Additive vs Subtractive

There have been a lot of exchanges over additive and subtractive systems so I thought I would try to clarify a few things about these two methods of producing color images.

In a subtractive system, you use 3 dyes to combine and cancel all light giving black. Absence of dyes gives white. This assumes a white light source is being used as illuminant. The dyes are cyan, magenta and yellow. The dyes may be stacked one on another.

All current analog color photographic systems use this method.

In an additive system, you use 3 dyes to allow passage of all light. This assumes white light as illuminant. These dyes are red, green and blue. Something must be used to block the light to form black.

Only two analog systems that I know of used this method. One was Autochrome, and the other was Dufay. Color television uses this method as well.

These dyes must be placed side-by-side and give rise to a 'digital' look due to the 'pixellization' of the color into dots, and also it causes 'aliasing' which, in simple terms, is blocking or passing a color that should not be allowed due to position. It often causes moire patterns in objects such as a wire screen seen on television or the like.

In all of these additive systems, passage of light is limited to the brighness of the illuminant - imperfect absorption of the dyes. In TV it is limited to the fluorescence of the screen. In TV, a black is produced by absence of illumination, and in analog film systems it is produced by blocking the light with a silver image.

No analog system has ever been devised to use additive color reproduction in a reflection print material. Since additive imaging relies on transmission of light, this is an impossible or unallowed method of use due to physics.

Digital sensors use subtractive methods for producing a color image. The sensors are side by side at the present time due to limitations in making a good transparent sensor set that can be used. However, this has been accomplished and such stacked sensors are available.

Hope this helps those asking questions.

PE

Wasn't Polachrome also an additive screen system as you describe?

I am not sure on that one. I had forgotten all about it. I don't have any info either.

Thanks for that addition (no pun intended).

PE

Polachrome was one of those many regular aditive grid systems (in contrast to irregular grid systems as Autochrome or Agfacolor Old).
It was unique as it was the only instant grid system. Even instant additive system in general.

Yes, Polachrome instant was a grid film based on an additive grid. Sorry I forgot about that. I have one old roll of Polaroid One Film, and that was not and it kinda confused me. My bad! I'm really getting old.

Sorry.

PE

There were a few other additive films more or less contemporary with Autochrome. Finlay was one and I think Paget was another. They are mentioned (& illustrated) in a book on early colour photography published over 20 years ago. It's not to hand as I write but I've got it somewhere.

Yes, those did exist, but Autochrome and Dufay earned a degree of immortality if you will and that is why I omitted the others, and even forgot the Polaroid product.

They were less than footnotes in history but I appreciate your mentioning them. I have forgotten most of the minor players and again apologize.

PE

This is not to contradict anything that's been written, but there's another use of the word "additive" in color photography: It can be applied to additive color printing, in which red, green, and blue light sources are independently manipulated to produce the correct color balance in a print. This is logically equivalent, at least in a broad sense, to subtractive printing, in which cyan, magenta, and yellow filters are used to adjust the color balance (by removing red, green, and blue light, respectively, from a white light source).

The additive/subtractive distinction in this sense is basically one of the technology used to produce a colored light source used to expose the paper. The paper itself is the same in either case, and would be subtractive in the sense PE relates, even if it were exposed with an additive enlarger.

Very few additive enlargers have ever been marketed. One I know of is the Philips PCS series. It's possible to create additive exposures with other enlargers by making three exposures, one each with a red, green, and blue filter; however, this is inconvenient compared to using subtractive filters and can create problems because of small position shifts when the filters are changed.

srs5694 said:

This is not to contradict anything that's been written, but there's another use of the word "additive" in color photography: It can be applied to additive color printing, in which red, green, and blue light sources are independently manipulated to produce the correct color balance in a print. This is logically equivalent, at least in a broad sense, to subtractive printing, in which cyan, magenta, and yellow filters are used to adjust the color balance (by removing red, green, and blue light, respectively, from a white light source).

The additive/subtractive distinction in this sense is basically one of the technology used to produce a colored light source used to expose the paper. The paper itself is the same in either case, and would be subtractive in the sense PE relates, even if it were exposed with an additive enlarger.

