The impact on people of colour symbolism in animation
Animation movies present an unusual set of challenges and questions to academics examining films from a cognitive perspective. When the boundaries of the real world do not exist like do in live action movies, the film maker is challenged to create the complete narrative space of scratch. How do animators succeed this seemingly enormous task? This question certainly precedes making film life and space in visual art has been a subject of deep study by artists, photographers, historians and psychologists alike. While the intention may be to create a highly realistic visual space, the option given to visual artists and animatorsÂ Â is to abandon principle of realism in favour of another different perspective on the visual reality. Animation alone can bring to life inanimate objects, challenge and defy laws of physics, and create visual effects beyond the bounds of possibility in live action film.Animation and colour have evolved since their respective beginning. Colour has been both an obvious challenge as well as a field for exploration for animators throughout the animation history. Scientific discoveries in aspect to the perception of colour also influenced its use in art and animation, making colour an ideal goal for further exploration in a psychological context.
In this chapter three questions will be posed and answered. First, what is colour exactly and how is it defined? And second, how has colour been used by animators through the history of animated film? Finally, how does our cognitive sense of colour shape the viewers’ cognitive sense of a film? The final question will focus on a specific population of animated movies (animated movies adapted for children), and how the use of colour in these films strategically differs from other types of films.
WHAT IS COLOR?
Color is a concept that philosophers, artists, and scientists have historically spent a great
deal of time exploring and quantifying. Physiologically, our perception of color results from
varying wavelengths of light being reflected onto the retina, which in turn are processed by cells
called photoreceptors. The relative responses to light spectra by these cells are what generate our
ability to see and distinguish between colors. Anomalies in photoreceptor cells can cause deficits
in the ability of an individual to see color, though in some unique circumstances, these anomalies
allow individuals to more finely discriminate between colors (Neitz, Kraft, & Neitz, 1998;
Jordan & Mollon, 1992; Nagy, MacLeod, Heyneman, & Eisner, 1981).
Quantifying Color. Attempts to categorize color vastly predate our understanding of the
physiology of the eye, but Isaac Newton’s Opticks (1704) is pivotal in its introduction of his
color wheel for understanding color theory2. The ordering of the colors around his color circle
(and in subsequent iterations by other color theorists) is based on the order in which the colors
are refracted out from the prism, uniting violet and red to close the radial axis (see Figure 1).
Thus the ordering of the colors on the color wheel is not arbitrary, but based in the physics of
light. Newton also introduced the notions of primary and secondary colors, and notes that
opponent colors on the color wheel combine to create a neutral light color3.
Despite that it has evolved over time and exists in varying forms, the color wheel
continues to play an important part in both the artistic and psychological understanding of color.
Notably, it is useful for defining several metrics of color, namely hue and saturation. Hue refers
generally to named colors, and corresponds to the sectors of color into which color wheels are
typically divided. Examples of hue-based descriptions include ‘blue-green,’ ‘red,’ and ‘pink.’
Saturation is another important color variable, and generally refers to how bright or potent a
color is. Pastel colors (which are closer to the center of the color circle) are relatively
unsaturated. Very saturated colors (which are referred to as ‘bright red’ or ‘bold blue,’ for
example) lie along the outer edges of the color circle. Luminance is another variable important
for discussing color. Luminance refers to how light or dark something is; when discussing color
in particular, it refers to how much black is contained within a particular color. Unlike hue and
saturation, luminance can be independent of color; in other words, black-and-white images
contain no hue or saturation information but do contain luminance information. Because
luminance is not a variable unique to color stimuli, it is not represented on the color wheel4, but
nonetheless it is an important variable when discussing color. These terms, including how they
are mathematically quantified, will be revisited later with data.
