Spatial and temporal aspects of chromatic adaptation and their functional significance for colour constancy

Adaptation under dichromatic illumination

Over the years, many CATs (chromatic adaptation transforms) have been developed, such as CMCCAT97, CAT02 and CAT16, to predict the corresponding colors under different illuminants. These CATs were derived from uniform simple stimuli surrounded by a uniform background with a single illuminant. Although some mixed adaptation models have been proposed in literature to predict the adaptation under more than one illuminant, these models are typically limited to a certain scene and exclude the impact of spatial complexity. To investigate chromatic adaptation under more complex conditions, an achromatic matching experiment was conducted with (simultaneously) spatially dichromatic illumination for three illumination color pairs and various spatial configurations. Spatial configuration was found to have an impact on both the degree of adaptation and the equivalent illuminant chromaticity, which is the chromaticity of a single uniform adapting illumination that results in the same corresponding colors as for the dichromatic lighting condition. A preliminary CAT model is proposed that considers the spatial and colorimetric complexity of the illumination.

Developmental changes in colour constancy in a naturalistic object selection task

When the illumination falling on a surface change, so does the reflected light. Despite this, adult observers are good at perceiving surfaces as relatively unchanging-an ability termed colour constancy. Very few studies have investigated colour constancy in infants, and even fewer in children. Here we asked whether there is a difference in colour constancy between children and adults; what the developmental trajectory is between six and 11 years; and whether the pattern of constancy across illuminations and reflectances differs between adults and children. To this end, we developed a novel, child-friendly computer-based object selection task. In this, observers saw a dragon's favourite sweet under a neutral illumination and picked the matching sweet from an array of eight seen under a different illumination (blue, yellow, red, or green). This set contained a reflectance match (colour constant; perfect performance) and a tristimulus match (colour inconstant). We ran two experiments, with two-dimensional scenes in one and three-dimensional renderings in the other. Twenty-six adults and 33 children took part in the first experiment; 26 adults and 40 children took part in the second. Children performed better than adults on this task, and their performance decreased with age in both experiments. We found differences across illuminations and sweets, but a similar pattern across both age groups. This unexpected finding might reflect a real decrease in colour constancy from childhood to adulthood, explained by developmental changes in the perceptual and cognitive mechanisms underpinning colour constancy, or differences in task strategies between children and adults. HIGHLIGHTS: Six- to 11-year-old children demonstrated better performance than adults on a colour constancy object selection task. Performance decreased with age over childhood. These findings may indicate development of cognitive strategies used to overcome automatic colour constancy mechanisms.

Understanding insect colour constancy

Colour constancy is the ability to recognize the colour of objects despite spectral changes in the natural illumination. As such, this phenomenon is important for most organisms with good colour vision, and it has been intensively studied in humans and primates. Colour constancy is also documented for several species of insects, which is not surprising given the ecological importance of colour vision. But how do insects, with their small brains, solve the complex problem of colour vision and colour constancy? In an interspecies approach, this review reports on behavioural studies on colour constancy in bees, butterflies, moths and humans, corresponding computational models and possible neurophysiological correlates. This article is part of the theme issue 'Understanding colour vision: molecular, physiological, neuronal and behavioural studies in arthropods'.

Colour constancy failures expected in colourful environments

Colour constancy refers to the constant perceived or apparent colour of a surface despite changes in illumination spectrum. Laboratory measurements have often found it imperfect. The aim here was to estimate the frequency of constancy failures in natural outdoor environments and relate them to colorimetric surface properties. A computational analysis was performed with 50 hyperspectral reflectance images of outdoor scenes undergoing simulated daylight changes. For a chromatically adapted observer, estimated colour appearance changed noticeably for at least 5% of the surface area in 60% of scenes, and at least 10% of the surface area in 44% of scenes. Somewhat higher frequencies were found for estimated changes in perceived colour relations represented by spatial ratios of cone-photoreceptor excitations. These estimated changes correlated with surface chroma and saturation. Outdoors, the colour constancy of some individual surfaces seems likely to fail, particularly if those surfaces are colourful.

