Are printer lights an additive method of color timing or a subtractive one?

[Owen A. Davies]

I’m pretty confused on this fact about contact printers, color timing, and printer lights, so hopefully someone will be able to shed a little light on this subject. A number of professionals on this sight have referred to the color timing process done with printer lights as a "subtractive color process”, yet I’ve read quite a few guides and manuals from both Kodak as well as Bell and Howell that explicitly says the opposite. The Bell and Howell Model C manual says that, prior to 1961, contact printers utilized a subtractive lamphouse which required the use of a vast number of ELM (Eastman Lamphouse Modification) filters if one wanted to make scene to scene color corrections. This was a very tedious and time consuming process. 

A subtractive printer uses a white light source to produce properly balanced prints. Combinations of color filters control the amounts of red, green and blue light. Printing that requires a lot of scene-to-scene color corrections is very difficult on a subtractive printer because a new color-correcting filter pack must be physically inserted between the light source and the printing aperture for every correction. Overall light changes (intensity changes) are made by using a variable aperture or a neutral density filter.

This problem was eventually solved with the advent of the additive lamphouse.

"Modern additive printers use a set of standard dichroic mirrors to separate or combine the light from a tungsten-halogen bulb into its red, green and blue components. The mirrors have a multilayer coating of dielectric material that reflects a specific wavelength region while transmitting other regions. Adjustments are made to the mirrors to achieve the desired color balance on the final print".

Over the decades, subtractive color timing was phased out and eventually went extinct, replaced entirely by the much simpler, speedier, and more direct additive model. From what I’ve seen, all of the contact printers still in use today are built with an additive lamphouse, exclusively utilizing an RGB color model as opposed to a CMY model. From searching I’ve done, I cannot seem to find a single contact printer still in use or operating in 2024 which uses a subtractive lamphouse. Yet even to my eyes, so much of the final print work that I’ve seen timed with a B&H Model C additive lamphouse bears a very strong resemblance to a more traditional subtractive color finish in terms of density, saturation, hue, and tonal rendition.

This whole revelation now seems all the more confusing to me as I first learned about printer lights through DaVinci Resolve, and that whole hotkey model operates on seven distinct levels: red, green, blue, cyan, magenta, yellow, and the master. And now I’ve come to learn that actual printer lights only utilize three, red, green, and blue (with the master being controlled by raising or lowering those three values by the exact same amount). 

Stupid Question:

Why couldn't one simply use the additive timing method with the dichroic mirrors, only with CMY taking the place of its corresponding RGB counterparts. I’m not sure if that would require a 4,800K white bulb instead of a 3,200K Tungsten one that additive lamphouses need. What is it about the subtractive timing process that absolutely requires the use of all those filters instead of simply the mirrors and deflectors? 

[David Mullen ASC]

In terms of the final result, the colors in the print (which is subtractive color), it doesn’t make a difference whether the colored light used to make the printer corrections was colored by filtering a single white light or by mixing colored lights, either way, the two pieces of film (let’s say, negative and print film in contact) are receiving a colored light that alters the color of the final print compared to if white light had been used.

It’s still a subtractive color process because you see the results by shining light through a piece of film containing YCM dyes that filter out certain wavelengths. If you had three b&w prints and shined red, green, and blue light separately through each print containing that information, and overlapped the three colored images on the screen to create a single full color image, then that would be additive color.

They are calling the printer itself additive versus subtractive depending how the colored light is generated but the color film process is subtractive.

[Mark Dunn]

I think there's a contributor here working at a co-op in France that still uses a Debrie printer with subtractive filters. Can't recall his name but you could try a search.

But I don't think you could possibly tell just from a print how the light had been filtered to make the exposure. The print stock can't tell the difference.

Brian Pritchard's website may help you

http://www.brianpritchard.com/FAOL/contents/2604200faol/Foncd/TEXTS/sect_7/analy7.html

Edited July 6 by Mark Dunn

[Simon Wyss]

Additive printing with Red, Green, and Blue is preciser than subtractive where you have filters in play that let pass broader bands, so to say. The early (and still used) additive heads of printers split RGB from an incandescent lamp. That works out narrower bandwidths allowing for the finer differentiation. Bell & Howell introduced 50 steps or points of RGB lights control.

When you look at the older colour filter systems you oftentime find Wratten filters of 0.025 log density graduation. Basically, that gives 120 steps (between log 0 and log 3) but since you have three light colours only 40 remain for each one of them. It’s even more complicated because of the side densities of the filters. Also, timers often used log 0.05 density variations. All in all about 20 steps were common.

