Tutorial: Character Animation

Early in animation’s history, when superman animators first proved that a single person could produce the thousands of drawings necessary to make a cartoon, they used a technique now called “straight ahead animation”, where every frame was drawn as a minor adjustment on the previous frame. Though the process is simple to comprehend, it had no timing – it was difficult to know how much time a move would take, and every move flowed evenly into the next. Watch some early (pre 1930) cartoons to see examples.

Modern “keyframe” animation instead requires the animator to time the movement first, draw the frames that are the extremes of the motion, then fill in the in-between frames that make the movement smooth. As keyframe animation developed, master animators adjusted the inbetween frames to produce a visual impression of acceleration and deceleration simply by changing the spacing. Other master animators created distinctive styles by exaggerating the acceleration and deceleration and carrying the exaggeration into the keyframes. For example, a character’s head may turn so quickly that it snaps back, and probably even distorts the shape.

As you develop as an animator, your own style will emerge. You will become very proficient at certain moves and find yourself adding nuances that make your “look” different and recognizable. However, this is many thousands of frames in your future. For now, learn the basics.

Production

Character design, script writing, storyboards, and dialog recording are defined as pre-production. What is the story all about and what will it look like when it is done? What does the character look like? What will the character act out? Will there be interaction with other characters? Is the skeleton setup (bones) capable of the animation that you have planned? All of these things are important questions which must be answered before you begin your animation.

Animating is part of production.

In computer animation, lighting is usually just before rendering, which is post production. Recombining the dialog track, credits, and adding sound effects and music finish the process.

Storyboarding

Thumbnail sketches of every scene, called “storyboards” provide continuity for everyone working on the project. In essence, storyboards provide the guidelines that will assist in planning, designing sets, determining types of lighting, and many other factors necessary to the completion of the job. Most often, you will be involved in small productions that might involve just yourself or maybe another animator. As is often the case, there are no storyboards available, other than maybe an overall idea of what the production is about. Somebody is going to have to at least block the camera shots: what’s visible and for how long. You don’t have to hand drawn anything, Animation:Master is an excellent tool to create the storyboards using stand-in characters from the CD ROM.

Frame Rate

Traditional film animation occurs at the rate of 24 frames per second, with two exposures per frame, which requires 12 pictures (or cels) per second, but most of the animation that you will be doing will end up on videotape, and your reference action will probably be videotaped, so we will be discussing timing for 30 frames per second. (European television plays at the rate of 25 fps, but you should get used to 30 fps timing). Animation:Master allows you to change the frame rate after the animation is finished, no matter what frame rate it was created at.

Cel Mode

Computers use their mathematics excellence to help you, but one of their more subtle idiosyncrasies is their preference for starting frame numbering from “0” (rather than “1”). It has to do with their ability to calculate movement at fractional frame numbers, and to change frame rates. However, those esoteric examples can be ignored, so if you have long time animation experience that makes it more comfortable to work with traditional “starting at 1” numbering, select the “Cel Number (1)” option in the “Show time as” section of the “Units” tab of [Tools][Options…]; but if you are just learning animation, select the “Frame Elapsed (0)” option. This tutorial starts are “0” so if you want to use Cel mode, each time a frame number is indicated, add one to it. ( A third option, “SMPTE (00:00:00)”, is for people who are used to videotape editing).

Skeleton

Animation:Master has a suggested hierarchical layout and bone naming convention, called a “skeleton”, but others exist. It is very important to know the skeleton of your character because that determines what existing actions can be used with it. (The “Notes” tab on the Model and Action Properties panels should identify the skeleton type.

Extra Constraints

Constraints are a common tool you use to computer animate. Constraints are so named because they “constrain” something: feet are constrained to the floor; eyes are constrained where to look; knees are constrained to point forward. Constraints often use invisible reference objects, called “Nulls”, to target to, but constraints also use bones, lights, cameras, splines, etc.

Since eye targets are so commonly used, oftentimes a model builder will put them permanently into the skeleton. Similarly, feet and knee targets are permanently in the skeletons of many models. Most constraints, however, are part of the action. For example, several unique constraints are needed to do the “hula”, but when then character is finished gyrating, the constraints go away.

Which Window?

