Month: June 2025

Project Conclusion and Reflection

This project has been a rigorous exploration of combining traditional artistic principles with modern 3D animation techniques, pushing the boundaries of painterly art styles within a digital workflow. Throughout, I experimented extensively with tools like Blender, Cinema 4D, Marvelous Designer, Cascadeur, and AI-assisted modeling platforms such as Meshy to develop a cohesive visual narrative and a strong stylistic identity.

Key Learnings and Achievements

• Painterly Style Integration:
  The pursuit of a painterly aesthetic influenced every major decision—from shader design and brushstroke layering to the texturing and animation of characters and environments. Through trials with brush filters, gradient quantization, and layered opacity effects, I successfully created a dynamic, expressive visual language that balances abstraction with readability.
• Character and Environmental Design:
  I designed a cast of characters whose visual identities, clothing, and elemental powers reflect their personalities and narrative roles. Although I recognize that earlier designs relied heavily on baggy clothing and lacked sufficient variation, this informed my intention to revisit and deepen these designs in future iterations. The environments—especially the mountainous terrain—were crafted with both storytelling and technical feasibility in mind, balancing dramatic forms with flat planes to support fluid animation.
• Dynamic Effects and Simulation:
  Experiments with procedural effects like ice freezing, lava flows, wind attacks, and fire damage demonstrated my commitment to integrating visual effects that serve narrative impact without sacrificing stylistic coherence. While some experiments, such as complex Alembic imports and geometry node texturing, faced technical challenges, others—like the dynamic lava simulation and the plume of smoke—successfully merged procedural and artistic methods.
• Cross-Software Workflows and Technical Challenges:
  Moving assets and animations between programs revealed bottlenecks and prompted adaptations in pipeline strategies. The use of AI-assisted tools like Cascadeur proved invaluable in overcoming traditional retargeting limitations. Additionally, I learned the importance of file formats (e.g., FBX vs. Alembic) for maintaining flexibility during editing and compositing.
• Animation and Timing:
  Throughout the project, I iteratively refined pacing and animation flow, particularly in synchronizing movements with music and adjusting complex sequences for clarity and impact. This process highlighted the delicate balance between creative ambition and technical constraints.

Reflections and Future Directions

While the project achieved many artistic and technical goals, it also highlighted areas for growth:

• Design Variation:
  Future projects will benefit from greater diversity in character clothing and silhouettes to avoid visual redundancy and enhance narrative clarity.
• Technical Pipeline Optimization:
  Improving interoperability between software and embracing real-time painterly rendering will streamline the workflow and reduce time spent troubleshooting.
• Enhanced Shader Development:
  Further refinement of painterly shaders, including motion-aware brush textures and more responsive lighting, will deepen the immersive quality of the visuals.
• Expanded Use of Effects:
  Building on promising trials with trails, ice, and elemental attacks, I plan to explore more nuanced, interactive environmental effects that evolve with the story.

Final Thoughts

This project has been an invaluable journey through the intersection of traditional art sensibilities and cutting-edge 3D technology. The balance between painterly expression and technical execution shaped not only the final animation but also my evolving approach to storytelling and design. Moving forward, I am committed to refining these methods and pushing creative boundaries to realize even more compelling visual narratives.

Environmental Fire Damage Concept – Geometry Nodes Experimentation

This section details an ambitious environmental effect initially designed to give the character a strong narrative and visual personality—setting the ground ablaze as they run and causing significant environmental destruction.

• Concept Overview:
  • The idea was to have fire spread organically beneath the character’s feet, visually communicating raw power and impact on the world.
• Technical Approach:
  • Used Blender’s Geometry Nodes to procedurally remove and delete ground geometry dynamically.
  • The geometry deletion was driven by a Voronoi pattern combined with Perlin noise:
    • Perlin noise was chosen for its organic, natural randomness, perfectly simulating how fire might spread unevenly over terrain.
    • Voronoi helped define the clustered, fractal-like fragmentation of burnt earth.
• 2D Fire Sprites:
  • Since the painterly paint filter did not register planes (P-L-A-N-E-S) used to generate 2D fire sprites, I created custom hand-animated fire effects to overlay and complement the geometry destruction.
  • These sprites were designed to grow out of the ground as flames.
• Outcome:
  • Despite the scale and innovation of this effect, it was ultimately not used in the final animation.
  • The complexity and resource demands, along with narrative shifts, led me to explore other, more focused visual storytelling methods.

---

Reflection

This experiment expanded my understanding of procedural environment destruction and highlighted the challenges of integrating 2D effects within a painterly 3D style. Although unused, the techniques and lessons learned will inform future projects aiming for large-scale dynamic environments.