Human color constancy based on the geometry of color distributions

The physical inputs to our visual system are dictated by the interplay between lights and surfaces; thus, for surface color to be stably perceived, the influence of the illuminant must be discounted. To reveal our strategy to infer the illuminant color, we conducted three psychophysical experiments designed to test our optimal color hypothesis that we internalize the physical color gamut under various illuminants and apply the prior to estimate the illuminant color. In each experiment, we presented 61 hexagons arranged without spatial gaps, where the surrounding 60 hexagons were set to have a specific shape in their color distribution. We asked participants to adjust the color of a center test field so that it appeared to be a full-white surface placed under a test illuminant. Results and computational modeling suggested that, although our proposed model is limited in accounting for estimation of illuminant intensity by human observers, it agrees fairly well with the estimates of illuminant chromaticity in most tested conditions. The accuracy of estimation generally outperformed other tested conventional color constancy models. These results support the hypothesis that our visual system can utilize the geometry of scene color distribution to achieve color constancy.

Themes of advanced information processing in the primate brain

Here is a review of several empirical examples of information processing that occur in the primate cerebral cortex. These include visual processing, object identification and perception, information encoding, and memory. Also, there is a discussion of the higher scale neural organization, mainly theoretical, which suggests hypotheses on how the brain internally represents objects. Altogether they support the general attributes of the mechanisms of brain computation, such as efficiency, resiliency, data compression, and a modularization of neural function and their pathways. Moreover, the specific neural encoding schemes are expectedly stochastic, abstract and not easily decoded by theoretical or empirical approaches.

Temporal dynamics of daylight perception: Detection thresholds

Temporal changes in illumination are ubiquitous; natural light, for example, varies in color temperature and irradiance throughout the day. Yet little is known about human sensitivity to temporal changes in illumination spectra. Here, we aimed to determine the minimum detectable velocity of chromaticity change of daylight metamers in an immersive environment. The main stimulus was a continuous, monotonic change in global illumination chromaticity along the daylight locus in warmer (toward lower correlated color temperatures [CCTs]) or cooler directions, away from an adapting base light (CCT: 13,000 K, 6500 K, 4160 K, or 2000 K). All lights were generated by spectrally tunable overhead lamps as smoothest-possible metamers of the desired chromaticities. Mean detection thresholds (for 22 participants) for a fixed duration of 10 seconds ranged from 15 to 2 CIELUV ΔE units, depending significantly on base light CCT and with a significant interaction between CCT and direction of change. Cool changes become less noticeable for progressively warmer base lights and vice versa. For the two extreme base lights, sensitivity to changes toward neutral is significantly lower than for the opposite direction. The results suggest a “neutral bias” in illumination change discriminability, and that typical temporal changes in daylight chromaticity are likely to be below threshold detectability, at least where there are no concomitant overall illuminance changes. These factors may contribute to perceptual stability of natural scenes and color constancy.

Specular highlights improve color constancy when other cues are weakened

Previous studies suggest that to achieve color constancy, the human visual system makes use of multiple cues, including a priori assumptions about the illumination ("daylight priors"). Specular highlights have been proposed to aid constancy, but the evidence for their usefulness is mixed. Here, we used a novel cue-combination approach to test whether the presence of specular highlights or the validity of a daylight prior improves illumination chromaticity estimates, inferred from achromatic settings, to determine whether and under which conditions either cue contributes to color constancy. Observers made achromatic settings within three-dimensional rendered scenes containing matte or glossy shapes, illuminated by either daylight or nondaylight illuminations. We assessed both the variability of these settings and their accuracy, in terms of the standard color constancy index (CCI). When a spectrally uniform background was present, neither CCIs nor variability improved with specular highlights or daylight illuminants (Experiment 1). When a Mondrian background was introduced, CCIs decreased overall but were higher for scenes containing glossy, as opposed to matte, shapes (Experiments 2 and 3). There was no overall reduction in variability of settings and no benefit for scenes illuminated by daylights. Taken together, these results suggest that the human visual system indeed uses specular highlights to improve color constancy but only when other cues, such as from the local surround, are weakened.

Exploring the Determinants of Color Perception Using #Thedress and Its Variants: The Role of Spatio-Chromatic Context, Chromatic Illumination, and Material-Light Interaction

The colors that people see depend not only on the surface properties of objects but also on how these properties interact with light as well as on how light reflected from objects interacts with an individual's visual system. Because individual visual systems vary, the same visual stimulus may elicit different perceptions from different individuals. #thedress phenomenon drove home this point: different individuals viewed the same image and reported it to be widely different colors: blue and black versus white and gold. This phenomenon inspired a collection of demonstrations presented at the Vision Sciences Society 2015 Meeting which showed how spatial and temporal manipulations of light spectra affect people's perceptions of material colors and illustrated the variability in individual color perception. The demonstrations also explored the effects of temporal alterations in metameric lights, including Maxwell's Spot, an entoptic phenomenon. Crucially, the demonstrations established that #thedress phenomenon occurs not only for images of the dress but also for the real dress under real light sources of different spectral composition and spatial configurations.