It is your preference whether you want to animate in an Action window or a Choreography window. Many animators prefer the Choreography window because the character is surrounded by the rest of the scene, which oftentimes makes it easier to interact with and react to. However, redraw times can be slow, slowing the animation process, but parts of the scene that are not immediately relevant can be hidden (which speeds up drawing times, so the animators that animate primarily in a Choreography window usually switch visibility as needed). Actions created in a Choreography window can even be saved as “reusable”, which is usually the domain of an Action window.

Suppose you had an action of a bipedal character sitting down. The action may have come from a library of actions created some time ago, maybe by some other animator (or motion capture), but now you want to use it. When you apply this libraried action to your character, you might have to use the Action List to make it go faster, or you may need only a piece of the action. Certainly you will want to modify it slightly to perfectly fit the situation. Any changes like these do not modify the original reusable action, instead they are simply adjustments called “Choreography actions”, and are not reusable. Actions created in an Action window are usually intended to be reused on the same or different characters.

Cycled actions, like “run”, are another example of actions that need to be created in an Action window. Cycled actions cannot be made in a Choreography window because the character’s relative movement is driven by an Ease channel in the choreography, each step speeding up or slowing down as needed. In an Action window, you work with a consistently timed step, and tools, like “stride length”, are provided to help you make the action reusable.

Multiple Cameras

When you animate in an Action window, the character stands facing you from the Front view (called an ordinal view: like Back, Left, Right, Top, and Bottom). Usually, no matter what you have the character doing, you can animate from an ordinal view that best shows the movement. You may even have several windows open at once, all showing the same action from different ordinal views.

When you animate in a Choreography window, however, the character could be facing any which way after it has been positioned in the scene; and during the animation it often changes its orientation dramatically. The ordinal views are usually of little help. To remedy this, Animation:Master lets you animate (or model, or whatever) from a Camera’s view, and the camera can be placed anywhere. The cameras may be part of the choreography where they can be constrained to point at a particular part of the character, or they may be part of the character itself (hierarchical cameras). Regular cameras distort what they see due to perspective, which affects animating, but a special type of camera, called an “Orthographic” camera has no distortion, (the Camera’s “Type” is selected on the “Attributes”tab of the camera’s Properties panel).

Face Camera

A particularly useful suggestion is to constrain a camera to always point at a character’s face. Have a separate window containing the Face Camera’s view, then no matter what the character is doing, you can easily animate the face.

Anticipation and Overshoot

If the viewer is watching another part of the screen, it takes about a fifth of a second (6 frames) after a new object moves before the viewer can refocus on that movement. Essentially, the viewer misses the beginning of the movement, so you, as the animator, should make a preliminary movement before the main movement (called “anticipation”) to attract the viewer’s attention. An example of anticipation is drawing back a character’s foot before a big kick; by the time the viewer is watching, the leg will begin its swing. Movement that is important to the narrative especially requires anticipation, however if the viewer is expecting the next action then the anticipating movement can be shortened.

For emphasis of important action (action critical to the narrative), the movement should “overshoot”. An example of overshoot is an arm stretching too far while pointing a finger, then snapping back.

Follow Through (Secondary Action)

We see the results of dynamics all around us in our everyday life. Secondary action due to dynamics is called “follow through” in animator parlance. For example, a character’s hair whipping around when it turns its head quickly is follow through. A rabbit’s ears, coat tails, hat plumes, cigarette smoke, etc., all require follow through. These parts of the character specifically cannot use the same keyframes as the rest of the character – they must be delayed several frames depending on the amount of follow through required.

Balance and Weight

As humans, we immediately recognize if someone loses their balance and is about to fall over. We intuitively know what is “balanced”. One of the things that makes animation funny is when it violates these balance expectations BUT, other than a storyline gag, characters MUST remain balanced! If a character appears off-balance, yet never falls down, the narrative is ruined because you have confused the viewer’s senses. Because of our expectations of what happens when something is out of balance, balance hints of upcoming action. If a character’s pelvis is shifted to the right, then its shoulders surely must be shifted to the left to maintain balance. If the character leans forward and points, he certainly should stretch his other arm behind him to even things up. You should exaggerate weight shifts to accentuate it.

A “balance” line is an imaginary vertical line that bisects the character between the two furthest spaced points of support. If the mass of the character is not evenly distributed on both sides of the line – the character appears unbalanced.

