JONATHAN BRYANT, CORI JACKSON, AND JAE YONG SUK
Lighting is essential to visual performance but also human health, mood, and overall well-being. A growing body of evidence underscores the importance of both the quantity and quality of light in shaping these outcomes. Traditional energy programs, such as those in California’s Building Energy Efficiency Standards or federal ENERGY STAR specifications, have historically emphasized luminous efficacy. While this approach has driven substantial improvements in energy performance, it often overlooks color fidelity and visual comfort, which are essential for user satisfaction, positive functional outcomes, and sustained use of LED lighting in homes and businesses.1,2
While conventional LEDs often prioritize efficacy at the expense of color quality, recent advances in full-spectrum, high-fidelity LED lighting offer a compelling alternative. These next-generation LEDs more closely mimic natural daylight, delivering improved color rendering and higher TM-30-20 fidelity scores. The result is better visual comfort, enhanced performance in color-critical tasks, and even psychological benefits linked to perceived naturalness.
One key advantage is that full-spectrum LEDs may enable occupants to feel visually satisfied at lower light levels, opening the door to energy savings without compromising experience. To explore this opportunity further, the California Lighting Technology Center (CLTC) at the University of California, Davis, in partnership with Seoul Semiconductor, conducted a controlled study to evaluate whether high-fidelity lighting can reduce preferred light levels while preserving visual performance. By isolating the effects of spectral quality on both subjective preference and functional color discrimination, the study aimed to provide actionable insights into strategies for improving occupant lighting experience while supporting energy efficiency goals in residential and commercial settings.
The study included two experiments combined with a set of cross-sectional surveys. The first experiment addressed lighting preference using a within-subjects design to determine if high-fidelity, full-spectrum lighting produced a significant effect on preferred lumen output. The second test executed a color sorting task under the reference source and then the full-spectrum source to capture impacts of the lighting treatments on color discrimination.
Following these experiments, participants completed three cross-sectional surveys to capture their subjective experiences, perceived visual comfort, and comparative preferences between the two lighting treatments at the time of the study.
This experiment took place in a controlled laboratory environment designed to simulate a typical residential vanity space, such as in a bathroom or bedroom (Figure 1). The room measured 6 ft by 9 ft by 9 ft with neutral white walls used to minimize any influence of surface color on lighting perception. A high-quality mirror was installed to replicate a standard vanity set-up and reduce color shift due to reflected light.
Photo: Veronica Then, California Lighting Technology Center
Figure 1. The experimental test space.
The study used two LED sources: a high-fidelity, full-spectrum LED with a TM-30-20 Fidelity Index (Rf) of 97 and a conventional LED with an Rf of 84 (Table 1). Both were mounted behind the vanity’s acrylic diffusers to ensure uniform illumination across the vanity surface. Researchers, located outside the test room, controlled source switching to maintain consistency.
Table 1. LED Performance Characteristics
Although the spectral distributions of the full-spectrum and conventional LEDs were different, it was critical that both light sources appeared identical in color to participants to eliminate any bias associated with perceived light color preference—common in human visual perception. The experiment controlled CCT at 3100K for both light sources. This was done using the 10-deg color space, where Xn,F,10 and Yn,F,10 chromaticity coordinates were calculated from integrating sphere measurements. Researchers iteratively refined both sources until their chromaticity coordinates were closely matched, ensuring minimal perceptible difference in hue or warmth.
In addition, to isolate the effects of spectral quality rather than illuminance level, both lighting systems were calibrated to deliver an equal illuminance of 35 footcandles (~377 lux) on the task surface during the color sorting experiment. This ensured that performance differences could be attributed to spectrum fidelity rather than brightness.
To evaluate the impact of high-fidelity, full-spectrum lighting on preferred light levels, a paired-samples t-test was completed. The intervention involved a change in ambient lighting conditions used during task performance. In the reference condition, participants were exposed to the conventional LED lighting. The lighting treatment replaced the conventional LED source with the high-fidelity, full-spectrum LED alternative.
The test compared the mean preferred lumen output recorded before and after the lighting intervention across 50 participants (25 females and 25 males evenly distributed in different age groups from 18 to over 60). Participants completed a series of interactive lighting preference trials designed to measure individual perceptions of ideal lighting conditions. The order of exposure to each light source was determined in advance and randomized to minimize sequence bias.
