JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 2021, VOL. 20, NO. 5, 556–565 https://doi.org/10.1080/13467581.2020.1799801 ENVIRONMENTAL ENGINEERING LED lighting system for better color rendition space: the effect of color rendering index SuJung Lee and Heakyung Cecilia Yoon School of Architecture, Hongik University, Seoul, Republic of Korea ABSTRACT ARTICLE HISTORY Received 6 May 2020 This study was conducted to investigate Color Rendering Index (CRI) recommendations for Accepted 16 July 2020 light emitting diode (LED) lighting in spaces needing better color rendition. Four LED spectra, that is, two spectra at Commission Internationale de l’Eclairage (CIE) Ra < 80 (79 and 76) and KEYWORDS two spectra at CIE Ra ≥ 80 (83 and 84) were used in this study. The CIE Ra lacks accuracy in Light-emitting diode (LED); predicting perceived color quality, and the Fidelity Index and the Gamut Index were used to color rendering index (CRI); complement the weakness of CIE Ra. These metrics were combined to quantify the average color preference; color increase or decrease in chroma. One hundred ninety-four participants evaluated retail-related quality; spaces tasks under different lighting conditions. The results showed that the CIE Ra had significant effects on preference, brightness perception, liveliness, and arousal. LED lights with CIE Ra < 80 received higher scores for preference, brightness perception, liveliness, and arousal than LED lights with CIE Ra ≥ 80, indicating that CIE Ra < 80 is the proper level for LED lighting in spaces needing better color rendition. 1. Introduction shortcomings of the CIE Ra, such as the Color Quality Scale (CQS) (Davis and Ohno 2010), Gamut Area Scale The Commission Internationale de l’Eclairage (CIE) (Qg), Color Preference Scale (Qp), Feeling of Contrast General Color Rendering Index (CRI) Ra (CIE 1995) has Index (FCI) (Hashimoto et al. 2007), Gamut Area Index been the most widely used color fidelity measure for (GAI) (Rea and Freyssinier-Nova 2008), and Memory specifying the color rendition properties of light Color Rendering Index (MCRI) (Smet et al. 2010). sources. However, with the emergence of white light However, it has not been possible to cover the differ - sources that have continuous spectra, such as found in ent aspects of color fidelity and color quality with light emitting diodes (LEDs), the problem of inaccu- a single metric (Houser et al. 2013), and consequently, rately measuring the color rendition features of a white these new measures have not been able to replace light source using the CIE Ra became apparent (Ohno CIE Ra. 2005). The color rendition features of LED lights have In 2015, the Illuminating Engineering Society of been divided into color fidelity and color quality. The North America (IESNA) published TM-30-15, IES color fidelity of LED lights quantifies the accuracy with Method for Evaluating Light Source Color Rendition which the color of illuminated objects appears. Color (IESNA 2015), in which is presented a two-measure quality quantifies the appearance of different psycho- system described by the Fidelity Index (Rf) and the logical weights according to the differences in color or Gamut Index (Rg) to determine a wide range of mea- brightness. The CIE Ra cannot predict both perceived sures that quantify different aspects of color rendi- color fidelity and color quality. Many studies have tion. The IES Rf is conceptually equivalent to the CIE proposed new measures to overcome the Ra, but it is different in that it is calculated using CONTACT Heakyung Cecilia Yoon firstname.lastname@example.org School of Architecture, Hongik University, Seoul, Republic of Korea © 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group on behalf of the Architectural Institute of Japan, Architectural Institute of Korea and Architectural Society of China. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 557 a different color difference. The IES Rg quantifies the values of 76, 86, and 96 were applied. For restaurant color gamut relatively under a reference illuminant and retail display lighting, participants preferred LED (the average increase or decrease in color saturation). lights with CIE Ra 86 (CIE Ra ≥ 80) to LED lights with Measurements from both IES Rf and IES Rg are being CIE Ra 76. Szabó et al. (2016) evaluated product color developed as a supplement to the CIE Ra (CIE 2017; appearance under 16 lighting conditions. They found Royer et al. 2017; Royer, Wilkerson, and Wei 2018; that CIE Ra ≥ 80 was preferred to CIE Ra < 80 for three IESNA 2018). The CIE Ra for most interior lighting categories of test samples: color textiles, meat, and was proposed by the International Organization for bakery products. Standardization (ISO) and the CIE in 2002 (ISO 2002). The impact of the CIE Ra of LED lighting on human The standard value for the CIE Ra is categorized emotion and cognitive activity is not well- according to the type of space and the task or activity documented. The CIE Ra can be affected by LED conducted within. Most light sources for architectural SPDs. Several researchers (Vogels 2008; Erp 2008; space such as retail interiors needing better color Steidle and Werth 2013; Wang et al. 2014; Wei et al. rendition have Ra values greater than 80 (ISO 2002; 2014; Islam et al. 2015; Li et al. 2019) have examined IESNA 2011). It is not clear how the