Traditional animators stress weight above all other aspects. Does the movement look as if it’s controlled by gravity? Is there appropriate follow through animation? Ease certainly contributes to the feeling of weight, but weight must be considered for each keyframe whether anything is moving or not.

Holds

The character must be balanced before holding.

A “look” off screen needs a half second of hold (15 frames) before cutting to what the character sees.

Allow two thirds of a second (20 frames) for each word of a title card for reading time.

The “freeze” position at the end of a fast throw is held for half a second (15 frames).

After a character stands up quickly, the pose is held for a third of a second (10 frames) before the next movement.

Avoid having all parts of a character start or stop at the same time.

Further Reading

For an exhausting explanation on the topic of timing, read:

“Timing for Animation” by Harold Whitaker and John Halas

(FOCAL PRESS LTD. ISBN 0 240 50871 8).

“How To Draw Animation” by Christopher Hart

(WATSON-GUPTILL PUB. ISBM 0 8230 2365 6)

Video Reference

Every animator needs guidelines in order to create an animation sequence. If you are a beginning animator working for a big animation company, your instructions will come from rough sketches of the movement of each character from a optimum angle called, “Action Boards”, and an “inbetweening scale”. The Action Boards are generally only the extreme poses of the character, with an indication as to where the inbetweens go. For small productions, where Actions Boards are a luxury, you can videotape a friend doing the proposed action. There are some inherent shortcomings in using video: actors simply cannot emulate a cartoon character. Consequently, a literal translation of the action from the video must be tempered by the nature of the character and the difference between the real world and the animation world that the character lives in. Obviously, the video actor cannot float, exaggerate movement, stretch and squeeze, or even effect some of the possible anatomical positions of a cartoon character. Therefore, the animator must be the liaison between the real world and the cartoon world.

When selecting the actor, consider his stature as compared to the stature of the cartoon character. Try to select people that are close enough to your cartoon character to effectively relate the actions. Find a location without many distractions in the background. (Try to find a blank wall to videotape in front of).

After completing the taping, the next step is to turn the video into an AVI or Quicktime movie. Upon close frame by frame examination of the movie, you will find that the body, head, and extremity movement do not all start at the same time. In some cases, the head turn may occur two or more frames before or after the body turn. The same may apply to the hands and arms, or feet. For example, you may find that as the actor shifts to the right, he automatically moves his leading foot first, plants it for balance, then the trailing foot follows as the weight of the actor shifts, and finally the pelvis completes the turn.

You want to have the movie opened in the background as a rotoscope for reference (though you usually will not be trying to physically align it to the character). The actual animation will move along very quickly. With a little practice it will be easy to determine the extremes as well as the ease frames. Note how few frames each movement actually takes, and how long it holds. Because the actor will not be in costume, you will have to determine the secondary actions after completing the primary action, which means that actions of a coattail or umbrella swinging will be animated later, and probably on different keyframes.

In all likelihood, the action will not represent exactly what you had in mind. Use the timeline to redistribute keyframes to better interpret the character’s nature. The timeline is especially useful for making changes in the ease. By adjusting the keyframes you can completely change the feeling of the action.

Animating Strategy

All animation is actually small bites from a much bigger project. Concentrate on short, one or two second (30 – 60 frame) sections at a time. Some animators start at the beginning of the story and continue through piece-by-piece to the end. Other animators like to pick and choose sections of the story in the order they think is important (or more fun). Whatever your preference, pick four or five consecutive Extreme keyframes and position the character (or characters), then go back and add Ease keyframes for each moving part. The Ease keyframes for each part do not have to be, and probably should not be, on the same frame – all parts of a character’s body do not usually travel at the same speed. This is where you can add your own flourishes as your skills increase, (when all Ease keyframes for all parts of the body occur on the same frame, it is symptomatic of a beginning animator).

Animation movement is divided into tenths of seconds. Most natural movement occurs over two tenths of a second which is 6 frames at 30 fps. Computer animation is described by the terms, “keyframes” and “inbetweens”. The animator establishes the “keyframes” and the computer calculates and performs most of the “inbetween frames”. You usually position at least one in-between between Extreme keyframes to control the ease between those Extremes.

THE STEPS

Pick [Open] from the “File” menu. Browse the CD to the Data\Tutorials\Character Animation folder and open “Character Animation.prj”.

