Understanding the Spectral Signature of Pigments

Summary: Understanding color bias is a key to successfully mixing the color an artist desires.  However, explaining bias is a more difficult task. The use of spectral data unlocks some of the mystery surrounding how color mixing behaves and how we approach color mixing.

I have what I consider to be an above-average interest in understanding how colors function. Some of this is evident by the hundreds of color swatches I have captured, as well as a large number of spectral readings I have taken of color paint-outs from both 19th-century and contemporary paint-making companies.

Example of Spectral Signature of Cadmium Green pigment

So, let's look at color bias in greater detail by explaining how a pigment is analyzed through the use of a spectrophotometer.  A spectrophotometer captures the spectral properties of a color and records a vast amount of data from the sample, including its position in color space (Lab), its lightness versus darkness, and a reading of how much color is reflected in 31 segments of the visible light spectrum. Why 31?  In the case of the spectrophotometer I am using, the instrument reads 31 segments of reflected light starting at 400 nanometers (nm) every 5 nm until it ends at 700 nm, where the human eye stops seeing visible light energy.  From 400 nm to 700 nm in 10-nm increments, it can effectively map the range of colors from violet to blue, green, yellow, orange, and ending in red.  Below 400 nm (violet) is ultraviolet light, and beyond 700 nm, the infrared spectrum begins.

Bear in mind that every color we see has a percentage of all the colors.  Just because they are not visible to our eyes does not mean they do not influence the attributes a pigment has when mixed with another pigment, which also has its own qualities.

But, you might wonder, why do we need 31 data points to analyze a color?  When we see green paint, wouldn’t we only need the spectral segment readings that represent the range from 500 nm to 600 nm to indicate a green hue? The answer is “No,” because like all colors, green has some violet, blue, yellow, and red components in it. However, they are not visible to us because the 500 nm to 600 nm range has a higher percentage of reflectance, or in simple terms, the green wavelength dominates the other colors in a sample. (Side note: if laser light is spectrally analyzed, it is so pure that it occupies a single nm of visible light.)

We could convert each segment of spectral data into musical notes, with low notes starting at 400 nm, progressively going higher in pitch, ending at 700 nm, coupled with the higher the percentage of reflectance, the louder the sound would emanate from that portion of the spectrum.   While aesthetically and musically, a color converted into sound might make little sense, as different colors are converted to sounds, patterns would emerge.

Generally, violet and blue hues would have low bass sounds, greens would have middle range sounds, and yellow and red hues would have high-pitched sounds. I am using this example of converting spectral segments into sounds and softness/loudness of sound for a reason.  This conversion would auditorily indicate color bias because many pigments have a combination of two or more fairly dominant spectral peaks, thus making a confusing, clashing acoustic sound. No harmonious sound would emanate from many pigments, that have a strong color bias.

Now let’s examine the visual aspects of color bias. Breaking down color spectrally allows us to look “behind the curtain,” so to speak, and see the percentages a color reflects.  While most artists understand color bias using knowledge based on thousands of times they mix paint to achieve a desired color, novice artists can get stymied and frustrated when they mix colors that fall short of the desired outcome.

An artist will never get a vibrant hue when two colors they select to mix together have a high degree of opposite (complementary) colors in them that are not apparent to the naked eye.  The complements cancel each other out, neutralizing the mixture, thus falling short of the hue that was intended.

Selecting a palette is a personal preference.  The Internet is filled with folks who argue about what colors are needed to mix a broad range of hues. People will insist that a palette should be anchored with cyan (phthalocyanine blue), yellow (hansa yellow), and magenta (quinacridone) along with white.  Note that these are all transparent organic pigments.  Other artists will select ultramarine blue, cadmium yellow, and cadmium red as their choice of primary colors. Every choice has positive and negative aspects. It will be hard to mix muted hues with the organic pigments and equally difficult to mix vibrant colors with the more traditional ultramarine and cadmium colors. Again, color bias plays a key role in how pigments behave when mixed.

So now that I have channeled both Philip Glass and Jacob Collier in a convoluted mashup of music and color to provide an example of color bias, I hope you have gained something from this essay on the inner workings of spectral reflectance.  Be sure to check out the Syntax of Color page titled Color Reflectance to see visual examples of primary and secondary colors to get an understanding of color bias in paints readily available to artists.

Syntax of Color

Key Words: Spectral properties, color, pigments, color bias, paints, spectrophotometer, artworks, artists, hues, color mixing, palettes.