By Sarah Sands

Introduction
On our Web site, for every color we make, an artist will find a wealth of information running in parallel next to the much more visual presence of a virtual drawdown. If asked what the color looked like, or told to mix a similar hue, most would instinctively go to the image for guidance. However, some might rely on the name of the paint, or what the pigment is. Still others might look it up using the Munsell Notation, or even enter the CIE L*a*b* values into Photoshop®. But which of these is the most accurate way to truly pinpoint just what color this color is? What follows is an attempt to answer that question, or at least to challenge and inform it with some additional tools and definitions.

Labeling Color
ASTM
While a Rose by any other name might mix as sweetly, if it’s actually a Quinacridone Red PV 19 and not labeled as such, it’s not in compliance with the American Society for Testing and Materials. Known more commonly by its acronym, ASTM, this is the organization largely responsible for setting the minimum standards for the testing and quality of artists’ materials. But their work goes largely unheralded. If artists are aware of them at all it is usually from reading the small print on the back of paint tubes, where the phrase, “Conforms to ASTM D 4236”, accompanies disclosure of chronic health hazards, or perhaps from their widely used Lightfastness ratings. Far less known, however, are how the ASTM requirements for labeling artists’ paint (D 5098-03) play such a critical piece when we speak about colors. First published in 1984, these standards literally provide the only assurance, outside of Federal and State mandated health warnings, that paints are accurately labeled. Whenever a tube or jar of paint does not clearly state it is in conformance with these guidelines, the artists can be left largely in the dark about what they are buying. Unfortunately, only a few companies outside of GOLDEN have ever followed these standards completely. True, compliance has always been voluntary, but keep in mind these are minimum guidelines and providing artists with accurate and consistent information should be an obvious part of every manufacturer’s mission and responsibility.

The standards themselves are fairly straightforward and, beyond issues of placement, cover three major requirements: the label must include the Lightfastness rating for the color, specify each pigment contained in the paint by both its official Common Name and Color Index Generic Name, and the word “Hue” must be included in the name of the paint whenever another pigment is substituted for one normally referenced. As defined in the standards, the term would apply not only to well-known examples such as Cadmium Red Hue, signaling that no actual cadmium pigment is present, but a long list of discontinued, historical colors like Indian Yellow, VanDyke Brown, and Sap Green. Sadly, because the standard is not strictly followed and the word has been used so loosely, “Hue” is often seen as a marker for inexpensive or cheaply made substitutes when often the actual intent is to genuinely provide safer, more permanent, or otherwise unavailable colors.

Color Index
If one is going to speak about color, pigments, and paint, it won’t be long before you need to refer to the Color Index. This thick compilation of endless rows of entries is something of a holy grail for anyone wanting to know exactly what constitutes that color lurking in the tube. First published in 1925 by The Society Of Dyers and Colourists of the UK, and currently managed in collaboration with the American Association of Textile Chemists and Colorists, it forms the official index of all commercially available colorants. Each pigment any artist might ever use today can be found there, along with those for every other industry. Organized by color, each listing is assigned a C.I. (Color Index) Generic Name, Constitution Number, and a listing of Common Names associated with the dye or pigment. However, by the time all this makes its way to the label, it will usually take on the more cryptic form of a code: PY 43, PB 29, etc.

Deciphering most of the information in a complete listing is actually not difficult, if you know a few basics:
• C.I. Generic Name, which is what you will likely see on a label, always has three components:

– Colorant Type
This is designated by the initial letter. For our purposes the most important for artist paints are solid pigments, designated by a P. Other possibilities, like D (Dye) or S
(Solvent Dye), are more common in other industries.