The color wheel is not the only color quantification system to define colors using the
metrics of hue, saturation, and luminance. One of the most noted color-classification systems,
and the one still most reliably used in psychophysiological testing, was originally developed by
Albert Munsell, and also uses these color parameters5. Munsell compiled and organized a
tremendous set of finely-grained discrete colors now known as Munsell colors or Munsell chips
(Munsell, 1912; Munsell, 1919). One important component of the Munsell color system is that it
emphasizes that color perception is dependent on the physiology of the human eye. For example,
humans can more identify many more discrete levels of yellow than blue at high values, whereas
the reverse is true at low values. In other words, one can argue more light yellows exist than dark
yellows, whereas color wheel representations suggest that all color values exist equally in our
While the study of how we physiologically perceive color is important, perhaps more
critical in studying art and film from a cognitive perspective is the question of how we
psychologically respond to color. Our preferences for and biases toward particular colors have
the potential to influence how we respond cognitively and emotionally to art.
Color Preferences. Artists across visual domains recognize how the use of color
affects viewers’ perception of their work. Deliberate and comprehensive choices regarding the
use of color permeate all types of visual art, including intentional choices to omit color from
artwork6. Unsurprisingly, people tend to have strong predelictions for particular colors. While it
might intuitively seem like individuals each have their own color preferences that are unique, the
psychological research on color preferences reveals a surprising amount of concordance across
people in terms of color preferences.
In terms of specific colors, research has consistently demonstrated a cross-gender and
cross-cultural preference for blue hues above other hues (Eysenck, 1941; Granger, 1952;
McManus, Jones, & Cottrell, 1981; Komar & Melamid, 1997). People also tend to consistently
rate yellow and brown hues as being least pleasant, especially in their darker forms (Palmer &
Schloss, 2010). Biases across populations are not limited to hue; people consistently tend to
favor colors in more saturated forms as opposed to more washed-out or pastel counterparts of the
same hue (Granger, 1952).
Naturally, the consistency in color preference drove psychologists to posit theories on
how color preference develops. Some have proposed that color preference is an innate artifact of
human evolutionary history, which developed to facilitate our early survival in hunter-gatherer
societies (Hulbert & Ling, 2007). While some biological evidence supports this idea, if color
preferences are present at birth, infants and adults should show similar color preferences, when
in fact they do not. Data collected from infants and young children suggest that color preferences
change over time, and that while children eventually match adults on their color preferences
later, they are not born with those preferences. Infants tend to prefer colors that adults classify as
unpleasant, namely dark yellows, yellow-greens and reds (Adams, 1987). Children also have a
preference for very high saturation that gradually diminishes to match the adult preference level
for saturation (Child, Hansen, & Hornbeck, 1968).
Since preferences for color dimensions seem to be dynamic over the lifespan, it is
unlikely that color preferences are built-in. This is not to say that color preference is purely nonfunctional;
in fact, the ecological valence theory of color preference suggests that the early
associations humans build with colored objects facilitate their color preferences (Palmer &
Schloss, 2010). For example, our early preference for dark yellows in infancy may come from
consistent positive exposure to caregiver skin tones and hair color; it is only later that we learn
the association between dark yellows and rotten food or excrement, at which point this
preference changes direction. Conversely, as we increase our exposure to stimuli like clean water
and fresh food, our preferences for blues and slightly-saturated hues begins to dominate color
Yet another theory, which is particularly relevant for the use of color in an art space, is
that we learn strong associations between emotion and color, and color can consequently be used
to evoke particular states of emotion. Specific colors have been shown to correlate with arousal
(Valdez & Mehrabian, 1994) and scales of emotional valence (Kaya & Epps, 2004; for a detailed
review on color-emotion literature, see also Steinvall, 2007). This theory is not necessarily at
odds with other theories on color preference; in fact, it may simply supplement the idea that
gaining positive associations with a color increases our preference for that color, which is an
assumption that guides most current theories on color preference.
The question that remains from our understanding of color preference is whether or not
art mimics life; in other words, how do animated filmmakers instill color in an artificial world,
and do filmmakers exploit our color preferences in order to make their films more engaging?
HOW IS COLOR USED IN ANIMATION?