recommendation LED lighting results with very different user emotional of CIE Ra ≥ 80 was determined for general lighting. responses and cognitive activity assessments. Wei et al. However, a number of different people and organiza- (2014) compared blue-pumped LED (BP-LED) lamps tions have been involved in the setting of specifica - with BP-LED lamps that had diminished yellow emis- tion color rendition criteria of CIE Ra for interior sion (YD-LED), evaluating brightness and preference. In lighting (e.g. ISO 2002; DOE 2001; IESNA 2011). With the study, the YD-LED lights were rated as brighter and the development of compact fluorescent lamp light- preferred than pure BP-LED lights. Islam et al. (2015) ing technology in the 1990s, Energy Star reported on conducted an experiment in which participants rated their energy efficiency improvement, and they used brightness and preference for a lit environment under the recommend CIE Ra ≥ 80 (DOE 2001). This is also six LED lights and two fluorescent lights with different adopted by ISO/CIE. People’s preferences vary with SPDs. The participants preferred the LED lights that light sources of different spectral power distributions they perceived as brighter. Mehrabian and Russell (SPDs). With the development of LED lighting, the first (1974) first proposed the Pleasure-Arousal-Dominance full-scale experiment was conducted under LED light- (PAD) scale, which measures the emotional effects of ing for evaluating color rendition by Miller et al. lighting. Park and Farr (2007) surveyed the influence of (2009), and many studies are being carried out to CCTs on PAD in a store-like setting in a laboratory evaluate whether the recommendation of CIE Ra environment, and this showed that subjects assessed should be considered based on experimental evi- 3000 K lighting as more pleasurable than 5000 K light- dence (Davis and Ohno 2010; Royer et al. 2017, ing. A study by Quartier, Vanrie, and Van Cleempoel 2020).It is thus important to investigate the criteria (2014) examined the effects of lighting on PAD. Their of CIE Ra for LED lighting recommendations in spaces study revealed that subjects rated a high-quality set- needing better color rendition based on evaluation of ting as more pleasurable than hard-discounter setting. light source color rendition and user acceptance. Many studies (Vogels 2008; Erp 2008; Wang et al. 2014; Many studies (Wei et al. 2014; Szabó et al. 2016; Lin Li et al. 2019) have developed PAD model. Li et al. et al. 2017; Royer et al. 2017) have examined the CIE (2019) investigated atmosphere perception under 36 Ra recommendation based on experimental evidence LED SPDs, and extracted four atmosphere factors: cozi- for LED lighting in spaces needing better color rendi- ness, spaciousness, liveliness, and warmth. In their tion, but the findings show mixed results. Royer et al. study, brightness for the lit environment was assessed (2017) conducted an experiment to characterize the with the term dim–bright. They found that brightness effect of a CIE Ra range of 43–95 on color preference. is associated with heightened coziness, spaciousness, Their study showed that participants preferred LED and warmth. However, none of these studies investi- lights with a CIE Ra value below 73, suggesting that gated at the effects of CIE Ra. CIE Ra ≥ 80 will lose popularity in lighting practice. In other research, Baron, Rea, and Daniels (1992) Wei et al. (2014) investigated color preference under discussed the effects of illuminance on task perfor- LED lights with a CIE Ra of 86 and 78. A room with LED mance, namely word categorization. Subjects reported lights with CIE Ra 78 was rated preferable to a room that 150 lux lighting improved task performance when with LED lights with CIE Ra 86. The results showed compared with 1500 lux lighting. A study by Knez that a light source should consider CIE Ra < 80 for LED (1995) discussed the effects of indoor lighting on cog- lighting in spaces needing better color rendition. nitive performance. Females performed better with However, other studies contradict those findings. Lin warm lighting than with cool lighting, while males et al. (2017) studied color preference for different performed better with cool lighting than with warm lighting environments: a restaurant, a retail display, lighting. In addition, cognitive performance (problem- and a supermarket. For each environment, CIE Ra solving tasks) improved, under which they found the 558 S. LEE AND H. C. YOON lighting condition evoked more positive feelings. 2. Method Steidle and Werth (2013) focused on the effects of 2.1. Experimental setup lighting on creativity. Their experiment involved fluor - escent light environments of 150 lx, 500 lx, and 1500 lx. Our experiments were carried out in a furnished office It was found that low illuminance improved creativity room at Hongik University, Seoul, Korea. Figure 2 and freedom. All these studies found that light source shows the experimental set up. The room was approxi- SPD affects human emotion. However, none of these mately 3.5 m by 3.5 m, with a ceiling height of approxi- studies looked at the effects of CIE Ra. mately 2.45 m. Black plastic films were placed over the Bodrogi et al. (2017) investigated the preference windows to block out daylight. The films were then of object color between Chinese and European sub- covered by white blinds. The