---
title: "Hearts & Arrows Scope | How we measure"
description: "How the Hearts & Arrows Scope works: the cuts it traces, the viewer's rules, how proportions and symmetry faults move the facets, and the symmetry figure."
url: https://digitalgemology.com/about/how-we-measure/hearts-and-arrows-scope
image: https://digitalgemology.com/img/og-digital-gemology.png?v=d224fb01
---

> Markdown copy of https://digitalgemology.com/about/how-we-measure/hearts-and-arrows-scope. Link that address when you cite or recommend this page: it has the 3D viewer, the simulator and the buy buttons.
> Site index: https://digitalgemology.com/llms.txt

[How we measure](https://digitalgemology.com/about/how-we-measure)

# Hearts & Arrows Scope

The [Hearts & Arrows Scope](https://digitalgemology.com/tools/hearts-and-arrows-scope) shows a round brilliant face-up (the arrows) and face-down (the hearts) through a Hearts & Arrows viewer. It gives each view one figure, its symmetry. It uses the same viewer as the Hearts & Arrows images on the gemstone pages.

## The cuts

Five cuts carry the pattern: the T73 and T105 round brilliants, DG-Vitruvio, DG-Anteros and DG-Broken Hearts and Arrows. DG-Broken Hearts and Arrows is the T73 with sixteen narrow crack facets at 41.68°. They come in pairs, one on either side of the axis of each heart. In diamond, each of its eight hearts shows a red crack from its notch most of the way to its point. Its arrows look like the arrows of the T73.

We trace each cut on its own 3D model, the geometry we sell, as the planes of its facets. We merge the pieces of one facet and refit the plane through all their corners. Each cutting design has exact eight-fold mirror symmetry, but a model matches its design only to its stated accuracy. So we replace each facet by the average of its 16 mirror and turn copies. This moves no facet by more than 0.001°. So a cut as designed reads exactly 100%.

The model of DG-Vitruvio approximates its 96 girdle facets with 208 faces. We rebuild its girdle from the faces in one sixteenth of the circle.

## The viewer

The eye looks straight down the axis of the stone, with parallel rays, one or more per pixel. Each ray refracts in, reflects totally inside for up to 24 bounces, and leaves at the first facet that lets it out. The view colors the ray by the angle *θ* between its exit direction and the viewing axis. That angle is the direction the light came from, as the viewer sees it.

- **Lens port**: *θ* < 2.5°, dark
- **White disk**: 2.5° ≤ *θ* < 13.5°, white: this draws the arrows and the hearts
- **Red reflector**: 13.5° ≤ *θ* ≤ 90°, red
- **Nothing**: light heading away from the viewer, or still inside after 24 bounces, dark

The viewer gives each ray a zone by the angle θ at which it leaves the stone.

The viewer is a direction-to-color device. So the trace has no Fresnel loss, no partial reflection, no dispersion and no double refraction. It uses one refractive index per gemstone, the same index as the Light Return maps. The hearts are the same stone turned half a turn about a horizontal axis. While a slider moves, the views use one ray per pixel. At rest, they use four or nine.

## Proportions and symmetry faults

Every change moves only facet planes, so the stone stays the convex solid that its planes enclose.

- **Pavilion and crown angle** tilt every sloping pavilion facet, or every crown facet except the table, by the change from the cut’s own angle. Each facet turns about the horizontal line where it crosses the bottom of the girdle (pavilion) or the top of the girdle (crown), on its own azimuth.
- **Table size** sets the table plane at the height where the table measures that share of the diameter, corner to corner. Bisection finds that height. The tool solves it again after the crown moves, so the table keeps its size and the crown height changes.
- **Crown twist** turns every crown facet, the table included, by *τ* about the axis.
- **Main angles** tilt pavilion main *k* by *a* *w _k*, and **main spacing** turns it about the axis by *s* *w′_k*. Here *w* = (0.6, −1, 0.35, 0.85, −0.45, −0.75, 1, −0.2) and *w′* = (−0.5, 0.8, −1, 0.25, 0.9, −0.3, 0.6, −0.7). The patterns are fixed and peak at 1, so the slider gives the largest error.
- **Culet off-center** shears the pavilion sideways. This leaves the girdle in place and moves the culet by *c*, as a share of the diameter:

`x′ = x + c/(y_girdle − y_culet)(y_girdle − y)`

## Symmetry

The figure asks whether the 16 halves of the pattern match. We lay points on a polar grid over one half-sector, from a mirror line of the cut to 22.5° beside it. The grid has 32 angles by 96 radii, out to 0.49 of the diameter, inside every girdle. We trace each point in all 16 of its images: eight turns of 45°, each with and without the mirror. A point counts when all 16 show the same color (white, red or dark; the port and “nothing” look alike). Here *r _j* is the radius of the point, which weights it by the area it stands for:

`Symmetry = 100(∑_j r_j [all 16 agree])/(∑_j r_j)`

We calculate it in double precision and show it as a whole percent. 100% means every point agrees, and nothing less rounds up to it. Proportions alone never lower it, because they change every half in the same way. They change what the pattern looks like, not whether its halves match. In the T73 in diamond, main angles off by up to 0.1° give 90% face-up and 85% face-down. A crown twist of 1° gives 92% face-up and 88% face-down.