– Hue
There are ten possible categories: R (Red), O (Orange), Y (Yellow), G (Green), B (Blue), V (Violet),
Br (Brown), W (White), Bk (Black), M (Metallic)

– Index Number
Pigments are assigned the next available sequential number, within each of the above color categories, at the time they are added. Gaps can occur in the series as pigments become obsolete and are removed over time.
• Constitution Number
Rarely found on labels it is included in more complete listings such as our Pigment ID Chart or the color drawdowns found on our Web site. It is an assigned five-digit number based on the chemical structure of a colorant, when made available by the manufacturer.
• Common Names
A list of the common, generally accepted names for the pigment. This is different than a list of the often proprietary names paint and pigment manufacturers might give.

So, taking for a moment one of our earlier examples, PB 29 would stand for Pigment Blue 29, or the 29th entry for blue pigments. Its Constitution Number, 77007, would refer to its chemical composition as a Polysulfide of Sodium-Alumino-Silicate. Its Common Name would be the familiar Ultramarine Blue.

So how does this help? Where is the path connecting code to color? By far the greatest strength and value of this system is that it provides the artist with an internationally recognized, standardized, and dependable way of knowing precisely which pigments are in a paint. Even if a product uses a familiar color name bearing no relationship to the actual ingredients, and therefore not in compliance with ASTM standards, the Color Index information can hopefully be relied on. Why ‘hopefully’? Because even with standards in place, the accuracy of the labels remains very much a matter of trust as the ASTM has no enforce-ment role. While for some these codes and regulations will appear to wring out any last vestige of poetry from the tubes they buy, it allows the artist to know exactly what they are getting when they get the goods.

Some Limitations
As essential and reliant as the Color Index is, there are also limitations. The Index is certainly useful for cross-referencing which pigments are in a particular tube of paint but by itself this does not necessarily tell an artist what the actual hue of the color will be. The reason for this is that a wide range of shades and qualities in a pigment can still share the same Color Index Name. Asking a pigment supplier for Ultramarine Blue can easily mean picking through 15 different shades or more, with a surprisingly wide variance, but in the end all authentically PB 29. Cadmiums are equally notorious. PR 108 could indicate a very warm Cadmium Red Light or a much cooler and deeper Cadmium Red Deep. PY 35 is equally Cadmium Yellow Primrose and Cadmium Yellow Deep – extremely different colors with identical Color Index Names and even Constitution Numbers. Both synthetic and natural iron oxides have surprisingly wide gamuts as well. PR 101, indicating a synthetic iron oxide, belongs to both the dense and opaque Violet Oxide and the very translucent Transparent Red Iron Oxide, while PBr 7, a natural oxide, covers anything from Raw Umber to Burnt Sienna. And the more modern organic pigments don’t escape easily either. PV 19 indicates both Quinacridone Violet and Quinacridone Red. To further complicate things, a single Common Name can belong to different pigments. For example, ‘Raw Sienna’ can be used with either PY 43 or PBr 7, and these happen to also share the same Constitution number since their underlying chemistries are essentially the same, the difference in hue coming from different levels of calcination or specific minerals. But don’t despair – the color equivalent of the cavalry is about to arrive.

Colorimetry and the Measure of Color
Because the Color Index’s primary focus and purpose is the cataloging of commercially available pigments, it does not provide the tools needed to really classify color in a very precise way. Needing a methodology that could do that, both industry and science have had to turn to various models of color space and methods of analysis drawn from Colorimetry, the field concerned with the quantitative measurement of color in general. Ultimately these technologies and systems have provided a neutral, objective language for people to accurately record, compare, and communicate the exact hue of any perceivable color. In the sections that follow we will describe the two models most widely used for these purposes, CIE Lab and Munsell, and go on to examine alternate systems such as the CMYK and RGB, which are widely used in the print and display industries respectively.