The project file contains an Action window that contains “Jim”, as well as an animated rotoscope that will be used as reference footage. The animated rotoscope was created by importing from a video camera and converting it to a QuickTime movie.

Before you begin animating, get a feel for the action you are about to create by double-clicking the “Reference Footage” item in the “Images” folder in the Project Workspace tree and play it through several times.

Keyframes for different parts of a character often are not the same – arms, hips, or head may have differing extremes. The skeletal hierarchical construction makes it essential to establish the hips’ extremes first, then the torso’s, then the arms’ and legs’, and finally the hands and head.

Figure 1

When you’re ready to begin, click inside the Action window. Manipulate the “Jim” model until he is positioned similarly to the actor in the reference footage, as shown in figure 1. If necessary, change to a side view to make sure that you have selected the correct bone in the model before moving it.

Continue to step through the frames until the actor’s foot goes back down, around frame 5. Reposition Jim to match the position. Work through the taps of the right foot that the actor performs using the reference footage for timing.

Figure 2

From viewing the footage beforehand, you know that the first part of this action involves the character tapping his right foot. Step through the frames until the actor in the footage reaches an extreme of motion. This will occur at frame 3, where the actor has raised his foot. Reposition Jim to match, as shown in figure 2.

Figure 3

After the right foot taps have completed, the actor shifts his weight and turns to his left at frame 26. Position Jim to match the reference footage, as shown in figure 3.

The actor now taps his left foot several times. Using the reference footage for timing, create this part of the action. Remember to use the frames where the extremes of the motion being performed are reached (with the foot up or down).

The foot tapping process is repeated back to the right foot, then back to the left foot. A commonly used too for repeating actions is Copy/Paste. Oftentimes a single keyframe position is repeated several times, so simply copy the existing occurrence and paste when needed. Copy/Paste has many extensive possibilities so it requires robust controls. For example, it may just be the arm’s animation you are interested in, or maybe only the skeletal portion of the action (and not the muscle motion). The Key Mode toolbar, along with careful hierarchical selection of bones, allows you to make these distinctions. You can copy a whole body’s action, or just the selected bone’s, with or without constraints, etc., depending on the buttons selected on the Key Mode toolbar.

The Key Mode button selections can be different between the copy and paste operations. For example, you may copy a whole body’s action but choose to paste only onto an arm. Whatever is in the Copy buffer can be pasted many times, with different Key modes selected each time!

When part of the animation is intended to remain stationary for some time (a “hold”), copy from the beginning of the hold and paste on the frame the hold ends.

In this example, the “foot up” keyframes can be copied and pasted, as can the “foot down” keyframes.

Figure 4

At frame 115, the actor turns to face the camera. Figure 4 shows this keyframe.

Figure 5

At frame 139, the actor reaches up to take and tip his hat to the viewer. To make Jim grab his hat on frame 139, click on the bone in the hat, turn on Compensate mode, then right-click (Command-click on a Mac) the bone and pick [New Constraint][Translate to]. Use the picker tool to select the hand bone. This will cause the hat to stay with the hand as Jim lowers his arm to tip his hat to the viewer. Figure 5 shows the keyframe created at frame 139.

Figure 6

As the tip of the hat is performed, the actor’s arm comes down, reaches its extreme, then is raised back up to place the hat back on his head. Frame 152 is the extreme of the hat tip motion where the arm is extended with the hat in the hand. Figure 6 shows this keyframe.

Figure 7

Figure 7 shows frame 162, where the character places the hat back on his head. This is done by selecting the “Translate To” constraint on the hat bone and changing the Percentage field on the Properties panel to “0”.

Continue through the frames to frame 182. The actor has moved back into a neutral position, facing the camera. On frame 186, he flips the umbrella into the air by raising his arm quickly. At frame 200, the actor has lowered his arm, and the umbrella is placed vertically on the ground next to the model.

At frame 223, the actor has shifted his weight back to his right, and is looking down at the umbrella model standing on the floor beside him. At frame 236, the actor has straightened his weight out, and placed his hands on his hips while still staring at the umbrella model.

Figure 8

Frame 275 shows the actor reaching out with his left hand to grasp the handle of the umbrella (use the “Translate to” constraint with Compensate mode on for this), he has also turned stare at the camera, and crossed his legs, as shown in figure 8.

