Month: June 2025

AI Animation Transfer and Stylization Experiment

This scene is a direct continuation of my earlier AI mesh experimentation. I used the same animation but migrated the project from Cinema 4D to Blender, continuing to develop the visual identity of the brushstroke style and elemental effects.

• 2D Lightning Effects
  As with the Cinema 4D version, I attached two-dimensional animated lightning to the character’s feet. This maintains visual consistency while allowing further experimentation with the energy-based aesthetic I’m pursuing.
• Brushstroke Stylization (Single Layer)
  In this version, I did not split layers, and that’s visibly apparent. It’s a very early iteration where I’m starting to play with how much layering and separation actually contribute to the final feel.
• Blurry Echo Effect
  One of the more experimental ideas I tried here was duplicating the mesh, offsetting it slightly, and lowering the opacity:
  • This created a motion trail effect or visual blur behind the character.
  • It helps hint at speed or displacement without relying on traditional motion blur tools.
  • This will likely be refined in future iterations.
• Environment Setup (Sky, Plane & Sphere)
  • I placed a plane in the background with a sky texture.
  • Behind that, I added a large sphere to round out the depth of the backdrop.
  • However, the result was grainy, especially in the background layers, due to either sampling limitations or the shader setup.
  • I plan to troubleshoot and refine this to reduce noise in future renders.

---

Reflection

Although this is clearly a rough test, I’m quite happy with the visual direction. The blurry mesh duplication and 2D lightning animations serve as promising leads for stylization. Going forward, I’ll be focusing on refining the brushstroke fidelity, experimenting with opacity layering, and optimizing render clarity—particularly when using painted backdrops or procedural skies.

Let me know if you’d like this written in a more academic tone or broken up visually for layout purposes (e.g. with subheaders for FX, Shading, and Environment).

In this section, I generated a 3D model through an online software called MESHY. I used it to quickly prototype a character and place them into a basic environment to test early concepts. The focus here was not refinement, but rapid iteration and experimentation with workflow enhancements.

• Environment Setup:
  A simple scene was constructed to simulate a mountainous environment. Although it’s very early-stage, it represents the initial direction I wanted to take for my world-building—especially the concept of ascending a mountain, which remains central to the broader project.
  
  
• Electric/Lightning Effects:
  The lightning around the character’s feet is two-dimensional animation. I achieved this by attaching animated planes to the feet of the character.
  • These effects are currently asynchronous—they don’t yet match the character’s footsteps.
  • That said, this still serves as an excellent early proof-of-concept and visual exploration of how elemental effects might interact with motion later on.
    
    
• Why I Used MESHY:
  This was part of my broader strategy to experiment with AI-generated meshes. My rationale for doing so is as follows:
  • It significantly speeds up the modeling pipeline, giving me more time to spend on composition, lighting, texture, and animation.
  • It allows for more creative flexibility, because I can quickly generate and iterate on multiple character concepts or props without getting bogged down in topology.
  • Even if the output isn’t perfect, it’s a solid foundation I can clean up, re-texture, and shape to my liking.
    
    
• Lighting and Direction:
  I also started experimenting with dramatic lighting—sharper shadows and dynamic highlights—to explore how I might accentuate the movement and atmosphere of the mountain climb in future shots.

Investigation into Painterly Art Styles in Animation

Painterly styles bridge the gap between traditional painting and digital animation, employing brush-like visuals, expressive color, and atmospheric composition to convey emotion and narrative depth. They emphasize mood over realism, often allowing less detailed models to evoke stronger emotional resonance.

---

1. Loving Vincent (2017) – Fully-painted feature film

Background: The world’s first fully oil-painted feature—65,000 hand-painted frames recreating Van Gogh’s impasto style .

