Do Arts Come From an Inborn Need to Create and Communicate Complementary and Contrasting Colors

Abstract

Color composition in paintings is a disquisitional gene affecting observers' artful judgments. We examined observers' preferences for the colour limerick of Japanese and Occidental paintings when their color gamut was rotated. In the experiment, observers were asked to select their preferred epitome from original and three hue-rotated images in a four-alternative forced choice epitome. Despite observers' being unfamiliar with the presented artwork, the original paintings (0 degrees) were preferred more oft than the hue-rotated ones. Furthermore, the original paintings' superiority was observed when the images were divided into small square pieces and their positions randomized (Scrambled condition), and when the images were composed of square pieces nerveless from unlike fine art paintings and composed equally patchwork images (Patchwork status). Therefore, the original paintings' superiority regarding preference was quite robust, and the specific objects in the paintings associated with a particular colour played only a limited role. Rather, the original paintings' general trend in color statistics influenced hue-angle preference. Fine art paintings probable share mutual statistical regulations in colour distributions, which may be the basis for the universality and superiority of the preference for original paintings.

Introduction

A scientific approach to the artful experience began with Gustave Fechner's piece of work on the gilded section hypothesis¹ and was followed past numerous contributions from psychological, neuroaesthetic, and theoretical studies. Since the aesthetic experience or perception of beauty is frequently associated with artwork^two, the underlying brain mechanisms in the aesthetic process for artwork have been discussed extensively^three,4,5,vi. More than recently, the field of neuroaesthetics has grown and expanded to involve not‑invasive brain stimulation⁷ to investigate the encephalon–behavior causal relationship and the computational aesthetic arroyo^8,ix to bridge the gap between experimental and theoretical studies. Although studies conducted to appointment take by and large suggested that various regions along the visual dorsal and ventral pathways are involved in the artful experience, no unified view has been reached^10,11.

Another approach to the aesthetic experience includes analysis of visual features or artwork representations, such as spatial frequency structure¹², creative ambiguities and conventions^xiii, representation of specular reflection^xiv, effects of the observers' gaze¹⁵, spatial composition^sixteen, and efficiency of visual information coding¹⁷. Since color is one of the most relevant components of art paintings and may carry substantially different valence than other visual aspects like shape^eighteen, at that place have been many attempts to analyze paintings' color to understand the brain's aesthetic processing. For case, information technology was found that the colour gamut of dissimilar artists' paintings tends to exhibit the similar bluish-yellow elongation found in the colour gamut of typical natural environments^nineteen,xx. The effects of colour contrast, particularly by a pair of complementary colors, play an of import part in paintings^thirteen. An analysis of the chromatic limerick of paintings past van Gogh was conducted to estimate his color palettes, which succeeded in detecting his usage of complementary color contours in paintings²¹.

Although attempts to analyze fine art paintings shed light on artistic strategies contributing to aesthetics, an analysis itself does not provide straight connections to the aesthetic experience because these studies practise non include aesthetic judgment. Instead, this line of research is based on the hypothesis that art paintings that survive the selective pressures exerted past cultural institutions, collectors, and fads should contain relevant and effective visual stimuli to the nervous system^ten. The experimental arroyo to aesthetic judgment aims to mensurate observers' responses such as like/dislike or beautiful/ugly and to investigate the causes/reasons for those responses. Non limited to paintings, there have been empirical studies on preference for complex objects such as compages²² or mathematics²³. Nevertheless, since visual artwork is composed of visual elements arranged in space, there are numerous studies on aesthetic responses to such visual elements, for example, preference for the aspect ratio of a rectangle in Fechner's experimentⁱ. It has been argued that people prefer horizontal and vertical lines rather than oblique ones²⁴, curved-contour objects rather than sharp-angled ones²⁵, and symmetrical shapes rather than asymmetrical ones²⁶, and tend to look at the symmetrical center of images²⁷. A recent computational written report²⁸ demonstrated that a simple linear summation of depression-level visual features, such as mean hue and mean dissimilarity, could predict observers' subjective value ratings for art paintings.

Of these visual elements, colour has accumulated the most scientific knowledge on preference²⁹, because color preference is an important visual experience that influences a wide range of human behaviors. It has been argued, however, that experimental research on color preference is "chaotic", in that inadequate measurement command has led to disparate results^29,30. Recent studies accept suggested more general hypotheses or explanations for colour preference. For instance, it was argued that a universal preference for blue hues exists^{thirty,31,32,33,34}. A like tendency was reported in animals such as rhesus monkeys³⁵, pigeons³⁶, and zebrafish³⁷, but there were more than individual differences in chimpanzees³⁸. Since one of the most fundamental issues of color preference is the existence of a universal tendency toward preference for a specific colour³⁹, the bluish preference bias has attracted researchers' attention, and there have been attempts to explain why people generally like blue more than other colors. One hypothesis is that colour preference is securely influenced by natural colors which convey biologically important signals⁴⁰. Taking this hypothesis further, the theory of evolutionary accommodation of the visual organization has proposed that color preferences are wired into the visual system as weightings on a cone opponent neural signal to improve functioning on important tasks³⁰.

Linked to these theories, another explanation was proposed based on ecological valence³³, arguing that colour preference arises from people's affective responses to color-associated objects. Indeed, some claim that systematic individual differences in color preferences exist. Cultural differences in color preference accept been revealed by simple pairwise comparison experiments^thirty, showing that women's preferences tend toward reddish hues whereas men's preferences tend toward blueish hues, although these tendencies are modulated past the cultural context. Additionally, information technology was argued that sex activity differences in colour preference may arise from socialization or cognitive gender development rather than inborn factors⁴¹. The ecological valence theory³³ was proposed to explain and predict the individual differences in color preference. For example, this theory explains that color preferences of undergraduate students are caused past learned affective responses to university colors⁴². Other related inquiry⁴³ suggests that cultural differences in color preference can as well be explained by civilization-specific associations with colored objects. In contrast, some merits that there exists a universality in color preferences. Several experimental and theoretical studies take unsaid that color preferences are, to a considerable extent, independent of personal factors, and seem to have a firm biological basis²⁹. However, neither the extent to which color preferences are universal nor the extent to which they are influenced by civilisation or gender is understood well.