By frame 321, the actor falls down. There is an important concept to be learned in this deceptively simple motion. Notice how when the actor begins to fall, he moves slowly then accelerates as the fall continues.

This type of movement is referred to as Ease, which either starts fast and slows or starts slow and ends fast. The ease keyframe should be two-thirds of the way from the fast extreme keyframe.

Timing is extremely important in creating the illusion of life. Extremes and the placement of inbetweens determine the force and effectiveness of the animation and whether it is believable or not. Because Animation:Master uses cubic channels (curvy graphs) to define motion, some amount of ease is automatic, but you the animator will be expected to add the exaggerated ease that typifies animated storytelling.

For the falling motion, start slow and end fast. Go to frame 306, which is thirty one frames from the start of the fall keyframe. Jim’s body will be at an in-between orientation generated by the computer. You want to lean the body more than the computer has. When you do this, a new keyframe will be created. The more you lean the body at this Ease keyframe, the more exaggerated the ease will be overall. Ease in-betweens are not spaced evenly. The slow to fast ease has closely spaced in-betweens when the fall starts, then larger spacing as the fall ends.

While you’re refining your actions, remember that every keyframe for every icon in the Project Workspace tree is shown in the adjacent timeline, (Initially, the timeline is barely visible, so you will need to drag it wider when you need it). When a folder is expanded in the Project Workspace tree (the “+” icon is clicked), the timeline expands to show the keyframes of the newly visible icons in the folder. For example, if the “Hand” bone in a character’s hierarchy is expanded, the timeline will then include keyframes for the fingers, (adjacent to the finger bones).

The timeline is of paramount importance to animation because it allows you to simultaneously move or delete a group of keyframes. For example, the upper and lower arm bones may both have a keyframe on “10” that needs to be moved to “11”: group the two keyframe markers for these two bones (drag a bounding box around them or <Shift> click them into a group), then drag the group to “11”. Or, as often happens during the creation of an animation, you have inadvertently created a keyframe that you do not want or need, simply group the offending keyframe markers of the appropriate bones and pick [EDIT][DELETE] from the main menu, (or <Delete> on the keyboard).

So far we have been concerned with the overall movement of our character. Once you are satisfied that the Extreme and Ease keyframes are complete, then you can go back and create what is called “Finish Action”. Finish Action is the eye blinks, lip-sync, hand gestures, and other ancillary movement required to bring the character to life.

EYES

The most important finish actions are on the eyes. When you listen to another person speak, you usually look into their eyes, and they look into yours. Whenever two characters converse in an animation, they must appear to be looking into each other’s eyes. When the camera cuts to a close-up, you are assuming the role of listener, so the character’s eyes should stare straight out of the screen as if they were looking at you. If you violate this simple rule of eye contact, your animation’s narrative will break down.

The reason we as humans look into another person’s eyes when we speak derives from long ago survival tactics. A speaker’s internal feelings are usually expressed in raised eyebrows for surprise, or lowered eyebrows for anger. Quizzical speakers have one eyebrow raised with the other lowered, while suspicion causes the eyelids to half close. These simple eye formulas are required for successful storytelling.

Eyes also are the visual clue that an animated character is “alive”. If an actor never blinked, you would think him a robot. Alive things blink at least once every two seconds, sometimes more if they are moving (eye-blinks are the physical response to wind pressure, dust, and changing light levels). So, if a character turns its head – blink. If a character bends over – blink. Even if a character is looking intently through a close-up, blink randomly every second or so.

Sometimes the blinks are fast but every time the character is supposed to be thinking, the eyelids stay in the halfway position for several frames. Surprised character initially flutter their eyelids. Sleepy characters have half-closed eyelids all the time until they abruptly awaken.

Keep in mind that many things in animation repeat over and over, like eye blinks and mouth shapes. These kind of movements can be controlled by poses. A pose is a feature in Animation:Master that allows a single slider to control a complicated but common movement. A pose usually has a middle “rest” position, and two opposite “extremes”, like a frown at one end, a calm mouth in the middle, and a smile at the other end. Depending on the slider’s position along this scale, the face is in different degrees of frowning or smiling. Eye blinks only vary between open and closed, where a half-lidded stare has the slider halfway between the middle (eyes open) and one end (eyes shut). Usually, each eyebrow has its own pose slider because many reactions are not symmetrical.