The youngest generation of additive printer light controls is based on RGB from light emitting diodes. These can put out light as narrow as within 10 nm (± 5 nm). Still more accurate is Laser. There you are within less than a nanometer, practically just one colour.

[Gautam Valluri]

4 hours ago, Mark Dunn said:

I think there's a contributor here working at a co-op in France that still uses a Debrie printer with subtractive filters. Can't recall his name but you could try a search.

That's me and yes we used a Debrie Matipo printer and used to physically make cardboard bands and punch holes of different sizes to control the contact printer light, and even stick Kodak colour filters on it to colour-correct. We have since developed a system where we use an LED lamp in the contact printer that we can dial in different colour values to colour correct without physical filters.

[Gautam Valluri]

11 hours ago, Owen A. Davies said:

I’m pretty confused on this fact about contact printers, color timing, and printer lights, so hopefully someone will be able to shed a little light on this subject. A number of professionals on this sight have referred to the color timing process done with printer lights as a "subtractive color process”, yet I’ve read quite a few guides and manuals from both Kodak as well as Bell and Howell that explicitly says the opposite. The Bell and Howell Model C manual says that, prior to 1961, contact printers utilized a subtractive lamphouse which required the use of a vast number of ELM (Eastman Lamphouse Modification) filters if one wanted to make scene to scene color corrections. This was a very tedious and time consuming process. 

A subtractive printer uses a white light source to produce properly balanced prints. Combinations of color filters control the amounts of red, green and blue light. Printing that requires a lot of scene-to-scene color corrections is very difficult on a subtractive printer because a new color-correcting filter pack must be physically inserted between the light source and the printing aperture for every correction. Overall light changes (intensity changes) are made by using a variable aperture or a neutral density filter.

This problem was eventually solved with the advent of the additive lamphouse.

"Modern additive printers use a set of standard dichroic mirrors to separate or combine the light from a tungsten-halogen bulb into its red, green and blue components. The mirrors have a multilayer coating of dielectric material that reflects a specific wavelength region while transmitting other regions. Adjustments are made to the mirrors to achieve the desired color balance on the final print".

Over the decades, subtractive color timing was phased out and eventually went extinct, replaced entirely by the much simpler, speedier, and more direct additive model. From what I’ve seen, all of the contact printers still in use today are built with an additive lamphouse, exclusively utilizing an RGB color model as opposed to a CMY model. From searching I’ve done, I cannot seem to find a single contact printer still in use or operating in 2024 which uses a subtractive lamphouse. Yet even to my eyes, so much of the final print work that I’ve seen timed with a B&H Model C additive lamphouse bears a very strong resemblance to a more traditional subtractive color finish in terms of density, saturation, hue, and tonal rendition.

This whole revelation now seems all the more confusing to me as I first learned about printer lights through DaVinci Resolve, and that whole hotkey model operates on seven distinct levels: red, green, blue, cyan, magenta, yellow, and the master. And now I’ve come to learn that actual printer lights only utilize three, red, green, and blue (with the master being controlled by raising or lowering those three values by the exact same amount). 

Stupid Question:

Why couldn't one simply use the additive timing method with the dichroic mirrors, only with CMY taking the place of its corresponding RGB counterparts. I’m not sure if that would require a 4,800K white bulb instead of a 3,200K Tungsten one that additive lamphouses need. What is it about the subtractive timing process that absolutely requires the use of all those filters instead of simply the mirrors and deflectors? 

I think David explained your question very well above.

The way I can explain it is that anything that concerns light is additive (lamps, digital projection etc) and anything that concerns dyes (paints, film prints etc) are subtractive.

I feel whats confusing you further is when working with positive prints, you have to reverse the colour filters. For example, if you want to remove cyan from your camera negative, you'd place an orange filter on your lens but in a contact printer, to remove cyan you'd have to add a cyan filter to your image. It's because a positive film inverses the colours of a negative film.

[Willy Dharmawan]

I'd like to clarify something and hope it aligns with the poster's question.

 

I read on several posts that the median or printer points that are considered "normal", theoretically, are 25-25-25 for the respective R-G-B. And, if the negative is overexposed, the printer points would be higher than the median because the negative is denser, thus require more light to shine through to make a print. Is this correct?

 

In regard to subtractive vs additive, if (for example) the printer points used to make a print is 29-33-29, does that mean the resulting print will be more green or more magenta? I assume it would be more magenta or maybe I am misunderstanding?

 

Sorry if this sounds like a very basic question, because I don't grow up in film days and are just currently trying to understand the photochemical color timing process.

[Phil Rhodes]

12 hours ago, Gautam Valluri said:

That's me and yes we used a Debrie Matipo printer and used to physically make cardboard bands and punch holes of different sizes to control the contact printer light, and even stick Kodak colour filters on it to colour-correct. We have since developed a system where we use an LED lamp in the contact printer that we can dial in different colour values to colour correct without physical filters.

