Goniochromatic 3D Printing

3D printing is no longer just about geometry, it’s increasingly about appearance. As additive manufacturing technologies continue to advance, there’s growing interest in full-color printing and appearance-driven design. Applications range from product design and prototyping to visual effects and functional parts where the visual outcome matters. Yet, the tools for controlling surface detail, material appearance, and light interaction remain limited. While color printing has seen steady progress, most systems still treat appearance as a surface texture or color. This approach ignores how light interacts with the geometry and materials within the object. These factors are essential for reproducing effects like translucency, gloss, or direction-dependent color. In practice, designers have limited ways to preview or predict how a print will look under a given lighting or from different viewpoints. To realize the potential of appearance in 3D printing, we need better ways to model how shape and materials interact with light, better tools to simulate what the final print will look like, and more control over how to physically reproduce those effects using existing hardware. This thesis addresses these gaps by developing a set of techniques that work together, from geometric detail encoding, to supporting accurate and printable appearance in 3D prints, to predicting material optical properties.
One particular visual effect is goniochromism, where an object appears to change color depending on the viewing angle. This effect is found in nature and is widely used in industries like automotive coatings and anti-counterfeiting. However, reproducing this effect with 3D printing has received little attention. This thesis brings together a complete workflow that makes it possible to design and print goniochromatic effects using commercial multi-material printers.
The first contribution is a method for representing fine surface details using Displaced Signed Distance Fields (DSDF). It allows for a detailed surface modulation on low-resolution meshes, without requiring surface subdivision. The method supports large displacements, avoids common issues like self-intersections, and runs in constant time per voxel - scaling well to large prints. The second contribution introduces a method for fabricating spatially varying goniochromatic effects using procedural meso-facets. These are small geometric structures added to the surface, each with its own directional color. We use near-isometric surface parameterization to maintain a consistent facet size across complex shapes, and apply angular color blending to achieve clear directional color changes, without extra post-processing or multiple fabrication steps. The third contribution focuses on predicting the intrinsic optical properties of printing materials - specifically scattering, absorption, and refractive index. These properties are essential for simulating the print’s appearance, which is crucial for designing reproducible goniochromatic effects. We simulate a spectrophotometer using Monte Carlo path tracing, generate synthetic training data, and train neural networks to predict material properties from measured reflectance and transmittance. The method is validated against real PolyJet printing materials and serves as a practical tool for appearance softproofing. Together, these contributions form a complete pipeline that connects geometry, appearance modeling, and fabrication. The techniques enable more controllable and accurate appearance in 3D printing, which goes beyond color. This work helps build the tools needed for future improvements in appearance reproduction, inverse material design, and rendering-aware fabrication.