A MODIFICATION OF THE MESOPIC LUMINANCE EQUATION IN THE CIE MESOPIC PHOTOMETRY SYSTEM FOR A BETTER REPRESENTATION OF THE MESOPIC VISION IN THE WHOLE VISUAL FIELD

Kyota Morimoto, Dong Nie, Yukio AkashiJournal of Environmental Engineering (Transactions of AIJ) 2017 · DOI

Abstract

The CIE mesopic photometry system (CIE, 2010) was based on target detection performances at an eccentricity angle of 10 degrees. It has not been confirmed whether the MES2 system can be applied to other eccentricity angles in the peripheral visual field. To verify the validity of the CIE mesopic photometry system and modify the system if necessary, this study conducted a target detection experiment by employing six subjects. This experiment measured the subjects' reaction times for targets presented at eccentricity angles of 5, 10, 15, and 25 degrees from the center of the visual field. This experiment used light levels between 0.005 cd/m² and 2 cd/m² and two lamp spectral power distributions of yellow LED (S/P ratio=0.10) and white LED (S/P ratio = 1.86). Based on the results of the experiment, this study found that the visual field within 5 degrees from the center of the visual field can be treated as the central visual field where only cones function even under mesopic light levels. The CIE mesopic photometry showed good correlations between mesopic luminances and reaction times for the two SPD conditions at eccentricity angles of 10 and 15 degrees. However, the luminance borderline between the photopic vision and the mesopic vision at the eccentricity angle of 25 degrees seemed to be higher than 5 cd/m², which the CIE MES2 defined. Therefore, this study attempted to modify the mesopic luminance formula only for the eccentricity angle of 25 degrees.

Introduction

Most lighting standards are traditionally categorized into photopic vision (daytime, cone-dominated) or scotopic vision (darkness, rod-dominated). However, many critical real-world scenarios—such as night driving or street lighting—occur in the mesopic range, where both rods and cones are active. In 2010, the CIE introduced the MES2 system to calculate mesopic luminance, but this model was primarily calibrated using data from targets at a 10° eccentricity (offset from the center of vision).

Distribution of rod and cone photoreceptors across the human retina relative to visual angle.

As shown in Figure 1, the density of rods and cones changes drastically as one moves from the fovea to the periphery. Because human retinal sensitivity varies so widely across the eye, a model tuned for 10° may not accurately represent visual performance at other angles, such as 5° or 25°. This study investigates whether the current international standard holds up across the entire field of view or requires adjustment for the far periphery.

Methods

The researchers conducted a target detection experiment with six subjects to measure reaction times (RT) across a range of conditions:

  • Eccentricity: Targets were positioned at 5°, 10°, 15°, and 25° from the visual center.
  • Luminance Levels: Background brightness varied from 0.005 cd/m² to 2 cd/m².
  • Light Spectra: Two light sources with distinct Scotopic/Photopic (S/P) ratios were used: a "Yellow" LED (S/P = 0.10) and a "White" LED (S/P = 1.86).
  • Performance Metric: Reaction time served as the proxy for visual performance, with faster times indicating better visual representation by the luminance model.

The study generated nearly 7,000 data points (1,152 trials per subject). The team utilized Vaughan’s equation to model the relationship between reaction time and luminance, comparing the experimental results against the existing MOVE and USP models that informed the original CIE MES2 standard.

Results and Discussion

The experimental data revealed significant insights into how visual performance shifts across the retina under mesopic conditions.

The 5° "Cone Zone"

At a 5° eccentricity, the visual system was found to behave almost entirely like photopic vision, even at lower light levels. The results suggest that within this central 5° circle, rod contribution is negligible enough that standard photopic photometry remains sufficient.

Robustness of MES2 in the Mid-Periphery

The current CIE model demonstrated excellent correlation ($R^2 > 0.9$) for the 10° and 15° eccentricities. This validates the robustness of the standard for "near-peripheral" vision, which is commonly used in street lighting design.

The 25° Shift in the Far Periphery

In the far periphery (25°), the transition from "daytime" to "twilight" vision occurs at higher light levels than the standard assumes. Because rod density is significantly higher in the periphery, rods begin to influence vision even when the environment is relatively bright.

Reaction times as a function of background luminance for white and orange LED light sources.

As illustrated in Figure 4, reaction times decrease as background luminance increases. Notably, the performance gap between the White LED (higher S/P ratio) and the Orange LED widens at lower luminance levels, highlighting the increased rod contribution. For the 25° eccentricity, the researchers found that the "borderline" where mesopic vision begins is likely closer to 10 cd/m², rather than the current standard of 5 cd/m².

Conclusion

This study demonstrates that while the CIE MES2 system is effective for the mid-periphery, it underestimates the transition to mesopic vision in the far periphery. To address this, the researchers proposed a modification of the mesopic luminance parameters for 25° eccentricity, changing the standard coefficients ($a=0.767, b=0.333$) to $a=0.697, b=0.303$. Future work with larger sample sizes and a broader range of light spectra and higher luminance levels (up to 10 cd/m²) will be necessary to further refine these boundaries and ensure lighting standards accurately reflect human visual performance across the entire visual field.

Summary updated

Abstract

The CIE mesopic photometry system (CIE, 2010) was based on target detection performances at an eccentricity angle of 10 degrees. It has not been confirmed whether the MES2 system can be applied to other eccentricity angles in the peripheral visual field. To verify the validity of the CIE mesopic photometry system and modify the system if necessary, this study conducted a target detection experiment by employing six subjects. This experiment measured the subjects' reaction times for targets presented at eccentricity angles of 5, 10, 15, and 25 degrees from the center of the visual field. This experiment used light levels between 0.005 cd/m2 and 2 cd/m2 and two lamp spectral power distributions of yellow LED (S/P ratio=0.10) and white LED (S/P ratio = 1.86). Based on the results of the experiment, this study found that the visual field within 5 degrees from the center of the visual field can be treated as the central visual field where only cones function even under mesopic light levels. The CIE mesopic photometry showed good correlations between mesopic luminances and reaction times for the two SPD conditions at eccentricity angles of 10 and 15 degrees. However, the luminance borderline between the photopic vision and the mesopic vision at the eccentricity angle of 25 degrees seemed to be higher than 5 cd/m2, which the CIE MES2 defined. Therefore, this study attempted to modify the mesopic luminance formula only for the eccentricity angle of 25 degrees.
BibTeX
@article{morimoto2017modification,
  author = {Kyota MORIMOTO and Dong Nie and Yukio Akashi},
  title = {{A MODIFICATION OF THE MESOPIC LUMINANCE EQUATION IN THE CIE MESOPIC PHOTOMETRY SYSTEM FOR A BETTER REPRESENTATION OF THE MESOPIC VISION IN THE WHOLE VISUAL FIELD}},
  journal = {Journal of Environmental Engineering (Transactions of AIJ)},
  year = {2017},
  doi = {10.3130/aije.82.113},
}
Figure