The first offset lithographic press for printing on paper was invented by Ira Rubel in 1903. Yet, offset lithography did not overtake letterpress in publications printing until the second half of the 20th century. It took many decades for the industry to fully understand offset printing technology, such as the chemistry of ink-water balance, plate coatings, and color reproduction. Offset lithography in the early 20th century did not seem like it would be a threat to the letterpress. Stephen Horgan wrote in the Inland Printer in 1922, citing the The British Printer, “Craftsmen need not feel that [three- or four-color offset lithography] is going to replace three and four color typographic printing, nor will it wholly replace the lithographic stone press or the lithographic direct rotary.”1 Printers at the time could not have foreseen the future of printing. Research into the science of lithography by organizations, such as the Lithographic Technical Foundation (LTF), and the wide-spread availability of technologies such as multi-function scanning systems and photo type-setting, gave offset lithography the market boost it needed. Among the important innovations in offset lithography were techniques for the accurate and efficient reproduction of colors.
Manual Methods
In the late 19th century and early-to-mid 20th century, the color reproduction process began with the creation of photographic color separations and screen negatives. An object was photographed through three color separation filters—red, green, and blue—to create separation negatives. Each negative was then copied to another negative in a special camera with a crossline screen placed a short distance from the second negative, creating a screen positive. Finally, the screen positive was copied to another negative, creating a screen negative, referred to as the “printer.” In four color printing, you would end up with cyan printer, magenta printer, yellow printer, and black printer negatives used to produce plates for those inks. Though variations on this process were used, this was how screen negatives were created. An illustration of the three-color printing process is shown below.
The process of creating letterpress plates was labor intensive. The photosensitive plate was exposed through the screen negative and etched to remove the areas that did not receive exposure. To achieve the desired shape and depth of highlight dots, the etching process had to be repeated multiple times, called staging and etching. According to Mertle and Monsen,
The reason for [manual etching] becomes clear when it is realized that a relief etching has to have a certain depth in order to print properly, or to be reproduced by electrotyping. In order to reach this depth, the actual surface of each printing dot also is reduced, and this has to be taken into consideration when making the negatives…In other words, we must furnish the etcher with negatives which have enough tone in them to allow for this loss of dot area.2
The screen negative contained large highlight dots that were than “tuned” to the correct size during the stage and etch process. Resist was applied to certain areas of the plate where etching would not have an effect. Over multiple etchings, the tones, from mid-tone to highlights, would take their correct shape and the tone reproduction curve would be formed. This was repeated for each separation.
Correct color reproduction in any process depends both on the color of the process inks (cyan, magenta, yellow, and black) and on the tone reproduction (how the different gray levels in the original object relate to the gray levels in the print). An advantage of letterpress printing was that proofs of the plates could be made after each etching to fine-tune the tone reproduction, so the correct color was achieved. The labor-intensive process required great skill and took many days to complete. The result was a set of color plates that provided a reasonable color reproduction.
Color reproduction was limited by the ink set, despite the amount of manual work done in photoengraving. Cyan, magenta, and yellow inks (or blue, red, and yellow, as they were referred throughout much of color history), were not ideal subtractive primaries. Regardless of how much platework was performed to compensate for the imperfect absorptions of the three primary inks, there were certain colors that could never be reproduced using a combination of the three colors
The photographic separations filters were red, green, and blue. The red filter absorbed green and blue light. The green filter absorbed red and blue light, and the blue filter absorbed red and green light. The ideal primary printing inks are complementary to those filters. Cyan (or “blue-green”) absorbs red. Magenta (or “red”) absorbs green, and yellow absorbs blue. Accurate reproduction of an object made through red, green, and blue filters is only possible if the inks are equal to those ideal primaries. However, the inks in the early 20th century were not pure (nor are they today, though the chemistry may be a little different). Cyan absorbed some green and blue, as if it was “dirtied” by magenta and yellow. Magenta absorbed a lot of red, and a little blue, as if it was dirtied by cyan and yellow, and the yellow ink is pretty close to ideal. These less-than-ideal primaries limited the color gamut and made it difficult to achieve true-to-life color reproduction.
