No. 1 Synaesthetic colour experiences are perceptually real

Christoph Witzel1, Kevin O’Regan2 & Nicolas Rothen3

1Allgemeine Psychologie, Justus-Liebig-Universität, Gießen, Germany;
2Laboratoire Psychologie de la Perception, CNRS and Université Paris Descartes, Paris, France;
3Department of Psychology, University of Bern and Center for Cognition, Learning and Memory, Bern, Switzerland

No study has unambiguously shown that synaesthetic colours are perceptually real rather than just conceptually represented in memory. Here we tested for the perceptual reality of synesthetic colours using an achromatic adjustment task. This task is particularly sensitive for measuring perceived colour, and it has been successfully used in the past to show the influence of memory on colour perception. The results showed that in contrast to non-synaesthetes, synaesthetes saw inducers in their concurrent colour even when they were shown without colour, i.e. greyscale. Although these induced colours were not as strong as real colours, they were statistically very reliable and specific for the inducers and concurrent colours of each individual observer. Additionally, we conducted a synaesthetic variant of the Stroop task to replicate previous findings about semantic effects of synaesthesia. Results confirmed that semantic colour associations are automatically triggered by synaesthetic inducers, but not by control stimuli and not in non-synaesthetes. We also included a control stimulus that reproduced the known effects of memory on perceived colour in both groups of participants, and another control stimulus, for which participants had learnt an arbitrary colour-shape association. This latter control stimulus also produced perceptual effects in synaesthetes, but not in non-synaesthetes. Taken together, the present results show that grapheme-colour synaesthesia involves real perception of concurrent colours. Furthermore synaesthetes are more susceptible to memory effects on perceived hue than normal subjects. Acknowledgements: This research was supported by ERC Advanced Grant “FEEL”, number 323674 to J. Kevin O’Regan. Nicolas Rothen is supported by the Swiss National Science Foundation (Grant PZ00P1_154954).

No. 2 Kandinsky or me? How free is the eye of the beholder in abstract art?

Doris Braun & Katja Dörscher

Allgemeine Psychologie, Justus-Liebig-Universität, Giessen, Germany

Research on the perception of art has focused often on aesthetic ratings. Here we take a different approach and investigate abstract art perception by assessing (i) how the artist’s color palette of a painting influences observers’ choice of color for one element in the same painting (ii) if observers prefer their color choices over the original and (iii) how the composition of the painting affects its perceived balance. Participants (i) adjusted, starting with a neutral gray, the color of a single shape selected in 24 abstract paintings of Baumeister, Hoffmann, Delauney, Kandinsky and Klee, and indicated (ii) their preference between the adjusted and original color. To measure the perceived balance (iii) in a painting we asked participants to indicate the center of gravity for each artwork by adjusting the location and size of a black circle on a corresponding adjacent white rectangle. Our results show that, color settings are frequently not in agreement with the artist’s choice but differ significantly from the distribution of randomly picked CIE-LAB samples. Notably, the color palette of a painting influences color choices in two ways: it either elicits harmonious or contrasting settings and when compared, usually the artist’s setting is preferred. Unlike for color settings observers’ exhibited a remarkable consistency in their perceived center of gravity. Having seen a painting once influences quite effectively whether the original or its mirror image is preferred.

No.3 Influence of hue on the subjective evaluation of color samples under various light sources

Yuki Nakajima

Graduate School of Advanced Integration Science, Chiba University, Japan

Past studies have showed that the subjective evaluation of color feelings under various lighting conditions depend on the hue of color patches (e.g. Nakajima and Fuchida 2015). In this study, psychophysical experiments were performed to investigate the effect of influential color on the overall feeling of samples including the color patches and pictures under the eight kinds of test light sources. We used a haploscopic viewing method. A reference lamp was a warm white fluorescent lamp with high fidelity (Ra96). Test lamps were two fluorescent lamps (Ra55 and Ra96), and six LED lamps (ranged from Ra21 to Ra86). The color samples were a mosaic color sheet composed of 20 color patches (red, yellow, green, blue, and gold leaf) and the seven kinds of pictures with a main color: red, orange, yellow, green, yellow green, blue, purple, and complexion. Observers were asked to evaluate the feeling of the color samples under the test illumination as comparing with that of the same samples under the reference illumination by a Semantic Differential method. They were also asked to report the top three hues which most influenced to their color evaluation. It was shown that the most influential color for the subjective evaluation of color feelings was red regardless of mosaic or pictures. On the other hand, the influence of blue, yellow, purple, and yellow green on the evaluation was little. Our results showed that color rendering evaluation under different light sources changed significantly depending on the existence of red in the samples. It was suggested that “red” would be the most important for color evaluation under different illuminations.

