Visually matching daylight fluorescent lamplight with two primary sets

Kevin W. Houser, Xin Hu

Research output: Contribution to journalArticlepeer-review

19 Scopus citations


Thornton reported discrepancies in chromaticity and power content in Maxwell-method visual matches when using different primary sets. He found particularly large chromaticity differences when a broadband white light was matched with what he refers to as an antiprime (AP) primary set [blue-green, yellow, deep red, as, for example 497-579-653 nm, mean full-width half-maximum (FWHM) = 15.6 nm] and comparatively smaller chromaticity differences with a primecolor (PC) primary set (blue-violet, green, orange-red, as, for example, 452-533-607 nm, mean FWHM = 15.6 nm). In related work, Ábrahám et al. found similar chromaticity discrepancies when subjects reported a visual match, partially attributing the errors to age-related changes in the specular transmittance of the lens. This present study was carried out to explore these findings and to investigate individual differences in color matching. Thirty-nine subjects made two matches in a horizontally bisected 10° field. The bottom field was illuminated with daylight fluorescent lamplight and the top field was illuminated with one of two primary sets that were selected to be similar to those used by Thornton and Ábrahám et al.: (1) PC = 453-533-619 nm with a mean FWHM of 18.3 nm and (2) AP = 493-581-657 nm with a mean FWHM of 23.7 nm. Statistical data screening revealed two outliers that were removed from further analysis. The PC primary set required 5% less radiant power per unit area to match the daylight fluorescent reference and the AP set required 117% more. Thirty-six of the 37 matches with the PC primary set were within a 1O-step MacAdam ellipse that was centered at the daylight fluorescent reference. In contrast, just 2 of the matches with the AP set were within a 10-step ellipse, and 24 of the matches were outside of a 20-step ellipse. These trends are similar to those reported by Thornton and later by Ábrahám et al. The primary set and the interaction between primary set and age were statistically significant factors. Gender and subject knowledge were not significant.

Original languageEnglish (US)
Pages (from-to)428-437
Number of pages10
JournalColor Research and Application
Issue number6
StatePublished - Dec 2004

All Science Journal Classification (ASJC) codes

  • Human Factors and Ergonomics
  • General Chemistry
  • General Chemical Engineering


Dive into the research topics of 'Visually matching daylight fluorescent lamplight with two primary sets'. Together they form a unique fingerprint.

Cite this