Color is arguably one of the most salient features of even the earliest animated films. This
is not to imply, however, that the techniques involved in creating an animated space with
dynamic color is a simple process. In fact, some of the biggest obstacles in moving animation
forward as an art form arose from the complications of colorization.
Cel Animation. Often referred to as ‘traditional animation,’ the cel animation approach
dominated the animated film landscape from very early in film’s history to the relatively recent
advent of computer animation. Cel animated films composite a meticulously painted background
layer with transparent celluloid (or ‘cel’) layer containing foreground information. Each layer
carries with it important implications for how color is ultimately represented and rendered in the
The background layer, while usually created first, must work reciprocally with the cel
layers in order for the colors to appear natural together and for the layers to appear integrated.
The overuse of color, in particular colors that are heavily saturated, tends to overwhelm cel
forms placed overtop the background; instead, the background ideally consists of more muted
colors to complement the component cel forms. This led to the Disney animated film signature
‘watercolor effect’ of its background layers (Thomas & Johnson, 1995).
The cel layer presents significantly more challenges where color is involved, and these
challenges were originally addressed by Disney’s larger-budget animation studios. The physical
properties of celluloid itself have implications for color; the thicker the cel, the darker the
resulting colors layered onto the cel layer (Thomas & Johnson, 1995). Thus, paint color had to be
balanced in such a way that the resulting cel painting did not clash with the watercolor
appearance of the background layer. Colors high in saturation were often difficult to achieve
because they also ultimately darkened when photographed from the cel. Disney’s animators
found that muted colors in the cel layer often were the best complement for a variety of
background layers. When designing a character or a cel-layer object, animators were often
limited by the expense of cel paint colors, and thus character design was in a sense limited by
color. Adding to this complication, cel artists and color keys also had to adjust the color palettes
of characters depending on the implied lighting of a background, to avoid a character looking
overly-red or overly-saturated in a nighttime scene, for example (Thomas & Johnson, 1995).
Color in the cel layer also contained some complications for maintaining realism in the animated
scene. For example, outlining characters in black often made their appearance visually heavier
and detracted from their integration with the background layer. Disney first introduced colored
inking to replace universal blank inking, and colored inking was also integrated with cel
Xeroxing technology as that emerged (Thomas & Johnson, 1995). Another color problem dealt
with creating depth in the cel layer: textures in hair and fur could be created via airbrushing and
drybrushing, but this created a flicker effect when the individual cels were captured in sequence.
Animators ultimately decided this depth was worth a certain small-scale amount of flicker
tradeoff (Thomas & Johnson, 1995).
Computer Animation. The cel approach dominated animated films for decades, and the
interest in streamlining the cel animation process led to the initial involvement of computers in
animation. The first film to be digitally composited was Disney’s 1990 film The Rescuers Down
Under (Prince, 2012). Computer involvement in animation was also prioritized as a means of
film restoration and improving film resolution; that same year, digital paint techniques allowed
Disney to fix flaws in the original print of Fantasia for reissue, and in 1993, Snow White and the
Seven Dwarves was completely ‘restored’ to create a higher-resolution version of the film
(Bordwell, 2012). Computer-based coloring was particularly valuable because it generated more
freedom to alter independent components of an image. Prior to computer involvement, color
correction had to be done on a whole-frame basis; the process of digitally compositing and
altering films meant that color-correction could be done on an individual object or character
without the need to alter the entire frame image (Prince, 2012).
The involvement of computers in animation continued to grow as the technology became
more inexpensive and accessible, and animators experimented with new computer-based
techniques for animating (such as crowd-generation in Mulan(1998))7. By the mid-90’s, the vast
majority of cel animated films employed computers to streamline the once-arduous tasks
involved in hand-animating films, including colorization. Because animators no longer had to
rely on physical paint or hand-calibrate background and cel layers, the colorization and
texturization processes became much easier, and artists in turn were able to work with more
degrees of freedom in their animating.