temperature inside the jects under seven LED lightings. Their study found room was kept at 21–23°C. A table (l 2.1 m x w 0.9 m x h differences in the preference for perceived illumina- 0.75 m) and four chairs were centrally placed under the tion chromaticity between the two subject groups. ceiling light. During the experiment, a restricted num- Park and Farr (2007) investigated the perception of ber of supposed retail objects including brightly pleasure in Korean and American subjects under colored fruit (tomatoes, bananas, and green grapes) four lighting conditions. Their study showed differ - were placed on the table. The objects were replaced ences in perception of pleasure between Americans every day. They were selected to provide a reasonable and Koreans. Subjects of different cultural back- distribution within dimensions of enhanced red-green grounds could therefore have a different preference saturation, considered with reference to colors at of lighting environment. This study included sub- opposing ends of the color spectrum. jects of Korean cultural background who all lived in South Korea. The main goal of this study was to investigate 2.2. Lighting settings CIE Ra recommendations for LED lighting in spaces A multi-channel LED lighting system (LED Cubes, needing better color rendition. Four LED lights were Thouslite Ltd.) was used to create four different types designed at two CIE Ra (Ra < 80 and Ra ≥ 80) by of LED. The luminaire had eleven independently con- considering the two systems of IES Rf and IES Rg. trolled LED channels. The luminaire was controlled via We used conflicting values for the IES Rf–Rg combi- software on a laptop connected to the luminaire. Four nation index, which is a new evaluation method LED lights were used for the experiments, organized as that complements the CIE Ra. Two correlated color a 2 × 2 factorial design, with two levels of CIE Ra (CIE Ra temperatures (CCTs) (3000 K and 6500 K) were used < 80 and CIE Ra ≥ 80) and two levels of CCT (3000 K and in this study. Participants assessed the colored 6500 K), as shown in Table 1. Figure 3(a,b) shows the object, brightness, and cognitive activity. Figure 1 spectral power distribution (SPD) of each LED lights shows the CIE Ra evaluation model for LED lighting and TM-30 Colour Vector Graphics. Houser et al. (2013) in spaces needing better color rendition. This study used two measures to characterize overall color rendi- will provide basic data for the CIE Ra recommenda- tion: a reference-based measure to quantify the fidelity tion when using LED lights in spaces needing better of light sources and a gamut-based measure to quan- color rendition. tify the quality of light sources. IES Rf is an improved version of the reference-based measure, and IES Rg is Figure 1. CIE Ra evaluation model for LED lighting in spaces needing better color rendition. JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 559 Figure 2. The experimental environment. Table 1. Color characteristics of the LED light sources. 6500 K for cool lighting condition. For the CCT, LED CRI CCT Illuminance the two levels (3000 K and 6500 K) cover the lighting condition (Ra) (Rf) (Rg) (K) (lux) range currently used in spaces needing better A 79 83 103 3007 300 B 83 83 100 3000 300 color rendition. The average room illuminance was C 76 83 112 6495 300 maintained between 290 and 310 lux in the room D 84 83 103 6490 300 with LEDs. A Gossen Mavolux 5032 C Meter GO CRI: color rendering index; CCT: correlated color temperature 4056 illuminance meter was used to measure hor- izontal illuminances. a developed version of the gamut-based measure. Two studies used IES Rf in the range of 63–93 and IES Rg in the range of 80–120 for evaluating the color rendition 2.3. Participants properties of LEDs (Royer et al. 2017; Esposito and One hundred ninety-four participants from a private Houser 2019). Their studies showed that object color university in South Korea (all students; 105 males, 89 preferences were higher in LED lighting with higher IES females; aged 18–27) were recruited for the experi- Rg values than those with lower IES Rg values. We ment. Results from One hundred ninety-three partici- therefore selected LED A (CIE Ra of 79, IES Rf-Rg of pants were used in the final colored object evaluation 83–103) and LED C (CIE Ra of 76, IES Rf-Rg of 83–112) (one was invalid due to insincere answers). None of the for the more preferred lighting condition and LED participants were on courses related to the study of B (CIE Ra of 76, IES Rf-Rg of 83–112) and LED D (CIE lighting. All participants had normal color vision, as Ra of 84, IES Rf-Rg of 83–103) for the less preferred tested using the 24 Plate Ishihara Color Vision Test. lighting condition. The participants were asked not to wear hats, watches, The American National Standards Institute (ANSI) and accessories. Each participant received around USD used six nominal levels to specify the chromaticity 10 in cash immediately after their participation in the of fluorescent lamps. (i.e. 2700 K, 3000 K, 3500 K, experiment. The participants were randomly allocated 4000 K, 5000 K, 6500 K) (ANSI 2001). The chromati- to different lighting conditions. city of solid-state lighting has ten levels from 2200 K to 6500 K (NEMA 2015). Energy Star uses the CCT criterion of 2200–6500 K for fluorescent 2.4. Questionnaire lamps and LEDs (U.S. EPA