• Color & Composition: Thick, vibrant brushstrokes replicate Van Gogh’s dynamic color palette—yellows, blues, and greens merge to evoke emotional turbulence.
• Model Detail: Rough, stylized models embrace abstraction; character likenesses emerge through color planes and strong silhouettes.
• Effect on Story & Movement: The style enhances the emotional weight of each scene—brushstrokes swirl during turmoil, remaining static in melancholy, guiding viewer emotion.
• Pros: Unmatched expressive power, unique visual identity.
• Cons: Immense resource requirements and slower production.

---

2. Tangled (2010) – Disney’s painterly CGI approach

Background: Disney used Non-Photorealistic Rendering (NPR) to mimic Rococo painting influences in 3D .

• Color & Composition: Soft pastel lighting layered with brushtexture shaders gives forest scenes a lush, storybook appearance.
• Model Detail: High-detail CG models combined with textures that kind of imitate brush strokes looks surprisingly appealing, but not glaring at all.
• Effects & Motion: Brushstroke textures subtly animate with light, reinforcing movement and depth, yet remain invisible.
• Pros: Combines rich visuals with strong character animation; high production polish.
• Cons: Balancing painterly aesthetics with photoreal requirements is technically demanding.

---

3. Lilo & Stitch / Ghibli-Inspired Indie Games

Background: Games like Crescent County and The Red showcase hand-painted textures in real-time engines https://shahriyarshahrabi.medium.com/creating-painterly-3d-scenes-preparing-assets-for-npr-8d6c726cc34f

• Color & Composition: Inspired by Studio Ghibli, they use vibrant yet muted palettes, hand-brushed textures, and layered depth. https://www.cartoonbrew.com/cgi/pixar-makes-painterly-cg-new-research-could-change-the-look-of-their-films-95205.html
• Visual Traits: Stylized filter retains visible brushstroke orientation and texture over time engadget.comyoutube.com.
• Color & Composition: Dull warfronts contrasted with vibrant highlights—brush direction draws focus to characters and dialogue.
• Model Detail: Block-form characters textured with painterly strokes create emotional atmosphere over fidelity.
• Special FX: Light bloom and subtle canvas overlays add painterly finish; motion blur integrates shading into stroke flow.
• Pros: Emotional resonance through atmosphere; simple models reduce workload.
• Cons: Texture repetitiveness can reduce visual novelty.

---

Color, Composition, and Models

• Color Usage:
  • Painterly art prioritizes emotional palettes—warm reds signal climax (Loving Vincent), pastels evoke calm (Tangled), earth tones create nostalgia (11-11 Memories), while greens evoke serenity (Ghibli-inspired).
  • Color transitions often align with narrative shifts—palette wars or dawn-to-dusk style changes.
• Composition Techniques:
  • Defined by broken brush textures—silhouettes are drawn into scenes using directional strokes that follow character motion.
  • Depth cues rely on color over detail—foreground held crisp, backgrounds soft and blurred.
• Model Geometry:
  • Low-detail geometry is common: stylistic abstraction replaces realism (Loving Vincent, indie games).
  • Some solve motion with high geometry but simplified shaders (Tangled), balancing animation quality and texture fidelity.
• Special Effects:
  • Use of procedural shaders to mimic brush wobble or streaks responding to character movement.
  • Explosion, wind, and smoke effects styled with animated stroke patterns that complement the painted aesthetic.

---

Pros and Cons of Painterly Styles

Pros

• Emotionally powerful and evocative—painting textures engage viewers at a visceral level
• Simplifies geometry workload—modeling needs are often minimal
• Unique visual identity—distinguishes projects in saturated markets
• Natural integration of FX and shading—brush style serves both texture and motion

Cons

• High labor intensity—frame-by-frame consistency is challenging (Loving Vincent)
• Shader complexity—real-time painterly requires bespoke tools (Crescent County)
• Potential loss of readability—overworked textures can obscure actions
• Risk of visual fatigue—static brush patterns become heavy or redundant over time

---

Conclusion

Painterly art styles in animation and games offer a powerful synthesis of traditional art and digital performance. Case studies from Loving Vincent to Crescent County show how palette, composition, and texture coherence can elevate storytelling and mood. While resource-intensive, the end results can be visually stunning and emotionally resonant—provided pipelines and shader systems are properly managed. If used thoughtfully, painterly styles empower expressive visuals without sacrificing performance or narrative clarity.

