RGB separations for gum bichromate?

In my dilettante way I've been looking into gum printing in 3 or 4 colours.
I've produced CMYK seps using digital means but it occurred to me that the original gummists were more likely to have used RGB filters in camera to produce their negatives as I guess colour photography and electronic methods were unavailable.
Any current reference to colour printing flags up dire warnings about using RGB.
So, did early practitioners have CMY filters and if not what colours did they print through each of their separated RGB negs to reproduce anything resembling true colours?
I'm sure this is basic stuff but if the idea is to print complementary colours through each separation where did blue printing come from without a yellow filter?
No wonder I've largely kept to black and white.
Basically what I want to know is did the Victorians attempt colour and is there anywhere I can find information about how they managed it. I've tried the usual searches without success.
I understand autochromes used RGB dyed starch grains and I have a couple of autochromes that look ok colour-wise but I suppose they were transparencies.
I suppose hand tinting was popular because of the problems but surely someone must have tried alternatives?

BJC0000 said:

did the Victorians attempt colour

Click to expand...

Yes, but not a whole lot has survived and most of it isn't very compelling. So it's not directly the area I'd look into if you're interested in striking or realistic colors, or even much color to begin with. Nevertheless, it's an interesting question that should have considerable artistic potential today.

BJC0000 said:

it occurred to me that the original gummists were more likely to have used RGB filters in camera to produce their negatives

Click to expand...

Do you realize that when you talk about RGB filters during exposure (in-camera separations) and CMY separations in digital space, it's really the same thing? The former yields 3 negatives that are ideally printed with the complementary colors of the original filters - so CMY. In the latter case, you make separations also aimed at the same subtractive primaries.

Now, the question which actual colors people used 'back in the day' may give some interesting answers, since the whole CMY concept as we know it today went through its own evolution. An interesting and recent testimony of this is @Katayounpd's recent work carbon printing M. Yevonde's color photos. In doing so, she went down the rabbit hole of getting the colors period-correct and found out that the pigments used back then strayed quite far from what we today would consider a common CMY set. But this is still within the for us familiar concept of CMY subtractive printing.

Other approaches have been around; see e.g. early color work in the motion picture industry, which achieved (crude) color reproduction using only two primaries (red and green). Undoubtedly, there's documentation about (reflective) printing processes that experimented along similar lines. However, it appears that as soon as the photographic community figured out the tandem of RGB + CMY and (especially) the technical means with which to make this work, the de facto standard established itself pretty swiftly. Only fairly recent developments associated with digital printing seem to have really rocked the boat at a large scale, although the skeleton approach in modern printing still remains based on CMY.

RGB Tricolor photography was the earliest form of predictable color imaging. There were various printmaking options. Autochrome plates were one of the earlier commercial processes. Some lovely work was done, but more in the "Pictorial" mode popular in those days. You're really in the Edwardian era, however. Actual color photography wasn't possible until panchromatic film was invented. Prior to that, a handful of color Daguerrotypes actually "happened", but no one knows quite how.

What was possible quite soon were stunning color slide shows using three aligned carbon arc projectors for sake of RGB exposed pan glass plates. But hot and slow would have also described that experience, long before air conditioned theaters.

There are entire website dedicated to past methods of color photography and even the revival of numerous of those. "Gum" printing has undergone numerous reiterations. But generally speaking, it's not an ideal path for realistic color.

Koraks, Thanks for your usual comprehensive reply.
I do appreciate that printing is based on subtractive colour and light is additive colour and that the filters RGB produced negatives which were printed in complementary colours and therefore the red filter is printed to green, the green to red and the blue to yellow - but what produces blue colour in the print? I can't see any combination of green, red and yellow producing blue (whereas I can understand cyan, yellow and magenta producing a realistic spectrum)
I'm not interested in producing 'realistic' colour renditions - just in reproducing the effects early users would have achieved.
Of course early cinema work would have been based on projected light so RG and later B would have worked - right back to Maxwell's ribbon.
I think Drew Wiley has made the point that none of these techniques would work without panchromatic film - no point in putting a red filter over colour-blind or orthochromatic film.
Thanks again for taking the time and any suggestions for further investigation?

Drew. Thanks also for the reply and I appreciate the point about needing panchromatic film for separations - some of the illustrations in The Light Farm illustrate this really well although not really intended to.
I don't have a problem with the 'Pictorial' school - actually I quite like it! However I've clearly missed examples of early colour (with the exception of Autochrome) and if you could point me in the direction of any websites I'd be grateful. I assumed coloured Daguerrotypes were just hand tinted?