and U.S. DOE 2017). The CCTs of 2200 K and 2500 K are limited to filament- The participants had to complete three types of eva- type lamps. ISO (2002) advises that lamps with low luation for each lighting condition: colored object eva- CCT values (below 3300 K) provide light that luation (preference, saturation, and vividness), appears warm, while lamps having high CCT values brightness evaluation (brightness perception, tem- (above 5300 K) provide light that appears cool. We perature perception, preference, atmosphere percep- selected 3000 K for warm lighting condition and tion, and PAD), and cognitive activity (creativity and 560 S. LEE AND H. C. YOON Figure 3. (a) Relative SPDs of LEDs. (b) TM-30 colour vector graphics. freedom). For the colored object evaluation, preference tasks for about 6 minutes. The participants were given (1 = dislike, 7 = like), saturation (1 = dull, 7 = saturated), 2 minutes to solve each task. The researcher mea- and vividness (1 = desaturated, 7 = vivid) were used. sured the time using a stop watch. After the creative For the brightness evaluation, brightness (1 = dim, tasks, the participants completed the questionnaire 7 = bright), color temperature (1 = cool, 7 = warm), about freedom. Finally, the researcher read the and preference (1 = dislike, 7 = like) were used. The instructions for the colored object evaluation to the atmosphere perception questionnaire of Li et al. participants. The participants rated an object’s color (2019), which consists of 22 questions, was used. as it appeared on the desk. This experiment used two Mehrabian and Russell’s PAD scale, which includes six research designs; 2 (CIE Ra) x 2 (CCT) within-subjects questions about pleasure and six questions about design for colored object evaluation and between- arousal, was also used. For cognitive activity evalua- subjects design for brightness and cognitive activity. tion, three questions by Dow and Mayer (2004) and Each participant carried out one judgement for every Schooler, Ohlsson, and Brooks (1993) were used as one of the four light sources in within-subjects a creative measure, eliciting verbal, spatial, and math- design. Each participant performed one judgment ematical responses. Questions about freedom included for one of the four light sources in between-subjects a question/response from Ryan and Deci (2000) design. Between-subjects design controlled for the (1 = externally controlled, 7 = self-determined) and effect of subject expectancies formed. The experi- a question/response from Jacobs and Scholl (2005) ment took about 50 minutes. (1 = inhibited, 7 = self-assured). Most questions were rated on a seven-point Likert scale. Table 2. Atmosphere perception dimensions. 2.5. Procedure Coziness Liveliness Alertness Spaciousness Agitated- Plain- Dim-bright Mysterious- Two to four participants took part in each experimen- tranquil luxurious familiar tal session. All lighting was initiated 30 minutes prior Tense-relaxed Lifeless- Drowsy-alert Personal-open lively to the start of the experimental sessions. Upon arriv- Hostile-friendly Boring- Depressing- ing at the experiment location, the participants interesting inspiring Dangerous-safe Unromantic- Crowded- adapted to the lighting environment for 3 minutes. romantic spacious The researcher read to them simple instructions Uncomfortable- cozy about the experiment, and the participants wrote Cool-warm down their personal information such as their age Distant- intimate and gender before starting the experiment. After Informal-formal evaluating the dim–bright, warm–cool, and dislike– Unpleasant- like of the LED lights, the participants rated the atmo- pleasant Terrifying- sphere of the room using the PAD scale. After the delightful brightness evaluation, participants were instructed Detached- attached to perform a cognitive activity using three creative JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 561 Table 3. Means and standard deviations (SDs) of colored SDs at low and high CIE Ra values as well as the object evaluations as a function of LED lighting condition. p values. Evaluations Questions LED A LED B LED C LED D N. 193 N. 193 N. 193 N. 193 Colored Dislike-like Mean 0.30 −1.06 0.38 −0.06 3.2. Brightness evaluation object SD 1.35 1.24 1.47 1.36 A one-way ANOVA revealed a significant effect of CIE Dull-saturated Mean 0.72 −0.47 0.85 −0.38 SD 1.29 1.38 1.32 1.31 Ra on the question regarding dim–bright (F Not vivid- Mean 0.76 −0.26 1.01 −0.18 (1,94) = 6.70, p <.012). The participants in the low vivid CIE Ra environment rated the room as brighter than SD 1.26 1.30 1.29 1.26 N: number of responses. did those in the high CIE Ra environment. Another one-way ANOVA on liveliness revealed the intended significant effect of CIE Ra (F(1,94) = 7.50, p < .050). The participants in the low CIE Ra environment felt 3. Results livelier than did those in the high CIE Ra environment. For the data analysis using the seven-point Likert scale, A one-way ANOVA on the arousal affect showed the answers were converted to −3 and 3, with 0 as the a significant effect of CIE Ra (F(1,94) = 4.01, p < .048). median value. For the creativity questions, 1 was given The participants in the low CIE Ra environment felt for right answers and 0 for wrong answers based on more arousal than did those in the high CIE Ra con- existing studies. A one-way analysis of variance