Who first discovered CMYK?

93

25

It has been used for colour printing for many, many years, but who (and when) first discovered that a bright pink colour (magenta), a light blue (cyan), yellow and black can be mixed to create almost all colours? How did they discover this?

(Or more correctly "CMY" - the Black (K) isn't strictly necessary)

Digital Lightcraft

Posted 2016-06-07T09:03:11.930

Reputation: 3 197

5It should be noted (unless I missed it in the answers) that CMY are simply the negatives of RGB. If you take a black and white photograph of a scene using a red gel (where the red is calibrated to the RGB system), the resulting negative is going to be the C value. To derive a 0-255 value for R using a 0.0 - 1.0 value of C, you can calculate R=(1-C) * 255Yorik 2016-06-08T17:07:05.307

@Yorik that would be a good answer to "how was CMY discovered?", if you pick up that and join Stan's reference to Herbert E. Ives below, imo you'll have the best answer.

Armfoot 2016-06-11T13:28:09.537

1@Armfoot that would make it Newton. Problem was that the technology wasn't available until hundreds of years later. If you go with Newton, you'll have to acknowledge Alhazen who predated most all of Newton's [re]-discoveries by almost a millennium! Alhazen discovered complementaries by the afterimages that persist when you stare at and then look away from a saturated hue. Try it.Stan 2016-06-11T22:20:32.880

@Yorik Please see comment to Armfoot re early discovery and notation of the complementary colour phenomenon of the nature and appearance of "afterimages"Stan 2016-06-11T22:25:38.407

@Yorik and onlookers That's how colour reversal film work, in effect, by coupling a dye image that is the inverse colour of the colour sensitivity of the film layer in each of the three layers of (old) colour film or (new) electronic image sensors.Stan 2016-06-11T22:30:10.993

Please consider upvoting/following the StackExchange Color Theory site: https://area51.stackexchange.com/proposals/110687/color-theory

Adi Shavit 2017-06-22T08:09:08.103

Answers

87

According to Joe Scout the first company to use CMYK in printing was Eagle Printing Ink Company and the year was 1906. It was not until 1956 that it became a standard as a result of Pantone trying to streamline the workflow.[1]

This however does not really answer who really invented/discovered the choice of colors, the first scientific literature to mention this appears to be published in 1908. Going deeper it appears that we have a significant problem. It seems that there is a certain ambivalence in the naming of colors. In certain later sources we see that cyan and magenta refereed to as blue and red even as late as 1950's. So language no longer works quite well searching for earlier sources and it becomes hard to follow the lead. Secondly pigments fade so we can not compare easily what they mean. Certainly lot of work of finding the primaries has been done from 18ths century. [2]

It has been fairly long known that the human eye was sensitive to 3 color wavelengths (called tristimulus values in scientific lore). Some sources attribute this as far back as 1613 [3]. But certainly it was known and verified by mid 1800's see Young–Helmholtz theory. So the sensory primaries were known for quite some time. RGB primaries are chosen to closely match the 3 sensors in our eye. However, a paper does not transmit light, it reflects light. So a paper is usually white, representing maximum reflection. Each ink then subtracts one color off the palette since you need to be able to independently manipulate each RGB channel. This fact was known since Newton [2]. So you need to find the inverse of RGB which simply is CMY. The black color is there mostly, and originally, to avoid registration problems as black is the most common body text color ink. Though black does make it easier to mix certain dark tones.

Of course CMYK printing predates RGB monitors by a comfortable margin of 60 years. CMYK is not derived from RGB but the intuition needed is the same. The authors must have known or indirectly inferred the inner workings of the eye's sensory apparatus. But its not nearly as simple as this as finding good pigments for the job is a journey on its own. Knowing what to look for is not the same as knowing what is there.