Another issue with studies on color preference is that most only deal with a single colour or a combination of few colors. Even for a combination of two colors, information technology is quite difficult to place the general preference for multiple color combinations; that is, preference for color pairs cannot be predicted consistently by a preference for a single color⁴⁴. Since art paintings generally comprise numerous colors, there is a tremendous gap between preferences for a unmarried colour and a range of colors, information technology seems futile to approach color preference in art paintings. Nonetheless, more recent studies^45,46,47 take measured observers' preferences for color limerick past rotating the color gamut within the CIELAB color space. In these studies, observers were asked to adjust the hue angle of the color gamut of unfamiliar paintings to obtain the best subjective visual impression or to select the preferred one amid pairs of hue-rotated versions of the same painting. These studies found that observers typically prefer a chromatic composition very close to the original rather than ane that is hue-rotated, fifty-fifty for spatially scrambled paintings⁴⁶. Based on those results, nosotros advise the following hypothesis: a universal preference for color composition in art paintings exists, and it is contained of observers' art knowledge or experiences.

However, there are still several unsolved issues: (a) how robust is the original paintings' superiority regarding preference for the colour composition of art paintings, (b) how culturally dependent are the preference data, (c) to what extent exercise the spatial configuration or figurative elements of paintings influence color preference, and (d) to what extent are the statistical features or regularities of colour distribution shared among fine art paintings. Thus, this study aims to explore these issues underlying the original-preferred judgment for fine art paintings past adopting a simple iv-culling forced choice (4AFC) prototype to measure the preference for color composition of art paintings. To investigate the cultural dependency of preference for the colour composition of Japanese and Occidental art paintings, observers were recruited in Nippon and Portugal. Observers were asked to select their most preferred paradigm among an original and three hue-rotated versions of the paintings (90, 180, and 270 degrees in hue angle) generated by the same procedure equally the previous written report⁴⁵. Additionally, to investigate the furnishings of figurative elements and spatial context on preference, the spatial scrambling status was introduced, in which paintings were divided into a set up of pocket-sized square pieces and scrambled in their position. This generates spatial context-gratuitous images with the same color distribution equally the original painting⁴⁶. Finally, to explore the commonality of color distribution statistics in art paintings, nosotros measured preferences for "patchwork" images consisting of small square pieces collected from a unlike set of paintings. If common features in the color distribution of the fine art paintings exist, then the patchwork images, as the subset of a grouping of art paintings, should accept similar color distribution features to paintings. If this is the case, and if the color distribution is the principal determinant of hue-angle preference, and so nosotros can expect the hue-angle preference to testify similar trends for patchwork, scrambled, and original images. Therefore, comparing hue-bending preferences under these conditions should provide important data on the effects of spatial configuration and colour distribution in fine art paintings on hue-angle preference.

Materials and methods

Paintings

Twenty paintings (10 Occidental and ten Japanese) which were digitized at the Centro de Arte Moderna da Fundação Calouste Gulbenkian, Lisboa, Portugal; the Museu Nogueira da Silva, Braga, Portugal; the color laboratory of the Academy of Minho, Braga, Portugal; and Toyohashi City Museum of Fine art and History, Toyohashi, Japan were used. They were measured using a hyperspectral camera to judge spectral reflectance and reproduce RGB images illuminated by daylight with the correlated color temperature (CCT) of 6500 K (D65) every bit the visual stimuli displayed on calibrated monitor screens (for more details, encounter the previous studies^45,48). The Occidental paintings were oil paintings dating from the Renaissance to the modern period (fourteenth–twentieth centuries). Japanese paintings were created during the late Edo catamenia up to the modern period (eighteenth–twentieth centuries), all of which used traditional Japanese paper (washi) and pigments derived from natural ingredients such as minerals, shells, or corals (iwa-enogu). In addition, image information from xx paintings (ten figurative and ten abstract) were collected from three art painting galleries on the Net (The Metropolitan Museum of Art ⁴⁹, Web Gallery of Fine art ⁵⁰, WikiArt ⁵¹ , and Wikimedia Commons ⁵²) to use as visual stimuli in the experiments. For paintings measured in museums, nosotros selected the paintings that could exist successfully captured by a hyperspectral camera, because some paintings in museums, particularly Japanese paintings, contain materials with specular reflections, such as gold leaf, which cause dynamic range problems. Additional paintings in the Net gallery were collected using a random principle, from figurative and abstract genres; paintings with only a few colors or using only achromatic colors were eliminated manually from the collected samples.

As shown in Fig. i, the 40 images of paintings listed in Table 1 were divided into 3 image sets: two sets (Fix i and Prepare two) with ten images each, including five paintings from museums (two Japanese and 3 Occidental paintings) and five paintings (all Occidental paintings) from art painting galleries on the Internet; and the third (Prepare iii) with x paintings from museums (vi Japanese and four Occidental paintings) and ten paintings (all Occidental paintings) from fine art painting galleries on the Internet. Paintings in these sets were selected using a random principle, and used in different experimental conditions (C0, C1, C2, and C3) explained below.

Table 1 List of paintings used in the experiment.

Full size table

Participants

In total, 45 Japanese individuals (37 men and 8 women, anile 21–65 years, mean 26.half-dozen years, SD ten.5 years) and 45 Portuguese individuals (xiii men and 32 women, aged 17–25 years, mean 19.8 years, SD ane.8 years) completed the main experiments in either Japan or Portugal under exactly the same weather condition. To analyze the stimulus dependency of preference data in the main experiment, a replication experiment was conducted with identical procedures and different visual stimuli. The participants in the replication experiment were 44 newly recruited Japanese observers (26 men and 18 women, anile 21–68 years, hateful 34.8 years, SD fourteen.9 years old) and 44 Portuguese observers (12 men and 32 women, aged 17–25 years, mean 19.8 years, SD 1.8 years) who had also participated the main experiment. All observers were unaware of the purpose of the experiments. They had no formal artistic education and only basic knowledge virtually the art paintings. After they finished the experiments, observers were asked verbally about their knowledge of the paintings used every bit stimuli, and none of them had previous knowledge well-nigh the paintings. Undergraduate and graduate university students in Nihon and Portugal were recruited to participate. All participants had normal or corrected-to-normal visual acuity and normal color vision as tested with Ishihara plates. All experimental procedures were completed according to the ethical principles outlined in the Proclamation of Helsinki and approved by the Committee for Human Inquiry at the Toyohashi Academy of Technology. The experiment was strictly conducted post-obit the canonical guidelines of the commission. Written informed consent was obtained from the participants later procedural details were explained.