Solving the challenges described above—reducing the amount of retouching required in plate making and accurate color reproduction—drove the advancement of color reproduction technology until the start of WWII.
The solutions for achieving accurate color reproduction, for both letterpress and offset lithography, often involved a lot of manual retouching. The manual re-etching in letterpress printing involved many iterations of staging and etching. If manual color correction was not performed, Mertle and Monsen wrote, “color printing with commercially available process inks would be typically muddy and false in hue. Millions of man-hours of handwork are spent annually in accomplishing such operations and the future of color printing most certainly lies in the direction of achieving required corrections more automatically and minimizing or eliminating their basic causes.”2 However, due to the need to achieve depth in etching, there was little photo-engravers could do with perfectly color-corrected separation negatives or positives.
In the early 20th century, some felt that selecting a different set of primaries, or adding additional primaries, would help reduce the amount of manual color correction. While the process for producing a black printing plate was known, its use was not standard progress. In 1922, Stephen Horgan wrote in the Inland Printer, citing William Gamble, “some thoughtful people were coming to the conclusion that the three-color idea had had its day, and there were some who doubted its validity, believing that four colors must be used to render all the hues of nature…This seemed to foreshadow a revolutionary change in color printing methods, and sooner or later some such change was likely to be brought about.”3 C. G. Zander notably promoted his idea for the addition of a green primary.4 Zander envisioned his process as both extending the gamut and removing the need for the laborious staging and etching of letterpress plates. While he did not successfully demonstrate the latter point, he was one of the first to show the viability of extended gamut printing. It was commonly believed the deficiencies in color reproduction were due to limitations in the filters and plate sensitivity.
According to Murray, Zander notably took the position that “the line of progress was to develop a process that would get the highest efficiency out of the best printing colors they had at the time.” Another proponent of printing with four chromatic colors was Michel Jacobs, who Murray mentions as having proposed the use of “yellow-yellow-orange, blue-blue-green, crimson-scarlet, and blue-blue-violet” hues. However, this process still required fine-etching. Even Murray himself explored the use of a purple, rather than a green, fourth color to extend the printable gamut. Murray, among other scientists at the time, sought a process that would produce color-accurate reproductions without fine-etching. Like letterpress, manual retouching was required to achieve correct color in early offset lithography. One lithographer, Louis Moeller, of New York, asked in a letter to the Inland Printer, “how to correct the errors in color separation in a set of halftones on grained zinc for offset printing. The three-color process engraver does it by reëtching the halftone engravings on copper. This, as you know, cannot be done on zinc.” In letterpress, the dots were physically etched into the surface of the copper plate. In offset lithography, a zinc or aluminum plate was coated with a layer of photo-sensitive gelatin. Halftone negatives were placed in direct contact with the plate and exposed to light. Exposed areas hardened and became ink-accepting areas. Non-exposed areas washed off and became ink-repelling areas. After exposure, little could be done to further color-correct on the plates themselves. Color correction was performed by manual retouching on the negative. Responding to Moeller, Stephen Horgan, noted that it was most common to make color corrections “by manipulating the lighting of the copy and by retouching the negatives, but the most successful way of revising the correction is called the ‘submarine method.’” However, their method of manual adjustment was still a labor-intensive, manual process.
After the best color-separation negatives possible are secured, then albumen, or glue-albumen-prints are made from these negatives on the several pieces of grained zinc or aluminum to be used in printing the colors. After these prints are inked they are developed in shallow trays under water. Hence the name “submarine.” The trained color artist who is developing them with tufts of cotton, or camel’s-hair brush, leaves color where he wishes or reduces the halftone dots where he pleases, and if he wants to remove the dots entirely he can do it with a touch of strong potash applied with a brush while the plate is held near a tap of running water so as to stop the action of the potash instantly.5
Another method of manual adjustment was written about in The National Lithographer. Essentially, a thin copper or zinc coating was deposited on a glass plate via electrotyping.
