Splash

Alexander Tullo - Director, Background Painter

Xing Ye - Character Rigger

Alan Wyatt - Pipeline Engineer

Overview

"Splash" was one of the most technically complex pieces that I created for my thesis, taking nearly three months to complete. It was an advanced water simulation scene with my filter running way beyond what Blender's geometry nodes could handle without crashing. This piece was created in response to the difficulty of preserving artistic integrity when depicting natural elements - something Sonya Shannon and Douglas Cooper have described as a significant limitation of animation. By abstracting and stylizing the water simulation through my filter, I addressed the aesthetic dissonance between realistic and more stylized elements.

The development of "Splash" involved careful planning and iterative workflow, with each iteration refining both the technical and aesthetic aspects of the scene, such as enhancing fluid dynamics and visual coherence. As noted by Aderito Fernandes Marcos in his research, achieving stylistic coherence in complex animations often requires innovative use of technology. This combination of fluid simulation and NPR techniques is proof of how procedural tools can combine to form a common visual language.

Artistic Intent

Artistically, Splash sought to imitate an impressionistic quality in water abstraction using non-photorealistic rendering. The painterly effect on the water was applied to create a natural yet expressive rhythm of movement that focuses on the scene's fluidity and energy while retaining emotional resonance. I wanted to abstract the water using brush strokes to suggest dynamism and continuity to the viewer rather than photorealistic accuracy that would distract from the expressive Intent of the piece.

Artistically, "Splash" is an excellent example of how to stylize characters and environmental effects with the filter. The painterly effect on the water simulation lends the scene an impressionistic quality that can not be achieved with traditional animation techniques. The fluid motions of water are complimented by a filter which resembles brush stroke, giving the piece rhythmic quality.

The brush strokes used to depict water also serve as a metaphor for the emotional upheaval of the characters in A Home for Swallows. The moving water reflects the protagonist's internal struggle, and the filter plays on the rapidity and fluctuation of these emotions. Such a visual metaphor is consistent with how Colleen Case analyzed how animated elements might be interpreted as representing states of mind.

Minor mistakes, like slight stroke pattern misalignments, are almost invisible to viewers but were good learning experiences for me as a creator. These imperfections also contribute to a more handcrafted look to the piece, in keeping with Gregory Garvey's argument that such minor errors can make non-photorealistic animated pieces feel more natural and emotional. Thus, The final product speaks to procedural techniques in 3D animation and produces an aesthetically pleasing, visually cohesive product combining technical innovation with artistic expression.

Technical Challenges

It took planning and a long iterative workflow to make "Splash" happen. I divided the scene into several elements to avoid overloading the system and crashes, including the background, the water simulation, and the character. By rendering all these layers separately, I could reduce the computational load and apply the filter much more efficiently without compromising quality. This also gave me more creative control during compositing as I could fine-tune each layer's interaction with the others.

Another challenge was balancing rendering time with the difficulty of the water simulation. Fluid dynamics is a notoriously resource-intensive process, and the filter added complexity. However, by baking out the simulation and optimizing the stroke density of the filter, I could keep rendering times within an acceptable bound and achieve the effect I was after. This optimization is reinforced by Tina O 'Hailey's exploration of hybrid techniques that balance traditional artistry with digital efficiency. Reducing detail where it was not critical helped maintain the stylistic integrity of the scene. This also ensured rendering stayed within the software and hardware limitations. These bakes many times took upwards of 30-60 minutes alone, but it was still far faster to seeing the result then working with the raw simulation data.

A further important step in the workflow was the temporal coherence of the water's painterly effects. In early tests, the fluidity of water caused inconsistent brush strokes throughout the animation. I ended up using a point distribution simulation that instead of using the points from the initial simulation that would have random point IDs each frame, I would distribute my own points and have them follow the velocity and voracity of the fluid simulation using vector math. This kept the water looking consistent, and avoided abrupt transitions that would break immersion.

Conclusion

By abstracting the fluid simulation with procedural geometry nodes, what would be a hyperrealistic representation of water became a visually expressive element supporting the emotional narrative of a scene. This approach demonstrates that procedural tools can solve technical problems and shows what animation can do visually. Therefore, animation's role is not to reproduce reality but to interpret and enrich it in ways that trigger more profound emotional responses. In this way, "Splash" illustrates how this philosophy can be implemented using new technologies to produce an innovative and emotionally resounding animation.