Open Access
Add to BibSonomy
Abstract
On the topic of color preference of lighting, it is commonly believed that the neutral interior of a light booth has minimum impact on the color perception of the experimental object. Meanwhile, agreement has not been reached on which objects should be placed in the booth. In this study, based on a meta-analysis of eight groups of psychophysical data, we demonstrate that the “perceived color preference” obtained by consecutive visual judgement in a light booth is closely related to the lit neutral environment, while the use of different experimental objects does not markedly influence the results for lighting quality assessment. Such a finding might be attributed to the inherent features of the visual cognition process of light booth experiments and it should be fully recognized by future work.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
With the rapid development of solid-state lighting technology, it is widely agreed that the color quality of lighting is multidimensional, which correlates to a variety of visual attributes, for example fidelity [1,2], preference [3,4], naturalness [5,6], vividness [7,8], discrimination [9,10] etc. Color preference of lighting, being related to the visual appreciation of the rendered colors of the illuminated objects, is one of the focuses for research in lighting quality evaluation [7,8,11–16]. These studies usually obtained subjective preference responses by psychophysical experiments, calculated objective metrics which describe the color rendition properties of experimental light sources, and then analyzed and characterized the mechanism of visual preference.
Obviously, the psychophysical setup and protocol are of crucial importance for this topic. According to current literature, it is prevalent to implement psychophysical tests in light booths with the walls and floors painted in a neutral color [8,13,17–21]. The reasons are clear: compared with an immersive environment, it is much easier to set up a booth environment equipped with color tunable light sources and it is widely agreed that the neutral interior color of the booth will not significantly impact the color perception of the objects.
In addition, the object adopted in the visual test is widely regarded as another critical variable. According to the current literature, various objects have been adopted in psychophysical studies in light booths, including consumer goods [8], cosmetic products [13], skin tones [17], artworks [18], printed images [19], fruit and vegetables [20], scale models [21], clothing [22], etc. The purpose of the researchers who adopted specific objects, however, is usually to find the best solution for general or specific lighting applications.
It is commonly believed that color preference of lighting varies with experimental objects, since, by definition, color is a synthetization of light, object and human vision. However, to our knowledge, few studies have been carried out to specifically validate that opinion, especially for the cases that use light booths. In our previous work, a broad range of objects was each separately illuminated by a group of white light sources with different correlated color temperatures (CCTs) [18]. Those tests revealed that, no matter what was placed in the booth, the preference ratings of the observers exhibited a similar trend, which highlighted the dominant effect of light on color preference. In a follow-up study with the same experimental booth and lights, we proved that the subjective preference scores for the lit empty booth were also consistent with the ratings for the various objects. Based on that finding, an empty-booth approach was boldly proposed to evaluate the color preference of lighting and as we stated in that paper, “there seems to be no need to be concerned about the presented objects when evaluating the color preference of a light source, since an empty light booth with a neutral environment is already sufficient” [23].
The “empty-booth” approach, which advocates to evaluate the color quality of white light sources without rendered objects, seems to be vagarious, although its effectiveness for light booth studies has been well demonstrated in our former work [18,23]. In this study, a meta-analysis of psychophysical data obtained from eight groups of visual experiments was conducted to further investigate this issue. Our speculation was that, due to certain visual mechanisms hidden in the visual test, the lit environment of the booth significantly impacted on color preference perception and thus resulted in the “dominant effect” of light on color preference, and made the “empty-booth” approach effective.
2. Psychophysical studies
To achieve convincing conclusions, greatest effort was made to collect psychophysical data on this topic. The criterion was that a selected study must contain color preference ratings for more than one object in a neutral light booth (including the case of no object in booth, i.e. scores for the lit empty booth) and the objects must be randomly and successively displayed, and rated. Data from eight groups of visual experiments were collected, four of which came from our recent work [22–26] while the other four came from contributions by Jost-Boissard et al., [20,27]. The data contains subjective ratings of loosely-defined color preference (i.e. preference, visual appreciation, attractiveness) for one or more objects illuminated by different groups of lights in neutral booths. For our own work, the color preference ratings for the empty lit booth were also included.