The revolution in computer animation began with the first fully computer-based animated
film, Toy Story (1995). Moving from a two-dimensional animation space into a threedimensional,
digitally-constructed environment had a huge initial investment cost (both in labor
and finance), but ultimately gave the animated filmmaker a great deal of flexibility in
constructing visual narratives (for a review, see Lasseter, 1987). Constructing and coloring a 3D
environment and set of characters involves a great deal of initial time and planning, but the
ultimate outcome is a greater degree of control in colorization, in which every individual element
in the digital landscape can be fine-tuned in color space.
Film Stock. One important caveat worth noting when discussing animation is that the
color of the final product is always affected by the film stock. Even in contemporary computer
animation, where color design can be done on a very fine-grained scale, the final film is
ultimately rendered onto film stock. The choice of film stock, as evidenced especially by the
changes in stock availability and popularity over time, as well as advances in stock quality,
renders color variably (Bordwell & Thompson, 2004). Technicolor film stock was popular with
early Disney animated films, which exacerbated complications with cel painting by rendering
colors heavy in midtones. This forced animators and color keys8 into a particular spectrum of
colors when painting in order to achieve the desired final look on the Technicolor film stock
(Thomas & Johnson, 1995). Even in modern animated films, the change between the cel or
computer and the film stock accounts for some variability in coloring of the final product.
Indeed, this is not even the last step in color variance: the original camera negative is almost
always different from the colors displayed in theaters, on home televisions, or on computer
screens (Prince, 2012). Some of this variance can potentially be put to rest with the increasing
number of films being distributed as Digital Cinema Packages (rather than in 35mm form), but it
persists as a problem for those interested in studying pinpointing color in film scientifically
It is clear that artists have more freedom with color in animated films. Before digital
technology, live action films were confined by the natural color of objects in a scene as well as
by the limited amount of post-production work available to alter color (Prince, 2012). However,
from animation’s inception, animators have been able to select a wide range of colors to best suit
their needs, despite some of the early cost and technical constraints. The introduction of
computer animation allows for the greatest amount of freedom in color control, putting the entire
digital color environment under the direction of the artistic team.
The precise control of color in this setting not only has artistic consequences, but also
important implications for how films can evoke particular psychological responses from its
viewers. The rest of this chapter will examine work revolving around the use of color for a
particular audience of animated viewers: specifically, how filmmakers use color in animated
films intended for children.
CHILDREN’S ANIMATED FILMS:
ARE THERE DIFFERENCES IN COLOR USE?
In the introduction to her book A Reader in Animation Studies, Jayne Pilling (1997)
discusses how Disney, as the first company to invest heavily in animated features, eventually
became the model for animated films and subsequently marginalized animation into an art form
“somehow intrinsically only appropriate for entertaining children” (xi). Indeed, it appears that
the Disney model caused an aggressive bifurcation in the animated feature world, with heavy
emphasis being placed on the creation of child-oriented animated films, and a smaller contingent
of artists attempting to legitimize animation as an art form appealing to adults. While Pilling is
correct in that the latter set of films is certainly underrepresented in the film studies literature,
child-oriented animated features have a particular appeal for being studied from a cognitive
perspective. Filmmakers in this animation subset face a specific challenge in trying to engage
children in their visual narrative; there is ample evidence that the cognitive and attentional
capacities of children differ from those of adults considerably, so what changes must the director
of a children’s animated film make in order to captivate this unique audience? One potential shift
to accommodate this audience appears to take place in colorization of these films.
