Archeologists Examine a Color-Matching Device

Article: Human eye beats machine in archaeological color identification test
Source: University of Florida Museum of Natural History
Published: February 9, 2021

X-Rite Capsure device (right), along with
color chips (left) and clay samples (top)

Color perception, something most of us take for granted, is ultimately a subjective and individual experience. The ability to perceive and identify colors, however, becomes functionally relevant in some professions, such as interior design and cosmetology, or even archeology as in this case. Because color perception is subjective, scientists were excited when a handheld color-matching gadget called the X-Rite Capsure came on the market. Made by the same company that owns Pantone, they had hoped that it would offer a consistent way of determining color, free of human bias as well as variations in lighting, sample quality, and observer perspective. However, a study by archeologists, to whom the device was marketed to identify the color of artifacts and soil samples, found that the X-Rite Capsure often misreads colors readily distinguished by the human eye when tested against the Munsell color system, the current archeological standard for identifying colors using a binder of 436 unique color chips. The study specifically tested the Capsure's readings for all 436 color chips as well as 140 pottery briquettes for the three elements of Munsell’s system: a color’s general family (hue), intensity (chroma), and lightness (value). The details of how the X-Rite Capsure differed from human observers in determining color can be found in the articles. These differences are not functionally trivial, however, since identifying subtle differences in color gives information about a sample's composition, origins, and history of use. Highlighting a device that is functionally relevant within the field of archeology, the researchers argue, “We need to pay really close attention and record how we’re describing color in order to make good data. Ultimately, if we’re putting bad color data in, we’re going to get bad data out.” Because the device is nonetheless internally consistent in its color identifications, work remains to optimize the advantages of such devices and artificial intelligence to match human perception, arguably difficult for a quality as complex as color.

My rating of this study: ⭐

Bloch LC, Hosen JD, Kracht EC, et al. "Is It Better to Be Objectively Wrong or Subjectively Right? Testing the Accuracy and Consistency of the Munsell Capsure Spectrocolorimeter for Archaeological Applications." Advances in Archaeological Practice.  9(2):132-144.  8 February 2021. https://doi.org/10.1017/aap.2020.53