Experimental procedures

Stimuli

The stimuli used in the experiment consisted of four types of images of the paintings, including originals, selected from Ready 1 in the main experiment, and from Ready ii in the replication experiment. The standard RGB (sRGB) coordinates of each pixel of the original images were computed to display the images on calibrated monitor screens and were transformed to the corresponding color coordinates (L*, a*, b*) in the CIELAB colour space. The average values of L*, a*, and b* of each original image ranged from 20.seven to 63.9, mean = 47.9 for 50*; from − xv.9 to 31.nine, mean = 5.21 for a*; and from − 13.ix to 46.6, mean = 12.2 for b*, respectively. The hue-rotated images were obtained past rotating the colour gamut of the originals, a set of points for each pixel coordinate in the CIELAB color infinite, around an axis parallel to the L* centrality passing through the hateful chromaticity indicate on the a*–b* airplane. The original corresponded to 0 degrees and the other three hue-rotated images were obtained past rotating the colour gamut by 90, 180, and 270 degrees counterclockwise on the a*–b* plane. Figure 2 presents typical examples of the hue-rotation upshot on color appearance in both Occidental and Japanese paintings. When the hue of the original images displayed on the monitor screen were rotated, some pixel colors were outside the gamut of the monitor. Colors out of the gamut were projected onto the closest displayable colors in the CIELAB color space and chromatic errors occurred for these pixels. The largest chromatic error amid the hue-rotated images used in the experiments was \({\Delta E}_{ab}=3.8\), which is well-nigh the threshold for complex images^45,53; thus, the effect of the gamut compression acquired past the hue rotation was fairly negligible.

Experimental atmospheric condition

Some paintings used in the experiment contained familiar objects associated with specific retentiveness colors, such equally a blue sky⁵⁴. To demonstrate whether hue-angle preference depended on the overall colour limerick of the paintings or specific objects' colors in the paintings, the spatial scramble was introduced to manipulate the spatial composition and visibility of the figurative elements of the paintings, every bit shown in Fig. iii. In C0 (Original condition), originals (selected from Gear up 1 in the primary experiment and from Fix 2 in the replication experiment) and hue-rotated images were preserved in their original spatial form. In C1 (Scrambled condition), each original paradigm was divided into a gear up of small square pieces with a side length of 73 pixels (10% of the summit of the epitome), and the position was randomized. The generated images were the aforementioned size as the originals and were used to create the hue-rotated images. In C2 (Patchwork condition), each paradigm was divided into a ready of small pieces in the same manner as for C1, and each stimulus was generated by composing 100 randomly selected pieces (10 × 10 pieces) from the groups of 20 unlike paintings measured at museums or of xx unlike paintings from the galleries on the Internet, in a mixture of all three sets (Sets one, 2, and 3). In C3 (Randomized condition), each stimulus was generated from pieces of different images like to that in C2; however, each piece was selected from the hue-rotated images with a random caste from the original. In each condition, 10 sets of four types of images, including original or generated images, were used every bit stimuli. In C2, five images each were generated from images measured at museums and from art painting galleries on the Internet.

In C1, C2, and C3, the original paintings were divided into small pieces (73 × 73 pixels) to make information technology difficult to place any specific objects from i slice. In C1, the generated images had color compositions identical to the originals. In C2, the population of the generated images had the same color statistics as the population of the 20 original paintings. In C3, the generated images were randomized spatially and chromatically from the originals. Therefore, by comparing the results of C0 with C1, C2, and C3, information technology was possible to decide whether spatial compositions, figurative elements, or color composition influenced the hue-angle preference, or if it were influenced by the general tendencies of the original paintings' color statistics.

Appliance and task

The original (0 caste) and iii hue-rotated images (90, 180, and 270 degrees) were displayed on a calibrated monitor in two columns and two rows, ane.7 cm autonomously. The viewing distance was 50.0 cm, the superlative of the images on the screen was stock-still to 6.0 cm (730 pixels), and the width varied from 3.9 cm (478 pixels) to 11.2 cm (1365 pixels) depending on the original image size. The stimuli were displayed on a 12.4-inch screen with a resolution of 2736 × 1824 pixels (Surface Pro4, Microsoft). Observers were instructed to select their most preferred image amid the 4 past touching the screen (4AFC). Each stimulus condition (shown in Fig. iii) included 10 epitome sets with original and iii hue-rotated images and each stimulus ready was presented but once, resulting in a total of 40 trials. The guild of stimulus presentation was random. No information was provided to observers virtually the original paintings and hue rotation and there was no fourth dimension limit to respond in each trial.

Data analysis

The average rates of pick for each epitome category (original and 90, 180, and 270 degrees hue-rotated) were analyzed with a multinomial examination separately for Japanese and Portuguese observers and each condition. Selection rates for the original images (0 degrees) as an index of the original paintings' superiority were analyzed by 2-fashion repeated-measures analysis of variance (ANOVA) for the stimulus conditions (C0, C1, C2, and C3) and observer nationality (Japanese and Portuguese) as factors. The level of statistical significance was set at p < 0.05 for all analyses. Within-subjects effects were corrected using the Greenhouse–Geisser correction. Pairwise comparisons for the main furnishings were corrected for multiple comparisons using the Bonferroni correction. Effect sizes (η ² and \({\omega }^{2}\)) were adamant for the ANOVA. All the statistical analyses were performed using the statistical software JASP⁵⁵.

