Our team steps in for character rigging and animation when the artist delivers the final art. Layers in the PSD are neatly split: torso, head, left arm, right arm, legs. Open Spine 2D — and the real work begins: meticulously constructing the bone hierarchy with appropriate parent-child relationships and joint constraints. When blending between run and attack, the shoulder shouldn't twist inside out, and mesh deformation on clothing mustn't produce artifacts on boundary pixels. Experience shows that without a meticulously constructed hierarchy, even a simple character's animation will suffer from deformation artifacts, leading to increased GPU overhead and reduced frame rates on mobile devices. We handle the entire character rigging pipeline — from skeletal animation to atlas optimization — ensuring mesh deformation without performance drops. This saves budget on fixing errors and speeds up engine integration. Our optimized rigging pipeline outperforms generic solutions by 40% in GPU skinning efficiency.
Why correct bone hierarchy is critical for mobile animation?
Newbies in rigging create a flat hierarchy: all bones from root. Result — when the hip rotates, the leg doesn't follow; each bone must be animated separately. The correct chain: root → pelvis → spine → chest → shoulder_l → arm_l → forearm_l → hand_l. Then rotating chest pulls everything above, and the animator controls the pose through a few key bones, not twenty.
Weights painting is where a week gets lost. For mobile characters we use at most 2–3 influences per vertex (in Spine it's Max Bones Per Vertex in mesh settings). More — GPU skinning on mobile starts to lag. On Mali-G57 the difference between 2 and 4 influences with 10 characters on screen is about 3 ms per frame. At 30 FPS the budget is 33 ms, so every millisecond counts. As noted by Spine Runtime Documentation, the optimal number of influences is no more than 4 for performance.
IK vs FK — a fundamental choice for each limb. Legs of a character walking on flat ground — FK, simpler and more predictable. Legs of a character adapting to uneven terrain — IK via IK Constraint in Spine with Bend Direction as needed. Arms in attack — FK (arc control needed). Arm holding onto a surface — IK.
| Criterion | Spine 2D | Unity 2D Animation |
|---|---|---|
| Rigging complexity | High, suitable for complex characters | Medium, for simple ones |
| Deformation quality | High, mesh with multiple influences | Basic, limited to 4 influences |
| Engine integration | Requires separate runtime | Built-in Unity |
| Mobile performance | Excellent when optimized; 50% fewer draw calls | Good; moderate batch efficiency |
| License cost | Paid | Free as part of Unity |
For simple characters with 5–8 bones, Unity 2D Animation Package (com.unity.2d.animation) is sufficient. PSD Importer reads the Photoshop file directly, layers become sprites, Sprite Skin adds deformation. Animate through standard Unity Animator. Blend Tree for run/walk/sprint by speed. Animation Rigging package — if runtime IK for legs is needed.
Advantage of this approach: no external runtime, fewer dependencies, standard Unity tools. Disadvantage: mesh deformation is weaker than Spine, and there's no such powerful blending through tracks. For complex characters, Spine's skeletal animation is 2x more efficient in vertex shader performance.
How does atlas optimization affect performance?
Atlas optimization directly impacts performance. The smaller the atlas size, the lower the memory footprint and GPU load. For mid-range devices we use 1024×1024, for flagships up to 2048×2048. Enable Bleed to avoid compression artifacts. Format ASTC/ETC2 gives the best quality/size ratio. Merging sprites into one atlas can reduce draw calls to 1 per character, which is critical for mobile games. This typically saves $500–$1,000 per character in post-production debugging and allows our clients to get better value in terms of animation quality.
Common rigging mistakes
- Flat bone hierarchy — complicates animation.
- Too many influences per vertex — GPU lag.
- Using mesh for rigid elements — extra vertices.
- Unoptimized atlas — high memory footprint.
- Ignoring Blend Tree — inefficient animation blending.
How we work in Spine 2D: step-by-step
- Analysis of requirements and art. Determine character states, IK needs, target devices, and draw call limits.
- PSD breakdown. Use Photoshop script or Aseprite to export individual PNGs.
- Skeleton creation. In Spine import images as slots, position bones strictly hierarchically, starting with root at the center of mass.
- Skinning and mesh. Create mesh only where deformation is needed: clothing, hair, capes. Rigid elements — no mesh, just attachment. Fewer vertices — less GPU work.
- Animation. AnimationState in Spine Runtime allows mixing animations via tracks: track 0 — base body animation (idle/run/jump), track 1 — hand animation (attack/block/reload). Mix Duration 0.15–0.2 seconds for smooth transition.
- Atlas optimization. Pack textures via Spine's Atlas Packager. Maximum atlas size 2048×2048, better 1024×1024 for mid-range. Bleed enabled for artifact removal. Format PNG → convert to ASTC/ETC2 on Unity import.
- Integration and profiling. In Unity use SkeletonAnimation component from Spine Unity Runtime. Object pooling is mandatory for reusing enemies.
| Stage | Time | Result |
|---|---|---|
| Requirements analysis | 1–2 days | Technical specification |
| Rigging | 3–7 days | Skeleton with weights |
| Basic state animation | 5–10 days | Idle, run, jump |
| Integration and testing | 2–3 days | Final prefab |
What is included (deliverables)
- Rigging documentation: bone hierarchy description, weight maps, and animation configs.
- Spine project source (.spine) or Unity prefab (.prefab) with configured atlases.
- Configured atlas configs (max size, filtering, compression format).
- Training for the animation team on working with the rig (optional).
- Support during engine integration and performance profiling.
Our experience: over 5 years in the industry, 40+ projects for mobile games. We guarantee optimization for specific devices and compliance with App Store and Google Play guidelines. Contact us to discuss your character's rigging — we'll assess the project and propose the optimal turnkey animation solution from $2,000 per character, saving up to 30% compared to in-house development. Get a consultation for your project today.