Reproducibility comparison among multiangle spectrophotometers

New color‐measuring instruments known as multiangle spectrophotometers have been recently created to measure and characterize the goniochromism of special‐effect pigments in many materials with a particular visual appearance (metallic, interference, pearlescent, sparkle, or glitter). These devices measure the gonioapparent color from the spectral relative reflectance factor and the L*a*b* values of the sample with different illumination and observation angles. These angles usually coincide with requirements marked in American Society for Testing and Materials (ASTM) and Deutsches Institut Für Normung standards relating to the gonioapparent color, but the results of comparisons between these instruments are still inconclusive. Therefore, the main purpose of this study is to compare several multiangle spectrophotometers at a reproducibility level according to ASTM E2214‐08 guidelines. In particular, we compared two X‐Rite multi‐gonio spectrophotometers (MA98 and MA68II), a Datacolor multi‐gonio spectrophotometer (FX10), and a BYK multi‐gonio spectrophotometer (BYK‐mac). These instruments share only five common measurement geometries: 45° × ?30° (as 15°), 45° × ?20° (as 25°), 45° × 0° (as 45°), 45° × 30° (as 75°), 45° × 65° (as 110°). Specific statistical studies were used for the reproducibility comparison, including a Hotelling test and a statistical intercomparison test to determine the confidence interval of the partial color differences ΔL*, Δa*, Δb*, and the total color difference ΔE*_ab. This was conducted using a database collection of 88 metallic and pearlescent samples that were measured 20 times without the replacement of all the instruments. The final findings show that in most measurement geometries, the reproducibility differences between pairs of instruments are statistically significant, although in general, there is a better reproducibility level at certain common geometries for newer instruments (MA98 and BYK‐mac). This means that these differences are due to systematic or bias errors (angle tolerances for each geometry, photometric scales, white standards, etc.), but not exclusively to random errors. However, neither of the statistical tests used is valid to discriminate and quantify the detected bias errors in this comparison between instruments. © 2012 Wiley Periodicals, Inc. Col Res Appl, 38, 160–167, 2013.