dition. The differences in the mean ratings under (ANOVA) (SPSS IBM version 26) was used with different CIE Ra were not statistically significant for a significance level of 0.05. Factor analysis was used cool–warm (F(1,94) = 1.160, p = .284), coziness (F to assess atmosphere perception. A principal compo- (1,94) = 1.065, p = .305), alertness(F(1,94) = 1.664, nents analysis with Varimax rotation was used to p = .200), spaciousness (F(1,94) = 0.001, p = .972) extract the dimensions. Four atmosphere factors were and pleasure (F(1,94) = 0.834, p = .363). These results extracted. Factor 1 was named coziness, factor 2 was named liveliness, factor 3 was named alertness, and factor 4 was named spaciousness. Coziness (α = 0.90), liveliness (α = 0.74), alertness (α = 0.79), and spacious- Table 4. Means and standard deviations (SD) at CIE Ra of low (76, 79) and high (83, 84), and p values for three factors. ness (α = 0.66) indices were created from a set of Evaluations Questions CIE Ra Mean SD p-Value eleven items from factor 1, four items from factor 2, Colored Dislike-like low 0.34 1.41 0.000 * four items from factor 3, and two items from factor 4 of object the atmosphere perception scale. Table 2 shows the high −0.56 1.39 Dull-saturated low 0.78 1.30 0.000 * atmosphere perception dimensions from this study. high −0.42 1.35 The pleasure (α = 0.89), arousal (α = 0.81), and dom- Not vivid-vivid low 0.88 1.28 0.000 * high −0.22 1.28 inance (α = 0.87) indices were created by averaging Brightness Dim-bright low 0.92 1.47 0.012 * a set of six items from each dimension of the PAD scale. high 0.17 1.40 Table 3 shows the means and standard deviations Cool-warm low −0.19 1.88 0.662 high −0.35 1.85 (SDs) of the ratings of one hundred ninety-three sub- Dislike-like low −0.13 1.77 0.121 jects for the colored object evaluations for different high −0.65 1.48 Coziness low 0.11 0.97 0.348 SPDs at CIE Ra < 80 or at CIE Ra ≥ 80. The results high −0.08 1.04 from brightness evaluations and cognitive activity eva- Liveliness low −0.73 0.99 0.050 * high −1.11 0.88 luations were analyzed for ninety-six subjects. Alertness low 0.00 1.16 0.279 high −0.25 1.09 Spaciousness low −0.39 1.37 0.969 high −0.38 1.28 3.1. Colored object evaluation Pleasure low 0.21 1.01 0.405 high 0.03 1.07 The visual dataset consisted of 193 (subjects) x 4 Arousal low 0.00 1.02 0.048 * high −0.42 1.05 (light sources) x 3 (colored object evaluation: pre- Cognitive Horse trading (M) low 0.85 0.36 0.134 ference, saturation, vividness), giving 2316 data activity points. A one-way ANOVA revealed the CIE Ra sig- high 0.73 0.45 Triangle (S) low 0.71 0.46 0.518 nificantly affected one of the three dependent vari- high 0.65 0.48 ables (F-test; p < 0.000). The participants rated the Antique coin (V) low 0.42 0.50 0.067 high 0.60 0.49 light with the lower CIE Ra 80 (79, 76) as more Inhibited-self-assured low 0.33 1.29 0.516 preferable for colored objects than the light with high 0.17 1.21 Controlled-self- low 0.13 1.53 0.464 the higher CIE Ra 80 (83, 84). The same result for determined the scaled color saturation and vividness were high −0.13 1.79 found in this study. Table 3 shows the means and * Significant with p-value < 0.05. 562 S. LEE AND H. C. YOON indicate that being exposed to CIE Ra < 80 was pre- higher ratings for saturation and vividness of colored ferred to CIE Ra ≥ 80. objects to LED lights with CIE Ra < 80 than LED lights with CIE Ra ≥ 80. In order to accept the current recom- mended CIE Ra (≥ 80) for the spaces needing better 3.3. Cognitive activity evaluation color rendition, the results should have shown that the participants give higher ratings for the three scales of No statistically significant differences of cognitive preference, saturation, and vividness of objects under activity evaluation were found in the ratings between LEDs with CIE Ra ≥ 80. However, LEDs with CIE Ra < 80 CIE Ra < 80 and CIE Ra ≥ 80. However, in most cases, received higher scores for preference, saturation, and the mean ratings were slightly higher when under vividness of object colors. These studies showed that lighting with CIE Ra < 80 than CIE Ra ≥ 80. These results participants preferred LED lights of higher chroma. indicate that the subjects’ cognitive activity improved The third approach involved exploring the effects of under lighting environments with CIE Ra < 80. a LED light source on human emotion. For the bright- The results of the statistical analysis for different CIE ness evaluation, the results showed that the CIE Ra had Ra values are presented in Table 4. a significant effect on dim–bright, liveliness, and arou- sal affect. The room under LED lights with CIE Ra < 80 4. Discussion was perceived as brighter, livelier, and provoked a greater arousal affect than the room under LED lights This study was designed to investigate CIE Ra recom- with CIE Ra ≥ 80. No statistically significant differences mendations for LED lighting in spaces needing better were found in the question responses regarding color rendition and to carry out user preference eva- warm–cool, dislike–like, coziness, alertness, spacious- luation for four LED lights with CIE Ra < 80 or CIE Ra ≥ ness, and pleasure between CIE Ra < 80 and CIE Ra ≥ 80. The LED lights with CIE Ra < 80 had IES Rf–Rg 80. However, the means of these items were higher combinations that increased in chroma, and the LED under LED