A character is waiting. Their shoulders rise, their weapon shifts, they look left, and they adjust their clothing. Then the whole sequence starts again.
There is plenty of movement. But what are they actually doing?
For a character to feel specific, it may help to design not just an idle animation, but the circumstances in which that animation belongs. The useful question is less “How do we stop this character looking frozen?” and more “What can this character notice or do while nothing urgent is happening?”
Separate the loop from the little performance
Here is a practical way to divide the work. This is a proposed design structure, not a universal animation standard.
The base loop maintains the stance: breathing, balance, a relaxed grip. It needs to tolerate repetition because it may remain visible for a long time.
The occasional fidget is a small event: checking a strap, brushing something from a sleeve, or stretching a hand. It has a beginning and an end. Putting it inside a short breathing loop also commits you to its repetition rate.
The contextual response has an external reason: turning toward a nearby sound or watching someone approach. Its usefulness depends on whether that reason is still present.
These can share animation assets, but they should not automatically share scheduling rules.
Christopher Totten describes adding fidgets and small moving details to enemy and NPC animations in his firsthand account of work on Little Nemo and the Nightmare Fiends. That is evidence of a production approach to visual interest, not evidence that every character needs more movement or that fidgeting causes attachment. Source: “12 principles for game animation,” Secondary Action section
Give a gesture a small contract
Before adding another clip, write down four things:
- Eligibility: When does this gesture make sense?
- Attention: Is the character attending to an object, a person, their own body, or nothing in particular?
- Suppression: What should prevent it from starting?
- Repetition: What makes it eligible again after it finishes?
For example, a fictional guard might check a loose glove while waiting in a safe area. The gesture is suppressed while aiming, speaking, or tracking an immediate threat. It becomes eligible again after a waiting interval, rather than at every loop boundary.
That does not make the guard intelligent. It makes one authored performance less likely to contradict the scene.
Random timing can reduce an obvious rhythm, but randomness alone cannot supply a reason. Randomly inspecting a glove while someone is addressing the guard can still be the wrong performance.
Make the personality visible through selection
Imagine three characters with the same quiet standing loop.
One keeps their gaze on the doorway. Another occasionally examines the room's decorations. A third remains almost motionless until a noise interrupts them.
Those are different choices about attention, not simply different quantities of animation. They are hypotheses about how to communicate watchfulness, curiosity, or reserve. Context can change the reading, so the labels should not be treated as guaranteed outcomes.
Riot's Teemo update is a useful reminder that expression has to fit its viewing conditions. The animation team describes reinforcing his carefree explorer identity while preserving his characteristic waddle. Separately, they made his emotes short and punchy because his quick transition into stealth made them harder to see. These are emotes and locomotion decisions, not documentation of an idle scheduler. Source: Riot's Teemo ASU development breakdown, Animation section
The transferable question is: how much of this performance will actually be visible, in this situation?
Review the waiting, not just the clip
An isolated animation preview cannot show whether a gesture happens too often or at an inappropriate moment.
Put the character into a representative scene. Let them wait longer than a single loop. Then introduce a conversation, a moving object, or the kind of interruption the game actually supports. Watch at the real camera distance.
Look for specific problems: repeated gestures becoming synchronized, attention directed toward something that has disappeared, a tiny face movement that cannot be seen, or a fidget competing with important information elsewhere on screen.
A small playtest can ask what viewers thought the character was attending to. That is a more focused question than asking whether the animation made the character “more alive.” It still does not establish long-term liking or attachment.
Stillness is also an option
This approach is not an argument for a complex behavior system on every NPC. A small sprite may communicate enough through its stance. A tense scene may benefit from restraint. A stylized comic character may deliberately repeat a conspicuous gesture.
The aim is to spend animation where it expresses something useful. Sometimes that means an extra clip. Sometimes it means playing an existing clip less often. Sometimes it means letting the character wait.
This article was researched, drafted, and edited by an AI assistant. The suggested design structure and review exercise are proposals, not reported player experiments.
Top comments (0)