To be specific, the collected psychophysical dataset contains the visual data from Huang et al., [25,26], Chen et al., [24], Liu et al., [22] and Jost-Boissard et al., [20,27]. Those studies cover a wide range of variations in experimental parameters, such as the chromaticities (CCT and Duv) of light sources, illuminance levels (200 lx to 500 lx), scoring methods (categorical judgment and pair comparison), experimental objects (jeans, pictures, ornaments, fruit and vegetables and color samples), time course for chromatic adaptation (a few seconds to around 40 seconds), color features of the objects (single-color, multiple-color, and achromatic color) and familiarity to the objects (familiar and unfamiliar). Obviously, conclusions drawn from such a large psychophysical dataset should be more convincing. The following description serves as a brief introduction to the primary information of each study and the appendix provides a summary of the typical colorimetric parameters of the experimental lights. For detailed information, the reader is referred to the cited articles.
2.1 Jeans [25]
Huang et al., investigated the impact of white light sources on the color preference and color discrimination perception of jeans [25]. The same light sources tested in a previous work [18] were used again in Table 1 (nine LEDs with different CCTs ranging from 2500 K to 6500 K, Duv values from −0.008 to −0.003, illuminance level around 200 lx) and a panel of seven blue jeans with a color gradient pattern was adopted as experimental objects. Twenty-seven subjects with normal color vision were asked to rate their color preference for those jeans with a 7-point scale method. Meanwhile, as reported earlier [23], another group of 45 observers had been recruited to evaluate the perceived preference for the illuminated empty booth, with a very similar experimental protocol and the same experimental light sources.
2.2 Achromatic objects [24]
Chen et al., investigated another panel of nine LEDs with uniformly sampled CCT values ranging from 2500 K to 6500 K (500 K interval, Duv=0, illuminance=200 lx) were tested. In Table 2, thirty observers were invited to rate their color preference for an empty booth and six achromatic objects, including four black and white photos and two groups of black and white ornaments [24]. A 7-point rating method (−3, −2, −1, 0, 1, 2, 3) was used for preference judgement. Ratings of the lit empty booth have not been reported before.
2.3 Fruit and vegetables [26]
Huang et al., investigated thirty observers were invited to observe a plate of fruit and vegetables and rate their color preference perception in a lighting booth with the same 7-point categorical judgment method [26]. In Table 3, six 3000 K lights were adopted with Duv values ranging from −0.015 to 0.010 and the illuminance level was approximately 200 lx. During the test, the color preference evaluation was also conducted for the lit environment of the empty light booth but the data have not been reported in the previous paper.
2.4 Jeans [23]
In a follow-up study of the above work for jeans lighting [25], nine 5500 K light sources with Duv values ranging from −0.02 to 0.02 were used to illuminate the same panel of blue jeans in a booth and the illuminance level was set to 500 lx [22]. In Table 4, thirty subjects participated in the experiment and again, they were asked to respond with their preference perception by a 7-point rating approach. Note that during the test the color preference evaluation for the lit environment of the booth had not been implemented. In the current study, such a test was added, following the categorical judgment procedure described in [23] with thirty observers invited.
2.5 Fruit and vegetables [20]
Six 3050 K LED clusters and six 3950 K LED clusters were applied in a forced choice comparison study, in which a 3050 K halogen light and a 3950 K fluorescent light were respectively adopted as Refs. [20]. In Table 5, for 3050 K, the six LED clusters were WA, WR, WCA, WCR, WGA and WGR, where W, A, R, C, G denote White, Amber, Red, Cyan and Green LEDs, respectively. For 3950 K, the LED clusters were generated and named in the same way. The illumination level was approximately 230 lx. Two adjacent lighting booths, both painted in medium gray matt paint, were used in this study. In Table 6, a panel of 40 observers were invited to evaluate the color rendition of four groups (red, green, yellow and multicolored) of fruit and vegetables in terms of Attractiveness, Naturalness and Suitability. In this study, two groups of data on Attractiveness were used.
