This database was created in 2012 and has been developed and curated by Barbara Flueckiger, professor at the Department of Film Studies, University of Zurich to provide comprehensive information about historical film color processes invented since the end 19th century including specific still photography color technologies that were their conceptual predecessors.
Timeline of Historical Film Colors is started with Barbara Flueckiger’s research at Harvard University in the framework of her project Film History Re-mastered, funded by Swiss National Science Foundation, 2011-2013.
In 2013 the University of Zurich and Swiss National Science Foundation awarded additional funding for the elaboration of this web resource. 80 financial contributors sponsored the crowdfunding campaign Database of Historical Film Colors with more than USD 11.100 in 2012. In addition, the Institute for the Performing Arts and Film, Zurich University of the Arts provided a major contribution to the development of the database. Many further persons and institutions have supported the project, see acknowledgements.
Since February 2016 the database has been redeveloped in the framework of the research project Film Colors. Technologies, Cultures, Institutions funded by a grant from Swiss National Science Foundation, see project details on SNSF grant database.
Follow the links “Access detailed information ›” to access the currently available detail pages for individual processes. These pages contain an image gallery, a short description, a bibliography of original papers and secondary sources connected to extended quotes from these sources, downloads of seminal papers and links. We are updating these detail pages on a regular basis.
In June 2015, the European Research Council awarded the prestigious Advanced Grant to Barbara Flueckiger for her new research project FilmColors. Bridging the Gap Between Technology and Aesthetics, see press release of the University of Zurich and short abstract on the university’s research database.
Subscribe to the blog to receive all the news: http://filmcolors.org/ (check out sidebar on individual entries for the “follow” button).
The development of the project started in fall 2011 with stage 1. Each stage necessitated a different financing scheme. We are now in stage 3 and are looking for additional funding by private sponsors. Please use the Stripe interface to pay conveniently online or transfer your financial contribution directly to
Account IBAN CH2509000000604877146
Account holder: Barbara Flueckiger, CH-8005 Zurich, Switzerland
SWIFT Code / BIC: P O F I C H B E X X X
Bank: PostFinance AG, Mingerstrasse 20, CH-3030 Bern, Switzerland
Clearing Nummer: 09000
Read more about the financial background of the project on filmcolors.org.
The author has exercised the greatest care in seeking all necessary permissions to publish the material on this website. Please contact the author immediately and directly should anything infringe a copyright nonetheless.
It is easy to show that the permissible errors of colour reproduction are affected by so many factors that it would be practically meaningless to lay down any tolerance values. At the same time it is not very helpful to ignore the problem, and we can at least consider some of the factors which do affect the noticeability of errors.
First, there is the familiarity with the true colour of the object being portrayed. Take the Union Jack, for example. If the hue of the red is in error by being too orange or too purple it will be objectionable; if the red is too dark or too desaturated, it will be less serious as this can correspond to a flag being dirty or faded, provided the blue and white are similarly affected. For this type of problem, the data recorded on the noticeability of colour differences across the chromaticity chart by MacAdam and by the author3 may be applied fairly directly so far as the relative noticeability of errors in different parts of the chart is concerned. However, we may note that the sensitivity to colour differences is greatest when two large areas of colour are being simultaneously-compared, with a sharp dividing line between them. If the areas are small and are widely separated either in space or time, then larger differences will pass unnoticed.
In most problems, though, we are less concerned with the accuracy of two or three clearly defined colours as in a flag, than with the realism of an object perceived as an entity although built up from a varying pattern of light, hue and saturation. In the case of the daffodil, previously discussed, it may not be important if the hue of both leaf and flower are in error, provided they are in error in the same direction; or an error of saturation of the greenest part of the leaf may not matter if the whole of the leaf is affected by an error of the same type. To specify tolerances in such a case from laboratory data on discrimination is manifestly impossible; instead, the effect of different errors on the reproduction of the whole object must be studied as a problem on its own. Obviously, this could lead to a vast programme of work, but it could no doubt be kept within reasonable limits if carried out systematically on an appropriate selection of objects.
With some types of object, it might be quite acceptable to use a process which yielded relatively large errors in the average colour, provided the fine variations of colour over the surface of the object were successfully reproduced. This would apply particularly to the reproduction of the texture of matt surfaces; in the case of the texture of skin, for instance, the chromaticity variations over the cheek may not amount to more than 0.005 or 0.01 of x or y in the C.I.E. These differences set an exacting target in the discriminating power and constancy of the photographic process.
3 D. L. MacAdam. J. Opt. Soc. Am., Vol. 32, p. 247, 1942.”
(Wright, W.D. (1948): Colour Vision and the Film Industry. In: Journal of British Kinematography, 13,1, Jul., pp. 1–13, on pp. 7–8.)