References:

  1. Scout, Joe, The History Of The CMYK Colour Model, Club Ink Blog, 2016-05-02. http://www.clubink.ca/blog/print/history-behind-cmyk-colour-model/ referenced 2016-06-07

  2. Briggs, David. 2006. http://www.huevaluechroma.com/062.php

  3. Weale, R. Ophthalmic Physiol Opt. 2007 Nov;27(6):525-6.

joojaa

Posted 2016-06-07T09:03:11.930

Reputation: 36 627

12I would like to point out that:First commercial use is not the same as the invention by the way, so we still do not know who invented the combo. Still doing some research on the matter. We are now making a similar mistake as attributing Watt for invention of steam engines and light bulbs to Edison. Most likely the real situation is more complicated.joojaa 2016-06-07T09:46:29.767

1Aha seems like much more digging than i can commit is warranted.joojaa 2016-06-07T10:16:28.433

"So you need to find the inverse of RGB which simply is CMY." I guess the heart of this question is... How was that first deduced? (And by who, but those two answers would probably go hand in hand)recognizer 2016-06-07T15:29:46.030

1@recognizer ok i added the youg-helmholz theory and will add the fact that they knew about the difference between emitted and subtractive spaces already.joojaa 2016-06-07T15:53:23.833

16Point of information: Magenta was known as printer's red and cyan was known as printer's blue.Stan 2016-06-07T16:35:50.823

10The acronym for Eagle Printing Ink Company would be EPIC!Octopus 2016-06-07T17:23:24.153

6@Stan and, sadly, schoolkids are still taught "red, yellow, and blue" and are never taught any different as the get older. A huge number of adults will still think you're lying when you say "red and green light, mixed together, make yellow."Jamie Hanrahan 2016-06-07T21:25:42.897

7"RGB primaries are chosen to closely match the 3 sensors in our eye." Not quite. The cones loosely called "red" and "green" are actually centered on orange and on yellowish-green. This gives good discrimination among most of the colors of interest to early humans - i.e. colors seen in nature on the African plain. Since color TV needs to repro a lot of "non-natural" colors too, its RGB primaries were picked to give the best range of colors (with some influence from what phosphors were even available). http://hyperphysics.phy-astr.gsu.edu/hbase/vision/colcon.htmlJamie Hanrahan 2016-06-07T21:29:17.197

2@JamieHanrahan Mixing opaque pigments is what schoolkids are shown. They come around as soon as you put a R, a G, & a B filter on each of three flashlights and let them mix 'n match on a projection screen. : )Stan 2016-06-08T00:18:06.377

1@JamieHanrahan right, actually no system other than being able to show all spectrum colors can be perfect. Reason is that the overlap of the sensors is peculiar. Ive become increasingly sceptical of the 1906 figure. All explanations contain some lies to children i guess.joojaa 2016-06-08T04:16:52.253

I think the RGB system was developed to approximate our perception, rather than perfectly model the physicality of cones and rods. For example I seem recall that some Yellows are considered perceptually interpolated and not actually physically detected.Yorik 2016-06-08T17:11:08.767

@Yorik see my previous comment, its more of what the overlap conditions are. Anyway yes its a bit of a simplification, if i were to explain how it actually works we would be here for a loong time reading.joojaa 2016-06-08T17:27:48.193

1@JamieHanrahan: A typical blue paint (and I don't mean cyan), mixed with a typical yellow paint, will in fact yield green, though depending upon the types of paint the green may appear lighter or darker than the blue. Filters mostly behave more sanely than paints, but the effects of a filter on colored light cannot be predicted based upon the perceived hue of the filter. A didymium filter, for example, would look like a very light neutral gray filter in most cases, but will appear dark under sodium-yellow illumination.supercat 2016-06-08T18:03:06.320

1

When I mentioned blue + yellow I was thinking of mixing colored lights, not paints. Re the sodium-yellow lamps: Of course, b/c that filter doesn't pass much of the very narrow wavelengths those lamps puts out. It's like trying to use a colored filter in front of an LED. Even the "white" LEDs (which are blue LEDs with a fluorescent material that glows yellowish-white, combining to make sort-of white) don't filter the way we'd expect either. Related (see esp. the table at the end): https://www.rosco.com/technotes/filters/technote5.html

Jamie Hanrahan 2016-06-09T00:21:29.940

@JamieHanrahan: A more prosaic explanation for the similarity of the human "red" and "green" cone response curves may simply be that the genes encoding those receptor pigments arose relatively recently in our evolutionary history from a single ancestral gene via gene duplication and gradual mutations (most non-primate mammals only have one "red/green" receptor type) and have not yet had time to diverge further. (Of course, this also implies that the selective pressure can't have been too strong.) See e.g. https://en.wikipedia.org/wiki/Evolution_of_color_vision_in_primates for more information.

Ilmari Karonen 2016-06-10T03:54:26.647

Comments are not for extended discussion; this conversation has been moved to chat.