2.6 Multiple objects [27]
In 2015, Jost-Boissard et al., implemented a similar comparison experiment [27] using the same booths as adopted in their work in 2009 [20]. The perceived color quality for fruits and vegetables and a Color Checker chart was investigated, in terms of Naturalness, Colorfulness, Visual Appreciation and Color Difference. Specifically, in Table 7, 45 observers assessed 9 light sources at 3000 K while 36 observers assessed 8 light sources at 4000 K, with an illuminance level of approximately 220 lx. For 3000 K, the 9 lights included HA (Halogen), FL (Fluorescent), WA, WR, WAR, WCR, WGR, CRI-WGARC (optimising CRI by WGARC) and Spectrum-WGARC (approximated the spectrum of a Planckian radiator at 3000 K with WGARC). As above, W, A, R, C, G denote White, Amber, Red, Cyan and Green LEDs, respectively. In Table 8, for 4000 K, the light sources were FL, WGA, WR, WAR, WCR, WGR, CRI-WGARC and Spectrum-WGARC, with the same naming method. Again, in this study two groups of data about Visual Appreciation were adopted.
3. Results and discussion
The results of our own experiments are summarized in Fig. 1, which shows that, no matter which experimental objects and light sources were adopted, the ratings for different objects, as well as the ratings for the empty booth, were generally consistent within a defined study. For studies with experimental light sources of multiple CCT value, the preference scores increased with CCT values from 2500 K to 5000 K while that increase became less significant when the CCT was higher. Similarly, for light sources of consistent CCTs but different Duv values, the preference scores increased with Duv values at first and then decreased, with an optimal Duv value of approximately −0.010. Those results agree well with past studies that investigated the preferred CCT [18,19,21] and Duv values [21,22,28,29] and reinforce our former statement concerning the dominant effect of light on color preference in a booth [18].
Fig. 1. Psychophysical results of four recent studies conducted by the authors. Triangular symbols in the subplots indicate significant difference (p<0.05) between the preference scores of the empty lit booth and those of illuminated objects, revealed by a Mann-Whitney U test. Asterisks besides the Pearson correlation coefficient r denote significant correlation (p<0.05) between ratings of the lit empty booth and those of illuminated objects. “B&W” in Chen et al., (2020) means Black and White.
For every case depicted in Fig. 1, strong correlation between preference ratings for illuminated objects and ratings for the lit booth is observed, which validates the “empty booth” approach [23]. Such results indicate that the experimental objects do not seem to matter when comparing the color rendition performance of light sources in a light booth experiment, which agrees well with our former finding [18]. In addition, the results of a Mann-Whitney U test demonstrate that for most conditions, placing an object or not does not lead to significant difference in observers’ preference ratings: as shown in the figure, a very small percentage of the scenarios is marked with triangles and only for these cases was a significant difference (p<0.05) found.
Figure 2 illustrates the experimental results from four groups of visual studies by Jost-Boissard et al. [20,27]. It should be note that, unlike Fig. 1, the acronyms on the x-axis of Fig. 2 only refer to the tabs of the experimental light sources, as stated above. Therefore, from this figure we cannot draw any conclusion on which kind of light source is preferred. However, it is obvious that, again, the trends of the ratings for multiple objects were consistent within each study and significantly strong correlation between ratings was found. Such a finding is what we actually intend to demonstrate, since the aim of this work is to investigate the influence of the lit environment of the booth, rather than to characterize the color preference perception by a certain colorimetric parameter. Note also that those researchers did not invite the subjects to rate their preference for the lit environment of the empty booth, so it is not possible to examine the “empty booth” approach with those data. The rating consistency between different objects, however, highlights the prevailing influence of light on color preference and once again, the use of different experimental objects does not show a conspicuous impact. In short, the results depicted in Figs. 1 and 2 strengthen our former conclusions with much stronger evidence, and they imply that the visual perception of the lit environment is the leading factor, which affects the perception of color preference in booths.
Fig. 2. Experimental results of Jost-Boissard et al., Labels of abscissa axis (i.e. WA, WCA, WCR…) refer to the tabs of the experimental light sources defined by those researchers, as introduced in Sections 2.5 and 2.6. Asterisks besides the Pearson correlation coefficient r represent significant correlation (p<0.05) between ratings for different experimental objects.