Very few additive enlargers have ever been marketed. One I know of is the Philips PCS series. It's possible to create additive exposures with other enlargers by making three exposures, one each with a red, green, and blue filter; however, this is inconvenient compared to using subtractive filters and can create problems because of small position shifts when the filters are changed.

Click to expand...

Yep, exactly. That's what I thought this thread was about when I saw the subject title. When I started reading the thread, it blew me away. I didn't know there were actually films that used additive color mixing. Back in the 80's, I knew a guy that had done some printing with a Philips additive enlarger (at the time, he was trying to convince me that additive enlargers were far superior to subtractive).

Anyway, this chart gives a simplified visual example of the difference between additive and subtractive color mixing.

I would not even call this contradictory...
For example, if we look at a common print or a common transparency an a light table we speak of a subtractive image system though our eyes work the additive way.

But you are right, commonly the additive system is referred to as something producing images by means of three sub images (extractions/separations) in those additive colours being projected above each other the same time.
We all know these textbook illustrations of three separate slide projectors producing a white spot...

However, I always considered that Philips system of consecutively projected sub images as a kind additive system; a special kind of.


Really tricky this gets when referring to pigments (in paints etc.) and the way they produce an image. Some authors speak of hybrid systems.
Or think of a subtractive print in the actual way of printing. Might it be the offset or even the photo gravure process, those spots are never really lying over each other, thus not truly corresponding to the substractive system.

The original Kodak color printers, and the recommended printing method for Type "C" papers used additive color printing. The exposures were made using WR98, WR99 and WR70 (or 29) filters.

This was before Kodak was able to get good color separation all the time between layers and therefore additive printing was used to improve yellows. I have seen examples in which yellows were a 'pumpkin' color and then using separation filters the yellow was restored.

So, yes, additive printing can be used with subtractive materials, but they remain subtractive. Additive printing is extremely cumbersome and complex when done without the proper equipment. Since 3 exposures are involved, the image cannot move between exposures, and some exposures can be quite long.

BTW, some of the laser printers today use additive (R,G,B) printing.

PE

Hi PE
this is where I messed up with the Ciba RGB thing on the other thread, You are right the lasers on my Lambda are RGB but when we work the corrections there is a cmy and density adjustments for colour.
I worked on a Nord enlarger my last year in college and I was correcting in RGB on this unit.

Photo Engineer said:

The original Kodak color printers, and the recommended printing method for Type "C" papers used additive color printing. The exposures were made using WR98, WR99 and WR70 (or 29) filters.

This was before Kodak was able to get good color separation all the time between layers and therefore additive printing was used to improve yellows. I have seen examples in which yellows were a 'pumpkin' color and then using separation filters the yellow was restored.

So, yes, additive printing can be used with subtractive materials, but they remain subtractive. Additive printing is extremely cumbersome and complex when done without the proper equipment. Since 3 exposures are involved, the image cannot move between exposures, and some exposures can be quite long.

BTW, some of the laser printers today use additive (R,G,B) printing.

PE

Click to expand...

Burning and dodging were a nightmare in additive color printing. Imagine trying to dodge equally in three separate exposures . The only time that I have ever printed additive was while working at a place that used the Minolta additive head. They solved the problem by using rapidly alternating RG&B strobe flashes so a single dodge/burn affected all three exposures equally.
Interesting but I'll stick with the more ubiquitous subtractive method.

Having seen some of the comments on additive enlargers and what appears to be obvious problems compared to subtractive, what were the perceived advantages that manufacturers like Phillips had in mind when it decided to produce an additive enlarger?

My take on this is that additive enlargers v subtractive was not like the famous VHS v Betamax VCR contest where many say that sheer weight of numbers and marketing won when in fact Betamax was allegedly the superior system.

pentaxuser

pentaxuser said:

Having seen some of the comments on additive enlargers and what appears to be obvious problems compared to subtractive, what were the perceived advantages that manufacturers like Phillips had in mind when it decided to produce an additive enlarger?

Click to expand...

First, I'd like to make a small correction to an earlier post: The Philips PCS enlargers don't use sequential R, G, and B exposures; they've got three bulbs, each with an independent brightness control, and the light from the three bulbs is combined together so that you make a single exposure. (It's possible, though, to do it as three sequential exposures by turning on just one bulb at a time.)