CIE LAB
In 1976 the Commission Internationale d’Eclairage (CIE), an organization that writes the official standards for the scientific measurement of color, developed a color model known as CIE Lab (or, to be very precise, CIE 1976 L*a*b*). Its goal was to create a system for describing all perceptible colors in a manner that was both uniform and device independent. By ‘uniform’ is meant that equally measured distances in the color space ideally equate to equally perceived color differences as seen by a standard observer. The standard observer is a construct the CIE created from extensive research on what a person with normal sight might perceive. ‘Device independent’ refers to a color space independent of the limitations inherent in a specific media or device, such as a particular printer or monitor. Because CIE Lab is free from these restrictions it can represent colors as perceived by the human eye and act as an almost universal translator between the different color spaces native to those various devices. Consequently, as a result of these strengths, CIE Lab has become the most widely used and authoritative standard for all color management systems and wherever information needs to be calculated and communicated in a device-independent form. CIE Lab is also central to the formulas used to determine ASTM Lightfastness ratings and in GOLDEN’s own operations for color matching, color analysis, and to assure batch-to-batch consistency of both raw pigments as well as finished paint.

Conceptually the actual CIE Lab space is fairly easy to understand. As with all three-dimensional models, any point can be defined in terms of three coordinates plotted along their corresponding axis. The variables CIE Lab uses for locating each color are:
L*: lightness or value. The scale runs from 0 (Black) to 100 (White).
a*: red-green component. Negative number is greener, positive is redder.
b*: blue-yellow component. Negative number is bluer, positive more yellow.

These measurements are taken with a precisely calibrated spectrophotometer using standards set by the CIE. As an example, here is one typical reading taken from a GOLDEN Heavy Body Acrylic:
CIE L*a*b* Values: L*78.51, a*18.46, b*89.29

By simply looking at the numbers it would be difficult to divine exactly what the color is. The high L* speaks to the color being fairly light in value, while a* shows a slight degree of reddishness and b* a very strong yellow component. If a person went on to imagine it was a very bright, slightly warm yellow, they would be essentially correct. Even more importantly, give these numbers to anyone in the world with the appropriate software and tools, and they should be able to recreate, fairly accurately, the color for Hansa Yellow Medium. It is perhaps not the most lyrical way to communicate a color more associated with sunlight and lemons, but what it might lose in poetry it certainly gains in precision.

Munsell
Along with CIE Lab, the Munsell color system is perhaps the most widely disseminated and utilized in the world. Originally created by the American artist and educator, Albert H. Munsell, in 1905, it was certainly not the first attempt to arrange colors into a logical order but it has been the most successful at establishing a system for designating surface colors within a systematic space. He based his system around three attributes: Hue, or the quality that separates one color from another; Chroma, a concept similar to Saturation; and Value, or how light or dark a color was. These were then organized into an irregular three-dimensional model he termed a ‘color tree’ with a central value scale running from black through nine achromatic grays to white in the center, creating a vertical ‘trunk’ of perceptually equal steps in value. Around this core were arrayed five basic color groupings spaced at equal intervals: Red, Yellow, Green, Blue, and Purple, further divided by the five admixtures between them. Chroma was measured along a scale of likewise perceptually equal steps running from a particular color’s most saturated state to the corresponding grey of the same value.

The feature of consistently organizing the three major aspects of color into perceptually equal divisions in all directions was one of the most significant and recognizable features of Munsell’s model and profoundly impacted nearly every major system that followed. (Landa, Fairchild, 2003) By further subdividing these main divisions into ever-smaller intervals, any color within the space could be easily designated by its Hue, Value
and Chroma. As an example, the specification for our earlier example of GOLDEN Hansa Yellow Medium:
Hue 2.0Y, Value 7.75, Chroma 16.8 can also be written in the form H V/C, known as Munsell Notation: 2Y 7.75/16.8.

It is important to note that Munsell values are commonly determined by directly comparing a given color swatch to pre-made color chips organized into an atlas. This is a very different approach than the spectrophotometer readings CIE Lab relies on exclusively. Even if a great deal of care is taken to examine the colors against a neutral background, and under controlled lighting conditions, there are built-in limitations with this kind of system. A certain amount of subjective judgment and interpretation are inevitable whenever human perception is involved, so different operators can easily come to different conclusions. The Munsell system is also bound by the limited gamut of the colors used to generate the chips, rather than the range of human perception. And of course, the swatches need to be repeatedly and accurately manufactured to a tight standard, can degrade over time, and each person using this system must have a copy of the color atlas as a reference. But all that aside, it still remains one of the easiest models to grasp and, within its limits, is a valuable tool for anyone wanting a fairly accurate, flexible, and intuitive tool for assessing colors.