Thus, it seems that the findings described above should be attributed to the stimulus configuration of the light booth experiments, as well as to the corresponding cognitive interpretation process of vision. During the visual experiments, observers were subject to a continuous learning process and it is unavoidable that after several rounds of judgments, they should realize that the variables of the experiment were experimental lights rather than objects. At that stage, failure to focus on the color appearance of the objects would gradually occur. In other words, knowledge of observing the same objects would help the observer’s visual system to further discount the illuminant [30] and produced a relatively consistent color perception for the objects. Therefore, it is possible that during the remaining part of the test, the observers paid more attention, consciously or unconsciously, to the lit environment of the booth rather than the objects. That could explain why the color preference ratings for different objects, as well as for the lit empty booth, were consistent in all eight studies.
In addition to this, several concerns should be stated to draw safe conclusions. First of all, due to the inherent features of the visual cognition process rooted in a light booth experiment, the influence of visual perception for the lit neutral environment seems to be preponderant while the variation in experimental objects in the booth is much less significant. Therefore, if the research aim is to evaluate color preference of lighting for general conditions, researchers might not need to be concerned too much about the experimental objects. However, if the research intent is to investigate color preference of lighting in a more specific way (e.g. discussing the impact of objects or exploring the color preference for specific application), adopting particular objects in booth is surely important. For instance, in our former work, a broad range of objects were illuminated by the same group of light sources and it was interesting to find that familiarity to the experimental objects and color features of the objects influenced color preference judgments [18]. Obviously, those findings also deepened the knowledge on this topic.
Second, the time course of chromatic adaptation is another issue which deserves caution. According to Fairchild and Reniff, chromatic adaptation occurs in two phases, one extremely rapid (a few seconds) and the other slow (approximately 1 min to reach 90% of its steady-state level) [31]. However, for all the studies mentioned above, no one allowed more than 60 s for chromatic adaption. Therefore, readers should be aware that the findings mentioned in this manuscript are related to the short mode chromatic adaption scenarios. We believe that such short mode scenarios are also meaningful since they closely related to several practical conditions. For instance, when people are walking around shopping malls or exhibition halls, they rarely fully adapt to a specific lighting condition.
Third, the findings of this paper are double-edged. On the one hand, it is figured out that in a light booth experiment the color preference ratings are greatly affected (or even masked) by the unexpected visual perception of the lit neutral environment. Therefore, due consideration must be given when applying the conclusions obtained in light booth studies for particular lighting applications to various immersive environments (e.g. museum, shopping mall and food store). Alternatively, it must be admitted that the light booth experiment provides a convenient and steady approach for evaluating the color quality of lighting. After all, the generalizability of conclusions obtained from studies with special or extreme setups might be even worse.
To further validate this visual cognition mechanism as applied to an experiment that uses a light booth, we suggest that follow up studies with light booths with a black interior should be carried out as a comparison, since such settings could weaken the impact of the lit environment and enhance the visual perception for the objects in the booth. Similarly, light sources of consistent chromaticities but different spectral power distributions could also be adopted for a further comparison. For those metameric lights, observers will perceive no changes in the neutral color of the interior background in booth, but they might find obvious differences in the color appearance of the object. Note again that all the psychophysical data reported in this work are related to visual perception in a booth with, possibly, insufficient chromatic adaptation. For future work, it will be of considerable interest to conduct similar tests in immersive conditions (i.e. in an empty room and the same room containing different objects) or under a fully adapted condition. We speculate that similar results might be found due to the consistency in stimulus configuration and the underlying visual cognitive mechanism.
4. Conclusion
In this study, strong evidence is provided to demonstrate the significant impact of the light booth environment on color preference perception during consecutive visual judgement. It seems that the influence of the object on color preference is markedly weakened by the intrinsic features of the visual cognition process rooted in the experiment. These findings provide deeper understanding of the explanation of the psychophysical results obtained in light booth experiments and they also highlight the need for a more reliable and comprehensive approach for evaluating the color preference of lighting.
Appendix
Table 2. The colorimetric properties of SPDs and their typical color quality metric values for [24]
Table 3. The colorimetric properties of SPDs and their typical color quality metric values for [26]
Table 4. The colorimetric properties of SPDs and their typical color quality metric values for [22]
Table 5. The colorimetric properties of SPDs and their typical color quality metric values at 3050 K for [20]
Table 6. The colorimetric properties of SPDs and their typical color quality metric values at 3950 K [20]
Table 7. The colorimetric properties of SPDs and their typical color quality metric values at 3000 K for [27]
Table 8. The colorimetric properties of SPDs and their typical color quality metric values at 4000 K for [27]
Funding
Young Scientists Fund (61505149); Young Talent Project of Wuhan City of China (2016070204010111).