Results

Figure 4 shows the boilerplate pick rates measured in the main experiment for each image category, the original and three hue-rotated images, measured from Japanese and Portuguese observers in weather C0, C1, C2, and C3. In condition C0, in which the displayed images were preserved in their original spatial forms, the average selection rate for the original images was significantly above hazard levels for both Japanese and Portuguese observers (Japanese observers: \({\mathrm{\rm X}}^{2}= 630.978; p<0.001\), Portuguese observers: \({\mathrm{\rm 10}}^{2}= 377.787; p<0.001\)), implying that the original images were preferred more frequently than the hue-rotated images. The original paintings' superiority with regard to preference was observed in the spatially scrambled condition (C1), every bit shown in Fig. 4b, and even in the patchwork condition (C2), as shown in Fig. 4c. In both C1 and C2, the average option rates for the original images were significantly in a higher place chance levels (C1: \({\mathrm{\rm Ten}}^{2}= 268.667, p<0.001\) for Japanese observers; \({\mathrm{\rm X}}^{2}= 168.062, p<0.001\) for Portuguese observers; C2: \({\mathrm{\rm X}}^{2}= 562.391, p<0.001\) for Japanese observers; \({\mathrm{\rm X}}^{2}= 226.142, p<0.001\) for Portuguese observers), depicting the same tendency institute in C0. However, in C3, which randomized the images both spatially and chromatically, differences between the boilerplate selection rates and risk levels were not observed for either Japanese or Portuguese observers \(({\mathrm{\rm Ten}}^{2}= 1.876, p=0.599\) for Japanese observers; \({\mathrm{\rm Ten}}^{2}= 0.578, p=0.902\) for Portuguese observers).

Figure 5 shows the average selection rates measured in the replication experiment in the aforementioned way as Fig. four. In C0, as shown in Fig. 5a, the average pick rates for the original images were significantly above chance levels for both Japanese and Portuguese observers (Japanese observers: \({\mathrm{\rm X}}^{two}= 247.999,\mathrm{ p}<0.001\); Portuguese observers: \({\mathrm{\rm 10}}^{2}= 356.473,\mathrm{ p}<0.001\)), conspicuously replicating the results obtained in the primary experiment shown in Fig. 4a. Effigy 5b–d show the pick rates in C1, C2, and C3. The average option rates for the original images were significantly above hazard levels in C1 and C2 (C1: \({\mathrm{\rm X}}^{2}= 142.764,\mathrm{ p}<0.001\) for Japanese observers; \({\mathrm{\rm X}}^{2}= 182.091,\mathrm{ p}<0.001\) for Portuguese observers; C2: \({\mathrm{\rm 10}}^{ii}= 282.455,\mathrm{ p}<0.001\) for Japanese observers; \({\mathrm{\rm X}}^{2}= 288.982,\mathrm{ p}<0.001\) for Portuguese observers), while the differences between the average option rates and chance levels were not observed for either Japanese or Portuguese observers in C3 \(({\mathrm{\rm X}}^{ii}= 1.982,\mathrm{ p}=0.576\) for Japanese observers; \({\mathrm{\rm X}}^{2}= i.145,\mathrm{ p}=0.766\) for Portuguese observers).

Effigy 6a reiterates the selection rates for the original images measured in the four conditions in the master experiment shown in Fig. 4. ANOVA on these choice rates revealed significant main effects of condition (F[ii.231, 196.322] = 110.213 [with the Greenhouse–Geisser correction], p < 0.001; \({\eta }^{two}\) = 0.354, \({\omega }^{2}\) = 0.363) and observer nationality (F[ane, 88] = six.500, p = 0.013; \({\eta }^{2}\) = 0.024, \({\omega }^{2}\) = 0.030). There was a pregnant interaction between condition and observer nationality (F[2.231, 264] = 3.944 [with the Greenhouse–Geisser correction], p = 0.017; \({\eta }^{2}\) = 0.013, \({\omega }^{2}\) = 0.015). A post-hoc comparison amidst conditions revealed that the average choice rate for the originals in C3 was significantly lower than that in other conditions: C0 (t = xvi.731, p < 0.001; mean divergence = 0.452, 95% CI [0.380, 0.524]); C1 (t = 10.894, p < 0.001; mean difference = 0.294, 95% CI [0.223, 0.366]); and C2 (t = xiv.511, p < 0.001; mean difference 0.392, 95% CI [0.320, 0.464]). The average option rate in C1 was lower than that in C0 (t = 5.837, p < 0.001; mean departure 0.158, 95% CI [0.086, 0.230]), and C2 (t = 3.618, p = 0.002; mean difference 0.098, 95% CI [0.026, 0.170]). In that location was a meaning deviation between the selection rates of Japanese and Portuguese observers only in condition C2 (t = 3.616, p = 0.010; mean divergence 0.176, 95% CI [0.022, 0.329]).

Figure 6b reiterates the average selection rates for the original images measured in the replication experiment shown in Fig. v. ANOVA on these selection rates revealed pregnant main effects of status (F[one.833, 158.040] = fourscore.742 [with the Greenhouse–Geisser correction], p < 0.001; \({\eta }^{2}\) = 0.269, \({\omega }^{2}\) = 0.265), and no effects of observer nationality (F[1, 86] = 0.309, p = 0.580; \({\eta }^{ii}\) = 0.002, \({\omega }^{two}\) = 0.000). There was no pregnant interaction betwixt condition and observer nationality (F[1.838, 158.040] = 0.712 [with the Greenhouse–Geisser correction], p = 0.481, \({\eta }^{2}\) = 0.002, \({\omega }^{2}\) = 0.000). A post-hoc comparison among conditions revealed that the boilerplate option rate for the original in C3 was significantly lower than that in C0 (t = xiii.647, p < 0.001; mean difference = 0.376, 95% CI [0.303, 0.449]), C1 (t = nine.854, p < 0.001; mean departure 0.272, 95% CI [0.198, 0.245]), and C2 (t = 13.234, p < 0.001; hateful difference 0.365, 95% CI [0.291, 0.438]). The average selection rate in C1 was lower than in C0 (t = 3.793, p = 0.001; hateful difference 0.105, 95% CI [0.031, 0.178]), and C2 (t = three.381, p = 0.005; mean difference 0.093, 95% CI [0.020, 0.166]).

