A round-robin study of color measurement on the General Electric recording spectrophotometer was carried out by 15 participating laboratories, using transparent glass filters and opaque plastic and glass specimens. In terms of CIE Y and x, y for Illuminant C (daylight), 95% confidence limits for individual measurements averaged ±1.5% for Y and ±0.0165 for x and y if all the data were included. Elimination of results known or suspected to contain systematic errors reduced these values to ±0.57% for Y and ±0.0119 for x and y. The confidence limits for x and y are much greater for low-luminance than for high-luminance specimens, but those for Y showed relatively little dependence on luminance. The short-time repeatability of the spectrophotometer averaged ±0.09% for Y and ±0.0007 for x and y, while its reproducibility over a 14-month period averaged ±0.62% for Y and ±0.0028 for x and y.
The results of this round-robin study clearly show, as in the repeatability and reproducibility studies, that the GE spectrophotometer, when suitably calibrated and operated, can still be considered the referee instrument for accurate color measurement. Serious doubt is cast, however, on the suitability of the procedures generally practiced for the calibration and operation of this instrument.
© 1965 Optical Society of AmericaFull Article | PDF Article
OSA Recommended Articles
Fred W. Billmeyer
Appl. Opt. 8(4) 775-783 (1969)
Fred W. Billmeyer, Ellen D. Campbell, and Robert T. Marcus
Appl. Opt. 13(6) 1510-1518 (1974)
Eugene C. Rhodes and Fred W. Billmeyer
Appl. Opt. 8(4) 769-773 (1969)