Chromatic fading following complete adaptation to unique hues

Profound vision loss occurs after prolonged exposure to an unchanging featureless visual environment. The effect is sometimes called visual fade. Here we investigate this phenomenon in the color domain using two different experiments. In the first experiment we determine the time needed for a colored background to appear achromatic. Four backgrounds were tested. Each represented the observers’ four unique hues. This adaptation time was compared with time to recover after adaptation Hue shifts at the end of the adaptation period were also measured. There were wide individual differences in adaptation times and recovery times. Overall recovery was faster than adaptation (p < 0.02). There were minimal shifts in hue. In the second experiment the changes in saturation (Munsell chroma) and lightness (Munsell value) of the background were monitored at six time intervals during the adapting process. Again asymmetric matching with Munsell samples was used. There were two distinct components to both the adaptation and recovery phases; one fast with time constant <1s, the other slow with time constant between 40 and 160s.

The experiments show that the special case of visual fade involving color represents the sensory basis for many color-related effects involving adaptation.

Blue blood, red blood. How does the color of an emotional scene affect visual attention and pupil size?

The function of color in the processing of emotional scenes is not entirely clear. While there are studies showing that color matters in terms of the capture of covert attention by emotional stimuli, the impact of color on fixation patterns, reflecting overt attention, is unresolved. Studies on the role of color in evoking emotional response have also produced mixed results. Here, we aimed to explore how image color and content influence pupillary response and the engagement of overt visual attention. In the first experiment, we examined the pupillary reaction to neutral images (intact and phase scrambled) in three color variants (natural, abnormal, and grayscale). In the second experiment, we investigated the pupillary changes and fixation pattern in response to images of different valence (neutral, positive, and negative), again in three color versions. The results showed that pupillary responses were influenced by both content and the color of the images. The pupillary response to phase-scrambled images did not differ between the color versions. Intact neutral and positive images, but not negative ones, evoked smaller pupil responses if they were presented in abnormal colors rather than natural ones. The initial capture of attention by emotional content depended on the color version, whereas holding of attention was affected solely by the emotional valence. Thus, color changes the physiological response to images, particularly low-arousing ones, and modulates the initial engagement of attention by image content.

Disentangling simultaneous changes of surface and illumination

Retinally incident light is an ambiguous product of spectral distributions of light in the environment and their interactions with reflecting, absorbing, and transmitting materials. An ideal color constant observer would unravel these confounded sources of information and account for changes in each factor. Scene statistics have been proposed as a way to compensate for changes in the illumination, but few theories consider changes of 3-dimensional surfaces. Here, we investigated the visual system's capacity to deal with simultaneous changes in illumination and surfaces. Spheres were imaged with a hyperspectral camera in a white box and their colors, as well as that of the illumination were varied along "red-green" and "blue-yellow" axes. Both the original hyperspectral images and replica scenes rendered with Mitsuba were used as stimuli, including rendered scenes with Glavens (Acta Psychologica, 2009, 132, 259-266). Observers viewed sequential, random pairs of our images, with either the whole scene, only the object, or only a part of the background being present. They judged how much the illuminant and object color changed on a scale of 0-100%. Observers could extract simultaneous illumination and reflectance changes when provided with a view of the whole scene, but global scene statistics did not fully account for their behavior, while local scene statistics improved the situation. There was no effect of color axis, shape, or simulated vs. original hyperspectral images. Observers appear to be making use of various sources of local information to complete the task.

The impact of age-related cataracts on colour perception, postoperative recovery and related spectra derived from test of hue perception

Background

Cataract patients were always excluded from studies on ageing of colour vision; thus, effect of age-related cataracts on deterioration of colour perception has not been analysed. In present study, impacts of age-related cataracts on colour discrimination, postoperative recovery and related spectra were investigated.

Methods

In this cohort study, thirty age-related cataract patients scheduled for binocular surgery and 30 elderly volunteers were enrolled. Colour discrimination under photopic (1000 lx) and mesopic (40 lx) conditions was evaluated with Farnsworth-Munsell 100-hue test. The total error score (TES) and partial error score (PES) were calculated.