Vincent 2016-06-10T08:39:00.207

What detail would you like augmented for your bounty? My problem with web sites is there's no reference source to verify. I have half-a-wall of references to check. : )Stan 2016-06-22T17:48:28.123

@joojaa The web site by Joe Scout has no attributed authorized sources for the statements made there.Stan 2016-06-22T19:33:23.657

@Stan i am aware of it. However theres no clear cut answer as to who actually did get the cmyk idea first as it seems to have been popping into existence in several forms. Yours seems to be the last iteration so the situation is more like watt getting the name for steam engine despite the steam engine predating him by a few thousands of years.joojaa 2016-06-22T19:40:30.643

37

CMYK is an improvement over CMY which itself is improvement over RYB model, which has been used for centuries (if not millennia).

It's really hard to tell where one ends and the other begins, especially as some use words "red" and "blue" in more general sense. Eg. George Field's chart from 1841 lists "red, blue, yellow" but his red in our eyes looks closer to magenta than what we today consider as "red". His blue is also suspiciously pale, so his "RYB" chart may be viewed as early attempt of achieving CMY model. We also need to consider that people were working with dyes they had access to - so they could be aware that "proper" blue should be more cyanish, but it wasn't available to them. Move from RYB to CMY was a gradual process.

Adding black is mostly cost-cutting improvement (or workaround for imperfect CMY inks) so I personally would not make a clear distinction between CMY and CMYK at fundamental concept level. The usefulness of black ink varies heavily depending on the actual image being printed.

Wikipedia has nice historical article under RYB color model

Agent_L

Posted 2016-06-07T09:03:11.930

Reputation: 471

3Interesting answer. Can you back your explanations with some sources?Luciano 2016-06-07T11:50:11.833

1Yes this does quite well reflect what my research come up as well. It seems that the names magenta and cyan is taken into the use at the turn of the century. Certainly they were hunting for primary colors in the 18th century as indicated in my second source. So they might in fact have pinned this down quite well. But may have lacked abundant sources of pigment so didn't name them like this. It is almost certainly clear that red was used for magenta, but what is the situation of cyan is still a bit unclear to me. But the hunt for these pigments is actually relevant for anybody who does researchjoojaa 2016-06-07T12:12:40.723

Just some extra trivia: the K is critical in commercial printing as most text is printed in black (which would be really hard to achieve with CMY given misregistration issues on press)DA01 2016-06-22T16:20:16.370

@DA01 I'm holding an orange yogurt carton here and looking at the printing marks it's done without any black at all. But it has extra orange ink instead to make the main image pop. Text on the sides is dark blue - and there is a 5th dark blue ink too. This is what I said - other inks (except CMY) are special purpose addition depending on the content.Agent_L 2016-06-23T11:29:03.827

22

There is no single definable point when the CMYK Process Colour printing was discovered. High fidelity process colour reproduction printing has been a gradual series of technical refinements.

The persons responsible are, however, known.

Printed colour reproduction grew rapidly in popularity in the late nineteenth and early twentieth centuries to today, when (excluding newspapers) the great bulk of all pictorial reproductions is in colour. The basic principles of photomechanical colour reproduction remain the same today as there were in 1900, but there have been many important advances over the years that have led to improved quality and lower costs.

The Colour Process for Colour Reproduction depends on several different factors.

First, is the development of the necessary TRANSPARENT inks. As colour reproduction depends on the subtractive system, the overlap of the process colours cannot work with opaque inks that obscure the under-printed inks and the paper (usually) substrate.

The original pigments used in printing inks were mainly inorganic pigments that had a restricted gamut, and, in some cases, poor transparency. The development of organic pigments increased the available color gamut while still retaining reasonable permanence. The major developments were as follows:

  • The azo colors for ink manufacture developed between 1899 and 1912. Most Yellow pigments are of this class.
  • The discovery of the tungstated and molybdeated pigments in about 1914. The best process magentas fall into this class.
  • The discovery in 1928 of phthalocyanine pigments, which made possible the first really permanent brilliant cyan suitable for process-colour printing.

Other developments that led to use of process-colour printing included the four-colour printing press.

The first recorded use of a four-colour lithographic sheetfed press was by the Traung Label and Lithograph Company of San Francisco, California, during early 1932. This offset press was made by the Harris-Seybold-Potter Company of Cleveland, Ohio (now Harris Graphics). Four-colour web offset presses preceded sheetfed presses. In 1926 the Melbourne, Australia, daily newspaper The Argus installed the German-built Vomag web offset machine that had four perfecting (can print on both sides of the same press sheet) printing units. This press was used to print weekly colour supplements and magazines. The Berlin, Germany company of Messrs Dr. Selle and Company were reported as printing four-colour work by web offset in 1926.

