Technical Notes
More Detailed Discussion
Dye transfer is, of course, much more involved than the above description. One obvious complication is REGISTRATION, aligning all three colors perfectly, one on top of the other. Kodak publications contain directions for hand registering films at each stage of the operation, but most dye transfer prints are made with punches and pin registration equipment made or modified by Warren Condit in Sandy Hook, Connecticut. For serious purposes, Kodak directions for making Dye Transfer prints were outdated nonsense. The plastic base of the films is dimensionally stable. Holes punched on opposite corners of the separations fit onto pins on the easel and matching pins in the carrier. Holes along one side of the matrices fit into pins on the easel where they are exposed, and there are registration pins on the rolling stone. The enlargers are bolted into stable walls of a very solid building at several points; my matrix enlarger weighs half a ton. Vibration is kept to a minimum.
I have said that three separations are made, and this is true but, in practice, a separation is but one part of a more complicated sandwich. The contrast range of the slide is in the neighborhood of 2.9 log, but the print can only carry at most 2.1 log. A doubling of the amount of light is .3. If 2.1 is the range from the darkest shadow with detail still visible to the brightest highlight with detail the light intensity is doubled seven times or seven F stops. First one must shoot black-and-white images on a special film. These PRIMARY MASKS lower the contrast. Each slide film has known imperfections in its rendering of color. These can be reduced by switching around the principle masks in known configurations during this stage. With these principle masks on top of fresh separation film, we shoot the image again through the red, green, and blue filters on to black-and-white to make the separations.
For example, an Ektachrome cyan dye absorbs 100% of red light, just as it should do, but it also absorbs 30% of green light. This means that the cyan dye also contains 30% magenta that does not belong there. If we shoot a principle mask through the red filter and develop it to 30% of the contrast or gamma that is needed for a separation, then put that on top of the raw film when the separation shot through the green filter is made (the one used to expose the magenta matrix), then everywhere cyan and magenta both appear in the image 30% of the exposure will be blocked. This way the 30% false magenta in the original cyan dye in the slide can be eliminated. This is all very difficult to follow even for people who deal with this all of the time, but it is important to get the idea. By shooting through principle masks when making the separations, the original distortions in the slide dyes can be eliminated. This purifies the colors and makes the colors in the final print closer to those in the original scene.
Using special lithographic films, several other additional sets of three black-and-white images are made. These are very light images which record only highlight detail: HIGHLIGHT MASKS. These can be placed on top of the separations during part or all of the exposure of the matrices. With these highlight masks the lighter parts of the print are brightened and the colors of the highlights can be controlled. Remember, the straight line portion of any film's exposure curve is usually too short; the highlights need a boost to more nearly simulate reality. Leaving the highlight which goes with the magenta printing separation (made with a green filter) on for a little longer percentage of the total exposure will remove pink from a white turban, for example. With highlight masks, the highlights can be controlled separately from the mid-tones and shadows.
Similarly, I occasionally also make CONTRAST REDUCING MASKS. Through both the separations and highlight masks, another set of three black-and-white images on negative film is exposed; they are then positives. By using these contrast reducing masks on part of the matrix exposure, the shadow densities can be controlled. This way one can also control the color of the shadows.
Although they were not used for printing The Afghan Folio, later on in pictures which have a great deal of important shadow detail I began using SHADOW SEPARATIONS. This is yet another set of separations which is used to add contrast to the shadows. They are given more exposure and much more development than the regular separations. By exposing a sheet of matrix film through all three shadow masks at once one can isolate the black parts of the image and print black with an additional matrix. This way I put dark black detail into shadows opened already with contrast reducing masks.
When making primary masks, separations, highlight masks, and contrast reducing masks, graduated step scales are often placed on the edge of the original image. I use abbreviated scales placed so that they show through the sprocket holes of the slide. When read with a densitometer and plotted to a graph, these step scales give a curve for each film (density on one axis and log of exposure on the other). One must juggle different development factors to get the three curves to be parallel. For example, if the separation which is exposed to make the magenta matrix has a different contrast from the other two, the print will have green shadows and pink highlights, or the reverse. The color filters have different densities and the colors stimulate black-and-white film in slightly different ways. The exposures must be of similar durations, so neutral density filters are used to balance the times. The density measurements make it possible to monitor the separations before they are used. In dye transfer printing there are so many places for something to go wrong that consistency and repeatability are crucial.
Each print depends upon at least fifteen black-and-white film sheets. A variation in any one of these, which would present little or no problem to a black-and-white printer, will make a set of separations for dye transfer unusable, and it might be difficult to discover the one that is at fault. For the consistency demanded by this process, the developers and the tray have to be within a quarter of a degree of the proper temperature. Of the three images that work together when lone matrix is exposed (separation, highlight and contrast reducing mask), each is made on a different type of black-and-white film, and each of these films has different characteristics.