I'd love to see how you've retrofitted it with LEDs. Is that something you can go into more detail about? 

There seems to be a pretty healthy appetite for film handling services which is good to see.

[Gautam Valluri]

10 hours ago, Phil Rhodes said:

I'd love to see how you've retrofitted it with LEDs. Is that something you can go into more detail about? 

There seems to be a pretty healthy appetite for film handling services which is good to see.

Certainly! I'll ask my colleagues who worked on this system for details. Our lab's mission is free and open sharing of knowledge.

[Owen A. Davies]

On 7/6/2024 at 1:31 AM, David Mullen ASC said:

In terms of the final result, the colors in the print (which is subtractive color), it doesn’t make a difference whether the colored light used to make the printer corrections was colored by filtering a single white light or by mixing colored lights, either way, the two pieces of film (let’s say, negative and print film in contact) are receiving a colored light that alters the color of the final print compared to if white light had been used.

It’s still a subtractive color process because you see the results by shining light through a piece of film containing YCM dyes that filter out certain wavelengths. If you had three b&w prints and shined red, green, and blue light separately through each print containing that information, and overlapped the three colored images on the screen to create a single full color image, then that would be additive color.

They are calling the printer itself additive versus subtractive depending how the colored light is generated but the color film process is subtractive.

This might seem like a rather stupid question considering how concise and clear of an answer you gave to my previous question, but humor me for just a moment. In spite of the fact that the very process of printing a negative is inherently subtractive despite the RGB additive lantern, is there a way to color time your resulting print to bear a more traditionally "subtractive" color tonality?

For instance, color grading in DaVinci Resolve is also a process which operates within the RGB framework, yet in that process, there are steps that I usually take in my grade which allow me to do most of the color work by subtracting my RGB color values in either the offset, lift, gamma, or gain to yield a more subtractive color scheme from my resulting grade as opposed to simply manipulating the RGB values in an additive manner. 

There is work I have seen done with contact printers which has resulted in such a distinctly vivid and film like rendition of color subtraction that I am yet to seem done half as well through the digital process. I just saw Fear and Loathing In Las Vegas on an old worn down 35mm print at my local theatre recently. It looked unlike any color grade or version of the film I have ever seen before and I was really blown away and in love with it. However, as I have seen through other examples, simply utilizing the printer lights system alone is not usually enough to approximate this "subtractive" look that I have in my mind. 

In short, asking the question "how do you make your color look subtractive" in respect to a process which you've already clearly and helpfully confirmed is in fact subtractive does sound like a rather redundant question, but I'd still like to see if you can approximate what I am trying to get at. Thanks as always David. 

Edited July 13 by Owen A. Davies
typos

[David Mullen ASC]

As you make clear, color is a very subjective experience for the individual so I’m afraid it’s up to you to create the colors you are seeing in your mind.

If you said “I want this to looks like a dye transfer print”, all of us will have a different idea in our heads as to what that looks like, and you’d have to split-screen with dual projectors to show someone your reference versus your attempt to recreate it.

[Owen A. Davies]

On 7/13/2024 at 2:04 PM, David Mullen ASC said:

As you make clear, color is a very subjective experience for the individual so I’m afraid it’s up to you to create the colors you are seeing in your mind.

If you said “I want this to looks like a dye transfer print”, all of us will have a different idea in our heads as to what that looks like, and you’d have to split-screen with dual projectors to show someone your reference versus your attempt to recreate it.

https://share.icloud.com/photos/055l8PmtYj3WWMdHJqUj9mk3Q

These stills were all done on a Bell and Howell contact printer with printer lights. Very reminiscent of Eastmacolor to my eye. I am trying to learn more about the color timing process but there are so few still left in use and wokring condition. I found one in Brooklyn. Are there any still being used out in L.A?

[Mark Dunn]

On my monitor these look quite 40s Technicolor.

[Robert Houllahan]

On 7/14/2024 at 9:26 PM, Owen A. Davies said:

https://share.icloud.com/photos/055l8PmtYj3WWMdHJqUj9mk3Q

These stills were all done on a Bell and Howell contact printer with printer lights. Very reminiscent of Eastmacolor to my eye. I am trying to learn more about the color timing process but there are so few still left in use and wokring condition. I found one in Brooklyn. Are there any still being used out in L.A?

I have two model C Bell&Howell printers (one 35mm and one 16mm) and a Producers Service computer controlled optical printer at Cinelab in New Bedford about 45min south of Boston, c'mon down if you want to see them.

 

I made prints on the 16mm printer today.