Barry

The "accidental" color Daguerrotypes were probably due to some unknown chemical contamination; and their makers were unable to repeat the results. Since the examples of these are so extremely rare, nobody is likely to use any kind of destructive method of analysis as per the ingredients just to answer an arcane question. By contrast, hand-tinted examples are abundant. It was quite common to dub in the eyes with a paintbrush anyway, since the long exposures risked a lot of blinking.

Red, green, and blue are primaries. Cyan, magenta, and yellow are their respective complements, or opposites across the color wheel. The RGB primary filters are used for color separations, while CMY is used for the printing dyes or pigments, as the case may be (plus sometimes a "K" black printer). Modern inkjet is more complex, and modeled after architectural paint pigment programs, based on 4-axis CIE color mapping along with applied analytic geometry.

With gum printing you can choose from all kinds of artists pigments, and simply gradually layer them up any way you wish.
Early forms of color carbon printing and Fresson etc typically used alizarin crimson red, cobalt green, and cadmium yellow, which never were ideal in terms of either color repro or health implications.

Projected light (or backlit displays) are a somewhat different subject. For example, I use both true additive RGB enlargers, as well as traditional YMC subtractive colorheads.

BJC0000 said:

but what produces blue colour in the print?

Click to expand...

Generally cyan + magenta.

BJC0000 said:

I can't see any combination of green, red and yellow producing blue

Click to expand...

Well, for starters, RGB stands for Red Green Blue, so there's your blue.
Anyway, as we've established, RGB's not what's used in the print if you use RGB filters to make in-camera separateions. RGB filters during exposure = CMY pigments during printing.
Nevertheless, if you look at the work of Katayoun (see eg. here) you'll notice that the magenta is in fact quite red and the cyan is kind of muted, and so is the yellow. By today's standards, we might not considered that a 'proper' CMY pigment set. Still, it produced a wide gamut of hues - just not as wide as in the case where higher-chroma pigments would have been used.
If you think about this a little more, you'll realize that there's a 'tuning' that needs to take place between the transmission spectrum of the RGB taking filters and the reflectance spectra of the CMY pigments. In practice, this perfect fit would not have been the case in early work relying on in-camera separations; the cyan would not have been the perfect complement to red, etc. This creates particularities in how hues are rendered, and it's the reason why Katayoun spent so much time trying to figure out which pigments would have been used for 1930s Vivex prints - with modern pigments, she realized she didn't get the colors Yevonde created her work for/with!

If you want to understand better how mixing pigments works, I can recommend reading Bruce MacEvoy's excellent website on the topic. It's written for watercolorists, but the general principles apply to any form of color printmaking just the same. Start e.g. here: https://handprint.com/HP/WCL/color14.html
You'll notice that in a way, you could start with any set of pigments to mix your way towards a full gamut. There are benefits and drawbacks to whatever combination you'd choose. Note that this is also true for CMY, which is relatively poor at rendering orange, green and violet. The main reason we don't see it that way (literally) much of the time is that human vision adjusts quite flexibly to the gamut we're being offered. Only if we have a reference at hand that's more saturated, we'll notice the difference. Note also that the magenta's we generally use in (commercial) printing are in fact quite red, and the cyans are often quite blue (but not always). They're not fixed points on the color circle!
And yes, you could in fact make a workable palette using red, green and blue pigments. The 'saturation cost' (see MacEvoy's website) will be in the secondaries; i.e. it'll poorly render yellows, magentas and cyans. Of course, the separation negatives would have to be tailored towards this uncommon approach. How successful the palette would work, would depend on the mixing behavior of the pigments involved, which in turn relies heavily on their absorption spectra and transparency. It seems that only fairly recently, suitable pigments in the green, violet and orange ranges of the spectrum have become commonplace and hence, they are now being widely used in digital printing (think of expanded gamut CMYKOGV printing). Historically, to the best of my knowledge, green has always been a difficult color in terms of pigments with a relative lack of them altogether, while oranges tend to be low in chroma and violets are often both low in chroma and not particularly lightfast (I'm taking big & sloppy strides through a large landscape here, mind you).

BJC0000 said:

just in reproducing the effects early users would have achieved.

Click to expand...

They would have been all over the place, so I'd feel comfortable suggesting to just experiment happily away!

Thanks again to both.
Drew Wiley wrote "Red, green, and blue are primaries. Cyan, magenta, and yellow are their respective complements." That explains a lot - no art background explains the rest.
I've printed out the sites referred to - I'm still happier sitting reading paper, and I can add notes in the margin as I go along - so off to read and re-read.
My wife has recently returned to watercolours after a 40+ year gap (life gets in the way) so she may take an interest in these articles.
I meant it about being dilettante - I tend to flit from subject to subject as the whim takes me but usually return to the subject later - albeit with sometimes a long time gap.
I'm not trying to repeat old work or contribute to academic knowledge - just having fun and satisfying my curiosity.