The separation film (Kodak Separation Negative Film Type I) was designed to be run through roller-transport machines, so the development of this film by hand presents special problems. During the years when I was printing The Afghan Folio and Paris in Winter, I taped all three images from each set (i.e. all three highlights or all three separations, etc.) onto a large Plexiglas sheet and processed them as if they were one large sheet of film. Running a hand roller across them while they were in the developer made them process evenly as if in a roller-transport machine. Tape with a water soluble adhesive was used to hold them to the Plexiglas. A tape which was not water soluble was used to provide markers which could be felt in the dark.
The big labs that made the vast majority of Dye Transfer prints made many sets of separations every day. They purchased film sheets and chemicals in large quantities, and had workers that did only one part of the process over and over. It was possible for one print director to do this by feel; he made so many sets that he could make separations where all three colors had very close densities or at least the same contrast. They also had machines that did the development for the separations.
For an individual working alone or with one assistant, it was far more difficult to make good separations. Film and chemistry change over time, and hand processing is less consistent. I found that I averaged more than two sets of separations per image. Once I had the computer power to run a spread sheet program I began to make two sets of separations as a matter of course, one too contrasty and one too flat. Then with the spread sheet I could determine split exposures that would simulate one nearly perfect set of separations, and I could change the contrast of the separations without making them over again.
The exposure of a set of matrices is usually a complex procedure involving many exposures. The masks go on top of the separation for a short exposure, then the carrier is taken out of the enlarger and one or more masks taken away or a different one added. Then the carrier is placed back into the enlarger, and the matrix is given another short exposure. Three or four exposures for each of the three matrices is about average. For example, I might start with the same exposure for each color, and use the highlight and contrast reducing masks together with the separations for half of that exposure time. Many sets of test matrices and prints are usually required before a full-sized set of matrices is attempted, and these tests are made on a smaller size image (cut from the same emulsion as the final matrices) to reduce costs.
Each time a test print is made it is compared with the other test prints and with the exposure records. Once the results are evaluated, additional corrections are decided upon and computations with the calculator begin again for the next series of exposures. Major shifts in overall color balance and density are made by changing all of the times, or that of only one or two colors. Highlight and shadow colors and densities are altered by making changes in the exposure times for the various masks. There are only three colors, and they interact. As far as the color balance is concerned, increasing yellow is the same as reducing cyan and magenta. Every time a mask is given more or less time in the total exposure, one must take into account also the base plus fog density of the film as well as the change intended. Increases and decreases cannot be figured by merely adding or subtracting percentages of the former exposures, because light changes do not occur arithmetically. This can best be done with logarithms; precise records are essential.
Burning and dodging, the common tools used by black and white printers, are considerably more difficult to use with dye transfer, because one would have to do the same thing exactly at least three separate times. For dye transfer, dodging effects are achieved by painting with a dye called crocein scarlet on clear film sheets which are placed on top of the separations and the masks. Crocein blocks the light and can be painted where it is needed. For effects similar to burning one can cut windows from Amberlith or Rubylith, graphic arts materials which are orange or red and translucent. They block the blue light to which the matrix film is sensitive.
The developer for the matrix film is a two part pyro developer, and the parts of the developer can come together only at the last minute. Kodak gives directions for developing all three matrices together by interleaving the three continually while they are in the solutions. I have found this to be tedious and not very consistent. I develop each matrix individually on a rocker table, gently stack them in a fixer bath until all of the set are developed and fixed. Then I wash each with hot water individually. A single matrix can be remade rather than having to remake the entire set. Distilled water was used for mixing chemistry, so the fluctuations in the water would not alter the process. My distillation unit was arranged so that the hot water that came from the cooling tower could be adjusted to flow at exactly the right temperature to wash the gelatin off of the matrices. The water flowed at a certain rate and filled a measured container at a certain rate, also timing the rinses. Shortly after The Afghan Folio was completed each step in the process had been sufficiently tested to determine that the distilled water was unnecessary. The water in Middlebury, Vermont is quite clean, neutral, and consistent. I have removed the distillation unit, and my electric bills have gone down.
Rubber gloves are worn for all wet darkroom and rolling activities to keep oils from the skin from getting onto the matrices, causing tiny color spots on the final prints. Those of us who are allergic to rubber wear thin cotton gloves inside the rubber gloves. The gelatin surface of a wet matrix is soft and is ruined if touched or scratched. One learns in time to take large wet, fragile film sheets from moving trays of dye without touching the top surface.