(Quick note that I also hand animated 2d smoke too)

Flame Burst During Dodge – Voronoi Fracture Experiment

In this clip, I explored how to make the character burst into flames mid-dodge using an Alembic mesh generated in Cinema 4D with a Voronoi Fracture.

• Method:
  • Created a fractured mesh to simulate the character’s body breaking apart into fiery shards.
  • Attempted to have the flames move back toward the character, creating a dynamic flare effect.
• Technical Challenges:
  • The fire shards moved too far away, losing connection with the character.
  • Used Cinema 4D’s Push Apart feature combined with a plane effector and falloff just below ground level to prevent fire sinking underground.
  • Despite this, the flames’ movement appeared uncontrolled and unnatural, lacking the smooth float-up-and-settle behavior I wanted.
  • The flames moved in every direction without cohesive flow.
• Final Solution:
  • Abandoned the complex fractured simulation for a simpler, more controlled effect.
  • Made the character’s entire body glow red, using a Voronoi texture to gradually transition between normal and fiery states.
  • This approach achieved a clean, stylized visual that matched the painterly aesthetic and narrative needs.

Trail Effect Experiment Using Cuboids and Volume Meshing

In this clip, I experimented with creating dynamic trail effects that interact with the painterly filter by leveraging volumetric meshing and procedural cloning.

• Initial Setup:
  • In Cinema 4D, I attached a mesh to the character’s foot to serve as the trail emitter.
  • Used a Tracer object to follow select points on this mesh, simplifying the trail to about four or five paths for performance and clarity.
• Spline and Sweep:
  • The traced points were used to generate circular splines, which were then swept with Sweep NURBS to create volumetric ribbons.
• Volume Meshing:
  • These swept splines were input into a Volume Builder and Volume Mesher, converting the ribbons into volumetric shapes.
  • This gave the trails a solid, chunkier presence rather than just flat strips.
• Cloner Application:
  • Using a Cloner object in Volume Object mode, I cloned multiple flattened cuboids onto the volumetric trail shape.
  • The cuboids’ simplified geometry helps trigger the painterly paint filter effectively by providing polygon density and shape variation.
• Export and Result:
  • The volumetric trails were baked into an Alembic cache and imported into Blender for final compositing and rendering.
  • This workflow resulted in a visually interesting, painterly trail effect that follows the character’s motion dynamically.

Animation and Dynamic Clothing Experimentation

In this clip, I experimented with synchronizing the character’s movement to the rhythm of the music while exploring dynamic cloth simulation and animation flow.

• Character Design Iteration:
  • An earlier version of the character featured handcuffs on his wrists.
  • Due to technical issues and visual concerns, I ultimately removed the cuffs from the design.
• Technical Challenges:
  • The sequence was baked as an Alembic, which introduced limitations that affected further edits and retargeting.
  • Had it been an FBX file, these issues would likely have been avoided.
• Animation Flow:
  • Initially, the sequence was fully animated but the pacing was too fast and visually confusing.
  • To address this, I slowed the footage down significantly (by half or three-quarters), but this caused audio quality degradation and synchronization problems.
• Final Editing Decisions:
  • Due to these constraints, I removed approximately 70-80% of the sequence to streamline the animation and maintain coherence.
  • The remaining movement focuses on key moments that align better with the music and narrative pacing.
• Reflection:
  • This process underscored the importance of choosing appropriate file formats early in the pipeline and pacing animation to balance quality and timing.

Painterly Shader Experiment – Character Walk Cycle

In this experiment, I focused on how a character looks while walking with the painterly shader/filter applied. The goal was to explore how different brush strokes interact dynamically with the face and body to maximize the stylized effect.

• Trial and Error Process:
  • Spent hours testing various brush stroke patterns, densities, and directions.
  • Experimented with how strokes respond to motion, light, and facial features to enhance expressiveness.
• Interaction with Geometry Nodes:
  • Combined shader experimentation with in-depth study of Geometry Nodes in Blender.
  • This dual approach helped refine how the painterly effect integrates with mesh deformation and animation.
• Impact on Final Result:
  • The extensive experimentation informed key decisions in my final shader setup.
  • Helped me balance between clarity of form and painterly abstraction, essential for the project’s aesthetic.

Two-Dimensional Lightning Animation – Early Concept Work

This screenshot showcases the early 2D lightning effect I created during the initial stages of the project. Animated in Adobe Animate, this exercise was primarily a practice run to understand timing, flicker, and movement dynamics of electrical effects.