No. 4 The feeling of colors - semantic dimensions and topography of brain electrical activity

Wolfgang Skrandies

Institute of Physiology, Justus-Liebig University, Giessen, Germany

The semantic differential technique is used in order to define dimensions of connotative meaning. We investigated the affective meaning of color words. 13 different words were rated on adjective scales of opposite meaning by a total of 1865 healthy young adults. We found three dimensions that reflected ”evaluation” (E, friendly, good, nice, etc.), ”potency” (P, strong, big, heavy, etc.), and ”activity” (A, fast, noisy, lively). Different colors had different factor scores that were used to classify colors in six different classes (E+/E-, P+/P-, A+/A-). During ERP recordings, color words were presented in random order on a monitor. Attention was controlled by instructing subjects to could unrelated words appearing at random intervals. EEG was recorded from 42 healthy adults from 30 channels between the inion and Fz. ERPs were computed offline according to stimulus class. Repeated measurement ANOVA was used for comparing experimental conditions. Between 70 and 410 ms latency five components were identified by the oc- currence of maximal Global Field Power (GFP). Between 70 and 130 ms different semantic classes yielded significantly different GFP (F(2,82)=29.97; p<.00001): colors judged as ”active” or ”passive” (yellow / orange or brown / black) were followed by high GFP while colors related to potency (P) showed smallest response amplitudes (red / gold or pink / silver). A very similar effect was seen between 130 and 190 ms (F(2,82)=22.48; p<.00001). This compo- nent displayed significantly different latencies (F(2,82)=5.72; p<.0047) where color words related to activity (A) had smallest latencies. In addition, there occurred a number of significant topographical effects. Our results show that color words can be consistently classified according to their connotative, af- fective meaning. Such differences are reflected by ERP components occurring after about 100 ms latency when color words are read by healthy adults.

No. 5 Connotative meaning of color terms: The influence of sex and age

Wolfgang Skrandies

Institute of Physiology, Justus-Liebig University, Giessen, Germany

The semantic differential technique can be employed in order to define dimensions of connotative meaning. We investigated the affective meaning of color terms. 13 different color words were rated on adjective scales of opposite meaning by a total of 1865 healthy adults. Subjects were divided in three age groups between 11 and 76 years of age. Age groups as well as males and females were compared in their rations of color words. Principal component analysis (PCA) reduced the data to three dimensions that reflected ”evaluation” (E, friendly, good, nice, etc.), ”potency” (P, strong, big, heavy, etc.), and ”activity” (A, fast, noisy, lively). These factors explained 65.70% of the variance in the data. Different colors had different factor scores that could be used in a lin- ear discriminant analysis to classify successfully colors in six different classes (E+/E-, P+/P-, A+/A-). Thus, a given color word can be identified by its unconscious affective meaning that is quantified by different scores on Evaluation, Potency and Activity. In addition, there were effects of sex and age on the ratings suggesting that affective meaning is different in males and females, and it changes over life time.

No. 6 Swapping swatches: Adapting to and from an artist’s palette

Katherine E.M. Tregillus & Michael A. Webster

University of Nevada, Reno, USA

We describe a method for representing and manipulating the color gamuts used by different artists to explore how the color schemes employed by artists might appear to the artist or to others. The method involves modeling the visual response to color and then adapting that response to simulate how color percepts change across different states of adaptation. Analyses of paintings and nature photographs suggest that there are both important differences and regularities in the color palettes of artists and that these regularities reproduce prominent characteristics of the natural color environment. In particular, the works of many artists include a bluish-yellowish bias that is also a distinguishing feature of both the color statistics of natural images and of the neural coding of color. The algorithm adjusts the colors in an image so that they are equivalent to the colors that would be experienced by an observer adapted to a different environment, or for two observers with different spectral sensitivities but who are adapted to the same environment. This provides a novel method for visualizing how the colors in artwork are experienced by an artist or an audience, and could be generalized to explore similar questions for visual attributes beyond color.

