Reallocation of Narrow Bands of Spectral Energy: the Effect on Brightness Perception and Color Preference

Open Access
Wilkerson, Andrea Marie
Graduate Program:
Architectural Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
June 25, 2013
Committee Members:
  • Kevin William Houser, Dissertation Advisor
  • Richard George Mistrick, Committee Member
  • Jelena Srebric, Committee Member
  • Eric Loken, Committee Member
  • Robert Davis, Special Member
  • Light
  • Engineering
  • Perception
The research furthers the understanding of human psychophysical response to spectra and the role of spectral tuning in source optimization by examining the effect of spectral modification on the perception of brightness and color preference, building upon the work of William Thornton, Kevin Houser, and their associates. Thornton first theorized over forty years ago that the perception of brightness could be increased by placing energy in the spectral regions at and near 450, 530 and 610 nm, which he called the prime color spectral regions. He further identified the areas at and near 495 and 575 nm as having deleterious effects on color preference and termed these anti-prime color regions. Hurvich’s opponent signal sensitivity curves have a relative sensitivity of zero where the sensitivity switches between blue and yellow at 500 nm and switches between red and green at 480 and 580 nm. This dissertation focused on removal of optical radiation near anti-prime spectral regions, where the opponent signal relative sensitivity is zero, and resulting perceptions of brightness and color preference. This dissertation examined brightness and preference differences between a reference spectrum and three test spectra that had less optical radiation in the spectral regions near 500, 580, or 500 and 580 nm. Two experimental methodologies were employed, forced choice discrimination at fixed illuminance levels, and brightness matching. Energy was removed from the regions of 500 and 580 nm, with some energy added to other regions of the visible spectrum as necessary to maintain consistent chromaticity. Brightness discrimination and brightness matching experimental methodologies were both employed in order to examine if these methodologies would produce comparable results. The research aimed to answer the following questions: 1. Do spectra with energy reallocated from the regions of either 500, 580, or 500 and 580 nm appear brighter than a nearly continuous reference spectrum? 2. Do the experimental methodologies side-by-side brightness matching and side-by-side brightness discrimination (forced choice) produce comparable results? 3. Does the reallocation of energy from the regions of 500, 580, or 500 and 580 nm change preference for a person’s own skin tone? The test spectrum with energy reallocated near the spectral region of 500 nm (SPD500) and the test spectrum with energy reallocated near the spectral region of 500 and 580 nm (SPD500,580) appeared brighter than the reference spectrum (SPDRef) at equal illuminance, reaching statistical significance for both the brightness matching and brightness discrimination experiments. No statistical difference in brightness was found between the test spectrum with energy reallocated near the spectral region of 580 nm (SPD580) and SPDRef. The side-by-side brightness matching experimental methodology and the side-by-side brightness discrimination experimental methodology produced comparable results. The rank order of the reference and test spectra was similar for both experimental methodologies. The illuminance ratio of SPD500,580 versus SPDRef at equal perceived brightness was 94.2% for the brightness discrimination experiment and 95.6% for the brightness matching experiment. The ratio of SPD500 versus SPDRef was 95.5% and 98.1% for the discrimination and matching experiments, respectively. The ratios from the two experiments were both found to be significantly different from each other. The SPD580 versus SPDRef ratios resulting from the two experiments were not significantly different. The brightness matching ratios were closer to unity, unity indicating no difference, for every spectrum comparison versus the brightness discrimination ratios. SPD580 was preferred versus SPDRef for 72.8% of the comparisons between the two spectra. Similarly, SPD500,580was preferred versus SPDRef for 70.0% of the comparisons between the two spectra. SPD500,580 and SPD580 were both significantly different than SPDRef at an alpha level of 0.01. The rank order resulting from the preference experiment matched the rank order of the following metrics: Gamut Area Index, Color Quality Scale 9.0 Qg, Farnsworth Munsell Gamut, Color Preference Index, Color Discrimination Index and Feeling of Contrast Index. SPD580 had the highest preference rank, however it had the lowest CIE Color Rendering Index and R9 rank. The results of all three experiments show that SPD500,580 is the superior spectrum in comparison to SPDRef, SPD500 and SPD580. SPD500,580 appeared brighter at equal illuminance than SPD500 and SPD580 when compared to SPDRef. SPD500,580 was also preferred versus SPD500 and SPDRef, and was not significantly different than SPD580. The LER of SPD500,580 was 8.9% greater than SPDRef and SPD500,580 required an average of 5.1% less lumens than SPDRef at equal brightness appearance. When the difference in LER is combined with the results of the brightness experiments, SPD500,580 provides a 14% improvement in efficacy versus SPDRef at equal perceived brightness. The findings of this research clearly demonstrate, with comparable results from brightness matching and discrimination methodologies, that the spectrum with spectral energy reallocated in the regions of 500 and 580 nm was perceived as brighter at equal illuminance and was preferred for rendering a person’s own hand.