lights with CIE Ra < 80. (The room under lights with CIE Ra ≥ 80 had IES Rf–Rg combinations that LED lights with CIE Ra < 80 was perceived as cozier, decreased in chroma. The participants performed warmer, more alert, more spacious, and more pleasur- retail-related tasks and undertook colored object eva- able than under LED lights with CIE Ra ≥ 80.) The luation (preference, saturation, and vividness), bright- participants under LED lights with CIE Ra < 80 reported ness evaluation (brightness perception, color that the room looked warmer, cozier, more alert, more temperature, atmosphere perception, and PAD affect), spacious, and more pleasurable than under LED lights and cognitive activity evaluation (creativity and with CIE Ra ≥ 80. Our findings support the study by freedom). Islam et al. (2015) in which people preferred a lighting The first approach to establishing the CIE Ra recom- environment that seemed brighter and more spacious. mendation was to investigate metrics for evaluating Brightness was assessed using the atmosphere per- the color rendering properties of a LED light source for ception scale used in previous research (Vogels 2008; the spaces needing better color rendition. The statis- Erp 2008; Wang et al. 2014; Li et al. 2019). Four atmo- tical analysis showed that the CIE Ra significantly sphere factors were extracted: coziness, liveliness, affected preference, saturation, and vividness. LED alertness, and spaciousness. Some differences A (CIE Ra 79, IES Rf–Rg 83–103) and LED C (CIE Ra 76, between this study and Li et al. (2019) study on the IES Rf–Rg 83–112) with CIE Ra < 80 had higher scores effects of lighting on atmosphere can be observed. for preference, saturation, and vividness of objects Differences between the effects on atmosphere in the than LED B (CIE Ra 83, IES Rf–Rg 83–100) and LED two studies might be caused by differences in adapta- D (CIE Ra 84, IES Rf–Rg 83–103) with CIE Ra ≥ 80. tion time, experimental design (design within groups Both LED A and LED C had CIE < 80 and IES Rf–Rg vs. design between groups), stimuli, measurement combinations that increased in chroma, whereas LED method, and cultural differences. However, both stu- B and LED D had CIE Ra ≥ 80 and IES Rf–Rg combina- dies were similarly constructed, and the participants tions that decreased in chroma. This result demon- felt livelier under LED lights with increasing chroma. strated that the CIE Ra fails to specify the color The CIE Ra is not statistically significant for any rendering properties of a LED light source for the questions regarding cognitive activity evaluation. spaces needing better color rendition, while IES Rf–Rg However, small differences in the mean ratings were combinations, which complement the CIE Ra, provide found. Three creative insight problems were used for an effective tool. creativity assessment. More creativity was perceived The second approach was to examine color prefer- under LED lights with CIE Ra < 80 than LED lights ence. LED lights with CIE Ra < 80 were preferred more with CIE Ra ≥ 80. In addition, the participants under than LED lights with CIE Ra ≥ 80. The participants LED lights with CIE Ra < 80 felt freer than the partici- preferred red, orange and green under LED lights pants under LED lights with CIE Ra < 80. (Participants with CIE Ra < 80. In addition, the participants gave JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 563 gave higher ratings for horse trading, triangle, inhib- than CIE Ra ≥ 80. The participants preferred LED lights ited-self-assured, and controlled-self to LED lights with that increased in chroma rather than decreased in CIE Ra < 80 than to LED lights with CIE Ra < 80.) chroma. The results indicate, as expected, that the The participants preferred the colored objects and current recommendation of CIE Ra ≥ 80 for LED light- rated the space livelier and invoking more arousal ing in spaces needing better color rendition is under LED lights with CIE Ra < 80 than CIE Ra ≥ 80. incorrect. Although it was statistically insignificant, they rated Finally, for brightness evaluation, the CIE Ra had the space as cozier, more alert, and more spacious significant effects on dim–bright, liveliness, and arou- under the same conditions. They also rated a higher sal affect. The participants rated the room in our study sense of creativity and feeling freer under the same as brighter and livelier at CIE Ra < 80 than at CIE Ra ≥ conditions. The three evaluations in this study demon- 80. Moreover, the participants felt more of an arousal strated that CIE Ra < 80 is the proper level for LED affect at CIE Ra ≥ 80 than at CIE Ra ≥ 80. For cognitive lighting in spaces needing better color rendition. activity evaluation, no statistically significant differ - Taken together, the first approach showed that the ences between CIE Ra < 80 and CIE Ra ≥ 80 were CIE Ra was limited in predicting perceived color qual- found in creativity and freedom. However, CIE Ra < ity. IES Rf–Rg combinations were better measures as 80 promoted more creativity and heightened per- a supplement to the CIE Ra. The second approach ceived freedom when compared with CIE Ra ≥ 80. found that the participants preferred LED lights of These results support the finding in this study that increasing chroma. The third approach demonstrated the recommendation of CIE Ra ≥ 80 is wrong for LED that lighting