Results

Preoperatively, the TES in the patient group was 129.7 ± 59.5 at 1000 lx and 194.6 ± 74.5 at 40 lx, exhibiting worse discrimination than the volunteer group (TES_1000lux = 71.5 ± 37.5 and TES_40lux = 113.1 ± 38.8, p ≤ 0.001). Inferior perception were detected in the yellow to green-yellow (Y-GY), green-yellow to green (GY-G), green to blue-green (G-BG) and blue-green to blue (BG-B) colour bands (p ≤ 0.003), corresponding to the 470 nm–580 nm range of the visible light spectrum. Under mesopic conditions, the impact expanded to all colour bands except for yellow-red to yellow (YR-Y). Postoperatively, the TES in the patient group were 80.4 ± 62.4 at 1000 lx and 112.0 ± 85.2 at 40 lx, which were lower than those of the preoperative phase (p ≤ 0.001) but similar to those of the volunteer group (p ≥ 0.505). Postoperative improvement occurred in the Y-GY, GY-G and G-BG colour bands (490 nm to 580 nm) at 1000 lx (p ≤ 0.001) and shifted to the Y-GY, GY-G, G-BG and BG-B colour bands (470 nm to 580 nm) at 40 lx (p ≤ 0.001). Deterioration of hue perception for decrement of illumination was detected in the red to yellow-red (R-YR), Y-GY, G-BG, BG-B, blue to purple-blue (B-PB) and red-purple to red (RP-R) colour bands (450 nm to 500 nm) in the volunteer group (p ≤ 0.002) and the R-YR, G-BG, BG-B, B-PB, PB-P and red-purple to red (RP-R) colour bands (from the short-wavelength end to 500 nm) in the patient group preoperatively (p ≤ 0.001).

Conclusions

Phacoemulsification could effectively rebuild colour perception in patients with age-related cataract. The postoperative benefits were most significant in colour bands corresponding with spectrum from 470 nm to 580 nm.

Electronic supplementary material

The online version of this article (10.1186/s12886-019-1057-6) contains supplementary material, which is available to authorized users.

Serving Duplicates in a Single Session Can Selectively Improve Sensitivity of Duplicated Intensity Ranking Tests

Duplicating ranking tests can improve the power of preference and sensory intensity tests, and reduce the number of panelists required. With multiple-samples rankings, duplications could be served using different protocols: in two serving sessions (two sample sets, that is, 2SS) to allow a break period or jointly in one serving session (one sample set, that is, 1SS). Evaluating the duplicates in a single session improves statistical data analysis but increases concerns of sensory fatigue, adaptation, memory, and possible irritation. The extent to which each serving protocol for duplicated ranking affects detection of differences has not been reported. This study used panelists (n = 75) who performed both ranking test protocols on two sets of orange juice samples (k = 3). One set was designed to obtain higher similarity than the other set to investigate the effects of degree of difference. Sweetness and yellow color intensity rankings were performed separately for each set to compare the protocols between attributes. The magnitude of the differences was evaluated using Mack-Skillings (M-S) statistics, and the total and individual rank sum differences at varied n (10 to 75) values. With similar set samples, the 2SS serving protocol improved differentiation for yellow color intensity. Although in sweetness, using the 1SS serving protocol showed higher M-S statistics and higher sum of total rank sum differences. Paired comparisons followed the same pattern. With very different samples, both protocols had comparable performance. This study showed that serving duplicates in the 1SS can improve duplicated ranking's power, depending on the task difficulty and attribute, and should be considered before splitting replications into two sessions.

PRACTICAL APPLICATION

This study evaluated two alternative and not previously studied protocols for conducting a duplicated ranking test. One protocol required that each panelist received duplicates in the same single serving session (1SS), against the alternative of serving duplicates separately in two sessions to allow a break period (2SS). Two attributes: yellow color intensity and sweetness intensity of orange juice samples were studied. This study showed that serving duplicates in the same single session can improve duplicated ranking's power, depending on the task difficulty and attribute, and should be considered before splitting replications into two sessions.

The Verriest Lecture: Color vision in an uncertain world

The natural world is optically unconstrained. Surface properties may vary from one point to another, and reflected light may vary from one instant to the next. The aim of this work is to quantify some of the physical failures of color vision performance that result from uncertainty. In computational simulations with images of vegetated and nonvegetated outdoor scenes, it is shown that color provides an unreliable guide to surface identity. It is also shown that changes in illuminant may cause colors to no longer match and the relations between individual colors to vary. These failures are generally well described by a measure of the randomness of the colors in scenes, the Shannon entropy. Although uncertainty is intrinsic to the environment, its consequences for color vision can be predicted.