Other Color Models
Device Dependent – RGB and CMYK
While CIE Lab and Munsell are wonderful examples of global order and comprehensible structure, one doesn’t have to stray very far before bumping up against the far more unruly world of computer screens and printed pages. Monitors, televisions, printers, scanners, digital cameras, and even cell phones, all come with some form of native color space. Likewise, most of us have experienced the sometimes dramatic and unpredictable changes in color when the same image is printed on different devices or viewed on different screens. Each device seems to come with its own specs, its own way of interpreting the same color data, and its own particular limitations. Each manufacturer, we come to realize, sees color a little differently.

For the most part, all of these devices work with some variation of RGB or CMYK color models, which operate very differently from the CIE Lab and Munsell systems we first introduced. In those, a set of well-defined variables delineated a specific unique color. By contrast, RGB and CMYK values describe ratios between generalized inputs: additive primaries of light, in the case of RGB, or subtractive primaries for inks in CMYK. It’s as if, rather than designating a color, they provide something more akin to a recipe for mixing it. Since each device can only interpret those values based on its own parameters, the systems have eventually become known as ‘device-dependent.’ (Fraser, 1995)

RGB
RGB, which stands for Red Green Blue, is the color space most commonly encountered on display screens, such as computer monitors and televisions, and describes color in terms of differing amounts of red, green or blue light. Geometrically it is often depicted as a cube with Red Green and Blue occupying three vertices, white, black, and the secondary mixtures occupying the others. Each position within the space is defined by three values, often expressed by a range from 0-255, that represent the relative amount of the three primaries. For example, the purest red that can be specified in this system would be written: R:255 G:0 B:0 However, just having a set of RGB values will not tell someone which color they would see. For that, a person would need to know which particular version of the RGB system the values belong to and then display them on a screen calibrated to that specific space as well. Even in the example of a pure red, where it will always produce the most saturated red any particular device can deliver, the question remains – what red is that? Depending on whether it is a cell phone, HDTV, LCD Monitor, or a CRT, and adding in all the other variables one can imagine, there is simply no way to truly know.

CMYK
Widely known by anyone involved in graphics or printing, some version of this color space is used in nearly every color printing process – from the fanciest reproductions to the lowliest inkjets. In its most common form the system relies on Cyan, Magenta, Yellow and Black printing inks to achieve its range of colors. CMYK values are stated as percentages that give the degree of density for each ink. A color similar to Burnt Sienna, for example, could be written as: C25, M70, Y90, K35. However, just as we saw with RGB, these values only represent amounts of generalized colorants and do not describe a particular hue within an unvarying color space. In fact, in this aspect, CMYK can be even more fraught with uncertainty than RGB. The type of printer, which ink system is used, the substrate, and even the environmental conditions, all impact the ability of this model to reliably reproduce a specific color. Even the nature of the inputting device will have its effect on the outcome. Lastly, by being restricted to a relatively small set of inks, CMYK has by far the most limited gamut within the color models we have examined. Even when the number of inks are expanded, as in many of the current systems using six and eight colors, it still falls short of what can be expressed elsewhere.

Named Color Spaces
There is any number of proprietary systems, sometimes referred to as ‘ named color spaces,’ which are not actually color models in any true sense of the word. None of them attempt to generate a complete, general color space, rendering them ineffective for measuring or translating colors outside of their fixed systems. Rather, they are simply large compilations of precisely colored chips organized into numerous swatch books. They serve as a form of visual shorthand, a way for designers, graphic artists, and other professionals to quickly reference and specify pre-made standardized collections of solid colors. (Apple, 2005)

Although these systems might not be particularly useful for creating complex images they do allow for a tremendous amount of control since their formulations are specifically tied to a particular company’s ink system and even the type of paper stock and coatings that are used. By controlling the process to this degree of detail, they avoid the inherent problems we saw earlier with CMYK, where parameters were left open and variable.