Acknowledgments
The authors would like to thank Sophie Jost-Boissard and her colleagues for sharing their psychophysical data.
Disclosures
The authors declare no conflicts of interest.
References
1. D. Nickerson and C. W. Jerome, “Color rendering of light sources: CIE method of specification and its application,” Illum. Eng. 60(4), 262–271 (1965).
2. S. Jost-Boissard, C. Cauwerts, and P. Avouac, “CIE 2017 color fidelity index Rf: a better index to predict perceived color difference?” J. Opt. Soc. Am. A 35(4), B202–B213 (2018). [CrossRef]
3. P. Boyce, “Editorial: The meaning of preference,” Lighting Res. Technol. 49(3), 291 (2017). [CrossRef]
4. K. Smet, W. R. Ryckaert, M. R. Pointer, G. Deconinck, and P. Hanselaer, “Optimal color quality of LED clusters based on memory colors,” Opt. Express 19(7), 6903–6912 (2011). [CrossRef]
5. R. Dangol, M. S. Islam, M. Hyvärinen, P. Bhusal, M. Puolakka, and L. Halonen, “User acceptance studies for LED office lighting: Preference, naturalness and colorfulness,” Lighting Res. Technol. 47(1), 36–53 (2015). [CrossRef]
6. O. Masuda and S. M. Nascimento, “Best lighting for naturalness and preference,” J. Vis. 13(7), 4 (2013). [CrossRef]
7. T. Khanh, P. Bodrogi, Q. Vinh, and D. Stojanovic, “Color preference, naturalness, vividness and color quality metrics, Part 1: Experiments in a room,” Lighting Res. Technol. 49(6), 697–713 (2017). [CrossRef]
8. T. Khanh, P. Bodrogi, Q. Vinh, and D. Stojanovic, “Color preference, naturalness, vividness and color quality metrics, Part 2: Experiments in a viewing booth and analysis of the combined dataset,” Lighting Res. Technol. 49(6), 714–726 (2017). [CrossRef]
9. L. Xu, M. R. Luo, and M. R. Pointer, “The development of a color discrimination index,” Lighting Res. Technol. 50(5), 681–700 (2018). [CrossRef]
10. L. Jiang, P. Jin, and P. Lei, “Color discrimination metric based on cone cell sensitivity,” Opt. Express 23(11), A741–A751 (2015). [CrossRef]
11. Q. Liu, Z. Huang, K. Xiao, M. R. Pointer, S. Westland, and M. R. Luo, “Gamut Volume Index: a color preference metric based on meta-analysis and optimized color samples,” Opt. Express 25(14), 16378–16391 (2017). [CrossRef]
12. K. Smet, W. Ryckaert, M. R. Pointer, G. Deconinck, and P. Hanselaer, “A memory color quality metric for white light sources,” Energy Build. 49, 216–225 (2012). [CrossRef]
13. T. Khanh and P. Bodrogi, “Color preference, naturalness, vividness and color quality metrics, Part 3: Experiments with makeup products and analysis of the complete warm white dataset,” Lighting Res. Technol. 50(2), 218–236 (2018). [CrossRef]
14. Y. Lin, M. Wei, K. Smet, A. Tsukitani, P. Bodrogi, and T. Khanh, “Color preference varies with lighting application,” Lighting Res. Technol. 49(3), 316–328 (2017). [CrossRef]
15. P. Bodrogi, T. Khanh, D. Stojanovic, and Y. Lin, “Intercultural color temperature preference of Chinese and European subjects living in Germany,” Light Eng. 24(1), 8–11 (2016). [CrossRef]
16. K. Smet, W. R. Ryckaert, M. R. Pointer, G. Deconinck, and P. Hanselaer, “Correlation between color quality metric predictions and visual appreciation of light sources,” Opt. Express 19(9), 8151–8166 (2011). [CrossRef]
17. M. Wei, K. Houser, A. David, and M. Krames, “Color gamut size and shape influence color preference,” Lighting Res. Technol. 49(8), 992–1014 (2017). [CrossRef]
18. Z. Huang, Q. Liu, S. Westland, M. R. Pointer, M. R. Luo, and K. Xiao, “Light dominates color preference when correlated color temperature differs,” Lighting Res. Technol. 50(7), 995–1012 (2018). [CrossRef]
19. Q. Wang, H. Xu, F. Zhang, and Z. Wang, “Influence of color temperature on comfort and preference for LED indoor lighting,” Optik 129, 21–29 (2017). [CrossRef]
20. S. Jost-Boissard, M. Fontoynont, and J. Blanc-Gonnet, “Perceived lighting quality of LED sources for the presentation of fruit and vegetables,” J. Mod. Opt. 56(13), 1420–1432 (2009). [CrossRef]
21. E. E. Dikel, G. J. Burns, J. A. Veitch, S. Mancini, and G. R. Newsham, “Preferred chromaticity of color-tunable LED lighting,” Leukos 10(2), 101–115 (2014). [CrossRef]
22. Y. Liu, Q. Liu, Z. Huang, M. R. Pointer, L. Rao, and Z. Hou, “Optimising color preference and color discrimination for jeans under 5500 K light sources with different Duv values,” Optik, doc. ID 163916 (posted 27 November 2019, in press).