Possible influences on color constancy by motion of color targets and by attention-controlled gaze

We investigated the influence of motion on color constancy using a chromatic stimulus presented in various conditions (static, motion, and rotation). Attention to the stimulus and background was also controlled in different gaze modes, constant fixation of the stimulus, and random viewing of the stimulus. Color constancy was examined in six young observers using a haploscopic view of a computer monitor. The target and background were illuminated in simulation by red, green, blue, and yellow, shifted from daylight (D65) by specific color differences along L - M or S - (L + M) axes on the equiluminance plane. The standard pattern (under D65) and test pattern (under the color illuminant) of a 5-deg square were presented side by side, consisting of 1.2-deg square targets with one of 12 colors at each center, surrounded by 230 background ellipses consisting of eight other colors. The central color targets in both patterns flipped between top and bottom locations at the rate of 3 deg/s in the motion condition. The results indicated an average reduction of color constancy over the 12 test colors by motion. The random viewing parameter indicated better color constancy by more attention to the background, although the difference was not significant. Color constancy of the four color illuminations was better to worse in green, red, yellow, and blue, respectively. The reduction of color constancy by motion could be explained by less contribution of the illumination estimation effect on color constancy. In the motion with constant fixation condition, the retina strongly adapted to the mean chromaticity of the background. However, motion resulted in less attention to the color of the background, causing a weaker effect of the illumination estimation. Conversely, in the static state with a random viewing condition, more attention to the background colors caused a stronger illumination estimation effect, and color constancy was improved overall.

Improved color constancy in honey bees enabled by parallel visual projections from dorsal ocelli

How can a pollinator, like the honey bee, perceive the same colors on visited flowers, despite continuous and rapid changes in ambient illumination and background color? A hundred years ago, von Kries proposed an elegant solution to this problem, color constancy, which is currently incorporated in many imaging and technological applications. However, empirical evidence on how this method can operate on animal brains remains tenuous. Our mathematical modeling proposes that the observed spectral tuning of simple ocellar photoreceptors in the honey bee allows for the necessary input for an optimal color constancy solution to most natural light environments. The model is fully supported by our detailed description of a neural pathway allowing for the integration of signals originating from the ocellar photoreceptors to the information processing regions in the bee brain. These findings reveal a neural implementation to the classic color constancy problem that can be easily translated into artificial color imaging systems.

Quantitative studies of animal colour constancy: using the chicken as model

Colour constancy is the capacity of visual systems to keep colour perception constant despite changes in the illumination spectrum. Colour constancy has been tested extensively in humans and has also been described in many animals. In humans, colour constancy is often studied quantitatively, but besides humans, this has only been done for the goldfish and the honeybee. In this study, we quantified colour constancy in the chicken by training the birds in a colour discrimination task and testing them in changed illumination spectra to find the largest illumination change in which they were able to remain colour-constant. We used the receptor noise limited model for animal colour vision to quantify the illumination changes, and found that colour constancy performance depended on the difference between the colours used in the discrimination task, the training procedure and the time the chickens were allowed to adapt to a new illumination before making a choice. We analysed literature data on goldfish and honeybee colour constancy with the same method and found that chickens can compensate for larger illumination changes than both. We suggest that future studies on colour constancy in non-human animals could use a similar approach to allow for comparison between species and populations.

Relative colour cues improve colour constancy in birds

A ripe strawberry looks red to our eyes in sunlight and in the green light of a forest, although the spectrum of light reflected from its surface differs dramatically. This is caused by two effects: colour constancy and our ability to learn relative colour cues - the ripe strawberry remains relatively 'redder' than an unripe green strawberry. While colour constancy - the ability to recognize colours in shifted illumination - has been studied in many animals, the use of relative colour cues is investigated more rarely. In a previous study on chickens, we measured how large a shift in illumination their colour constancy mechanisms tolerate without reliable relative colour cues. Here, we show that chickens remain colour constant over larger illumination shifts, if they can use such relative colour cues. As relative colour cues are readily available in natural environments, we suggest that their use contributes strongly to colour constancy performance in nature.

Blue-Black or White-Gold? Early Stage Processing and the Color of 'The Dress'

Purpose

In Feb 2015 an image of a dress posted on Tumblr triggered an internet phenomenon: Is the Dress blue and black (BB) or white and gold (WG)? Many claim BB and others insist WG while the true colors are BB. The prevailing theory is that assumptions about the illuminant govern perception of the Dress with WG due to bluish lighting and BB due to yellowish. Our purpose was to determine if early stage optical, retinal and/or neural factors also impact perception of the Dress.