Pantone®
Pantone is by far the most widely known name in this field and its Pantone Matching System (PMS) is its exclusive set of over a thousand standardized, precisely printed colors. The entire system is available on coated, uncoated, and matte card stock, along with the corresponding specialized ink formulations to allow printers to accurately reproduce them. These color guides have become internationally recognized as a standard reference and theoretically anyone, anywhere, can reproduce each color by looking it up in the corresponding Pantone book and mixing a set of proprietary inks in the given ratios.
From time to time GOLDEN gets a request to match a specific Pantone number, which our Custom Lab is happy to do. You can also find the PMS designation corresponding to each of our colors by going to any of the virtual colorcharts on our Web site and clicking on a desired swatch. The large drawdown you will then see lists the Pantone Matching System value, such as ‘PMS 123′, which aligns with our Hansa Yellow Medium.

Other Terms in Color Language
Beyond these broad systems of classification are a host of other terms artists frequently use that are worth noting:

Organic / Inorganic
While these categories refer specifically to the chemistry of a pigment and not its color, artists will sometimes use this distinction when selecting which ones to buy. The organic pigments are synthetically produced from petroleum and natural gas and derive their name from the fact they are based on carbon-compounds, which is the defining characteristic of organic chemistry in general. By comparison the inorganics are generally composed of crystals of metal oxides, and while many are mined, they can just as likely be synthetically manufactured; Ultramarine Blue and the synthetic iron oxides being two such examples.

This last point is important because a common misperception is that the terms Organic / Inorganic also describe a division between synthetic, engineered pigments and more basic ones derived from natural sources. The truth is, the vast majority of pigments – even native earth colors – have been highly processed through large-scale industrial plants by the time they arrive to a manufacturer’s doorstep.

Transparent / Opaque
Currently there are no standards for measuring transparency or opacity and most ratings, including ours, are made through examining similarly prepared samples and rating them relative to one another. The difficulty here is that many pigments that are inherently transparent will seem quite strong and opaque if used full-strength from the tube, especially when made with a high pigment load. Phthalo Blue is an excellent example of this. In a 10 ml drawdown it was ranked on par with more commonly opaque colors such as Cobalt Blue, Pyrrole Red, and Cadmium Orange. However, when applied very thinly, mixed with a gel, or extended with a medium, Phthalo Blue shows another side and becomes a transparent and beautiful glazing color.

Masstone / Undertone
The masstone of a paint is simply its color when applied thickly enough to completely cover a surface. No other colors from below show through. Undertone, by contrast, is visible when we spread the color very thinly over a white surface. Certain colors, such as the Cadmiums and Cobalts, have similar masstones and undertones. With the transparent organic colors like the Quinacridones or Phthalos, the undertone can be quite different from what might be expected.

Tint Strength
This is the ability of a color to change the character of another color. We determine this by adding the same amount of Titanium White to each color and observing the resulting strength of the color mixture. Weaker tinting colors create light pastel mixtures. Stronger tinting colors create darker mixtures.

Other Resources:
The essential literature on color theories, nomenclature, and models is too vast to list here. A complete listing of books, Web sites, and other resources referenced in this article will be made available at www.goldenpaints.com/justpaint/index.php.

In addition, a few years ago GOLDEN was proud to be the force behind the publication of two books on color theory and traditions. These continue to be available from us and provide a wonderful place to start if you are interested in an approachable survey and introduction into the rich history of these fields.