In order to study the physical properties (including color) of children’s films, we
assembled a sample9 of G-rated children’s films made between 1985 and 2008 (Brunick,
DeLong, & Cutting, 2012; Brunick & Cutting, in prep). Films in the sample were the highestgrossing
G-rated theatrical films from each year in the range and also included some direct-tovideo
films10. The sample included live-action, cel animated and computer animated films geared
to a variety of ages11. We considered our entire sample of children’s films for our original
analyses; for the purposes of this chapter, only the animated films (both cel and computer) will
be discussed. This sample is contrasted with a subsample of adult-geared, non-animated films
from the same time period (see the 1985 through 2005 films from Cutting, DeLong, & Nothelfer,
2010). The following sections will (1) discuss how the color parameter in question was
mathematically quantified and (2) discuss the trends in the color parameter for the child- and
Saturation. As discussed earlier, saturation refers to the ‘brightness’ or ‘boldness’ of a
color. Saturation radiates outward from the center of the color circle: the center of the circle is
white, with no saturation, while the edges of the circle represent fully-saturated forms of a
particular hue. However, when analyzing color digitally, saturation is typically not discussed in
terms of a color wheel, but instead in terms of a digital color space known as the HSV cone. This
space is named for its dimensions: hue, saturation, and value. Value is roughly equivalent to
luminance, and this space is essentially constructed by adding this variable to the color wheel
(see Figure 2). The base of the HSV cone is a color wheel, and the height of the cone represents
value. As value decreases (as the colors become darker), colors are limited in their saturation.
Saturation is generally quantified on a scale from 0 (white, no saturation) to 1 (fully saturated).
Saturation levels for each pixel in a frame were digitally computed. The median
saturation level for all the pixels in each frame was computed, and an average of the frames was
obtained for the entire film. Within the children’s film sample, we found that cel animated films
use significantly more saturated colors than computer-generated animated films, independent of
the year that the films were made. Both live-action children’s films and the matched sample of
adult-geared films have been increasing in saturation over time; in other words, newer films are
more prone to be more saturated than older films. However, even with this trend, the live-action
children’s films and adult-directed films are dramatically less saturated than their animated
counterparts. This finding is both interesting and unsurprising for the same reason: the saturation
levels in children’s films likely reflect young children’s preference for bright colors. However, it
is unlikely that filmmakers are consciously making these choices based on the psychological
literature; filmmakers instead are likely intuiting this preference, perhaps based on their own
conceptions of how children respond to film or other parts of their visual environment.
Regardless of the basis of this intuition, it is important to note that the saturation trends in the
films appear to match the scientifically-established preferences of the target audience.
Luminance. Though it can be measured independently of color, luminance plays an
important part in color space and ultimately how a color is perceived on-screen. To assess
luminance, color was digitally removed from the film using a standard digital grayscale
conversion. Each pixel’s luminance value is computed, with values ranging from 0 (pure black)
to 255 (pure white). The mean of the pixels in a frame were averaged to create the mean
luminance for that frame, and the frames were subsequently averages to create whole-film
The trend in Hollywood films for adult audiences is a decrease in luminance; in other
words, films have steadily been getting darker throughout the studied period, which has
implications for directing eye-gaze and attention of the viewer during the film (Cutting, Brunick,
DeLong, Iricinschi, & Candan, 2011; Smith, 2012). Animated films for children, conversely,
maintain a steady level of brightness independent of year, the target age of the film, or what type
of animation (cel or computer) was used. While one could argue that consistent brightness is a
possible artifact of representing particular colors in animation, children’s live action films are
actually increasing in brightness over this period; this evidence instead supports an interpretation
that the intended audience is driving the brightness level, not simply that animated films are
Another question posed by these findings is the potential interaction between saturation
and luminance. As demonstrated by the HSV cone, colors with lower values are limited in their
saturation. Is it then possible that children’s films are more saturated only because they are
brighter? Or, perhaps, does the inclination of filmmakers to use saturated colors in children’s
films necessitate a certain luminance level? While this is certainly possible, it is unlikely that the
luminance findings are purely an artifact of the saturation levels, or vice versa. If this were the
case, one would expect the trends in both the children’s films and adult-directed films to be
complementary; in other words, both luminance and saturation should be increasing or
decreasing together in the samples. This is not what we find. In the children’s sample, saturation
levels hold steadily across time, while these films have increased in brightness over the same
period. Even more importantly, adult-geared films have gotten considerably darker, but have also
become steadily more saturated, not less. This evidence suggests that while luminance and
saturation have a reciprocal relationship, and while some of the variance in one accounts for
variance in the other, the findings reported here on the two metrics are largely independent.