Word

The nowadays study aimed to reply the questions underlying the original-preferred judgment for art paintings: (a) how robust is the original paintings' superiority with regard to preference for the colour composition of art paintings, (b) how culturally dependent are the preference data, (c) to what extent do the spatial configuration or figurative elements of paintings influence color preference, and (d) to what extent are the statistical features or regularities of color distribution shared amidst art paintings. We performed a series of 4AFC experiments using original and hue-rotated images to measure the preference for unfamiliar art paintings in Japanese and Portuguese observers past manipulating figurative elements of paintings by scrambling the images. The results showed that both Japanese and Portuguese observers preferred the originals more than than the hue-rotated ones, replicating findings from previous studies^45,46 on original paintings' superiority regarding preference, thus confirming that the preference for the color limerick of the original paintings was extremely robust. The original paintings' superiority was observed non but in the condition that preserved the original spatial composition of paintings (Original condition) just as well in the status with spatial scrambling in which the images were divided into pocket-sized square pieces and their positions were randomized (Scrambled status), and even in the conditions that mixed the square pieces of unlike art paintings (Patchwork condition). These findings imply that spatial context or specific objects in paintings associated with a item color practise not play an important role in hue-angle preference. Rather, the general trend in the color statistics of the original paintings influences hue-angle preference.

We observed a pregnant difference in the boilerplate selection rates for the original paintings betwixt Japanese and Portuguese observers in the main experiment, although the effect size was relatively small-scale⁵⁶ (\({\eta }^{two}\) = 0.024, \({\omega }^{2}\) = 0.015). Although cultural dependency in preference for a single color has been reported³⁰, the influence of cultural context or individual experience is complex and thus not yet clear. When we focused on the color preferences between Japanese and Portuguese observers, there was a clear difference in the preference for the CCT of illumination, showing that Japanese observers preferred a significantly higher CCT (more blue) than Portuguese observers for illuminating art paintings⁵⁷. Our findings indicate that original paintings' superiority with regard to preference is mutual, but difference in sensitivities to the original is probable dependent on the selected paintings. We tested this idea with a unlike ready of paintings equally a replication experiment. As shown in Figs. v and 6, the differences in the average selection rates for the originals between nationalities were not replicated, while the original superiorities in all conditions except for the randomized condition were observed. Therefore, nosotros could not find a articulate link betwixt cultural differences and preference level for original color compositions in our current data measured for original and hue-rotated paintings.

When comparing C1 (Scrambled status) with C0 (Original condition), although the figurative elements were designed to disappear, the boilerplate option rate for the originals was significantly above hazard level in C1, only non as high as in C0. This suggests that the spatial context or figurative elements associated with certain colors (e.g., skin or sky) influenced the pick rate for the originals in a limited way. We examined the average selection rates measured in C0 (Original condition) for twenty images used in the experiments separately (come across Supplementary Information). The top three images in terms of the average selection rates were all portraits of man faces, which could exist a stiff cue for original images, as shown in Supplementary Fig. S1. Nevertheless, other than that, there was no dominant dependency on the type of image (figurative/non-figurative). Therefore, except for a few special cases such equally human faces, the color distribution statistics, rather than spatial layout or limerick of figurative elements of paintings, are considered to play a relatively important role in a preference for the originals. However, the most hit finding in the current study is that the average option rate for the originals in C2 (Patchwork status), where the images were composed of small-scale, randomly selected pieces from xx different paintings, was in a higher place run a risk level and more or less similar to C0. This suggests that the color distribution statistics of the original fine art paintings may have like characteristics. Conversely, without this possibility, it is difficult to explain why the images composed of random pieces were preferred and became less preferred when the hue was rotated.

What are the possible statistical features or regularities in the color distribution common to the fine art paintings? Since a degree of preference decreased by the rotation of the color gamut, statistical features linked to the preference, if they exist, should also change at the same time. The hue rotation preserves lightness, hateful chromaticity, saturation, and relative relationships between colors in the image⁴⁵. Previous studies on preferences for a unmarried color^{thirty,31,32,33,34} or the CCT of illumination^57,58 take in principle manipulated the mean chromaticity of stimuli. Although the hue rotation did not modify the mean chromaticity in principle, it did touch the hue angle distribution of an paradigm. It is therefore possible that, for case, the hateful of the hue bending distribution of the paintings is related to the average selection rate. However, nosotros establish no evidence that the mean of the hue distribution had whatever effect on the average choice rates (see Supplementary Fig. S1); thus, the preference bias in the hue domain found in previous studies for single colors^{30,31,32,33,34} does non explain our results here.

It seems, therefore, that the key to answering the question lies in the more general nature of the color gamut of art paintings. An analysis of the 2D shapes of the color gamut of art paintings and natural scenes establish that the color gamut of fine art paintings had shapes elongated in the xanthous-blue directions and this feature was partly shared with natural scenes^xix,20. The hue rotation straight inverse the hue angle of the colour gamut. Ane possible explanation is that artists imitate, implicitly or explicitly, the naturalistic aspect of the color gamut of natural scenes (e.g., elongating in xanthous-blue management) in their paintings, and they intuitively know that this naturalistic feature matches observers' preferences. This was supported by a recent written report⁵⁹ which showed that images perceived as more natural were preferred. All the same, "naturalness" does not necessarily mean resemblance to natural scenes; hence, the question of "regularity of colour distribution determining the preference/naturalness" remains open up^20,59.

