A clear path
from frame to film.
An AI motion pipeline built around visual direction, controlled keyframes and careful finishing.
- Focus
- Motion & scene continuity
- Inputs
- Storyboard & reference frames
- Review
- Character, camera & timing
- Output
- Animation & finished edits
Character motion · Startup screens, assistant states and scene continuity
- My contribution
- I connected the visual brief with character behavior and the intended motion application.
- Key decisions
- Set behavior, camera and timing before iteration; compare alternatives against continuity and screen-context criteria.
- Delivery & outcome
- Dongfeng/Huawei Panda startup clips and a campfire sequence, six M-HERO Model B Lantern Lion studies, and X700 cat assistant-state motion.
Preserve the character.
Make the action purposeful.
Generated movement can change a character, shift the camera or introduce unwanted details. The challenge was to produce motion that preserved identity and scene continuity while fitting the startup screen, assistant state or story it needed to communicate.
Define the intended behavior.
Review against the same criteria.
I connected the visual brief with the behavior, camera and timing each sequence needed. Controlled reference frames made the intent clear before generation and provided a common basis for reviewing the result.
I used continuity, character identity, loops and screen context as review criteria. This made it possible to compare alternatives and refine the sequence around what the application needed to communicate.
- 01Storyboard
- 02Lock keyframes
- 03Generate motion
- 04Composite & deliver
Intent before iteration.
Behavior, camera and timing establish what each sequence needs to communicate.
A common review basis.
Continuity, character identity and screen context guide comparison and refinement.
Motion sequences
with a shared production method.
The outputs include Dongfeng/Huawei Panda startup animations and a ten-second campfire sequence, six M-HERO Model B Lantern Lion performance studies, and the X700 cat voice-assistant-state sequence.
The examples demonstrate character continuity across actions and contexts. The videos bring together the completed sequences and motion studies in their respective project groups.
Panda motion.
Startup screens & scenes.
Dongfeng/Huawei Panda motion spans startup greetings, driving, playful actions and the campfire scene. All examples carry the character into distinct screen experiences.
Good morning.
A greeting animation with a coffee cup.
Ready to drive.
A driving vignette with character motion and on-screen time.
Swing.
A playful swing animation paired with on-screen time.
Soccer.
Ball play and character motion paired with on-screen time.
Panda campfire
scene.
The Dongfeng/Huawei Panda campfire example connects a stable starting composition with character actions, scene continuity and a finished ten-second animation.
Lantern Lion.
Motion & performance.
Six motion studies bring the M-HERO Model B Lantern Lion into distinct poses and performances. Head movement, body language and timing extend the character’s visual identity into animation.
Grounded power stance.
A planted stance, a head turn and a forward-facing finish.
Festival lion dance.
Energetic steps, turns and raised poses.
Chinese lion dance.
A performance study in posture, head movement and expression.
Competition-style performance.
A second performance variation with changes in pose and direction.
Hero stance.
An upright character pose with controlled head and body movement.
Traditional lion dance.
A longer performance study with turns, lowered poses and expressive movement.
Voice assistant
stages.
The X700 cat character and on-screen face move through voice-assistant states. The two views make the relationship between character motion and interface feedback visible.
Voice assistant stages.
The X700 cat character and on-screen face feedback shown together.