Virtual facet patterns and diamond design
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Journal of Mathematics and the Arts |
| Published | 2026-09-16 |
| Authors | J. Conant, A. Prabhu, J. Quick, T. Blodgett, Ilene M. Reinitz |
| DOI | 10.1080/17513472.2026.2729158 |
Ray tracing of polished diamond geometries resolved light paths with at least 120 internal reflections. The authors use this depth information to connect cut geometry with apparent brightness, contrast, fire, and scintillation.
Virtual facets and the observed mosaic
Section titled “Virtual facets and the observed mosaic”The visible mosaic contains more polygons than the physical crown and pavilion show. Each virtual facet groups rays that share an ordered sequence of facet interactions. Reverse tracing from an observer follows those sequences back to the surrounding environment. When a beam meets a new surface, it divides, and each branch can divide again after another reflection.
For the illustrated round brilliant, light enters through one crown facet, reflects twice from the pavilion, and exits through the crown. For a convex gem under orthographic tracing with parallel rays, each virtual facet is convex and connected. Its boundaries correspond to the gem’s one dimensional edges.
Earlier virtual facet work focused mostly on round brilliants. The framework here gives varying visual depth a direct role in appearance. The authors regard parallel rays as a reasonable approximation at normal viewing distances.
Ray tracing and Fresnel assumptions
Section titled “Ray tracing and Fresnel assumptions”GIA’s custom ray and beam tracing software applies Snell’s law at entry and at any internal facet below the critical angle. A ray meeting a facet at a larger angle than $\arcsin(1/n)$ reflects internally through total internal reflection. The primary ray approximation follows the transmitted branch and neglects specular glare and higher order Fresnel reflections.
At 519 nm, an exiting ray typically carried a little over 80 percent of incident power. The reflected branch usually carried a little over 17 percent, but its share rose toward 100 percent at the critical angle. The authors defined the modified critical angle where transmitted and reflected branches carry equal power. Across 400 to 800 nm, this threshold differed from the critical angle by about 0.25 degrees.
That gap was smaller than the change caused by shifting wavelength, which supports the standard primary ray approximation.
Table 1 reports the calculated thresholds for light with randomly mixed polarization.
| Wavelength (nm) | Refractive index | Critical angle (degrees) | Modified critical angle (degrees) |
|---|---|---|---|
| 400 | 2.458 | 24.01 | 23.77 |
| 450 | 2.443 | 24.16 | 23.92 |
| 500 | 2.429 | 24.31 | 24.07 |
| 550 | 2.421 | 24.40 | 24.15 |
| 600 | 2.414 | 24.47 | 24.23 |
| 650 | 2.410 | 24.52 | 24.27 |
| 700 | 2.406 | 24.56 | 24.31 |
| 750 | 2.401 | 24.61 | 24.37 |
| 800 | 2.400 | 24.60 | 24.38 |
Diamond’s refractive index fell from 2.4709 at 380 nm to 2.3994 at 780 nm. Each wavelength therefore follows slightly different paths through a given geometry. Near the critical angle, part of the spectrum may reflect while the rest refracts, producing partial leakage. Polarization did not affect the primary trace, and one example intensity differed by about 0.16 percent.
Lighting maps and higher order rays
Section titled “Lighting maps and higher order rays”ASET sorts entry beam directions into blue, red, green, and below horizon regions. Blue covers 75 to 90 degrees from vertical, red covers 45 to 75 degrees, and green covers 0 to 45 degrees. ALGW retains five colors above the horizon, while Mono ASET uses a continuous grey scale. Below horizon light marks leakage or windowing in the maps.
The diffuse model uses a 30 degree obscuration disk at the celestial sphere’s north pole and a cosine power distribution above the stone. Secondary facets contribute little under diffuse lighting because stronger primary facets wash them out. Under sparse lighting, they can become prominent because the light overlaps their directions but not the corresponding primary directions.
A 5 degree spot source can therefore illuminate a secondary facet without illuminating its overlying primary facet. The authors also observed several fancy cuts in direct sunlight. The sun’s angular diameter is approximately 0.5 degrees, so it qualifies as a spot source. Most of each stone appeared dark, while facets returning sunlight were blindingly bright. Smaller colored sparkles came from secondary and tertiary facets.
The dual ASET map blends primary and secondary traces with Fresnel calculated relative brightness. On a research pear diamond, it broke up apparent windowing in the head that the primary map showed as dead areas.
Depth, focus, and pattern control
Section titled “Depth, focus, and pattern control”A k-flector has k internal reflections within a sequence of k plus 2 facet interactions. Most of a standard round brilliant’s surface area consists of 2-flectors. Fancy shapes can contain 20-flectors and beyond, with much longer internal paths. Longer paths produce smaller facets that appear farther away and tend to fall out of focus. These regions form much of the crushed ice seen in fancy shapes.