There's a Dead Link Removed that's dedicated to the Philips PCS (aka "Tri-One") enlargers. They've got a file section that includes manuals and Philips marketing material from the 1980s, so you can read how Philips was marketing their enlargers yourself. Some of their claims, which I present without endorsing their accuracy, include:

• Reduced heat in the enlarger head, hence no need for fans, hence reduced vibration and less image blur from that.
• Electronic adjustment of brightness of bulbs is more accurate than mechanical adjustment of subtractive filters.
• Subtractive filters block some light they shouldn't, leading to more need to adjust the one filter value when you adjust another one, and/or more need to adjust exposure with filtration. The PCS R, G, and B filters are supposedly very narrow-band and therefore superior in this respect.
• Less knowledge of color theory required when using an additive system.

I'm sure there are more claimed advantages; I just can't remember them offhand. I'm also quite skeptical of many of them, although I'm by no means an expert on the matter. I've also restricted the above list to features of the additive system specifically as implemented in the Philips products. The Philips enlargers themselves have specific features that could be considered advantages (and disadvantages, of course) because of design features that are unrelated to the enlargers' additive nature. For instance, they've got built-in timers.

As a user of a PCS130/PCS150, I can say that I like the enlarger overall, and I've produced prints I find satisfactory (both color and B&W). I've only used a couple of enlargers, though, so it's entirely possible that I'd prefer something else if I could use it for a while. The Philips enlargers have drawbacks that Philips didn't mention, of course. For instance, they need three bulbs, and today those bulbs are expensive (about $20 apiece). Because they're out of production, getting spare parts can be hard -- but that's true of most enlargers that are no longer made.

srs5694 said:

First, I'd like to make a small correction to an earlier post: The Philips PCS enlargers don't use sequential R, G, and B exposures; ...

Click to expand...

Sorry! My memory faded...

To be a bit clearer on some of the Phillips PCS-150's claimed advantages:

There is interaction between filters on a standard dichroic head as they pass in front of each other. The PCS-150 has three lamps, independently controlled by separate potentiometers (plus an on-off switch for each channel), and so changing the level of one lamp only affected one layer of the color paper emulsion, and didn't change the other channels. I believe one other claim was that the filter passbands were better matched to the color layers of the paper. I did once make a print with three parts (like a test strip), each exposed only to blue, red, or green light, and produced three sections with the image in yellow, magenta, and cyan. I also made another print from three different negatives of silhouetted vines and leaves, each printed with only R, B, or G light to make a composite print with a white background and Y, M, and C vines that gave other colors where they overlapped.

Because of this increased efficiency in exposure from using RGB dichroics, and the three 35W Halogen lamps that passed no unfiltered white light of changing color temperature, you got longer bulb life and a cooler head that didn't require a vibration inducing fan. My exposures for 8x10 type C prints in the f:5.6 - f:8 range with the PCS-150 head on an Omega D5 enlarger ran about 10-12 seconds from 35mm negatives. That's pretty good for 105W.

Lee

Lee,

This again causes questions:

I don’t understand this `interaction´ of filters issue. Or do you just mean the staggering of filters; light coming out of one filter being filtered further by the next one?

And I don’t understand that efficiency advantage of RGB filtering, most probably related to the above. I mean with a standard color enlarger (incandescant lamp, subtractive filtering) the majority of the lamp’s energy within the visible spectrum should be filtered out when you employ all three filters. Which would not be the standard case. In the RGB case at each lamp the visible spectrum will be cut off by 2/3, thus without using the potentiometers the net outcome of these three lambs would be 1/3. When using a CYM system with just two filters totally pushed in, the outcome would be less than 1/3. Is it this you are hinting at?

And regulating the outcome of any sort of incandescant lamp effects its spectrum, thus the net effective outcome of a regulated lamp in a filter system will be different (its efficiency too) than the use of a potentiometer lets one expect. Of course this can be coped by giving each of the potentiometers a calibrated scale.

I’m not in favour of the CYM system, but as you have seen I’m not accustomed to that Philips enlarger and would try to understand it.

AgX said:

Lee,

This again causes questions:

I don’t understand this `interaction´ of filters issue. Or do you just mean the staggering of filters; light coming out of one filter being filtered further by the next one?