affected human emotion and cognitive lighting in spaces needing better color rendition. activity. The findings reinforce that CIE Ra < 80 is good Overall, this study suggests that three approaches for LED lighting in spaces needing better color are important in establishing CIE Ra recommenda- rendition. tions. The CIE Ra has limitations, but combining with IES Rf–Rg provides a good measure for predicting perceived color quality. The participants preferred 5. Conclusions LED lights that increased in chroma; LED lights affected the appearance of objects’ color as well as One hundred ninety-four participants aged 19–26 took human emotion and cognitive activity. The results part in this study to investigate the optimal CIE Ra for from the three approaches demonstrate the current LED lights in spaces needing better color rendition. The recommendation for CIE Ra ≥ 80 is not appropriate for first approach to establishing a recommended CIE Ra LED lighting in spaces needing better color rendition was to investigate measures for evaluating the color and that CIE Ra < 80 should be considered instead for quality of LED light sources in spaces needing better LED lighting in spaces needing better color rendition. color rendition. The second approach involved exam- Further studies are required under different ranges of ining color preference, an important dimension of light CIE Ra with IES Rf–Rg combinations to establish source color rendition. The last approach involved improved CIE Ra recommendations. exploring the effects of lighting on human emotion and cognitive activity. In terms of key findings, firstly, the CIE Ra showed Disclosure statement limitations in evaluating LED lighting in spaces need- ing better color rendition. Based on the colored object No potential conflict of interest was reported by the authors. evaluation, LED lights with CIE Ra < 80 were rated as preferable to LED lights with CIE Ra ≥ 80. The results showed that the CIE Ra did not accurately predict Funding perceived color quality. In this study, IES Rf–Rg combi- This research was supported by the Basic Science Research nations were used to complement the CIE Ra, and the Program through the National Research Foundation of Korea IES Rf–Rg combinations were a good indicator of pre- (NRF-2015R1D1A1A01058577) and by the 2019 Hongik dicting perceived color quality. While the CIE Ra University Research Fund. remains an important metric as a standard for lighting practice, it seems advisable to complement the CIE Ra References with IES Rf–Rg combinations for LED lighting in spaces needing better color rendition. American National Standard Institute. 2001. “Specifications for the Chromaticity of Fluorescent Lamps.” ANSI C78.376- Secondly, for colored object evaluation, the CIE Ra 2001. Arlington, VA: National Electrical Manufacturers had significant effects on preference, saturation, and Association. vividness. The participants preferred colored objects Baron, R. A., M. S. Rea, and S. G. Daniels. 1992. “Effects of under LED lights with CIE Ra < 80 to those with CIE Indoor Lighting (Illuminance and Spectral Distribution) on Ra ≥ 80. It was also found that for the scale of satura- the Performance of Cognitive Tasks and Interpersonal tion and vividness, CIE Ra < 80 was again rated higher Behaviors: The Potential Mediating Role of Positive 564 S. LEE AND H. C. YOON Affect.” Motivation and Emotion 16 (1): 1–33. doi:10.1007/ Lin, Y., M. Wei, K. A. G. Smet, A. Tsukitani, P. Bodrogi, and BF00996485. T. Q. Khanh. 2017. “Colour Preference Varies with Lighting Bodrogi, P., Y. Lin, X. Xiao, D. Stojanovic, and T. Q. Khanh. Application.” Lighting Research and Technology 49 (3): 2017. “Intercultural Observer Preference for Perceived 316–328. doi:10.1177/1477153515611458. Illumination Chromaticity for Different Coloured Object Mehrabian, A., and J. A. Russell. 1974. “The Basic Emotional Scenes.” Lighting Research and Technology 49 (3): Impact of Environments.” Perceptual and Motor Skills 38 305–315. doi:10.1177/1477153515616435. (1): 283–301. doi:10.2466/pms.19126.96.36.1993. Commission Internationale de l’Eclairage. CIE 13.3-1995. Miller, C., Y. Ohno, W. Davis, Y. Zong, and K. Dowling 2009. 1995. Method of Measuring and Specifying Colour “NIST Spectrally Tunable Lighting Facility for Color Rendering Properties of Light Sources. Vienna: CIE. Rendering and Lighting Experiments.” In Proceedings of Commission Internationale de l’Eclairage. CIE 224: 2017. Light and Lighting Conference with Special Emphasis on 2017. Colour Fidelity Index for Accurate Scientific Use. LEDs and Solid State Lighting, Budapest 5. Vienna: CIE. National Electrical Manufacturers Association. C78.377-2015. Davis, W., and Y. Ohno. 2010. “Color Quality Scale.” Optical 2015. American National Standard for Electric Lamps—spe- Engineering 49 (3): 033602. doi:10.1117/1.3360335. cifications for the Chromaticity of Solid State Lighting (SSL) Department of Energy. 2001. ENERGY STAR Program Products. Arlington, VA: NEMA. Requirements for CFLs Partner Commitments, Version 2.0. Ohno, Y. 2005. “Spectral Design Considerations for White LED Washington, DC: DOE. Color Rendering.” Optical Engineering 44 (11): 111302. Dow, G. T., and R. E. Mayer. 