“In a three-color process, the accuracy of reproduction is greatly increased and the freedom of choice is greatly restricted” (Ball 1935), but gives no further details. Similarly Henry Willhelm, doyen of photographic conservators, devotes a chapter to “The Extraordinarily Stable Technicolor Dye-Imbibition Process” in which he remarks that “the dyes in the Technicolor imbibition process have better spectral characteristics than dyes available for current negative-positive processes” (Willhelm 1993, 355). Technicolor prints that have had a tough road life can show some slight fading of the cyan dye, for example, but Eastmancolor’s cyans fade radically and quickly.
Ball, J. A. 1935. The Technicolor Process of Three-Color Cinematography. Journal of Motion Picture Engineers XXV (2): 127–138. http://www.widescreenmuseum.com/oldcolor/ball.htm.
Wilhelm, Henry. 1993. The Permanence and Care of Color Photographs: Traditional and Digital Color Prints, Color Negatives, Slides and Motion Pictures. Grinnell, IA: Preservation Publishing Company.”
(Cubitt, Sean (2014): The Practice of Light. A Genealogy of Technologies from Prints to Pixels. Cambridge, Massachusetts, London: MIT Press, on p. 135.)
“As a mass medium, color photography relied on chromolithography, a medium that, similar to Technicolor, had a privileged relation to children’s media, a market that blossomed in the late nineteenth century. […]
The three-color Technicolor cameras introduced in May 1932 used a prism beam splitter to direct light from the aperture to two filmstrips. A third of the light passed direct from the aperture through a filter to the first panchromatic strip that recorded the green light. A further two-thirds were redirected 90 degrees by the reflective gold-flecked surface of the prism onto a bipack strip comprising a magenta filter, a blue-sensitive layer containing an orange-red dye that absorbed the blue light, allowing the remaining red light to pass through to a red-sensitive layer (Haines 1993, 21–23). Each negative was printed onto a thickly emulsion-coated stock that, when washed, became a relief “matrix” used to print to the blank release strip using Technicolor’s patent dye-transfer (or “imbibition”) processor. Manipulating the washing jets in the processor gave great control over the density of each dye. The first layer printed was a gray-scale image derived from the green negative, which enhanced shadows and enriched apparent resolution. Then, in order, yellow, cyan, and magenta dyes (the subtractive complementaries of the exposed blue, red, and green positives) were printed onto the substrate (Higgins 2007, 24–5). The Technicolor process allowed a great deal of dye to bond with the nitrate substrate by preparing it with a special mordant to anchor the dye in the emulsion. The matrices were notably long lasting; extra copies could be struck from the original negative and especially when the negative was lacquered, as it was in the British plant (Haines 1993, 32), the base itself was much less susceptible to warping and shrinkage.
The process used at Disney, from the period of The Band Concert, was subtly different: “the animation cels were filmed using the successive exposure of frames through filters of red, green and blue onto a single roll of black and white film. A step-printer could then derive the three matrices” (Higgins 2007, 26). In a slightly more technical description, Haines tells us that
on each roll of black and white negative, the animation cells were photographed on three successive frames filtered to emit the red, green and blue spectrum of color in silver densities. A step printer was then used to derive the three matrices, with each matrix exposing every third frame of the black and white negative. The red, green and blue matrices were dyed with their complementary colors – cyan, magenta and yellow – and transferred onto … blank stock. (Haines 1993, 18)
There are two issues to consider here. The first is the difference between the visible spectrum and “silver densities.” Each monochromatic layer was responsive only to values, as in any silver halide photography. Filters stopped any but the additive primaries registering: recording captured the values (brightness) in each of red, green, and blue light. The second is the nature of the dyes themselves, a topic curiously unaddressed in either Haines or Higgins. The first page of Technicolor founder Herbert Kalmus’s autobiography mentions “new dyes,” but the only other mention comes in his account of the mid-1920s: “There was still one difficulty – the tendency of the dyes to bleed, or spread sideways, during the transfer process. By this time our engineers had developed new dye formulas and improved control conditions and the problem was in the process of being solved” (Kalmus and Kalmus 1993, 56). His book offers a business and legal history of the company (with personal notes added by his widow), and does not expand on the dyes in use – most of all in achieving the saturated reds that were such a feature of the Technicolor process. A review of US patents taken out by Technicolor yields no data on dyes. Major supplier DuPont may have taken out such patents, but none of their late 1920s or early 1930s patents are expressly linked to Technicolor. It would appear that we are dealing not with legal protections but with trade secrets.
The prisms used in Technicolor cameras not only split the light but also absorbed or reflected a great deal of it. This loss of brightness required a balancing increase in illumination and set designs full of saturated colors, so manifest in Becky Sharp and other live-action Technicolor films of the 1930s (Misek 2010). The successive exposure technique used for animation could succeed because it was easy to pump sufficient light into the lens to allow for its diminution when reflected through the prism and filtered. However, as Thomas and Johnston note, there was “trouble with the cells themselves, especially when the only ones available were made of the highly flammable nitrate. One shipment would be yellow, one set would buckle, the next would warp, and all would shrink once they were cut to size” (Thomas and Johnston 1981, 279).