The Cottrell Company reportedly made a four colour common impression cylinder rotary letterpress sheetfed machine about 1912, but the thick letterpress ink films (layers) made wet-on-wet (ink "trapping") process-colour work impractical.

The first recorded three-colour (intaglio) prints were produced on a web machine at Siegburg in 1914. A common impression cylinder was on this machine.
Around the same time, a multiunit Goss intaglio press was installed at The Chicago Tribune. This machine had separate-unit type construction for each colour. However, it is thought that the first successful gravure process four-colour work on a multicolour machine was not produced until the late 1920s or early 1930s, probably on a machine made by the Albert company.

The specific answer to your question is in this entry:

The primary elements [for the CMYK process colour printing we know today] are a magenta red, a yellow and turquoise blue. These three basic hues were brought to perfection by Herbert E. Ives (1882–1953) and represent the minimum "primary" colours which, in combination, will produce a full array of fairly pure intermediates using average pigments. (Ives used the term achlor for magenta, zanth for yellow, and syan for turquoise blue.) Mixtures of magenta and yellow form reds and oranges. Mixtures of yellow and turquoise form greens. Mixtures of magenta and turquoise form purples. These three, in other words , are the fewest that can be employed to produce a satisfactory colour circle. For rich, powerful hues, however, more than three colours becomes essential.

Prior to, and inspiration for, the CMYK dot automated screen colour printing process as a means to create intermediate colours in a print was by manually using points of pigment. The artists that pioneered the effect known as optical mixture were French painters such as Georges Seurat and Paul Signac in 1886. From a distance we cannot clearly perceive the edges of small shapes. Similarly, from a distance a small areas of colour seem to blend into one another and are seen as a mixture rather than as separate hues. The effect occurs in the eye rather than in the pigments. Essentially, pointillism was the immediate precursor to the 4-colour dot screen used by the modern 4-colour process printing

This effect was the focus of experimentation for a group of painters called the pointillists, who theorized that optical mixture would provide a more dynamic expression of colour than the traditional pigment mixing.

A Canadian engraver, Phillip Desbarates, came up with a similar automated process using lines and dots to make halftones images in Montreal, Canada, who discovered the process for printing for the Montreal Star around the same time (1885). Desbaretes later took his concept to the Currier & ives Co. where porcelain china was decorated with scenes using the early "halftone" process with colour materials.

A collision of events near the turn of the 20th century heralded the 4-colour CMYK printing process we have today using transparent inks and dots.

Sources:

  1. Creative Color, Faber Birren, Lutton Publishing Compary, 1961
  2. Principles of Visual Perception, Carolyn M. Bloomer, Litton-Van Nostrand Reinhold, 1976
  3. Graphc Arts Photography: Color, Fred Wentzel and Ray Blair and Tom Destree, Graphic Arts Technical Foundation, 1983
  4. Color And Its Reproduction, Gary Field, Graphic Arts Technical Foundation, 1988.

Stan

Posted 2016-06-07T09:03:11.930

Reputation: 2 478

1It is interesting to note that black was most always used in any two-colour, three-colour, and process colour printing. The "K" in CMYK printing indicated the darkest "key" colour which in most cases was black. "Rich" black is the combination of black in addition to one or more other process colours in register (alignment).Stan 2016-06-07T23:02:15.510

18

In 1906, the Eagle Printing Ink Company incorporated the four-colour wet process inks for the first time. These four colours were cyan, magenta, yellow, and black (also known as key), hence the name CMYK. It was discovered that these four colours can be combined to produce an almost unlimited number of richer, darker tones.