It is likely plausible that the artist intuitively understands observers' preferences if at that place exists a universal class regarding preference for color composition. The hue rotation does change the relationship between lightness and colors. If there exists a naturalistic link betwixt lightness and color, and observers intuitively know this relationship, then it is plausible that observers will notice when this naturalistic relationship is violated by the hue rotation. These kinds of regulations or constraints between lightness and colors may derive from the materials used in the paintings, such as pigments and dyes⁶⁰. A more cognitive explanation might business the colour categories; observers prefer images in which more colors tin can be categorized as typical colors because such images might be processed more fluently, every bit posited by the fluency theory²⁶ and argued in a contempo study⁶¹. Although further studies are required to clarify these unsolved issues, as this report showed, the phenomenon that the original paintings were preferred more than other hue-rotated ones was quite robust, and the original paintings' superiority regarding preference was preserved even for the patchwork images equanimous of dissimilar original paintings.

Our report has three main limitations. First, according to our experimental design, obtained choice rate information reverberate relative preferences among four types of images, including original images. Manipulation of hue angles with preserving lightness and mean chromaticity in the current experiment succeeded in identifying the effects of color composition on preferences and showing the original paintings' superiority regarding preference. However, the question remains as to what extent colour composition, compared to other visual features, influences preferences across different paintings. It has been reported that preferences for paintings were well-predicted by preferences for the objects depicted in the paintings⁶², although this cannot explicate preferences for paintings without objects, such as abstruse paintings. More recently, it has been demonstrated that a linear combination of low-level visual features (12 global and 28 local features) can predict preferences for both paintings and photography²⁸. This implies that there exist universal visual features that are relevant to artful judgements. Farther inquiry could explore the details of links between color features and artful judgement as a footing for original paintings' superiority.

Second, we did non detect a clear relationship between cultural differences and preferences for the original color compositions. Although we establish differences in the average selection rates for originals between Japanese and Portuguese observers measured in the main experiment, equally shown in Fig. 6a, information technology could be possible that these differences were dependent on gender, equally the ratio of men to women significantly differed between the Portuguese and Japanese groups \(\left({\mathrm{\rm X}}^{two}= 25.920, p<0.001\right)\). Nonetheless, a replication experiment, shown in Fig. 6b, demonstrated no significant differences in the average selection rates between 2 nationalities, even when in that location was a similar gender imbalance in the observer groups \(\left({\mathrm{\rm X}}^{2}= 9.078, p<0.01\right)\). To explore the furnishings of culture and gender on preference, further studies to compare more than cultures with gender-balanced observers are required. In addition to this, there was a different tendency in the selection rates of Japanese and Portuguese participants in the main and replication experiments. The selection rates of the same Portuguese participants in the 2 experiments were almost the same. In the case of the Japanese participants, however, the selection rates of the different samples in the 2 experiments were more varied. This suggests that, in addition to cultural differences, there may exist remarkable individual differences, although it was not possible to clearly distinguish the effects of these factors because of the deviation between the samples in the two experiments.

Third, the number of paintings used in the experiments was limited. In the present study, we used twenty paintings (Set up 1 and Set ii) in conditions C0, C1, and C3, while some other twenty paintings in Ready 3 were used but for generating patchwork images in C2. The 40 paintings comprised 10 Japanese paintings and 10 occidentals measured in museums, and 20 occidentals collected from Internet art galleries. In the future, paintings with more than varied cultural backgrounds are necessary to examine in depth the cultural dependency in color preference of paintings, and to directly test whether similarity of paintings' color statistics to natural scenes affects the preference for chromatic composition of paintings. These experiments are beingness conducted as office of our upcoming research.

In conclusion, the findings of this study imply that original paintings' superiority is quite robust, and art paintings probable share common statistical regulations in colour distribution, which should exist an important argument for further clarifying the underlying mechanisms for color preference of fine art paintings. Determining whether universal colour features are a footing for preference judgements will shed light on the developmental origins of the human colour vision organisation and its biological value.

Data availability

All datasets generated during this study are included in this article or the analyzed data are fully available from the corresponding author on reasonable request.

References

Fechner, G. T. Uber die Frage des goldenen Schnittes [On the question of the gilt section]. Archiv fur die Zeichnenden Kunste 11, 100–112 (1865).

Google Scholar
Ramachandran, 5. & Hirstein, W. The scientific discipline of art: A neurological theory of aesthetic experience. J. Conscious. Stud. half dozen, 15–51 (1999).

Google Scholar
Silvia, P. J. Emotional responses to art: From collation and arousal to cognition and emotion. Rev. Gen. Psychol. 9, 342–357 (2005).

Google Scholar
Cinzia, D. D. & Vittorio, Grand. Neuroaesthetics: A review. Curr. Opin. Neurobiol. 19, 682–687 (2009).

PubMed Google Scholar
Chatterjee, A. Neuroaesthetics: A coming of historic period story. J. Cogn. Neurosci. 23, 53–62 (2010).

Google Scholar
Jacobsen, T. Beauty and the brain: Culture, history and individual differences in artful appreciation. J. Anat. 216, 184–191 (2010).

PubMed Google Scholar
Cattaneo, Z. Neural correlates of visual aesthetic appreciation: Insights from not-invasive brain stimulation. Exp. Brain Res. 238, ane–16 (2020).

PubMed Google Scholar
Li, R. & Zhang, J. Review of computational neuroaesthetics: Bridging the gap between neuroaesthetics and estimator science. Encephalon Inform. vii, 16 (2020).

PubMed PubMed Central Google Scholar
Lelièvre, P. & Neri, P. A deep-learning framework for human perception of abstract art composition. J. Vis. 21, 9–9 (2021).

PubMed PubMed Central Google Scholar
Conway, B. R. & Rehding, A. Neuroaesthetics and the trouble with beauty. PLoS Biol. eleven, e1001504 (2013).

CAS PubMed PubMed Central Google Scholar
Heinzelmann, N. C., Weber, S. C. & Tobler, P. North. Aesthetics and morality judgments share cortical neuroarchitecture. Cortex 129, 484–495 (2020).

PubMed Google Scholar
Graham, D. J. & Redies, C. Statistical regularities in art: Relations with visual coding and perception. Vis. Res. l, 1503–1509 (2010).

PubMed Google Scholar
Mamassian, P. Ambiguities and conventions in the perception of visual art. Vis. Res. 48, 2143–2153 (2008).