Hue. As discussed earlier, hue generally refers to named colors. In both the color wheel
and in the HSV cone, hue is represented around the radial edge. One major problem with this
representation of hue is that it is based in circular geometry, which makes mathematically
quantifying and comparing hues difficult and unintuitive. Fully isolating luminance from hue in
the HSV color space is also problematic; an ideal space for considering hue would allow for a
full spectrum of colors to be represented (1) in a more convenient mathematical space and (2)
independent of luminance.
Accordingly, we considered hue using the YCbCr color space, which meets these
important criteria. This color space takes the form of a rectangular prism on a diagonal axis (see
Figure 3). This color space is also named for its axes in the space: Y (on the vertical axis) refers
to luminance, while Cb and Cr refer respectively to chrominance-red and chrominance-blue. The
chrominance axes plot complementary colors from the color wheel (red-green and blue-yellow,
respectively) on opposite rectangular planes of the prism. The distinct advantage of YCbCr is
having luminance on its own axis; in this way, one could take a square slice through the prism to
get a square containing all colors at an isoluminant level.
Rather than examining whole-film hue, which is nearly impossible without reducing hue
on arbitrary dimensions, our research has examined the hue of particular characters in children’s
films. We asked independent coders to view children’s animated films in grayscale, and to
identify unambiguous protagonists and antagonists in the film. Frames containing these
characters were selected, and the characters themselves were extracted from their background.
The dominant hue of the protagonists and antagonists were plotted on an isoluminant slice of
YCbCr color space. The analyses showed that protagonists, defined as unquestionably positive
and morally-right characters, contained more blue and green hues. Antagonists, conversely,
contained more red and yellow hues (Brunick, DeLong, & Cutting, 2012). Unlike saturation,
where films mimic the preferences found in children, this analysis shows that the use of hue in
children’s films coincides with adult hue preferences. If children’s preferences were being
exploited, ‘good’ characters would likely contain more child-preferred hues, such as red and
yellow, when in reality precisely the opposite occurs. It is unclear why this trend is present, and
certainly merits further analysis. One possible explanation is that the shift in hue preferences
supposedly occurs earlier than the shift in saturation preferences; adults may not be as aware of
the hue preference in children because it shifts earlier, and thus adults and filmmakers have less
exposure to this cognitive facet of child color preference.
The implications for studying children’s animated films, and children’s films in general,
are vast. Researchers not only are able to gain insight into children’s cognitive capacities and
preferences, but they can also observe the early reciprocal relationship between filmmaker and
viewer. While films for adults are mostly classified as art or entertainment, film in a child’s
world also serves as an important tool for learning. Facilitating early learning from visual stimuli
is a major goal of both psychology and education researchers, and children’s films can serve as
an important medium for conveying both artistic and educational information to a young
Not only is it important to note that color in film is quantifiable on several dimensions,
but we must also recognize that these cinemetric data are accessible and worthy of further study.
Animation in particular presents a unique opportunity for quantifying and studying color;
animators have always had greater control of the colors used in their filmic world than directors
of live action films, and technology has only improved the level of detail and the deliberateness
with which the animator can control his color space. This is not to suggest that instantiating color
in an animated film is an easy task; whether through meticulous color selection and handpainting,
or through the arduous development of software for digital colorization, the creation of
the brilliant colors typical of animated films carries with it a high investment of both time and
The commonalities and differences in perception across people has always been a topic
of interest for psychologists, and a cognitive approach to film is yet another way to access and
interpret facets of this topic. The preponderance of children’s films among animated films is in a
sense very convenient because it allows research to observe the reciprocal relationship between
children’s perceptual biases and how filmmakers create art and media for this audience. The
developmental perspective affords us insight into the origins of our perceptions and preferences,
which will likely prove to be an important pillar in cognitive perspectives to film.
The impact on people of colour symbolism in animation