After a series of test matrices have paved the way, a full size set of matrices is made. When a print is made, a matrix is removed from its rocking dye bath and rinsed in one or more carefully timed rinses of 1% acetic acid. It is placed in another tray of 1% acetic acid to hold while paper is placed on a granite slab which has registration pins. The paper is rolled flat with a roller and squeegeed lightly. The matrix is taken with one hand from the holding bath and held vertically over the rolling stone. The other hand pushes the end of the matrix with the registration holes carefully onto the registration pins. While still holding the matrix off the paper, the free hand then takes the roller and rolls the matrix onto the paper.
There are three solutions which are used in order to make certain corrections while a matrix is in the first acid rinse. Very small amounts of these solutions are used and they effect the three colors at different rates. Amazingly tiny amounts of calgon clear dye from the highlights, sodium acetate washes dye out evenly across the entire density range, and acetic acid in the rinse increases the contrast. The pH of the dyes can also be changed slightly to alter the contrast. Acetic acid added to the dyes lowers the pH and raises the contrast; triethanalomine does the reverse. The control solutions can also be used on part of a matrix surface by holding part of the image out of the rinse as it is rocking, and diluted solutions can be squirted with squirt bottles. A dyed matrix can be dried off with a hair dryer and painted, ever so carefully, with a soft brush and one of the diluted solutions. A matrix can be painted in the same way with the dyes. Matrices can be made on which only portions of the image are exposed, and these can be dyed with blended colors at normal strength or with dyes that are diluted with water.
After a print is rolled using the three matrices, the matrices are rinsed in a series of warm baths and reused. A set of properly cared for matrices could possibly make as many as a hundred prints, just as a properly cared for human could possibly live as many as a hundred years. Many don't. As many as ten prints can be rolled during a long work day, but that would be well in excess of the average. Most of the time twenty or more prints have been made with a set of matrices belore the final prints before to appear. I would like to be less evasive concerning how long it takes to print a picture, but when I began with the process, questions about how long and how many always provoked chuckles from professional dye printers. It all depends upon how far one is willing to go with a given picture, the characteristics of that particular image, the feel for the process which the printer develops (as well as his knowledge), and minor changes in the materials, the water and the cosmos. Dye transfer is touchy. Sometimes even the more experienced men scratch their heads and do it over again. Also, the time that it takes to actually make the picture is only part of the story. Always there are weeks of other work which must be done, like balancing new tanks of dyes, realigning an enlarger, mixing chemistry, replacing a motor, or a faucet, and this time is also a part of the business of making prints.
Prints are examined under 5000 Kelvin fluorescent lights that simulate daylight, as well as under incandescent lights. This is an important factor when working in color. A print which looks right under incandescent light can look quite green to blue under light reflected from the sky. Painters in the past worked in reflected sun light; book printers copy the picture exactly (in theory), then we sit down under an incandescent light and look at pictures which are very red. It is quite likely that Rembrandt did like warm shadows, but his work undoubtedly needed warmer shadows when viewed in natural light. I tend to use both sources, and I switch back and forth. I compromise, and to a degree I assemble sets in which all of the prints are a little cooler or a little warmer. One must strive to remain centered. When the technology is under control, for the time being, the most difficult day to day task is trying to get the balance right, and there is no point in doing this unless it is done with the proper lights.
If the original slide was ever handled with bare hands there will be tiny holes in the emulsion layer that will leave small black spots on the final print. These are particularly noticeable in large light areas where flaws like to congregate. When slides are turned over to publishers, if they are returned at all, they always return badly mauled. Separation Negative Type I film has a tendency to develop with small pin holes in the emulsion. Small specks of dust or gelatin flakes can get into the dyes even if they are frequently filtered, and these leave small imperfections on the prints. Dark spots are removed by delicately scraping the speck with a razor knife point; light spots are retouched with a tiny brush and the original dyes. Dye transfer prints can be more satisfactorily retouched than prints made by other color processes, and they have a rather durable surface. I mount prints onto a pH-neutral but non-buffered board with a pH-neutral and reversible dry mount material.
The dry mount press takes two thousand watts of electricity, as does the distillation unit. The wet darkroom and the rolling area have motor driven rocker tables. There are sodium-vapor safelights in the darkrooms which give off a wavelength of light to which the matrix film is not sensitive. The enlarger that is used for the separations takes fifteen hundred watts, and it has its own voltage regulator. The dry darkroom also has cooling and ventilation fans, a vacuum pump to hold the film flat on the easels, voltage regulators for the separation enlarger and the homemade densitometer, and other room lights: ceiling lights, viewing lights, spot lights, and light table lights. When this much electricity is used in total darkness (especially when it is often used in close proximity with many liquids). a certain care must be taken in the construction of the premises in which they are used. Most of the chemicals used in this process are, however, fairly benign, especially when compaled with a process like Cibachrome. All darkroom activities should be approached with caution, but with dye transfer the percentage of time spent with the developers and fixers is fairly short. The actual making of a dye transfer print is best done in a well-lighted and ventilated room. The rolling room may smell like a salad bar, but 1% acetic acid is merely diluted vinegar. As long as proper precautions are taken, the process poses few serious health hazards aside from stress related diseases or personal problems resulting from bankruptcy. There are good reasons why few artists print their own dye transfer prints, but for the photographer who believes in his work and is willing to go to the trouble, dye offers significant advantages and much more satisfaction over other color processes.