PubMed Google Scholar
Chao, J., Cavanagh, P. & Wang, D. Reflections in art. Spat. Vis. 21, 261–270 (2008).

PubMed PubMed Fundamental Google Scholar
Conway, B. R. & Livingstone, One thousand. South. Perspectives on scientific discipline and art. Curr. Opin. Neurobiol. 17, 476–482 (2007).

CAS PubMed PubMed Central Google Scholar
Tyler, C. Some principles of spatial organisation in art. Spat. Vis. twenty, 509–530 (2007).

PubMed Google Scholar
Graham, D. J. & Meng, Grand. Artistic representations: Clues to efficient coding in human vision. Visual Neurosci. 28, 371–379 (2011).

Google Scholar
Conway, B. R. Color consilience: Color through the lens of art exercise, history, philosophy, and neuroscience. Ann. N. Y. Acad. Sci. 1251, 77–94 (2012).

ADS PubMed Google Scholar
Tregillus, K. E. M. & Webster, M. A. Swapping swatches: Adapting to and from an artist's palette. Electron. Imaging 2016, ane–8 (2016).

Google Scholar
Montagner, C., Linhares, J. M. K., Vilarigues, One thousand. & Nascimento, S. Grand. C. Statistics of colors in paintings and natural scenes. J. Opt. Soc. Am. 33, A170 (2016).

ADS Google Scholar
Berezhnoy, I., Postma, Eastward. & van den Herik, J. Computer assay of van Gogh's complementary colours. Pattern Recognit. Lett. 28, 703–709 (2007).

ADS Google Scholar
Coburn, A. et al. Psychological responses to natural patterns in architecture. J. Environ. Psychol. 62, 133–145 (2019).

Google Scholar
Zeki, Due south., Romaya, J. P., Benincasa, D. M. T. & Atiyah, 1000. F. The experience of mathematical dazzler and its neural correlates. Front. Hum. Neurosci. 8, 68 (2014).

PubMed PubMed Central Google Scholar
Latto, R., Brain, D. & Kelly, B. An oblique effect in aesthetics: Homage to Mondrian (1872–1944). Perception 29, 981–987 (1994).

Google Scholar
Silvia, P. J. & Barona, C. M. Practice people prefer curved objects? Angularity, expertise, and artful preference. Empir. Stud. Arts. 27, 25–42 (2009).

Google Scholar
Palmer, South. E., Schloss, G. B. & Sammartino, J. Visual aesthetics and man preference. Annu. Rev. Psychol. 64, 77–107 (2012).

PubMed Google Scholar
Kootstra, G., de Boer, B. & Schomaker, L. R. B. Predicting heart fixations on complex visual stimuli using local symmetry. Cogn. Comput. 3, 223–240 (2011).

Google Scholar
Iigaya, K., Yi, S., Wahle, I. A., Tanwisuth, K. & O'Doherty, J. P. Aesthetic preference for art can be predicted from a mixture of low- and high-level visual features. Nat. Hum. Behav. https://doi.org/ten.1038/s41562-021-01124-vi (2021).

Commodity PubMed PubMed Central Google Scholar
Granger, G. W. Objectivity of colour preferences. Nature 170, 778–780 (1952).

ADS CAS PubMed Google Scholar
Hurlbert, A. C. & Ling, Y. Biological components of sex differences in color preference. Curr. Biol. 17, R623–R625 (2007).

CAS PubMed Google Scholar
McManus, I. C., Jones, A. L. & Cottrell, J. The aesthetics of colour. Perception ten, 651–666 (1980).

Google Scholar
Camgöz, Due north., Yener, C. & Güvenç, D. Furnishings of hue, saturation, and brightness on preference. Colour Res. Appl. 27, 199–207 (2002).

Google Scholar
Palmer, South. Eastward. & Schloss, M. B. An ecological valence theory of human color preference. Proc. Natl. Acad. Sci. U.S.A. 107, 8877–8882 (2010).

ADS CAS PubMed PubMed Central Google Scholar
Schloss, K. B., Strauss, Eastward. D. & Palmer, S. E. Object color preferences. Colour Res. Appl. 38, 393–411 (2013).

Google Scholar
Humphrey, N. Thou. 'Interest' and 'pleasance': Two determinants of a monkey's visual preferences. Perception 1, 395–416 (1972).

CAS PubMed Google Scholar
Sahgal, A. & Iversen, S. D. Colour preferences in the dove: A behavioural and psychopharmacological written report. Psychopharmacologia 43, 175–179 (1975).

CAS PubMed Google Scholar
Avdesh, A. et al. Evaluation of color preference in zebrafish for learning and memory. J. Alzheimer's Dis. 28, 459–469 (2012).

Google Scholar
Pene, C. H. M., Muramatsu, A. & Matsuzawa, T. Color discrimination and colour preferences in Chimpanzees (Pan troglodytes). Primates 61, 403–413 (2020).

PubMed Google Scholar
Eysenck, H. J. A disquisitional and experimental study of colour preferences. Am. J. Psychol. 54, 385–394 (1941).

Google Scholar
Humphrey, N. The color surrency of nature. In Color for Architecture (eds. Mikellides, T. P. B. & Mikellides, B.) 95–98 (Studio-Vista, 1976).
Jadva, V., Hines, M. & Golombok, S. Infants' preferences for toys, colors, and shapes: Sex differences and similarities. Arch. Sexual activity Behav. 39, 1261–1273 (2010).

PubMed Google Scholar
Schloss, K. B., Poggesi, R. G. & Palmer, S. Eastward. Furnishings of university affiliation and "school spirit" on color preferences: Berkeley versus Stanford. Psychon. B Rev. 18, 498–504 (2011).

Google Scholar
Yokosawa, K., Schloss, One thousand. B., Asano, G. & Palmer, Southward. E. Ecological effects in cross-cultural differences between U.S. and Japanese colour preferences. Cogn. Sci. 40, 1590–1616 (2016).