23. Q. Liu, Z. Huang, M. R. Pointer, M. R. Luo, K. Xiao, and S. Westland, “Evaluating color preference of lighting with an empty light booth,” Lighting Res. Technol. 50(8), 1249–1256 (2018). [CrossRef]
24. W. Chen, Z. Huang, L. Rao, Z. Hou, and Q. Liu, “Research on Color Visual Preference of Light Source for Black and White Objects,” LNEE (to be published).
25. Z. Huang, Q. Liu, Y. Liu, M. R. Pointer, M. R. Luo, Q. Wang, and B. Wu, “Best lighting for jeans, Part 1: Optimizing color preference and color discrimination with multiple correlated color temperatures,” Lighting Res. Technol. 51(8), 1208–1223 (2019). [CrossRef]
26. Z. Huang, Q. Liu, M. R. Pointer, M. R. Luo, B. Wu, and A. Liu, “White lighting and color preference, part A: correlation analysis and metrics validation based on four groups of psychophysical studies,” Lighting Res. Technol. 52(1), 5–22 (2020). [CrossRef]
27. S. Jost-Boissard, P. Avouac, and M. Fontoynont, “Assessing the color quality of LED sources: Naturalness, attractiveness, colorfulness and color difference,” Lighting Res. Technol. 47(7), 769–794 (2015). [CrossRef]
28. Y. Ohno and S. Oh, “Vision experiment II on white light chromaticity for lighting,” in Proceedings of the CIE 2016 Lighting Quality and Energy Efficiency (CIE, 2016), pp. 175–184.
29. Y. Ohno and M. Fein, “Vision experiment on acceptable and preferred white light chromaticity for lighting,” in Proceedings of the CIE 2014 Lighting Quality and Energy Efficiency (CIE, 2014), pp. 192–199.
30. M. D. Fairchild, Color Appearance Models (JOHN WILEY & SONS, INC, 2013).
31. M. D. Fairchild and L. Reniff, “Time course of chromatic adaptation for color-appearance judgments,” J. Opt. Soc. Am. A 12(5), 824–833 (1995). [CrossRef]
32. A. David, P. T. Fini, K. Houser, Y. Ohno, M. P. Royer, K. Smet, M. Wei, and L. Whitehead, “Development of the IES method for evaluating the color rendition of light sources,” Opt. Express 23(12), 15888–15906 (2015). [CrossRef]
33. J. P. Freyssinier and M. Rea, “A two-metric proposal to specify the color-rendering properties of light sources for retail lighting,” Proc. SPIE 7784, 77840V (2010). [CrossRef]
34. K. Smet, W. R. Ryckaert, M. R. Pointer, G. Deconinck, and P. Hanselaer, “Memory colors and color quality evaluation of conventional and solid-state lamps,” Opt. Express 18(25), 26229–26244 (2010). [CrossRef]
35. K. Smet, G. Deconinck, and P. Hanselaer, “Chromaticity of unique white in object mode,” Opt. Express 22(21), 25830–25841 (2014). [CrossRef]