Link: http://www.clubink.ca/blog/print/history-behind-cmyk-colour-model/

Erik

Posted 2016-06-07T09:03:11.930

Reputation: 189

2

I don't think it is accurate to say that CMYK was "discovered". I am anything except an artist or an historian, so take what I have with a huge grain of salt. But, think about it. Ink is just paint for paper. Any question about ink must refer back to paint, and of course we've been using paint for 8 thousand years. So, the question is not productive. CMYK is an INVENTED system to quantify (industrialize) the color making process and description. Conflating CMYK with the fact (which most 5 year olds know) that colors can be mixed and that three plus black and white give you a huge pallet is not the most useful way to think about it. I'm a technologist, I think of it more like a standardization of color. As paper making got more sophisticated and quality (uniformity and color stability) of the paper improved, and as pigment production improved (but even today, two different batches of almost any pigment is of slightly different color! (titanium white and carbon black might be considered exceptions...), we were able to quantify and standardize the process of creating a multiple color picture. But if you dig in to the pigment (or paper) making process, you'll find that the artists of past centuries made their own paints, some their own pigments, and that uniformity was rare. HTH

Li Zhi

Posted 2016-06-07T09:03:11.930

Reputation: 29

2No it wasn't invented a s a singular act no. But there is a long progression of scientific discovery that predates it. But like some have said its hard to pinpoint the time and place also as we dont know what they used to call the colors. So yes tthe same thing as what you say basically what all the other answers amount to. But no painters rarely use a 4 color system they can mix much more paints.joojaa 2016-06-07T16:08:27.523

1I think the key here is that cyan, magenta and yellow are universal bases of subtractive color model. You can't get full coverage with any red, blue and yellow, they have to be particular shades. So there imaginably was a single moment when someone "discovered" the exact shades of cyan and magenta which yield best results.Agent_L 2016-06-07T16:49:03.670

1@Agent_L I agree with that but i dont think this answer adds much to your answer. But yes they were in search of the primaries even before they knew they were or did know of the theoretical underpinnings. and since you can deduce this by other means than the theory they most likely came very close. Like i said i have reference that the red was actually magentaish but im not so sure about the blue color. Its also possible that this was perfected and unperfected a few times since time immemorial. I would probably vote for this answer, if it didnt use the every five year old knows this.joojaa 2016-06-07T17:29:01.527

3Some choices of primaries give you more coverage than others but "full coverage" (producing every color the eye can see) isn't possible with a subtractive mixing system.Peter Green 2016-06-07T22:51:19.837

1@PeterGreen full range is not possible with any tritimulus system even rgb. Reason is that response curves overlap and there are areas in the senses where nothing bit showing a very specific wavelength suffices, such areas are present in greens and oranges.joojaa 2016-06-08T04:09:01.717

1Being able to mix CYMK to make all other colours is an intrinsic property of those inks. I would say that makes it a discovery not an invention.Wossname 2016-06-08T12:32:32.103

I almost commented that it was an invention not a discovery when the question was first asked, but there is definitely elements of both involved. The processes involved were certainly invented but the theory behind it is all based on scientific discovery.Cai 2016-06-09T10:26:01.517

1@Wossname Well, it's an intrinsic property of the way the human eye works, not so much a property of the inks themselves.Bacon Bits 2016-06-10T20:46:08.950

@BaconBits, good point. I would be interested to know if the color spectrum of light reflected by such an ink mixture can be used to verify this. Does green ink mixed using CYMK processes actually give off a predictable green spectral reflection or do you just get separate spikes at the wavelengths of the inks that went into it?Wossname 2016-06-13T07:16:31.113

2

It likely depends on what you qualify as 'CMYK'. Other answers have already established that the K is superfluous -- it's there to make printing easy, not because it's required for color.

So that makes us wonder who first established that we could create (most) colors using only three colors ... and I believe that at least in the West, that's credited to the work of Thomas Young who proposed in 1802 that humans had three color receptors, which was extended by Hermann von Helmholtz, and later proven by James Clerk Maxwell.

If we go back further, however, it was Isaac Newton who is credited in 1686 with the discovery of metamers -- that you could combine two colors of light to get something that is perceived as matching a third color even if they don't actually match spectrally.

So ... to summarize ... it's quite likely either the person who gave us Newton's Laws of Motion, Young's Modulus, or Maxwell's Equations ... or the namesake of the Helmholtz Association

Joe

Posted 2016-06-07T09:03:11.930

Reputation: 194

1

The three- and four-color process was invented by Jacob Le Blon around 1725. His original color model used RYB (red, yellow, blue) and RYBK (red, yellow, blue, and black). He wrote a book called The Coloritto, in which he explains his discoveries and results that lead him to use these set of particular colors and the reason for adding black to his original three-color process to produce more realistic natural colors. You can find the book in the Library of Congress.

James Belk

Posted 2016-06-07T09:03:11.930

Reputation: 11