What Is Animation? Twenty-Four Still Pictures a Second

The traditional explanation holds that the eye retains an image for a fraction of a second, so successive frames blur together. That accounts for not seeing gaps between frames and does not account for perceiving motion, which is a different problem: a retained image would produce a smear rather than a moving object. The current explanation involves two effects. The phi phenomenon, described by Max Wertheimer in 1912 and foundational to Gestalt psychology, is the perception of movement between two separated stimuli flashed in sequence, where the brain constructs a path that was never displayed. Beta movement is the related perception of an object actually travelling. Both are inferences made by motion-detecting neurons in the visual system, which respond to changes in position over time and do not care whether the intervening positions were shown. Animation exploits a perceptual system that is built to reconstruct motion from incomplete evidence.

The methods differ in how the frames are made and each has a characteristic look:

• •Traditional hand-drawn animation, in which each frame is drawn, historically on paper and then traced onto transparent cels and painted, with the cels laid over a separate background so that only the moving parts are redrawn, an efficiency invented in 1914
• •Stop motion, in which a physical object, puppet, model or clay figure is photographed, moved fractionally and photographed again, which is enormously laborious and gives the tangible quality of work in Aardman and Laika films
• •Cut-out animation, moving flat jointed pieces, which is cheap and was the original method of South Park before it moved to software imitating the look
• •Computer animation in three dimensions, in which models are built, given a skeleton, posed at key moments and the intermediate positions computed, with surfaces, lighting and physics simulated and the frames rendered, which now dominates feature production
• •Two-dimensional digital animation using vector puppets or drawn frames on a tablet, which has largely replaced paper without changing the principles
• •Rotoscoping, tracing over filmed footage, invented in 1915 and used whenever realistic human motion is required at low cost

Film runs at twenty-four frames a second and animators rarely draw twenty-four different images for each second. Working on twos, drawing each image and holding it for two frames, gives twelve drawings a second and is the standard for most hand-drawn work, because the difference is barely perceptible for ordinary movement and it halves the labour. Fast action is animated on ones and slow or held poses can go to threes or fours. Much Japanese television animation works on threes for economy and compensates with other techniques, including held frames with moving camera, limited animation of only the mouth and eyes, and the elaborate still compositions that became a stylistic signature rather than a limitation. Computer animation can trivially produce a different image for every frame and deliberately does not, since motion that is too smooth reads as weightless, which is why animators add motion blur and why the choice of frame rate in a game or a film remains an aesthetic decision.

Two Disney animators, Ollie Johnston and Frank Thomas, published twelve principles in 1981 that had been developed at the studio from the 1930s, and they remain the working vocabulary regardless of technique. Squash and stretch conveys weight and flexibility by deforming a shape while preserving its volume. Anticipation prepares the viewer for an action with a small opposite movement first. Follow-through and overlapping action mean that parts of a body continue after the main mass stops and that different parts move at different rates. Slow in and slow out reflects that things accelerate and decelerate rather than moving at constant speed, and is implemented in software as easing curves. Arcs recognise that almost nothing in nature moves in a straight line. Exaggeration pushes a pose beyond the literal to read clearly. Staging, timing, solid drawing and appeal cover the rest. The principles are about making motion legible rather than accurate, which is why a rotoscoped sequence of real movement frequently looks worse than an exaggerated drawn one.

The economics have driven the technology throughout. A hand-drawn feature requires tens of thousands of drawings, which is why studios developed the cel system, the multiplane camera, and, from the 1960s, the practice of outsourcing in-between drawing overseas, initially to Japan, Korea, the Philippines and later elsewhere, an arrangement that still underlies most television animation and that is largely invisible in credits. Computer animation moved the cost from drawing to setup, since building and rigging a character is expensive and reusing it is cheap, which is why computer-animated films have large casts of recurring characters and few crowd scenes with individual detail, until rendering became cheap enough that crowds became a selling point. Rendering itself is the industrial process nobody sees: a single frame of a modern feature can take hours of computation across a large cluster, which is why a studio's render farm is a substantial capital item. The current disruption is generative tooling, and the industry's response is an argument about authorship and employment that has already produced strike action.

Animation works because motion-detecting neurons reconstruct movement between separated images, an effect described in 1912 and distinct from the persistence of vision explanation still commonly taught. The techniques differ in how frames are produced, from drawn cels and physically posed puppets to computed three-dimensional models, and animators deliberately draw fewer than twenty-four images a second because the difference is imperceptible and the labour is halved. Twelve principles developed at Disney govern how to make motion legible rather than accurate, and the economics have driven every technical change.