The 12 animation principles are a great foundation for any animator to understand, and failure to do so will result in missing some of the underlying fundamentals of animation—visible in many a junior’s work. Ultimately, however, they were written with the concept of linear entertainment like TV and film in mind, and the move to 3D kept all of these elements intact due to the purely aesthetic change in the medium. Three-dimensional animated cartoons and visual effects are still part of a linear medium, so they will translate only to certain elements of video game animation—often only if the game is cartoony in style. As such, it’s time to propose an additional set of principles unique to game animation that don’t replace but instead complement the originals. These are what I have come to know as the core tenets of our new nonlinear entertainment medium, which, when taken into consideration, form the basis of video game characters that not only look good, but feel good under player control—something the original 12 didn’t have to consider. Many elements are essential in order to create great game animation, and they group under five fundamental areas:
1.Feel
2.Fluidity
3.Readability
4.Context
5.Elegance
Feel
The single biggest element that separates video game animation from traditional linear animation is interactivity. The very act of the player controlling and modifying avatars, making second-to-second choices, ensures that the animator must relinquish complete authorship of the experience. As such, any uninterrupted animation that plays start to finish is a period of time the player is essentially locked out of the decision-making process, rendered impotent while waiting for the animation to complete (or reach the desired result, such as landing a punch).The time taken between a player’s input and the desired reaction can make the difference between creating the illusion that the player is embodying the avatar or becoming just a passive viewer on the sidelines. That is why cutscenes are the only element in video games that for years have consistently featured a “skip” option—because they most reflect traditional noninteractive media, which is antithetical to the medium.
Response
Game animation must always consider the response time between player input and response as an intrinsic part of how the character or interaction will “feel” to the player. While generally the desire is to have the response be as quick as possible (fewer frames), that is dependent on the context of the action. For example, heavy/stronger actions are expected to be slower, and enemy attacks must be slow enough to be seen by the player to give enough time to respond.
It will be the game animator’s challenge, often working in concert with a designer and/or programmer, to offer the correct level of response to provide the best “feel,” while also retaining a level of visual fidelity that satisfies all the intentions of the action and the character. It is important not to sacrifice the weight of the character or the force of an action for the desire to make everything as responsive as possible, so a careful balancing act and as many tricks as available must be employed.
Ultimately, though, the best mantra is that “gameplay wins.” The most fluid and beautiful animation will always be cut or scaled back if it interferes too much with gameplay, so it is important for the game animator to have a player’s eye when creating response-critical animations, and, most importantly, play the game!
Inertia & Momentum
Inertia is a great way to not only provide a sense of feel to player characters, but also to make things fun. While some characters will be required to turn on a dime and immediately hit a run at full speed, driving a car around a track that could do the same would not only feel unrealistic but mean there would be no joy to be had in approaching a corner at the correct speed for the minimum lap time. The little moments when you are nudging an avatar because you understand their controls are where mastery of a game is to be found, and much of this is provided via inertia.
Judging death-defying jumps in a platform game is most fun when the character must be controlled in an analogue manner, whereby they take some time to reach full speed and continue slightly after the input is released. This is as much a design/programming challenge as it is animation, but the animator often controls the initial inertia boost and slowdown in stop/start animations.
Momentum is often conveyed by how long it takes a character to change from current to newly desired directions and headings. The general principle is that the faster a character is moving, the longer it takes to change direction via larger turn-circles at higher speeds or longer plant-and-turn animations in the case of turning 180°.Larger turn-circles can be made to feel better by immediately showing the intent of the avatar, such as having the character lean into the turn and/or look with his or her head, but ultimately we are again balancing within a very small window of time lest we render our characters unresponsive.
A classic example is the difference between the early Mario and Sonic the Hedgehog series. Both classic Mario and Sonic’s run animations rely heavily on inertia and have similar long ramp-ups to full speed. While Mario immediately starts cartoonishly running at full speed as his legs spin on the ground to gain traction, Sonic slowly transitions from a walk to a run to a sprint. While Mario subjectively feels better, this is by design, as Sonic’s gameplay centers on high speeds and “flow,” so stopping or slowing down is punitive for not maintaining momentum.