• Purpose:
  • To explore how lightning can be stylized in 2D before transitioning into 3D.
  • Focused on creating dynamic, flickering energy bolts with believable motion and impact.
• Benefits to Project:
  • This foundational work proved extremely valuable for my Advanced Body Mechanics project.
  • The 2D lightning was later translated into 3D by attaching animated splines to characters’ feet, bridging the gap between flat effects and volumetric animation.
• Impact:
  • Helped refine visual language and timing for elemental effects.
  • Established a basis for integrating traditional animation principles into modern 3D workflows.

Advanced Marvelous Designer Study and Workflow Refinement

This image reflects my continued exploration and refinement of Marvelous Designer. Over time, I improved my workflow by learning to adapt and customize existing clothing patterns.

• Learning Progression:
  • Initially, I modified pre-made patterns to fit my character models and project needs.
  • With practice, I progressed to creating original patterns, significantly increasing my creative control.
• Creative Freedom:
  • Designing from scratch allowed me to experiment with new shapes, fits, and fabric behaviors.
  • I gained insight into fabric draping, layering, and simulation performance.
• Reflection on Early Work:
  • The clothing created earlier in the project, while functional, is subpar compared to what I am capable of now.
  • My current skills enable more polished and dynamic garment construction.
• Future Direction:
  • I plan to revisit earlier character costumes to upgrade them with this improved skill set.

Track Matte Experiment for Focus and Saturation Control

In this screenshot, I experimented with track mattes to control saturation dynamically between the character and the environment.

• Layer Setup:
  • Two primary tracks: a green track (main footage) and a purple track (adjustment layer).
  • The purple track had a track matte key applied to it, functioning as a mask.
• Mask Details:
  • The mask is a black-and-white silhouette of the character:
    • Black represents the environment.
    • White represents the character’s silhouette.
  • By inverting the mask and cross-dissolving between the normal and inverted versions, I created smooth transitions in focus.
• Effect on Saturation:
  • Applied a saturation reduction to the adjustment layer controlled by the mask.
  • This allowed me to desaturate the environment while keeping the character vibrant during fast movement.
  • Conversely, when the character slows, the environment regains saturation, drawing viewer attention back.
• Purpose:
  • This subtle technique helps guide viewer focus dynamically based on narrative pacing and character motion.

Wind Attack Experiment with Painterly Filter

In this screenshot, I experimented with creating a wind attack effect that interacted beautifully with the painterly shader.

• Creation Process:
  • Started with a sphere as the base shape.
  • Applied a displacer modifier to twist and deform the geometry, creating dynamic holes and organic motion.
  • Removed some geometry selectively to enhance the swirling, airy feel.
  • Applied the painterly shader to maintain the project’s stylistic consistency.
  • Added a spin modifier to rotate and animate the sphere, simulating a whirling wind ball.
• Stylistic Details:
  • Adjusted the brush stroke settings in the shader by reducing brush width and increasing stroke length.
  • This produced long, continuous strokes that effectively communicated flowing wind motion.
• Inspiration & Influence:
  • The effect draws visual inspiration from “Avatar: The Last Airbender” wind bending and Valorant’s wind-themed abilities.
• Final Decision:
  • Despite the promising look, I chose not to use this effect in the final project, as no characters currently use wind powers.
  • I was unable to find a suitable narrative or gameplay use case for this effect.

In this stage, I attempted to weight paint an Alembic file to control environmental texturing. The original intention was to have grass appear on the lower elevations and rock textures on the upper elevations of the terrain. This layering would create natural variety and allow for more dynamic procedural texturing of the environment.

• Unfortunately, I discovered that weight painting isn’t supported directly on Alembic caches, which halted this approach.
• Despite this limitation, the plan reflects a desire to blend procedural animation techniques with environmental storytelling — enriching the landscape with organic, elevation-based detail.
• This method would have provided an efficient way to animate textural transitions dynamically, responding to terrain shape without manual mapping.

Although this workflow was unsuccessful, it helped clarify the constraints of Alembic workflows and guided me toward alternative approaches for dynamic environment texturing within Blender and Cinema 4D.

In this experiment, I animated the character Ardil and aimed to enhance the animation by creating dynamic trails. These trails were generated in Cinema 4D using procedural tools and then imported into Blender for further compositing and rendering.