On these prepared plates, prints are made from the halftone color negatives in fish glue by the enameline process, so that the prints will withstand subsequent etching. Prints prepared in this manner may be etched with nitric acid for zinc or ferric chloride for copper coatings, and leave opaque positive images on the clear glass. The prints may be re-etched or fine etched to obtain the desired color correction…When etching is complete, halftone negatives are made by contact, and these are printed down on to the press plates in the usual way.6
In 1925, William Gamble wrote that “where halftone illustrations are concerned offset methods can not give by present processes the same clean, sharp and precise results that is obtainable from relief block printed on letterpress machines. Colorwork for offset printing by the halftone screen process is costly to produce, owing to the amount of retouching work that has to be done and the difficulty of doing it, while at its best the offset color printing lacks the quality of letterpress work.”7
The examples above are just a few of the many methods proposed for manual color correction in offset lithography. Not surprisingly, these manual methods could not be developed into the standardized processes needed to compete with letterpress on an industrial scale.
In 1925, Ellis Bassist described his dream of improving the color correction process for offset lithography using “some method of making color extraction negatives which will eliminate the costly handwork necessary in making color-plate corrections,” noting that “it takes many hours of painstaking and arduous labor, as well as a thorough knowledge of color-tone values, to make up for the incorrect rendering of such values.”8 The main problem, he suggested, was that “at the very start of [the operator’s] work, for at the first step—the separation negatives—he can hope to obtain, at the very best, only about fifty per cent near correctness.”
The most important aspect of tone and color correction is that the results be repeatable. Any work done to the plate itself must be repeated perfectly across all plates, a feat not possible without a strictly standardized process. Bassist suggested that corrections should be done on the positive plate used to create the final negative. In this way new plates could always be made from the same source, much like the use of electrotyping to copy letterpress halftone plates. He proposed a method in which “First, a set of continuous halftone color-separation negatives are made from a colored original. Then from this a set of positives are made on fine grained glass by the wet plate process. These are then retouched and again rephotographed through the screen in the camera.” Notably, this method notably still suggests the use of manual retouching.
The Birth of Masking
By the 1920s, masking methods in both letterpress and offset lithography were not new, but they were still novel. Masking is, in general, a method by which a corrected negative is made from the exposure of an original negative overlaid with a positive “mask” of the same image. For example, the positive may be slightly out of focus to achieve a sharpening effect, or it may be a low contrast positive from a different color separation to make a color-corrected negative.
In 1923, A. C. Austin wrote in The National Lithographer that “masking, or blocking out, in color photography is recognized as a legitimate means to an end, a practical commercial method for color correction that is better and cheaper than correction by hand retouching.”9 Masking methods can serve many different purposes. Austin viewed masking as a means for correcting deficiencies in the plate materials and in the photographic separation filters, contrasting with predominate view a decade later that it was deficiencies in the inks that required correction.
In opposition to those promoting four-color chromatic printing, many felt that the fourth printing ink should be black, rather than chromatic. Masking was important for the production of a black separation. In three-color printing, black is present in the cyan, magenta, and yellow separations. Overprints of the three inks did not give a black with the richness of black ink alone. Wet-on-wet printing was still in its infancy and limited the maximum black density. Registration of three-color blacks was also problematic, especially where line-work was concerned. The removal of blacks and grays from three-color separations and their replacement with a black plate was known as under-color removal.
The person most often credited with inventing the use of color corrective masking is Dr. Eugene Albert, who in 1899 was granted a patent for his “Photographisches Farbendruck-Verhahren” (Photographic Color Printing Process)10 and in 1900 was granted a patent for his “Photographisches Mehrfarbendruckverfahren” (Photographic Multi-Color Printing Process).11 In the first patent, Albert claimed a process by which the black separation negative is made first. Then, the three-color separations are exposed. These separations are copied with an added positive of the black plate, thereby removing the black and dark gray areas from the three-color separations. In the second patent, Albert built on the first by claiming the layering of positives from different separations with individual negatives as a method of color-correction.