Visual Feedback
A key component of the “feel” of any action the player and avatar perform is the visual representation of that action. A simple punch can be made to feel stronger with a variety of techniques related to animation, beginning with the follow-through following the action. A long, lingering held pose will do wonders for telling the player he or she just performed a powerful action. The damage animation on the attacked enemy is a key factor in informing the player just how much damage has been suffered, with exaggeration being a key component here.
In addition, employing extra tricks such as camera-shake will help further sell the impact of landing the punch or gunshot, not to mention visual effects of blood or flashes to further register the impact in the player’s mind. Many fighting games employ a technique named “hit-stop” that freezes the characters for a single frame whenever a hit is registered. This further breaks the flow of clean arcs in the animations and reinforces the frame on which the impact took place.
As many moves are performed quickly so as to be responsive, they might get lost on the player, especially during hectic actions. Attacking actions can be reinforced by additional effects that draw the arc of the punch, kick, or sword-swipe on top of the character in a similar fashion to the “smears” and “multiples” of old. When a sword swipe takes only 2 frames to create its arc, the player benefits mostly from the arcing effect it leaves behind.
Slower actions can be made to feel responsive simply by showing the player that at least part of their character is responding to their commands. A rider avatar on a horse can be seen to immediately turn the horse’s head with the reins even if the horse itself takes some time to respond and traces a wide circle as it turns. This visual feedback will feel entirely more responsive than a slowly turning horse alone would following the exact same wide turn.
Much of the delay in visual feedback comes not from the animation alone, but the way different game engines handle inputs from the joypad in the player’s hands. Games like the Call of Duty series place an onus on having their characters and weapons instantly respond to the player’s inputs with minimal lag and high frame rates, whereas other game engines focused more on graphics postprocessing will have noticeably longer delays (measured in milliseconds) between a jump button-press and the character even beginning the jump animation, for example. This issue is further exacerbated by modern HDTVs that have lag built in and so often feature “Game Mode” settings to minimize the effect. All this said, it is still primarily an animator’s goal to make characters as responsive as possible within reason.
Fluidity
Rather than long flowing animations, games are instead made of lots of shorter animations playing in sequence. As such, they are often stopping, starting, overlapping, and moving between them. It is a video game animator’s charge to be involved in how these animations flow together so as to maintain the same fluidity put into the individual animations themselves, and there are a variety of techniques to achieve this, with the ultimate goal being to reduce any unsightly movement that can take a player out of the experience by highlighting where one animation starts and another ends.
Blending and Transitions
In classic 2D game sprites, an animation either played or it didn’t. This binary approach carried into 3D animation until developers realized that, due to characters essentially being animated by poses recorded as numerical values, they could manipulate those values in a variety of ways. The first such improvement that arrived was the ability to blend across (essentially cross-fading animations during a transitory stage) every frame, taking an increasing percentage of the next animation’s value and a decreasing percentage of the current as one animation ended and another began. While more calculation intensive, this opened up opportunities for increasing the fluidity between individual animations and removing unsightly pops between them.
A basic example of this would be an idle and a run. Having the idle immediately cancel and the run immediately play on initial player input will cause the character to break into a run at full speed, but the character will unsightly pop as he or she starts and stops due to the potential repeated nature of the player’s input. This action can be made more visually appealing by blending between the idle and run over several frames, causing the character to more gradually move between the different poses. Animators should have some degree of control over the length of blends between any two animations to make them as visually appealing as possible, though always with an eye on the gameplay response of the action.
The situation above can be improved further (albeit with more work) by creating brief bespoke animations between idle and run (starting) and back again (stopping), with blends between all of them. What if the player started running in the opposite direction he or she is facing? An animator could create a transition for each direction that turned the character as he or she began running in order to completely control the character’s weight-shift as he or she leans into the desired direction and pushes off with his or her feet. What if the character isn’t running but only walking? Again, the animator could also create multiple directional transitions for that speed. As you can see, the number of animations can quickly spiral in number, so a balance must be found among budget, team size, and the desired level of fluidity.
Seamless Cycles
Even within a single animation, it is essential to maintain fluidity of motion, and that includes when a cycling animation stops and restarts. A large percentage of game animations repeat back on themselves, so it is important to again ensure the player cannot detect when this transition occurs. As such, care must be taken to maintain momentum through actions so the end of the animation perfectly matches the start.It is not simply enough to ensure the last frame of a cycle identically matches the first; the game animator must also preserve momentum on each body part to make the join invisible. This can be achieved by modifying the curves before and after the last frame to ensure they create clean arcs and continue in the same direction. For motion-capture, where curves are mostly unworkable, there are techniques that can automatically provide a preservation of momentum as a cycle restarts that are described later in this book.