Upon entering the test space, participants sat at a vanity station facing a mirror. The vanity lighting was set at its minimum level for the first test when participants entered the room.
Using a handheld remote dimmer, participants adjusted the light level in fine increments, increasing or decreasing the output until they reached a self-determined “ideal” illumination level, which was defined for the participants as the light level that felt most natural and visually comfortable for grooming-related tasks.
Each participant completed five repetitions under the full-spectrum LED and five under the conventional LED. A 12-V DC pulse switch was used to record the time at which the participant finalized their lighting selection. This timestamp was synchronized with real-time power monitoring data from the vanity luminaire, allowing researchers to accurately capture participant preference, nominal power (watts), and lumen output for each repetition.
Photo: California Lighting Technology Center
Figure 2. A participant sorting the FM100 color tiles.
Participants completed a color discrimination test using the Farnsworth-Munsell 100-Hue Test (FM100), a well-established method for evaluating fine color perception and discrimination, whereby participants attempt to arrange a tray of colored tiles by increasing hue.3 Each participant sorted a complete set of hue tiles under one lighting condition, took a brief rest, and then repeated an identical sorting task under the alternate lighting condition (Figure 2). The order of the two lighting treatments was randomized across participants to minimize ordering effects. Participants were allowed up to five minutes per test to sort the tiles.
Upon completion, researchers calculated a total error score (TES) for each test using the Munsell hue number labeled on the back of each tile. TES is the sum of the differences between each tile and its two adjacent tiles. For example, a tile labeled 17 is placed after a tile labeled 19 and before a tile labeled 20. The difference value for tile 17 is |17-19|+|17-20|-2 = 3. Participants’ test scores were grouped into one of three standard FM100 color discrimination tiers: superior (TES = 0-16), average (TES = 17-100), and low (TES > 100).4
Participants in the lighting preference study had a mean score (preferred lumens) of 2,833 under the conventional, low-fidelity LED and a mean score of 2,323 using the full-spectrum, high-fidelity LED. This reflects a decrease of 510 lumens, with a standard deviation of the paired differences of 554 lumens.
The paired-samples t-test revealed that the lumen reduction was statistically significant, t(49) = 6.50, p < .00001 (two-tailed). The 95% confidence interval for the mean difference was [352.6, 667.4], indicating that the observed reduction was unlikely due to random variation. An effect size calculation yielded a Cohen’s d of 0.92, which is considered large according to conventional benchmarks. This suggests color fidelity and spectrum have a strong practical impact on people’s preferred light level for certain indoor applications.
On average, participants preferred 18% fewer lumens when exposed to high-fidelity, full-spectrum LED lighting as compared to the conventional, lower-fidelity LED reference source (Figure 3). Although the full-spectrum LED source is characterized by a lower luminous efficacy as compared to the conventional LED source, participants’ natural selection of lower light levels suggests potential energy saving opportunities for full-spectrum LED lighting in real-world applications.
The DesignLights Consortium grants a 5 to 10% efficacy allowance for LED products that demonstrate superior color fidelity, rewarding spectrum quality—not just efficiency.
Figure 3. Participants’ percent reduction in preferred lumen output for high-fidelity, full-spectrum LED lighting as compared to conventional LED lighting with lower color rendering.
No overall significant difference in color discrimination emerged between the full-spectrum LED source and the conventional LED reference when TES scores were aggregated across all participants and FM100 trays. However, certain trays—particularly those containing more subtle color transitions—produced modestly better scores under the full-spectrum LED treatment. Table 2 shows the final count of individual tests binned into the three standard FM100 tiers.
Other key observations include:
“Superior” test scores occurred twice as often under full-spectrum LED lighting as compared to scores resulting from tests conducted under conventional LED lighting.
On average, female participants produced fewer color sorting errors than males regardless of the lighting treatment, echoing research on potential tetrachromacy in a subset of the female population.4
Over the course of the study, participants completed three surveys designed to assess subjective evaluations of lighting quality across both experimental conditions. In general, respondents rated the two lighting conditions as comparable in terms of perceived brightness, visual comfort, and overall attractiveness. However, a notable subset of participants described the full-spectrum LED lighting as feeling more “natural” or “better suited” for visually demanding tasks.