Interestingly, while Albert was known as the pioneer of photographic masking, he was likely not the first to use such a method. One little-known British photographer, S. W. Bultz, proposed the use of an unsharp mask in 1856 to “subdue the normally excessive action of blue and white rays on the collodion surface—thereby allowing the less actinic rays (yellow, green and red) to record themselves on the color-blind collodion plate during camera exposure of the masked surface.”2 Collodion plates were overly sensitive to blues and whites. The negative, made from an out-of-focus normal exposure, would have been darkest where blues and whites predominated. Placed in contact with an unexposed plate inside the camera, this masking negative then reduced the illuminance of the blues and whites on the unexposed negative, thereby balancing the tones in the image. While his idea was novel, Bultz left it to others to determine how the procedure should be practically executed. He stated, “How long the negative proper is to be exposed to full action after removal of the screen plate, is a question equally difficult to answer. Another important, most important query, is the best way to place and remove the screen plate, and this must depend upon the ingenuity of the operator and the manufactures of the apparatus.”12
Early color correction masking methods were hampered by plate materials and halftone screens that resulted in negatives with suboptimal tone reproduction. Color correction masking could not advance until improved photographic emulsion and halftone screens became available.2
The science of masking was advanced by the focused efforts by large corporations, such as the Eastman Kodak Company, who developed entire product lines around masking. Alexander Murray, of the Eastman Kodak Research Laboratories, co-inventor of the Murray-Davies equation, helped to pave the way for scientific masking research. His goal was to “eliminate retouching by the masking process” with as few masks as possible.4 This is in line with the thoughts of Bassist, who stated, “If the process was to be used in the trade, the masks must be made by a simple procedure and kept down to a minimum in number.”8 A positive mask should not eliminate completely the masked colors, only reduce them. For example, a cyan positive is used to mask the magenta negative. The density of the cyan positive should be much less than the negative to reduce the amount of cyan ink in areas of the print containing magenta, but not eliminating it altogether. To be clear, the color correction masking did not correct the inadequacies of the inks, but rather adjusted the proportion of colors in the builds. Mertle and Monsen pointed out the paradox of masking, that individual inks are corrected not because of errors in that ink, but because of errors in the other inks. For example, yellow receives the most correction, and yet has a spectral absorption closest to the ideal. It must be, brought down to the level of inadequacy of the other inks.
Color correction masking had advantages beside correcting the hues. As they block unwanted colors, the masks effectively lighten black and gray areas, thereby increasing the saturation of the color regions of the print.2 The black are then filled in with the black plate, which replaces the brownish three-color neutral overprints that would otherwise appear in the shadows of three-color prints.
The density range of each mask was determined based on the science of sensitometry. We discussed the concept of characteristic curves in a previous article on the Murray-Davies Equation. The characteristic curve of a negative (or positive) is a plot of density versus exposure. The shape of the curve and the maximum density can be controlled by the exposure, development time, development chemistry, and development temperature (among other factors). It is here that scientific rigor was required, and not simply trial-and-error. In one such example, a positive was exposed, then developed, such that it was “equal to 40% of the density range of the negatives to be corrected” after a development time of “2 2 ¼ minutes in half-strength D-11 [a developer] at 68oF.”4
Each separation negative was masked by one or more positives from other separations. The cyan negative was not masked. The magenta negative was masked by the cyan, and the yellow negative was masked by the magenta negative (and sometimes the cyan, in some implementations). The precise characteristic curve of each masking negative and the process for making them had to be determined via scientific experimentation, but the end result was repeatable and could be standardized. Unfortunately, in 1934, according to Murray, the masking process required “rather expert direction and so has little chance of very wide adoption.” However, “it is capable, especially with complicated subjects, of effecting large savings in time and costs, when competently applied.” As with many new technologies, the development of competent operators took time.
Alexander Murray continued to develop masking techniques in the Eastman Kodak Research Laboratories. In 1939, the Eastman Kodak Company published a guide, understood to be penned by Murray, entitled, “The Modern Masking Method of Correct Color Reproduction.” This guide gave the general public insight into the masking methods Murray developed years prior (and helped promote Kodak’s line of masking products, such as negatives and formula guides).
The science of masking was further advanced by one of Murray’s colleagues at Kodak, John Yule, whose work focused on understanding the mathematics of subtractive color mixing. Several important developments lead up to Yule’s work, including the development of visual densitometers, the advancement of sensitometry, and the development of colorimetry (a model for human color perception). These developments, incidentally, were also spearheaded by the scientists at the Kodak Research Laboratories.