The visual influence score combines face up area as a fraction of the diamond footprint, compactness from an oriented bounding box, and average internal ray length in millimeters. The score gives greater weight to larger, more compact facets with shorter paths. The authors report that brightness metrics using visual influence consistently outperformed alternatives when tested against observations.
Bow-tie darkness comes from head obscuration. A study of a pear with 7 main facets varied crown angles from 30 to 35 degrees and pavilion angles from 38 to 44 degrees. Substantial bow-tie appeared at pavilion angles from 39 to 41 degrees, and pavilion angle had the stronger effect. A separate study held crown and pavilion angles at 32 and 39 degrees while varying length to width ratio from 1.2 to 1.8. Bow-tie behavior stayed similar across that ratio range.
A green marked girdle on a pear diamond showed why reflection matters. A leakage-only map captured light entering through the girdle. A reflection map also captured internal girdle reflections and matched the marked area more closely.
Scintillation and fire under spot lighting
Section titled “Scintillation and fire under spot lighting”Sparkle probability measures whether a light source of radius r illuminates at least one primary or secondary virtual facet. The calculation places the source randomly in the ASET red zone, from 15 to 45 degrees from vertical. Tertiary facets can be added to model sunlight scintillation. Sparkle area instead records the average illuminated facet area. Sparkle spread averages the convex hull area of illuminated facets as a fraction of the visible diamond.
Forward dispersion measures the separation of wavelength endpoints after a beam propagates from the environment. Reverse dispersion traces collinear wavelengths backward from the eye. They give different answers because the beam cross section can compress or widen asymmetrically. Across 180 fancy shape diamonds, forward dispersion tended to be about 85 percent of reverse dispersion. A linear combination over several tilts correlated best with human observation, but those details await a future paper.
Designing and evaluating new cuts
Section titled “Designing and evaluating new cuts”An enumeration algorithm listed crown and pavilion faceting arrangements with at least threefold symmetry. The team selected dozens of arrangements and built 3D models for them. A COBYLA optimizer then searched a set of stone parameters against a prototype cut grade. It used multiple random starting points because the optimizer does not require gradients.
ALGW grids varied crown angle, pavilion angle, and table size for each arrangement. Gilbertson selected models that appeared most appealing in his experience. Some top models were submitted to a manufacturer in Surat, India for cutting.
The resulting designs include the Jasmine, Poppy, Sunray, and Blossom cuts. Jasmine has an even distribution of virtual facet sizes. Poppy and the round brilliant had the most dynamic contrast, although Poppy developed more severe windowing at a 30 degree tilt. Sunray had the most fire potential, narrowly exceeding the round brilliant. Blossom kept its pattern stable under tilting. These cuts had little variation in facet depth and little crushed ice.
Limits and engineering implications
Section titled “Limits and engineering implications”Long path models compound geometric approximations and 3D scan errors, which reduces fine scale fidelity in crushed ice regions. Perspective tracing can also create curved, nonpolygonal boundaries near the critical angle. The authors did not pursue binocular vision, although cited work suggests that each eye may see a different virtual facet pattern. Sparse lighting also has no single canonical environment for evaluating scintillation.
The paper’s sparkle section places its red zone at 15 to 45 degrees, while Table 2 assigns red to 45 to 75 degrees. This definition needs clarification before the metric is reproduced.
The framework remains intended to support a future fancy shape grade system. It does not compare diamond growth methods or material grades beyond optical dispersion. Its guidance therefore applies to facet geometry and proportions, rather than growth material selection. Systematic exploration of oblong designs remains future work. The present output consists of a virtual facet analysis method, a prototype grade metric, and several new GIA cut geometries.
View Original Abstract
Diamonds used in jewellery represent the quintessential fusion of aesthetic appeal and technical expertise. Not only does it require extreme skill to cut a diamond, analyzing and predicting a diamond’s Visual PerformanceTM is essentially a technical mathematical problem. In particular, diamonds and other faceted gems have the property that they fragment the visual field into a set of polygons called virtual facets, which represent equivalence classes of light paths according to their facet interactions with the gem polyhedron. In this paper we explain how this mathematically elegant conception forms a framework from which to analyze a diamond’s appearance and beauty. We intend this framework to form the foundation for a fancy-shaped diamond cut grading system. Although virtual facet patterning has been studied before, mostly in the context of round brilliant diamonds, a significant new feature of our approach is to take into account the effect varying visual depths of virtual facets has on visual perception. We also showcase some novel designs created by our team at GIA using a mathematical enumeration algorithm for symmetric diamond-like polyhedra.
Electronic Grade Intrinsic Diamond Epitaxial Wafer H Terminated Superior 100 Mosfet Ready Substrate
View Specs →
High Concentration Quantum Grade Single Crystal Diamond
View Specs →