Bibliography for Article

Annual Book of ASTM Standards, Vol. 06.02, 2003, ASTM International, pp.475-484

Boker, S. “The Representation of Color Metrics and Mappings in Perceptual Color Space”, Department of Psychology, University of Virginia, Dec. 2006, http://people.virginia.edu/~smb3u/ColorVision2/ColorVision2.html

CIE, International Commission On Illumination, Dec. 2006, http://www.cie.co.at/

Apple, Color Management Overview, Developer Connection, 2005, Dec. 2006,
http://developer.apple.com/documentation/GraphicsImaging/Conceptual/csintro/csintro.pdf

Color Systems in Art and Science, ed. Klaus Stromer, tr. R. Cassada, Edition Farbe/Regenbogen Verlag, Klaus Stromer, 1999

Colour Index, 3rd Edition, 1982, Pigments and Solvent Dyes, The Society of Dyers and Colourists

Colour Index International, Dec. 2006,  http://www.colour-index.org/

Fraser, B., “Color Under Control”, Adobe Magazine, September/October 1995, Dec. 2006,http://www.adobe.com/products/adobemag/archive/pdfs/9509febf.pdf

Kuehni, R. G, Color Space and its Divisions, Color Order from Antiquity to the Present,  Hoboken, New Jersey: Wiley , 2003

Landa, E. T. and Mark D. Fairchild, “Charting Color From the Eye of the Beholder”, American Scientist, Vol. 93, September/October 2005, pp. 436-443

Lindbloom, B.J., Dec. 2006, http://www.brucelindbloom.com/

McEvoy, B., Dec. 2006, http://handprint.com/HP/WCL/water.html

Pascale, D., “A Review of RGB Color Spaces…from xyY to R’G’B’”, The BabelColor Company, 2003, Dec. 2006,http://www.babelcolor.com/download/A%20review%20of%20RGB%20color%20spaces.pdf

Poyton, C., “Color Management Technology for Workstations”, Sun Expo ’92, Manchester, U.K., September 10, 1992, Dec. 2006, http://www.poynton.com/PDFs/Sun_Color_Mgmt.pdf

Rodney, A., “The Role of Working Spaces in Adobe Applications”, Technical Paper, Adobe Systems Inc., 2006, Dec. 2006, http://www.adobe.com/digitalimag/pdfs/phscs2ip_colspace.pdf

Susstrunk, S., Robert Buckley, and Steve Swen. “Standard RGB Color Spaces”, In Color Science, Systems, and Applications, volume 7, pages 127–134, Scottsdale, Arizona, November 1999. IS&T’s 7th Color Imaging Conference.

Additional Publications of interest
Berns, Roy S., Billmeyer and Saltzman’s Principles of Color Technology, 3rd Edition, New York, John Wiley & Sons, 2000.

Gage, John,  Color and Culture: Practice and Meaning from Antiquity to Abstraction, UC Press, 1993

Hardin , C.L.,and Luisa Maffi , Eds., Color Categories in Thought and Language, Cambridge University Press, 2002

Hardin , C.L , Color for Philosophers: Unweaving the Rainbow, Hackett Publishing Company, 1988

Stromer, Klaus, Ed., Traditions and Colors, Edition Farber/ Regenbogen Verlag, 2000 (available from Golden Artist Colors)

Stromer, Klaus, Ed., Color Systems in Art and Science, Edition Farber/ Regenbogen Verlag, 1999 (available from Golden Artist Colors)

Online Resources

Color Space Conversion Tools
http://www.brucelindbloom.com/index.html?Calc.html
http://www.colorpro.com/info/tools/converters.html
http://www.cs.rit.edu/~ncs/color/a_spaces.html
http://www.easyrgb.com/calculator.php
http://www.efg2.com/Lab/Library/Color/

General Color Theory and Color Space
http://www.cis.rit.edu/mcsl/online/cie.html
http://www.colorsystem.com/?lang=en
http://www.efg2.com/Lab/Library/Color/
http://www.handprint.com/HP/WCL/wcolor.html