Click to expand...

Yes, you've got the sense of it. As one filter in a conventional enlarger slides in and out of the light path, it covers or uncovers another, changing the mix of light going through the unmanipulated filter.

AgX said:

And I don’t understand that efficiency advantage of RGB filtering, most probably related to the above. I mean with a standard color enlarger (incandescant lamp, subtractive filtering) the majority of the lamp’s energy within the visible spectrum should be filtered out when you employ all three filters. Which would not be the standard case. In the RGB case at each lamp the visible spectrum will be cut off by 2/3, thus without using the potentiometers the net outcome of these three lambs would be 1/3. When using a CYM system with just two filters totally pushed in, the outcome would be less than 1/3. Is it this you are hinting at?

Click to expand...

In a conventional CMY color head, only two filters are employed at once, not all three, and typically not close to 100% filtration, so you have significant "white" light, plus magenta and yellow light, with some of the magenta or yellow passing through both filters. With three simultaneous fully filtered RGB lamps on potentiometers, each lamp is targeted at a single emulsion layer. Yes, much more than 2/3 of the spectrum is cut off by each filter, but within the 50nm bandwidth of each filter, something like 90%-95% of the light output within that passband is targeted at the specific layer in the paper, yet doesn't affect the other two layers. It's also interesting to note that I got very long exposure times when printing with Cibachrome materials using the PCS-150 adapted to an Omega D5 condenser head.

AgX said:

And regulating the outcome of any sort of incandescant lamp effects its spectrum, thus the net effective outcome of a regulated lamp in a filter system will be different (its efficiency too) than the use of a potentiometer lets one expect. Of course this can be coped by giving each of the potentiometers a calibrated scale.

Click to expand...

The Phillips lamps are halogen, and color temp is less affected by the potentiometer setting than with straight tungsten lamps. The halogen lamps don't have the same problem with color shift over time due to evaporated tungsten deposits on the interior of the lamp. However, all this is moot, as 100% of the light is filtered into three 50nm wide passbands, so the color temp doesn't vary with the potentiometer setting. The PCS-150 does have calibrated scales, actually 2 scales on each dial, in both Kodak and Agfa standards. The PCS 2000 is calibrated in the same way.

Hope this is clear.

Lee

Lee;

Again an excellent answer.

In the absence of looking at the curves of the filters, here is another way to try to explain it.

If you look at the spectrophotometric curves, C, M and Y filters are broad and sloppy looking. You can place one on top of another to create mixed colors such as M + Y = R which inherits the sloppy broad nature of the two parents.

However, the R filter itself, such as a WR70 is very sharp and narrow. Much better at controlling the white light spectrum. Of course, you cannot stack R, G and B filters as any two give you black. So, you have to use 3 exposures or 3 bulbs and a chamber where the light can combine.

If you look at a wedge spectrogram of color paper, you will see that the spectral sensitivites are also broad and sloppy. This is what you get from the reaction of the sensitizing dye with the silver halide. By comparison of these new papers with old ones though the new papers are very sharp and narrow.

In current papers, by using CMY filtration, you almost get the result from RGB filtration 50 years ago from the first color papers due to the improvement in the emulsions. Therefore RGB filtration is not needed much now and is / was quite cumbersome.

Kodak developed a filter set which was 'anti' RGB and which could be placed into the coating to mimic use of RGB exposure with CMY filters. This material was coated and used in research for some years as a test material, but was never sold. IIRC, there were two problems. One was the expense of these two very special dyes and the other was the difficulty washing them out. They left a mild stain behind that was unacceptable.

Hope this helps more.

PE

Lee,

Thanks for your comprehensive reply.

But I take a different view on colour shifts when employing a potentiometer, which means running lamps on lower voltages than the lamp is rated for:
Halogen lamps, due to that halogen process and the resulting higher lifetime, can be run at higher voltages than the same filament would be capable in a normal bulb in vacuum or Krypton. (Or, other way round, a shorter filament could be used at the same voltage.) Both would result in an increase in light output which is over proportional due to an extension into the blue part of the spectrum. On the other hand, reducing the voltage of a lamp which is already running on lower voltage, and thus emitting in the rather red of the spectrum, will of course lead to a smaller outcome but also with less colour shift.
In such a RGB enlarger the halogen lamp responsible for the blue beam will suffer more from reducing voltage than the `red´ one. As stated before this could be overcome by calibrating the scale for that added reduction of output.