Methods

Thirty-nine subjects were categorized as BB or WG based on their initial perception of the Dress and their perception reported when viewing the Dress on iPhone 5, iPad, and 22” LCD displays. Macular pigment optical density (MPOD) measured with the QuantifEye™ MPS II and visual brainwaves (VEPs) in response to brief presentations of a transparency of the Dress illuminated by a flashing light were measured on each subject and compared between BB and WG groups. Additionally, CIE chromaticity (color) and luminance (brightness) were measured from multiple areas of the Dress image to determine cone stimulation and contrast.

Results

Mean MPOD was higher in the WG group (0.49) vs. the BB (0.41, p = 0.04) and median values were higher as well (WG = 0.46, BB = 0.36, p = 0.03). There was no difference in VEP amplitude between groups (p > 0.85) but mean VEP latency was longer in WG (130 msec.) vs. the BB group (107 msec., p = 0.0005). Colorimetry of the Dress showed significantly greater stimulation of blue cones (contrast = 73%) vs. red and green sensitive cones (contrast = 13%).

Conclusions

Our findings indicate that observers with denser MPOD may be predisposed to perceive the Dress as WG due to great absorption of blue light by the macular pigment. Moreover, the novel, substantial stimulation of blue cones by the Dress may contribute to ambiguity and dichotomous perception since the blue cones are so sparse in the retina. Finally, the delayed WG VEPs indicate distinct neural processing in perception of the consistent with fMRI evidence that the WG percept is processed at higher cortical levels than the BB. These results do not fully explain the dichotomous perception of the Dress but do exemplify the need to consider early stage processing when elucidating ambiguous percepts and figures.

Effects of surrounding stimulus properties on color constancy based on luminance balance

The visual system needs to discount the influence of an illuminant to achieve color constancy. Uchikawa et al. [J. Opt. Soc. Am. A29, A133 (2012) showed that the luminance-balance change of surfaces in a scene contributes to illuminant estimation; however, its effect was substantially less than the chromaticity change. We conduct three experiments to reinforce the previous findings and investigate possible factors that can influence the effect of luminance balance. Experimental results replicate the previous finding; i.e., luminance balance makes a small, but significant, contribution to illuminant estimation. We find that stimulus dimensionality affects neither the degree of color constancy nor the effect of luminance balance. Unlike chromaticity-based color constancy, chromatic variation does not influence the effect of luminance balance. It is shown that luminance-balance-based estimation of an illuminant performs better for scenes with reddish or bluish surfaces. This suggests that the visual system exploits the optimal color distribution for illuminant estimation [J. Opt. Soc. Am. A 29, A133(2012)].

[Excitement over the color of a dress: Blue-black or white-gold or else?]

BACKGROUND

The overexposed photograph of a dress caused a worldwide media hype in February 2015. The picture of the dress was posted on a social site with the question of which colour it was. Markedly different views on the color were voiced and strongly defended.

OBJECTIVE

The phenomenon is presented in this article, the actual color of the dress is clarified and a discussion on why such conflicting views were held is presented.

CONCLUSION

It becomes clear that in the perception of colors we unconsciously but appropriately take the lighting situation into consideration. Because we additionally name colors categorically, small differences in hue and illumination can lead to an assignment to very different categories.

Time-lapse ratios of cone excitations in natural scenes

The illumination in natural environments varies through the day. Stable inferences about surface color might be supported by spatial ratios of cone excitations from the reflected light, but their invariance has been quantified only for global changes in illuminant spectrum. The aim here was to test their invariance under natural changes in both illumination spectrum and geometry, especially in the distribution of shadows. Time-lapse hyperspectral radiance images were acquired from five outdoor vegetated and nonvegetated scenes. From each scene, 10,000 pairs of points were sampled randomly and ratios measured across time. Mean relative deviations in ratios were generally large, but when sampling was limited to short distances or moderate time intervals, they fell below the level for detecting violations in ratio invariance. When illumination changes with uneven geometry were excluded, they fell further, to levels obtained with global changes in illuminant spectrum alone. Within sampling constraints, ratios of cone excitations, and also of opponent-color combinations, provide an approximately invariant signal for stable surface-color inferences, despite spectral and geometric variations in scene illumination.