Yule introduced a mathematical masking model in a 1938.13 He stated that “the chief cause of imperfect color rendering in subtractive processes is the fact that the cyan and magenta dyes or pigments partially absorb the colors which they are supposed to transmit.” This reiterates the point discussed above that cyan and magenta were impure. Masking, he says, citing Murray’s 1934 article and Albert’s patents, is the “only known photographic method” of correcting for the impure cyan and magenta inks. However, even with masking, the results give “greatly improved, but still by no means perfect, color rendering.”
Yule also pointed out that masking was restricted to the reproduction of scenes containing “no colors beyond the gamut of the reproduction process and materials.” It is these cases in which additional chromatic inks, such as Zander’s green or Murray’s purple, would become necessary. The purest colors in any three-color (regardless of whether black is used) are the primaries themselves and pure color rarely appears in three-color reproductions. Yule defined requirements that must be met to achieve accurate color reproduction of a theoretical “color chart” made from the printable gamut (does not contain out of gamut colors), then derived a set of equations based on these requirements that could be used to calculate the effective density and contrast of masks for each separation: how much of a cyan mask for the magenta, how much of a cyan and magenta mask for the yellow, and so forth.
Different variations on the masking techniques proposed by Murray, Yule, and others, were proposed over the many years that followed. For a complete treatment on masking, I recommend John Yule’s book, “Principles of Color Reproduction,”14 which contains chapters on masking and the derivation of the masking equations, and Gary Field’s “Color and Its Reproduction.”15
While the manual masking methods described in this article are no longer used today, there are several notable techniques that remain in use. One of the requirements specified by Yule for his masking equations was that the “gray scale must be neutral and give accurate tone reproduction” such that “relative contrast and density of images must be adjusted to give neutral gray scale” and the “net result of all steps in the process must give a 45o straight line reproduction.” This is akin to the requirements for gray balance reproduction in the G7 process. In addition, the physical complexity of photographic masking and plate making techniques gave rise to the electronic scanner, which could scan a transparent or reflective original, divide the detected light into separations, then expose separation negatives automatically with color corrections programmed into the analog circuits. A patent for the first scanner was granted to Alexander Murray and Richard S. Morse in 1941,16 but many more followed. Scanning technology became widespread in the 1950s, and, employing the masking techniques developed in decades prior, helped lead to the rise of offset lithography in the second half of the 20th century.
An Illustrated Example of Masking
The color of an object is determined by the amount of light it reflects at each wavelength, the color of the light illuminating the object, and the visual system of the person viewing the object. Let us assume that the illumination and the person viewing the object are always the same, in which case the color of an object is determined solely by the percent of light reflected at each wavelength. For of simplicity, we’ll divide the visible spectrum into a blue spectrum, a green spectrum, and a red spectrum (a common method for illustrating color mixing). The color of an object is then determined by the amount of blue, green, and red components. White objects reflect an equal amount of blue, green, and red. This is illustrated in the figure below, where equal parts of blue, green, and red light are viewed as white.
In subtractive color mixing, the primary chromatic inks are cyan, yellow, and magenta. We already discussed above how these inks are not ideal. However, let’s assume we have perfect inks, where cyan perfectly reflects blue and green, and absorbs red; magenta perfectly reflects blue and red, and absorbs green; and yellow perfectly reflects green and red, and absorbs blue.
We can take two primary inks, cyan and yellow for example, and mix them together to get green. In the illustration below, light passes through the cyan ink first, then through the yellow. When white light passes through the cyan ink, red light is absorbed. When the cyan light passes through the yellow ink, blue light is absorbed. The remaining light is green because both blue and red light were absorbed, leaving green.
If we reduce the amount of cyan and yellow primaries—light cyan and light yellow—and mix them together, we get a light green. Light cyan reflects blue and green and absorbs only some red light. Light yellow reflects green and red light and absorbs only some blue light. The resulting light green reflects green and absorbs some red and some blue.