Table 2. Count of FM100 TES scores for color discrimination tests
Figure 4. The significant differences in participants’ overall preference, sorting preference, and appearance preference between full-spectrum LEDs and conventional LEDs.
While most survey items revealed no significant differences in preference, three questions produced significant results (Figure 4) and revealed sex-related trends for this participant sample. Female participants were more evenly divided in their overall preferences for full-spectrum and conventional LED lighting, whereas male participants demonstrated a significant inclination toward conventional LED as their preferred option. When asked which lighting condition they would prefer specifically for color discrimination tasks, both sexes trended toward the full-spectrum LEDs, indicating a shared perception of its enhanced suitability for tasks requiring precise color perception. In the context of personal appearance, female respondents preferred full-spectrum LEDs, while males showed a preference for conventional LEDs; however, this distribution was more balanced compared to responses on overall lighting preference. The consistent trend across sexes favoring full-spectrum LED for the color sorting task suggests that high-fidelity, full-spectrum LED lighting may be more effective when accurate color rendering is essential.
Color fidelity and spectrum matter: Color fidelity and spectrum appear to have a strong practical impact on people’s preferred light level for certain indoor applications.
Reduced illumination demand: A significant reduction in lumen output—quantified at an average of 18% for this study—may be possible when using high-fidelity, full-spectrum LED lighting, which may translate to meaningful energy savings.
Enhanced perceived brightness: Participants frequently noted that full-spectrum LED lighting felt brighter and more comfortable at lower light levels.
Color-critical tasks: Full-spectrum LED lighting showed small but notable advantages for tasks requiring color accuracy.
Biological sex: Sex may play role in color preference for certain types of activities.
This research shows that high-fidelity, full-spectrum LEDs can create a more visually satisfying environment without requiring higher lumen outputs, opening the door to lighting energy savings. Improved color rendering appeared to help participants perceive adequate brightness at lower light levels. While color sorting accuracy didn’t consistently favor full-spectrum lighting, specific conditions—especially those demanding “superior” discrimination—highlight its advantages.
TM-30-20’s Fidelity Index—which measures how closely a light source renders colors compared to a natural reference—was the primary metric used to characterize spectrum quality. An Rf of 100 represents a perfect color match. UL’s Color Rendition Rating system defines an LED with color fidelity index (Rf) higher than 95 as “Diamond,” while an LED with index score below 85 does not qualify for UL rating.
Spectral quality plays a subtle but significant role in both perception and energy use. Though full-spectrum LEDs may deliver fewer lumens per watt, participants frequently preferred lower illumination levels under these sources—which may potentially offset efficacy concerns. In real-world applications like offices, classrooms, bathrooms, or dressing areas, this self-selected reduction could rival or surpass savings from higher-efficacy but lower-quality light.
Industry leaders, including the IES and DesignLights Consortium, are increasingly recognizing the value of advanced metrics like TM-30. This research supports the case for efficacy allowances or added tiers for high-fidelity lighting products.
The positive response to full-spectrum LEDs in this study signals a shift: moving beyond one-dimensional efficacy metrics to prioritize human-centric, visually balanced illumination.
THE AUTHORS | Jonathan Bryant is a Research and Development engineer at the CLTC.
Cori Jackson is responsible for planning, budgeting, scheduling, and monitoring CLTC research projects.
Jae Yong Suk is the faculty co-director of the CLTC and an associate professor in the Department of Design at UC Davis.
1 Pacific Northwest National Laboratory, “Solid-State Lighting R&D Opportunities,” 2019.
2 M.S. Rea and J.P. Freyssinier, “Light and Human Health: LED Lighting and Beyond,” Lighting Research & Technology, vol. 47, no. 1, 2015.
3 Munsell Color, “What does your score on the Farnsworth–Munsell 100 Hue Test mean?” 2024. Available: https://munsell.com/faqs/what-does-score-farnsworth-munsell-100-hue-test-mean/
4 Gabriele Jordan and John Mollon, “Tetrachromacy: The mysterious case of extra–ordinary color vision,” Current Opinion in Behavioral Sciences, vol. 30, Dec. 2019.