The video above is a rough reconstruction of my animation progress so far. I’ve retargeted the original skeleton onto a mesh I previously developed for another module within the Experimental Animation unit. This allowed me to consolidate workflows across modules and experiment with character continuity between projects.

Retargeting & Rig Adjustments

• I had to shorten the character’s arms slightly to align with the proportions of the imported rig.
• The skeleton came from Autodesk Maya and included numerous deform bones, some of which caused unwanted mesh distortion.
• To resolve this, I learned how to unbind specific bones, allowing for a cleaner, more intentional deformation across the mesh.

Environment & Composition

• I initially considered using a more complex backdrop (like mountains, which featured in earlier iterations), but I felt this overcomplicated the composition.
• I opted instead for a minimalist environment: a flat plane with light rock variation, and a simplified skyline in the background. This keeps the focus entirely on the character and their motion.

Camera Logic & Movement

• The camera tracks the character dynamically, at first chasing them and then overtaking them as they reach the apex of their attack.
• This final camera position, ending in front of the character, is symbolic — it subtly reinforces the character’s dominance over the crystal enemy they just defeated.
• Throughout the animation, I focused on maintaining strong camera framing — ensuring the character remains in view for clarity, weight, and visual intent.

This version of the scene represents an important turning point in my project: it consolidates character performance, rig integration, and cinematic storytelling into one evolving sequence.

In the blocking stage of my dialogue animation, I began implementing key elements drawn from my reference footage and feedback I received. One of the first considerations was character orientation—I kept the character facing the camera more directly, using a maximum of a ¾ angle. This was to ensure that the head and body didn’t shift too far to the side like they did in my reference, which made facial expressions harder to read. Since the character is speaking to someone just off-camera, it made sense to have the upper body turned toward the viewer, prioritizing clarity over realism.

Initially, I had considered having the character’s body turned away to reflect their perceived superiority, but I quickly realized that in 3D animation, visibility and silhouette are key. I also applied George’s general class feedback: avoid disconnecting facial and body movement. Instead, the head and face should move in relation to the body. I began incorporating subtle upper body movement that matched the character’s expressions and gestures, which George praised for enhancing expressiveness—something I’ll continue refining.

In terms of critique, George pointed out that animating the eyebrows as if they’re connected in a unified curved motion reads much better than moving each brow separately. When he demonstrated this on my animation, I agreed—it cleaned up the motion significantly. I’ll be implementing this across future work.

Another issue was the mouth shapes. I initially had them too flat, which made expressions look stiff. Raising the corners and shaping the mouth more like a semicircle helped improve clarity, even if it strays slightly from realism. I’ll also be focusing on placing the teeth more intentionally going forward.

For the eyes, I was leaving the lids slightly open even when “closed,” which left a pixel of white visible—something that actually disrupted the look more than I expected. I’ve since corrected this to allow the eyelids to fully close and meet in a clean arc. I also had the eyelids slightly overlapping the iris, which gave the character an unintended “drugged” look. While I was trying to portray a regal, detached expression with heavy eyelids, I now understand that I can still lower the lids for that effect without obscuring the iris.

I’ve also started syncing eyebrow movement with the eyelids—when the eyebrows raise, the upper lids should follow, and when they lower, the lids come down too. This adds a more natural skin-pulling effect.

Finally, I received positive feedback on how I animated the jaw—there’s a good amount of movement, which adds a lively, stylized touch to the character. George said it reminded him of King Julian from Madagascar, which I take as a compliment—DreamWorks nails personality-driven facial animation.

Overall, this blocking phase has taught me a lot about how small adjustments can make a massive difference in clarity, appeal, and character.