No. 7 THEY CAN DISAPPEAR - Can the Panda illusion be used to test visual acuity?

T. Strasser1, H. Langrova2, L. Kühlewein1, A. Werner1, A. Kurtenbach1, E. Zrenner1

1Centre for Ophthalmology, University of Tuebingen, Germany;
2University Eye Hospital, Hradec Králové, Czech Republic

Beginning of 2016, the Russian artist Ilja Klemencov revealed the artwork “They can Disappear”, pointing out the danger of the extinction of the panda bear. The illustration presents a panda hidden behind of black-and-white zigzagged lines. Many people struggled to see the bear. However, stepping back or taking off glasses usually unveils the panda. This lead us to the question if the ability to see the panda is related to the observers' visual acuity and if the panda illusion can be used as a test of visual acuity. We developed a test which dynamically generates images of animals, similar to the original illustration, with changing spatial resolutions of the zigzag pattern, based on an adaptive staircase method. Here we present the results of 23 volunteers with best-corrected visual acuity and artificially degraded vision using plus lenses (+1D, +2D) and Bangerter occlusion foils (0.2, 0.6).

No. 8 Multiple mechanisms underlying color appearance and color naming

Kara Emery1, David H. Peterzell2, Vicki J. Volbrecht3, Michael A. Webster1

1University of Nevada, Reno, USA;
2John F Kennedy University, Pleasant Hill, USA;
3 Colorado State University, Ft. Collins, USA

The number and nature of the mechanisms underlying color appearance remain poorly understood. Standard models posit two opponent channels mediating red vs. green or blue vs. yellow sensations. However, a neural basis for this representation has yet to be identified. We took advantage of the large individual differences in color perception to analyze the processes underlying color percepts in color normal observers, using factor analysis to explore the patterns of variability in both hue scaling and hue naming. Twenty-three observers were shown 36 hues of roughly equal saturation spanning steps of 10 deg around the standard cone-opponent space. In the hue-scaling task observers judged the component proportions of red, green, blue, or yellow. In the naming task they used a fixed set of basic color terms to label each hue. A factor analysis of both data sets revealed multiple factors with each largely confined to a narrow range of contiguous hue angles. These were inconsistent with predictions for classical opponent channels varying in their tuning or relative sensitivity, which instead predict broad and bimodal patterns of factor loadings. The observed factors instead reveal that putatively opponent pairs (e.g. red and green) are constrained by independent processes, and moreover that binary hues (e.g. orange) vary independently of their constituent primaries (e.g. red and yellow). Our results suggest that different narrow angles of color space may each be governed by different mechanisms that can vary across observers, and these are manifest even in the standard hue-scaling task that requires observers to perceptually decompose each hue into red-green and blue-yellow components. Supported by EY-10834

No. 9 Colorcontex – Zusammenhänge zwischen Farbe und textilem Material

Jaqueline Satzinger

Textil Kunst Design, Kunstuniversität Linz, Linz, Österreich

In textilen Kollektionen sind in der Regel verschiedenfarbige Ausführungen des gleichen Designs zu finden. Dass sich in einer Kollektion stets einige Farbvarianten besser verkaufen als andere, mag auf den ersten Blick an aktuellen Trends oder persönlichem Gefallen liegen. Jedoch sehen Farb- und Wahrnehmungsforscher den Grund dafür im gegenseitigen Einfluss des Oberflächenmaterials und der jeweiligen Farbe, den es noch genauer zu erforschen gilt. Basierend auf den Zusammenhang von Farbe zur Tastwahrnehmung wurden Thesen bezüglich der Materialität entwickelt. Diese wurden mittels schriftlicher Befragung anhand von15 x 15 cm großen, gefärbten Mustergeweben aus Wolle, Seide und Baumwolle überprüft. Kriterien für die Farbauswahl waren die Eindeutigkeit in der Theorie sowie deren färbetechnische Umsetzbarkeit. Da die Umsetzung nur mit synthetischen Pigmenten möglich war, dienten verschiedene natürliche Färbungen als Vorlage um keine zu künstlichen Farbtöne zu färben. Verwendet wurden ausschließlich Naturfasern. Im Gegensatz zu den stark modifizierbaren synthetischen Fasern besitzen diese Eigenschaften und Oberflächen, die nur begrenzt verändert werden können. Dies führt zu eindeutigeren Ergebnissen, welche auch hinsichtlich synthetischer Fasern gedeutet werden können. Eine heterogene Gruppe von fünfzig Probanden beurteilte anschließend die subjektive Wirkung der 18 farbigen Gewebe anhand eines Polaritätsprofils. In der abschließenden statistischen Auswertung der Fragebögen zeigten sich tendenziell harmonische und dissonante Material- und Farbkombinationen sowie generelle Erkenntnisse über den Zusammenhang von textilem Material und Farbe. Die zu Grunde liegende Vermutung wird gestützt und eröffnet neue Ansatzpunkte, die Zusammenhänge von Farbe und Material genauer zu untersuchen. Mehr Wissen darüber ermöglicht einen sichereren und kompetenteren Umgang mit beiden Komponenten innerhalb des Gestaltungsprozesses.