During the last seven years since this Technical Notes essay was first written, several changes have come to dye transfer. First, in July of 1993 a dye transfer printer from Canada called and told me that Kodak had canceled all of the materials for the process. This was the first word to reach the dye transfer printers in the United States, though we have since heard that some Kodak representatives in Europe were able to tell their clients as long as half a year in advance. I spread the word immediately, and within a very few days Kodak had ceased to take orders and had worked out a system to ration the meager supply of remaining materials.
Kodak representatives had promised us for years that we would be given at least two years warning to buy in materials and make plans when and if dye transfer were to be cancelled. When any product becomes unprofitable it is expected that the company involved will discontinue production. Kodak, however, failed to give American printers any warning. In fact they repeatedly denied that the materials were soon to be discontinued for months after the production had stopped. Large phlotographic plates for astronomical observatories were made by Kodak since the beginning of the nineteenth century. Those too were stopped without warning, leaving many projects unfinished, people with jobs and dissertations in jeopardy.
For nearly half a century dye transfer was used primarily to make copy for the advertising industry, and its demise was inevitable once sophisticated, digitalized retouching became possible. Some fine art work had been done all along, and the percentage of dye materials used for art purposes was rising toward the end of the life of the process. Still, out of the thousands of tons of paper and chemicals produced for dye transfer, all but a tiny fraction was destined for the lanldfill or sewage treatment facility within a matter of weeks of its production. The few artists who still worked with the product at the end were so insignificant that they were not even told when production ended.
Another important development of the last few years was the publication of Henry Wilhelm's classic study The Permanence and Care of Color Photographs. For years Wilhelm had been testing all of the available films and papers to determine their archival stability, and finally his long-awaited results have become available. Dye transfer prints proved to be much more stable in the light than any Type-C papers, even more permanent than Cibachromes. In dark storage dye transfer prints should show little deterioration in six hundred years if properly stored.
The most sensational revelation from Wilhelm was that Type-C color print papers from Fuji are over four times as permanent as those printed on Kodak paper. Type-C papers are the conventional color print material, the kind used when color negatives are used to make ordinary color prints. If we had had today's papers for the last two hundred years, then pictures from the Civil War done on Fuji paper would be less faded than prints from the Kennedy administration printed on Kodak paper.
The Afghan Folio exhibition was glazed with UF-3 Plexiglas, because Kodak and the archival materials companies stated as fact that color photographs would be far slower to fade if screened from ultraviolet light. Wilhelm has proved that this is not the case. Color photographs are actually slightly better off without the UV coatings; the fading is brought on by visible light. Kodak's failure either to test its materials for fading under various glazing materials or to share the results of its tests with the public cost me several thousand dollars in unnecessary expense buying UF-3 Plexiglas for The Afghan Folio.
Most of the changes that have come to dye transfer printers during the last decade are essentially a tale of the frustrations of having to deal with a huge corporation that has lost its way. At the moment several other color print processes are emerging that promise to give beautiful and permanent results, such as Evercolor and Ultrastable; none of these are Kodak. However, Kodak has done one thing in recent years for which it deserves praise. It has a new color slide film callled Lumiere that is very beautiful. The color balance is lovely, it has a fine grain, and it pushes well.
Several changes unrelated to product change have come to my work with dye transfer since I printed The Afghan Folio. I now interleaf the three or four separations in each set in a small tray, and I give them different development times. It is very difficult to make perfect separations, especially when one is not doing it every day on a large scale. Now I regularly make two sets of separations, high and low contrast. With a spread-sheet program, I determine split exposures that simulate a better single set than I can usually hope to make straight off, and I can change the contrast as needed when I make test prints. All of my separation and matrix exposures are done now on the computer. This is much faster and more accurate than before.
Recently I have replaced my Apo Rodigon enlarger lens for making separations with an Apo El Nikkor lens. This has not only provided sharper separations, especially in the red, but it has significantly reduced the flair from the blue filter.
At the present time I am making prints for a new exhibition that contains landscape images from many countries. By the year 2000 I will have exhausted all of my Kodak dye transfer materials. Though some efforts may be made to revive dye transfer, my guess is that the process is simply gone for all practical purposes. It will be remembered, however, for dye transfer prints will be our best color record of the second half of the twentieth century. And, considering what has passed for art in our time, dye transfer prints may someday be regarded as a surprisingly large percentage of the most beautiful and enduring art from our age.