, which is also a layered or "sandwich assembly" color printing process, but with wonderful hue transparency and deep depths of pure black (but alas, less permanent color).