• Goal:
  To evaluate how well visual trail effects created in Cinema 4D could be integrated into a Blender pipeline while maintaining animation fidelity.
• Technical Challenge:
  The process involved exporting and importing Alembic caches between software packages. Unfortunately, this proved problematic:
  • Alembic data often breaks or corrupts when transferred between different programs.
  • Despite investing hours trying to resolve the issues and replicate the effect, I was unable to achieve a fully stable and usable result.
• Outcome:
  Although the current implementation did not succeed, the effect itself was visually compelling and showed strong potential for enhancing action sequences and accentuating character motion.
• Future Direction:
  I’m keen to continue exploring alternative methods for exporting trail effects and optimizing cross-software workflows, perhaps experimenting with other cache formats or real-time plugins to maintain the integrity of complex visual effects.

Retargeting Motion Capture – Early Attempts and Challenges

In this image, I’m showcasing my initial attempts at retargeting motion capture data using Autodesk Maya. This process involved adapting mocap animations onto my character rig.

• Workflow:
  • Imported mocap data into Maya.
  • Tried to align and retarget the animation to fit my custom character rig.
• Challenges:
  • The process was complex and unreliable.
  • The retargeting barely worked, with significant issues in joint alignment and movement fidelity.
  • Industry-standard software proved difficult to navigate and limited by technical hurdles.
• Solution:
  Due to these challenges, I transitioned to using Cascadeur, an AI-assisted tool that offered a more reliable and efficient way to retarget and clean up animations.
• Reflection:
  This experience highlights how emerging AI tools can sometimes outperform traditional software in specific animation workflows, especially when dealing with complex mocap retargeting.

Alembic Texturing Experiment Using Geometry Nodes

In this attempt, I tried to apply textures to an Alembic mesh using Blender’s Geometry Nodes system. The goal was to dynamically control the texturing process through procedural nodes, enabling more flexible and non-destructive workflows.

• Initial Outcome:
  At first, the technique seemed promising—textures appeared to map correctly and the procedural setup was flexible.
• Technical Limitation:
  Unfortunately, the structure of the Alembic mesh was not suitable for this approach.
  • The topology and data format of the Alembic caused issues with node-based texturing workflows.
  • Similar experiments on different meshes or projects yielded inconsistent results.
• Conclusion:
  Despite the potential of using Geometry Nodes for procedural texturing, this specific experiment did not yield successful results. Further research and alternative approaches may be necessary to overcome these limitations.

Cascadeur Retargeting in Action

In this screenshot, I am leveraging Cascadeur to its full potential during motion capture retargeting. After applying for a trial version, I gained the ability to import and export .fbx files, which greatly improved my workflow flexibility.

• Process Overview:
  • Importing mocap data into Cascadeur for cleanup and retargeting.
  • Exporting animations to multiple rigs across different software packages.
• Challenges Encountered:
  • Some bones experienced severe deformation, requiring manual correction.
  • Transferring animations between meshes necessitated significant hands-on adjustment.
• Outcome:
  Despite the hurdles, Cascadeur’s AI-assisted tools proved extremely beneficial.
  • It enabled me to reuse mocap data across varied character rigs efficiently.
  • The overall time and effort saved were substantial compared to previous attempts.

Marvelous Designer Exploration and Cloth Simulation

These two images illustrate my progression in learning Marvelous Designer and integrating its outputs into Blender for real-time simulation.

• Cloak Simulation:
  • Created a detailed cloak pattern in Marvelous Designer.
  • Simulated cloth physics within Marvelous Designer to achieve realistic draping and movement.
  • Exported the cloth mesh into Blender for real-time simulation alongside character animation.
• Weight Painting Workflow:
  • Used Blender’s weight painting tools (see screenshot below) to precisely control mesh influence.
  • The top of the cloak mesh remains fully attached to the character’s shoulders to prevent unnatural detachment.
  • The rest of the cloak is free to respond dynamically to physics, allowing for believable fabric motion and flow.
• Benefits:
  • This approach strikes a balance between character control and physical realism.
  • Helps maintain animation integrity while enhancing visual fidelity through cloth dynamics.

Gradient Quantization for Motion Path Stylization

In this screenshot, I experimented with quantizing a gradient, a technique that proved very beneficial for visualizing the motion path of Ardil’s spear.

• Purpose:
  • To create a stylized, posterized effect along the spear’s trajectory.
  • This approach breaks smooth gradients into stepped color bands, adding graphic clarity and emphasis to the movement.
• Effectiveness:
  • The quantized gradient clearly highlights key points along the spear’s path.
  • It enhances the visual storytelling by making the motion more readable and visually striking.
• Outcome:
  • The effect worked extremely well within the painterly aesthetic.
  • I was very pleased with the result and plan to use this technique in future animations to elevate dynamic action sequences.