2004. “Teaching Students to Solve doi:10.1117/1.2130694. Insight Problems: Evidence for Domain Specificity in Park, N. K., and C. A. Farr. 2007. “The Effects of Lighting on Creativity Training.” Creativity Research Journal 16 (4): Consumers’ Emotions and Behavioral Intentions in a Retail 389–398. doi:10.1080/10400410409534550. Environment: A Cross-cultural Comparison.” Journal of Erp, T. 2008. “The Effects of Lighting Characteristics on Interior Design 33 (1): 17–32. doi:10.1111/j.1939-1668.2007. Atmosphere Perception.” Unpublished manuscript for tb00419.x. Philips Research, Eindhoven, Netherlands. Quartier, K., J. Vanrie, and K. Van Cleempoel. 2014. “As Real as Esposito, T., and K. Houser. 2019. “Models of Colour Quality It Gets: What Role Does Lighting Have on Consumer’s over a Wide Range of Spectral Power Distributions.” Perception of Atmosphere, Emotions and Behaviour?” Lighting Research and Technology 51 (3): 331–352. Journal of Environmental Psychology 39: 32–39. doi:10.1177/1477153518765953. doi:10.1016/j.jenvp.2014.04.005. Hashimoto, K., T. Yano, M. Shimizu, and Y. Nayatani. 2007. Rea, M. S., and J. P. Freyssinier-Nova. 2008. “Color Rendering: “New Method for Specifying Color-rendering Properties of A Tale of Two Metrics.” Color Research and Application 33 Light Sources Based on Feeling of Contrast.” Color Research (3): 192–202. doi:10.1002/col.20399. and Application 32 (5): 361–371. doi:10.1002/col.20338. Royer, M. P., M. Wei, A. Wilkerson, and S. Safranek. 2020. Houser, K. W., M. Wei, A. David, M. R. Krames, and X. S. Shen. “Experimental Validation of Colour Rendition 2013. “Review of Measures for Light-source Color Specification Criteria Based on ANSI/IES TM-30-18.” Rendition and Considerations for a Two-measure System Lighting Research and Technology 52 (3): 323–349. for Characterizing Color Rendition.” Optics Express 21 (8): doi:10.1177/1477153519857625. 10393–10411. doi:10.1364/OE.21.010393. Royer, M. P., A. Wilkerson, and M. Wei. 2018. “Human Illuminating Engineering Society. 2011. The Lighting Perceptions of Colour Rendition at Different Handbook. 10th ed. New York: IES. Chromaticities.” Lighting Research and Technology 50 (7): Illuminating Engineering Society of North America. IES TM- 965–994. doi:10.1177/1477153517725974. 30-15. 2015. IES Method for Evaluating Light Source Color Royer, M. P., A. Wilkerson, M. Wei, K. Houser, and R. Davis. Rendition. New York: IESNA. 2017. “Human Perceptions of Colour Rendition Vary with Illuminating Engineering Society of North America. IES TM- Average Fidelity, Average Gamut, and Gamut Shape.” 30-18. 2018. IES Method for Evaluating Light Source Color Lighting Research and Technology 49 (8): 966–991. Rendition. New York: IESNA. doi:10.1177/1477153516663615. International Standards Organization. ISO 8995-1. 2002. Ryan, R. M., and E. L. Deci. 2000. “Self-determination Theory Lighting of Indoor Work Places—part 1: Indoor. Geneva: and the Facilitation of Intrinsic Motivation, Social ISO. Development, and Well-being.” American Psychologist 55 Islam, M. S., R. Dangol, M. Hyvärinen, P. Bhusal, M. Puolakka, (1): 68. doi:10.1037/0003-066X.55.1.68. and L. Halonen. 2015. “User Acceptance Studies for LED Schooler, J. W., S. Ohlsson, and K. Brooks. 1993. “Thoughts Office Lighting: Lamp Spectrum, Spatial Brightness and beyond Words: When Language Overshadows Insight.” Illuminance.” Lighting Research and Technology 47 (1): Journal of Experimental Psychology. General 122 (2): 166. 54–79. doi:10.1177/1477153513514425. doi:10.1037/0096-34188.8.131.52. Jacobs, I., and W. Scholl. 2005. “Interpersonale adjektivliste Smet, K. A., W. R. Ryckaert, M. R. Pointer, G. Deconinck, and (IAL).” Diagnostica 51 (3): 145–155. doi:10.1026/0012- P. Hanselaer. 2010. “Memory Colours and Colour Quality 19184.108.40.206. Evaluation of Conventional and Solid-state Lamps.” Optics Knez, I. 1995. “Effects of Indoor Lighting on Mood and Express 18 (25): 26229–26244. doi:10.1364/OE.18.026229. Cognition.” Journal of Environmental Psychology 15 (1): Steidle, A., and L. Werth. 2013. “Freedom from Constraints: 39–51. doi:10.1016/0272-4944(95)90013-6. Darkness and Dim Illumination Promote Creativity.” Li, B., Q. Y. Zhai, J. B. Hutchings, M. R. Luo, and F. T. Ying. 2019. Journal of Environmental Psychology 35: 67–80. “Atmosphere Perception of Dynamic LED Lighting over doi:10.1016/j.jenvp.2013.05.003. Different Hue Ranges.” Lighting Research and Technology Szabó, F., R. Kéri, J. Schanda, P. Csuti, A. Wilm, and E. Baur. 51 (5): 682–703. doi:10.1177/1477153517702532. 2016. “A Study of Preferred Colour Rendering of Light JOURNAL OF ASIAN ARCHITECTURE AND BUILDING ENGINEERING 565 Sources: Shop Lighting.” Lighting Research and Technology Wang, H. H., M. R. Luo, P. Liu, Y. Yang, Z. Zheng, and X. Liu. 48 (3): 286–306. doi:10.1177/1477153515573042. 2014. “A Study of Atmosphere Perception of Dynamic U.S. Environmental Protection Agency and U.S. Coloured Light.” Lighting Research and Technology 46 (6): Department of Energy. 2017. ENERGY STAR Program 661–675. doi:10.1177/1477153513506591. Requirements for Lamps, Version 2.1. Washington, DC: Wei, M., K. W. Houser, G. R. Allen, and W. W. Beers. 2014. EPA. “Color Preference under LEDs with Diminished Yellow Vogels, I. 2008. Atmosphere Metrics, 25–41. Dordrecht: Emission.” Leukos 10 (3): 119–131. doi:10.1080/ Springer. 15502724.2013.865212.
Journal of Asian Architecture and Building Engineering
– Taylor & Francis
Published: Sep 3, 2021
Keywords: Light-emitting diode (LED); color rendering index (CRI); color preference; color quality; spaces