The perfect green and light green, as mentioned above, can be produced only by the subtractive mixture of ideal cyan and yellow inks. However, true cyan inks are not ideal. Keeping with the above scenario, let’s assume we have an imperfect cyan that reflects green, absorbs a small amount of blue, and absorbs red.
If we mix a light imperfect cyan with the same light yellow as above, the resulting light green will also be imperfect because some additional blue light is absorbed by the imperfect cyan. The imperfect light green is a little warmer than the previous light green because it has less blue.
We can compensate for this problem and correct the imperfect light green using masking. The main culprit for the imperfect light green is the cyan ink. We cannot add light to the cyan in the blue to compensate for the poor absorption. Rather, we take away some of the yellow since yellow absorbs blue, thereby reducing the overall blue absorption. The reduction of yellow is accomplished using a low contrast positive mask made from the cyan separation negative which reduces the amount of yellow ink wherever it is mixed with cyan. Reducing the amount of any ink makes that ink a little more white than pure, meaning less light is absorbed. Therefore, reducing the amount of yellow ink reduces the amount of blue light absorbed. The result is a corrected light green appearing closer to the ideal light green.
Comparing the three light greens discussed above shows how masking can be used to effectively correct for unwanted hue shifts that result from the use of imperfect inks.
This illustration is only a simplified illustration of the problem surrounding the use of imperfect inks in printing. There are no printing inks that match precisely with the ideal inks. Yellow is the closest, but magentas and cyans have always been problematic. The contributions of the scientists involved in masking research made possible the accurate reproduction of colors and helped drive the use of color printing in the 20th century.
Disclaimer
This article was written by Brian Gamm in his personal capacity. The views, thoughts, and opinions expressed in this article belong solely to the author, and not necessarily to the author’s employer, organization, committee or other group or individual with which the author has been, is currently, or will be affiliated.
- 1.Horgan SH. Offset Printing: Offset Printing’s Future. Inland Printer. 1922;69:368.
- 2.Mertle JS, Monsen GL. Photomechanics and Printing; Practical Information on Platemaking and Presswork by Recognized Procedures. Mertle Publishing Co. ; 1957.
- 3.Horgan SH. Process Engraving: A Three-Color Pessimist. Inland Printer. 1922;69:239.
- 4.Murray A. The Masking Method of Color Correction Applied to Three- and Four-Color Reproduction. The Photo-Engravers Bulletin. Published online January 1934:13-21.
- 5.Horgan SH. Notes on Offset Printing: Color Correction on Offset Plates. Inland Printer. 1922;69:872.
- 6.An Improvement Process: Photo-Lithography in Color by an Ideal Method with Accurate Tonal Values. The National Lithographer. 1923;39:44. Accessed November 26, 2020. https://www.google.com/books/edition/_/ZTAyAQAAMAAJ?hl=en&gbpv=0
- 7.Gamble W. The Future of Letterpress Printing. Inland Printer. 1925;74:757.
- 8.Bassist E. Retouching for Offset Lithography in Colors. Inland Printer. 1925;74:756.
- 9.Austin AC. Photo-Lithography: Suggestions for the Lithographers Which Should Prove Beneficial to the Trade. The National Lithographer. 1923;30(11):25-26.
- 10.Albert E. Photographisches Farbendruck-Verhahren. Published online 1899.
- 11.Albert E. Photographisches Mehrfarbendruckverfahren. Published online 1900.
- 12.Bultz SW. New Method of Copying Paintings: To the Editor of the Photographic Journal. Journal of the Photographic Society of London. 1856;3(40):15.
- 13.Yule JAC. The Theory of Subtractive Color Photography I. The Conditions for Perfect Color Rendering. Journal of the Optical Society of America. 1938;28:419-430.
- 14.Yule JAC, Field GG. Principles of Color Reproduction. 2nd ed. GATFPress; 2001.
- 15.Field JAC. Color and Its Reproduction. 2nd ed. GATFPress; 1999.
- 16.Murray A, Morse RS. Color Photography. Published online 1941.
It´s really really good! Thanks for lovely reading. It also makes my thoughts wonder away and thinks about things like skin tones etcetera, but that´s the fun of reading things like this. Thanks again Brian!