After some trial and error, I’ve finally settled on a reference that really excites me—Monty Python and the Holy Grail. The scene I’ve chosen features King Arthur attempting to assert his authority over a group of peasants who challenge the legitimacy of his rule, delivering a satirical take on anarcho-syndicalist communes. What drew me to this scene was the rich variety of emotion: the character (Arthur) fluctuates between pride, confusion, and incredulity, all while maintaining a comedic undertone.

This reference provides an opportunity to animate nuanced emotional transitions—starting from superiority and indignation, moving through confusion, and landing in moments of visible frustration. I especially like how Arthur keeps his chin raised, conveying condescension, then briefly dips into bewilderment when his authority is questioned. These shifts in posture and expression offer a lot of material to explore through 3D animation.

George provided some helpful critique during our session: one key point was ensuring that the character maintains consistent eye contact with the other person in the scene. I tend to look away at times, which can reduce clarity in the facial performance—especially from a three-quarter angle. Keeping both eyes visible strengthens readability and emotional impact. I was also encouraged to include the little chuckle I do in the reference—it’s subtle, but adds humanity and sells the internal emotional state.

Another suggestion was to slow the timing down slightly. While I already have snappy movements for contrast, adding variety in pacing should give more room for micro-expressions to land and feel intentional. Lastly, I plan to trim the clip to focus on the most essential beats—the peasant’s reply at the end, for example, doesn’t add much to the core performance and can be left out.

Overall, I’m excited to dive into this one—it’s expressive, challenging, and gives me space to push character personality through facial acting and timing.

This reference was originally recorded for the Dialogue Animation module. Although I won’t be using it in my final submission, I still wanted to include it here to document my process and thought development. The purpose of this exercise was to explore the nuances of transitioning between emotional states—both internally (subtle facial shifts) and externally (spoken delivery and gestures).

The reference is based on a scene from Monsters, Inc., specifically the moment when Sully and Mike meet the Abominable Snowman in the Himalayas. What drew me to this clip was the dynamic emotional interplay between the characters: one emotes before speaking, the other speaks before emoting. This layering of emotional beats and verbal rhythm really helped me observe how character intention shifts from moment to moment.

However, while studying this clip was useful, I’ve decided not to use it for my final piece. Although we weren’t explicitly told to avoid referencing existing animations, it makes sense from an academic standpoint to animate original material or live-action references. Using pre-animated content—especially from stylised films—risks undermining the learning objective, which is to interpret and convey raw emotional beats through our own animation work.

Still, I found this study valuable for understanding how emotional expression isn’t just about exaggerated mouth shapes or eye movements—it’s about timing, contrast, anticipation, and subtlety. I’ll carry this insight into the rest of my project, using live-action or self-recorded references instead.

This second reference is taken from Whiplash, specifically a moment when Fletcher—the film’s intense and domineering antagonist—publicly confronts a student with thinly veiled contempt. What drew me to this scene was the emotional tension beneath Fletcher’s delivery: there’s an underlying aggression simmering below the surface, yet his facial expression remains controlled and composed. The subtlety of this interaction made it a compelling study for how rage can manifest through micro-expressions and tone rather than overt gestures.

My intention was to explore this emotional restraint and attempt to reinterpret it in a 2D animated format. However, I ultimately chose not to use this reference for a few reasons. Firstly, the shot involves swearing which I apparently shouldn’t use in my showreel, and if I censor it then whats the point really? Might as well use a better clip, the character will be swearing visibly as well so. Secondly, after reviewing the criteria, I realized that the emotional transitions within the clip are minimal. Fletcher moves from calm to contempt without much visible progression, and the shift happens quickly and with limited variation in facial expression.

Additionally, because this moment is already animated cinematically (in the live-action sense), recreating it one-to-one would offer little creative interpretation. The goal of this task is to explore the character’s internal state and show emotional evolution—something better achieved through original or live-action references where I can capture a broader spectrum of emotional change.

While I won’t be using this scene in my final submission, studying it still helped me reflect on the complexity of restrained emotion and the importance of contrast and build-up when animating subtle psychological shifts.