Disney uniquely stuck with nitrate negatives and matrices right up to the 1950s, by which time most studios had migrated to safety stock. Archivists wax lyrical about the clarity of resolution in nitrate, but it is dangerous, spontaneously combusting at very low temperatures and burning with the oxygen trapped in its own molecules, so that even tightly rolled filmstrips can blaze, as they have done in numerous theater and archive fires over the years. Working with such stuff under the intense lighting required for best-quality Technicolor must have been as risky as it was painstaking and hot.
Further problems arose when Walt Disney began to complain about the color-match between cells and film: “For years Walt battled with Technicolor to get them to give him the exact colors his artists had painted, until everyone began to realise it was the color system in the film itself that was too crude to control to such a fine degree (Thomas and Johnston 1981, 258).
One response may have been Kalmus’s “new dyes.” Another was the ingenious solution of studio color coordinator Phil Dike:
He asked Technicolor to print a scene as far to the red side as they could (what they called “out of line”); then, gradually, on successive prints, he had them come hack, one stop at a time, to normal. Then he ordered the same thing with the blue. This way he knew what he could expect, what hues were on his palette, and could work to their limitations. (Thomas and Johnston 1981, 274)
But what exactly were these colors? They were dependent on the dyes employed, and as we have seen the dyes were a source of complaint and development. J. A. Ball, then vice president and technical director of Technicolor Corp, wrote in 1935 that “in a two-color process many colors are compromised, so to speak, and there is considerable choice as to the manner and extent of compromise. In a three-color process, the accuracy of reproduction is greatly increased and the freedom of choice is greatly restricted” (Ball 1935), but gives no further details. Similarly Henry Willhelm, doyen of photographic conservators, devotes a chapter to “The Extraordinarily Stable Technicolor Dye-Imbibition Process” in which he remarks that “the dyes in the Technicolor imbibition process have better spectral characteristics than dyes available for current negative-positive processes” (Willhelm 1993, 355). Technicolor prints that have had a tough road life can show some slight fading of the cyan dye, for example, but Eastmancolor’s cyans fade radically and quickly. Willhelm (1993, 361) cites Thomas Tarr, formerly of Technicolor, suggesting that “the life of cellulose nitrate films with which the early Technicolor films were made increased as a result of treatment with the acidic dye solutions and/or the mordants used in the process.” Clearly significant for commercial exploitation of back catalogs as well as to archivists and scholars, even Willhelm does not specify the actual dyes themselves. When, in 1939, it became important to create a register of synthetic dyes, especially those used in drugs, the standard list contained 7,500 names: It is quite possible that these dyes do not appear among them.
According to the most chemically oriented account available, recipes for all three dyes incorporated acetic acid, the mordant that would bind the dye to the celluloid. Ryan suggests anthracene yellow, a coal derivative, for the yellow dye; a blend of two acid magentas and a fast red for the magenta; and a blend of pontacyl and fast acid greens for the green. But in a footnote he adds, “These dyes [are] given as an example of what could be used. There is no data as to the actual dyes used by Technicolor Corp” (Ryan 1977, 82). Wikipedia suggests that “the dye transfer process used stable Azo dyes” (Technicolor, retrieved April 12, 2011), the aryl form of azo compounds, industrial products with strong hues, especially in the yellow, orange, and red ends of the spectrum. Because azo dyes frequently change hue according to whether they are in their acid or base forms, the mordanting and developers would be critical components in the overall mix. My colleague Les Walkling reports that “the original dyes were manufactured by DuPont, and were supplied as powders that had to be mixed and boiled. Also prior to 1946 Bob Speck and Louis Cohdex invented and perfected the chemistry and dyes for the ‘new process’ and promptly joined Kodak” (e-mail correspondence, March 26, 2008). (I have also recorded from a source I have mislaid the unusual claim that three very different dyes were in use: a cyan azo dye, a selfmordanting yellow, and an aniline magenta.) In the absence of a reply from technicolor.com (“we reply within 2 working days”), and despite the elapse of over seventy years (which should end any patent period), the chemistry is still not popular knowledge, and I have to agree with Wittgenstein, whether, if Technicolor could speak, we would be able to understand it.
Chromolithography, Cartoons, and Childhood
Technicolor mattered so much to Disney because color no longer meant merely the colors of the world in daylight. It meant the extraordinary range of colors that had become available, and indeed ubiquitous, during the half-century when chromolithography was the pinnacle of commercial art.
Ball, J. A. 1935. The Technicolor Process of Three-Color Cinematography. Journal of Motion Picture Engineers XXV (2): 127–138. http://www.widescreenmuseum.com/oldcolor/ball.htm.
Haines, Richard W. 1993. Technicolor Movies: The History of Dye Transfer Printing. Jefferson, NC: McFarland.
Higgins, Scott. 2007. Harnessing the Technicolor Rainbow: Color Design in the 1930s. Austin: University of Texas Press.
Kalmus, Herbert T., with Eleanore King Kalmus. 1993. Mr. Technicolor. Absecon, NT: Magiclmage Filmbooks.