Anyway, those higher voltage halogen lamp would produce much more efficient the necessary blue light than a `plain´ incandescent lamp.
And as you said, the halogen process will ensure a constant level of output of those lamps until their very end.

Lee L said:

The PCS-150 does have calibrated scales, actually 2 scales on each dial, in both Kodak and Agfa standards. The PCS 2000 is calibrated in the same way.

Click to expand...

Actually, a small correction/elaboration: The PCS150 was released in at least two versions. (US and Europe, maybe?) One has calibration for both Kodak and Agfa standards; the other is calibrated only in Kodak units. Mine (bought secondhand in the US) has only Kodak units. Not that this is even remotely important for the main issues under discussion....

AgX said:

I would not even call this contradictory...
For example, if we look at a common print or a common transparency an a light table we speak of a subtractive image system though our eyes work the additive way.

But you are right, commonly the additive system is referred to as something producing images by means of three sub images (extractions/separations) in those additive colours being projected above each other the same time.
We all know these textbook illustrations of three separate slide projectors producing a white spot...

However, I always considered that Philips system of consecutively projected sub images as a kind additive system; a special kind of.


Really tricky this gets when referring to pigments (in paints etc.) and the way they produce an image. Some authors speak of hybrid systems.
Or think of a subtractive print in the actual way of printing. Might it be the offset or even the photo gravure process, those spots are never really lying over each other, thus not truly corresponding to the substractive system.

Click to expand...

Okay, I think I'm getting a little confused here now after the reference to printing via pigments. I spent some time in the flexographic printing industry, and I want to make sure I understand PE's initial post correctly. It really became a battle for me to uderstand the difference between additive and subtractive color after the advent of electronic pre-press, where software driving color computer displays utilizing additive methods was employed to produce the same colors that would ultimately be created on press in a subtractive system...I think, anyway. It's getting murky again now.

I don't think it's the distance between or overlap of individual dots on a pigment-printed image that dictates it's subtractive nature. Even though the dots don't often lie directly on top of one another, the pigments in the inks used to create them are in fact absorbing light and reflecting all wavelengths *except* the ones they reflect, with these reflected wavelengths producing the perceived color(s). By carefully positioning the individual dots and varying the relative number of the diffent colored dots, it's possible to create many other colors. That's why in printing via a 4-color process using CMYK inks, color separations become necessary. The separations allow the printer to control the density of a particular color in any given printed area.

I also remember the huge upsurge of hexachrome process printing, where hexachrome green and hexachrome orange were added to the 4 traditional colors, allowing printers to cover a significantly larger portion of the visible light color spectrum. The same subtractive system properties still applied, however.

Could someone please clarify? PE, perhaps you might be willing to straighten me out--for my sanity's sake! :-D It's been awhile...

Well, for one, in the printing industry they use C/M/Y/K inks which are Cyan, Magenta, Yellow and Black. The filters are R/G/B/Amber respectively for those inks. So the exposures are additive but the inks are subtractive just as they are in color photography.

The layers are sensitive to R/G/B light and the dyes are C/M/Y. It is just that the dyes in color photography are selected to not need the Amber/Black combination used in the print industry, which to some extent is dictated by both the dye set and the dot method used for printing.

All exposure is, in this sense goverend by additive filters or additive dyes.

For example, the blue sensitive layer of any analog product in use today has a yellow sensitizing dye and forms a yellow subtractive image, and so on.

I am leaving out the screen based products, now discontinued. However, reproducing the Polaroid or Dufay method is quite easy. One can do it by printing the screen as sets of R/G/B lines in a crosshatch manner on a piece of transparent film using a digital printer. Sandwich this with a piece of 4x5 sheet film, and expose in-camera and process.

Then laminate with the screen again, totally in register with the original image and you can project a full color image. I've seen it done.

PE

Someone mentioned AGFAcolor old. I've not heard about that process. Does anyone have any further info?