No. 10 Of mice and men: Factors underlying normal individual differences in photopic and scotopic spectral sensitivity functions

David H. Peterzell1, Donald I.A.MacLeod2, Vicki J. Volberecht3, Michael A. Crognale4, Kara Emery4, Michael A. Webster4

1John F Kennedy University, Pleasant Hill, USA;
2University of California, San Diego, USA;
3Colorado State University, Ft. Collins, USA;
4University of Nevada, Reno, USA

Normal observers vary in judgments of how luminous monochromatic lights must be to match the intensity of a broadband (white) light. We used factor analyses of individual differences to confirm and explore mechanisms underlying spectral luminosity functions. Separate analyses of covariance matrices were performed on classic photopic and scotopic data from humans, and on photopic data from wild type (dichromatic) and transgenic mice (human L-cone transgene added) (Coblenz & Emerson, 1918; Gibson & Tyndall, 1923; Crawford, 1949 [series 1]; Kraft & Werner, 1994; Shabaan et al. 1998). In man and transgenic mouse, two Varimax-rotated factors from photopic data were bipolar. Their loadings coincided with absorption spectra for M and L photopigments; they had zero-crossings at deuteranopic and tritanopic confusion points (equal M or L responses in test and standard). A third factor was consistent with variability in the S photopigment. Similarities in results between man and transgenic mouse help confirm that factors from humans are L and M photopigments. From scotopic data, four obliquely-rotated factors coincided with (1) lens density, (2) macular pigment density, (3) rhodopsin density, and (4) Rhodopsin λmax factors. (1) and (4) intercorrelated, presumably accidentally. Variability estimates for (3) slightly exceeded those for cones (Webster & Macleod, 1988), while (4) was slightly less than reported by Bowmaker et al. (1975). Factors seem to confirm existing theory.

No. 11 Monochromatic Equivalence in Natural Colour Stimuli using Opponent Coding

Thomas Bangert, Ebroul Izquierdo

EECS, Queen Mary University of London, London, England

Monochromatic stimuli (light of just one wavelength) within a specific range of the electromagnetic spectrum may be accurately coded by just its wavelength (assuming constant amplitude). Natural colour stimuli are, however, rarely monochromatic. Flowering plants are a very common source of natural visual stimuli and their colour varies across a broad range. The colour of flowers is typically in the range of yellow – orange – red. Expressed as a monochromatic wavelength (or dominant wavelength if set within the CIE colour space) the range of colour is typically between 580nm and 650nm. A spectral power diagram (SPD), shows that flowers are able to reflect light across much of the visible spectrum, but for coloured flowers the most common and by extension the most important colour is yellow. The spectrum of a yellow flower is, however, very different from a monochromatic yellow. A common yellow flower reflects light in the range of 520-650nm, and absorbs light in the range of purple – blue – cyan. We measured the SPD of the flower petals of a broad range of flowering plants and found that there is a simple underlying principle in how plants reproduce colour. We found that flowering plants control colour by changing the spectral position of the transition zone between absorption and reflection. Flowers are yellow by default and shift the spectral transition point to change colour, with the upper boundary for this shift being red. Colour can therefore in many cases be coded by a single value; just as if it were monochromatic.