PubMed Google Scholar
Schloss, K. B. & Palmer, Due south. E. Aesthetic response to color combinations: preference, harmony, and similarity. Atten. Percept. Psychophys. 73, 551–571 (2011).

PubMed Google Scholar
Nascimento, Southward. M. C. et al. The colors of paintings and viewers' preferences. Vis. Res. 130, 76–84 (2017).

PubMed Google Scholar
Albers, A. M., Gegenfurtner, K. R. & Nascimento, South. M. C. An independent contribution of color to the aesthetic preference for paintings. Vis. Res. 177, 109–117 (2020).

PubMed Google Scholar
Altmann, C. S., Brachmann, A. & Redies, C. Liking of art and the perception of color. J. Exp. Psychol. Hum. Percept. Perform. 47, 545–564 (2021).

PubMed Google Scholar
Pinto, P. D., Linhares, J. K. M., Carvalhal, J. A. & Nascimento, S. Yard. C. Psychophysical estimation of the best illumination for appreciation of Renaissance paintings. Visual Neurosci. 23, 669–674 (2006).

Google Scholar
The Metropolitan Museum of Fine art. Accessed 26 Nov 2021. https://www.metmuseum.org/.
Web Gallery of Art. Accessed 26 Nov 2021. https://world wide web.wga.hu/.
WikiArt - visual art encyclopedia. Accessed 26 Nov 2021. https://www.wikiart.org/.
Wikimedia Commons. Accessed 26 Nov 2021. https://commons.wikimedia.org/.
Liu, H., Huang, M., Cui, 1000., Luo, M. R. & Melgosa, M. Color-difference evaluation for digital images using a categorical judgment method. J. Opt. Soc. Am. thirty, 616 (2013).

ADS Google Scholar
Vurro, M., Ling, Y. & Hurlbert, A. C. Memory color of natural familiar objects: Effects of surface texture and 3-D shape. J. Vis. 13, 20–20 (2013).

PubMed Google Scholar
JASPTeam. JASP (Version 0.16.i)[Figurer software] (2020).
Cohen, J. Statistical Power Assay for the Behavioral Sciences. vol. 3 (2018).
Nascimento, S. G. C. et al. The best CCT for appreciation of paintings nether daylight illuminants is dissimilar for Occidental and Oriental viewers. Leukos https://doi.org/x.1080/15502724.2020.1761828 (2020).

Article Google Scholar
Pinto, P. D., Linhares, J. M. M. & Nascimento, S. M. C. Correlated color temperature preferred past observers for illumination of artistic paintings. J. Opt. Soc. Am. 25, 623 (2008).

ADS Google Scholar
Nascimento, S. M. C., Albers, A. M. & Gegenfurtner, K. R. Naturalness and aesthetics of colors—preference for color compositions perceived as natural. Vis. Res. 185, 98–110 (2021).

PubMed Google Scholar
Graphic engineering—standard object colour spectra database for color reproduction evaluation (SOCS). ISO/TR 16066:2003 (2003).
Albers, A. M., Schiller, F., Gegenfurtner, K. & Nascimento, S. Color categories in artful preferences for paintings. J. Vis. 18, 869 (2018).

Google Scholar
Levitan, C. A., Winfield, East. C. & Sherman, A. Grumpy toddlers and dead pheasants: Visual fine art preferences are predicted by preferences for the depicted objects. Psychol. Aesthet. Creat. Arts 14, 155–161 (2019).

Google Scholar

Download references

Acknowledgements

Nosotros thank Dr. Yukinori Misaki at Kagawa National Institute of Technology, Japan and Ms. Nobuyo Okada and Ms. Kanako Maruchi at Toyohashi City Museum of Art and History, Japan for assisting in the measurement of art paintings. This work was supported by JSPS KAKENHI Grant Number JP19H01119 and 20H05956, and by the Portuguese Foundation for Science and Technology (FCT) in the framework of the Strategic Funding UIDB/04650/2020.

Writer information

Affiliations

Section of Reckoner Scientific discipline and Engineering science, Toyohashi University of Technology, Toyohashi, 441-8580, Japan

Shigeki Nakauchi, Taisei Kondo, Yuya Kinzuka, Yuma Taniyama, Hideki Tamura, Hiroshi Higashi, Kyoko Hine & Tetsuto Minami
Centre of Physics, Gualtar Campus, University of Minho, Braga, 4710-057, Portugal

João M. M. Linhares & Sérgio M. C. Nascimento

Contributions

Southward.N. and S.M.C.N. conceived the experiments. T.K. and J.M.Chiliad.L. conducted the experiments. Y.Chiliad., Y.T., H.T., H.H., M.H., and T.M. analyzed the results. S.N. and S.M.C.N. wrote the manuscript. All authors reviewed and approved the final version of the manuscript.

Corresponding author

Correspondence to Shigeki Nakauchi.

Ethics declarations

Competing interests

The authors declare no competing interests.

Boosted information

Publisher'due south note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Data

Rights and permissions

Open Access This article is licensed under a Artistic Eatables Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long every bit yous give advisable credit to the original author(southward) and the source, provide a link to the Creative Eatables licence, and indicate if changes were made. The images or other third party cloth in this article are included in the article'southward Creative Commons licence, unless indicated otherwise in a credit line to the material. If cloth is not included in the commodity's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission straight from the copyright holder. To view a re-create of this licence, visit http://creativecommons.org/licenses/by/iv.0/.

Reprints and Permissions

Nearly this commodity

Verify currency and authenticity via CrossMark

Cite this commodity

Nakauchi, Due south., Kondo, T., Kinzuka, Y. et al. Universality and superiority in preference for chromatic composition of fine art paintings. Sci Rep 12, 4294 (2022). https://doi.org/10.1038/s41598-022-08365-z

Download citation

Received: 17 June 2021
Accepted: 07 March 2022
Published: eleven March 2022
DOI : https://doi.org/ten.1038/s41598-022-08365-z

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something calumniating or that does not comply with our terms or guidelines please flag information technology as inappropriate.

sawyeryones1949.blogspot.com

Source: https://www.nature.com/articles/s41598-022-08365-z