Settling
This kind of approach should generally be employed whenever a pose must be assumed at the end of an animation, time willing. It is rather unsightly to have a large movement like an attack animation end abruptly in the combat idle pose, especially with all of the character’s body parts arriving simultaneously. Offsetting individual elements such as the arms and root are key to a more visually pleasing settle.Notably, however, games often suffer from too quickly resuming the idle pose at the end of an animation in order to return control to the player to promote response, but this can be avoided by animating a long tail on the end of an animation and, importantly, allowing the player to exit out at a predetermined frame before the end if new input is provided. This ability to interrupt an animation before finishing allows the animator to use the desired number of frames required for a smooth and fluid settle into the following animation.
Settling is generally achieved by first copying the desired end pose to the end of an animation but ensuring some elements like limbs (even divided into shoulder and forearms) arrive at their final position at different times, with earlier elements hitting, then overshooting, their goal, creating overlapping animation. Settling the character’s root (perhaps the single most important element, as it moves everything not planted) is best achieved by having it arrive at the final pose with different axes at different times. Perhaps it achieves its desired height (Y-axis) first as it is still moving left to right (X-axis), causing the root to hit, then bounce past the final height and back again. Offsetting in the order of character root, head, and limbs lessens the harshness of a character fully assuming the end pose on a single frame—though care must be taken to not overdo overlap such that it results in limbs appearing weak and floppy.
Readability
After interactivity, the next biggest differentiator between game and traditional animation, in 3D games at least, is that game animations will more often than not be viewed from all angles. This bears similarity to the traditional principle “staging,” but animators cannot cheat or animate to the camera, nor can they control the composition of a scene, so actions must be created to be appealing from all angles. What this means is when working on an animation, it is not enough to simply get it right from a front or side view. Game animators must take care to always be rotating and approving their motion from all angles, much like a sculptor walking around a work.
Posing for Game Cameras
To aid the appeal and readability of any given action, it is best to avoid keeping a movement all in one axis. For example, a combo of three punches should not only move the whole character forward as he or she attacks, but also slightly to the left and right, twisting as they do so. Similarly, the poses the character ends in after every punch should avoid body parts aligning with any axes, such as arms and legs that appear to bend only when viewed from the side. Each pose must be dynamic, with lines of action drawn through the character that are not in line with any axes.
Collision & Center
of Mass/Balance As with all animation, consideration must be given to the center of mass (COM; or center of balance) of a character at any given frame, especially as multiple animations transition between one another so as to avoid unnatural movements when blending. The COM is generally found over the leg that is currently taking the full weight of the character’s root when in motion or between both feet if they are planted on the ground when static. Understanding this basic concept of balance will not only greatly aid posing but also avoid many instances of motions looking wrong to players without them knowing the exact issue.
This is especially true when considering the character’s collision (location) in the game world. This is the single point where a character will pivot when rotated (while moving) and, more importantly, where the character will be considered to exist in the game at any given time. The game animator will always animate the character’s position in the world when animating away from the 3D scene origin, though not so if cycles are exported in place. Importantly, animations are always considered to be exported relative to this prescribed location, so characters should end in poses that match others (such as idles) relative to this position. This will be covered in full in the following chapter.
Context
Whereas in linear animation, the context of any given action is defined by the scene in which it plays and what has happened in the story up to that point and afterward, the same is impossible in game animation. Oftentimes, the animator has no idea which action the player performed beforehand or the setting in which the character is currently performing the action. More often than not, the animation is to be used repeatedly throughout the game in a variety of settings, and even on a variety of different characters.
Elegance
Game animations rarely just play alone, instead requiring underlying systems within which they are triggered, allowing them to flow in and out of one another at the player’s input—often blending seamlessly, overlapping one another, and combining multiple actions at once to ensure the player is unaware of the individual animations affording their avatar motion.If not designing them outright, it is the game animator’s duty to work with others to bring these systems and characters to life, and the efficiency of any system can have a dramatic impact on the production and the team’s ability to make changes further down the line toward the end of a project. Just as a well-animated character displays efficiency of movement, a good, clean, and efficient system to play them can work wonders for the end result.