Introduction

Talk of cut almost always starts with proportions. Crown angle, pavilion depth, table, symmetry — a convenient language: figures can be checked, entered on a certificate and set next to the price. The market needs that discipline: without it, comparison of stones quickly becomes arbitrary.

But in live jewellery trade an awkward remainder always stays. Two stones with tidy parameters look different. One seems calm in a photograph and comes alive under window light. Another gives a striking still and loses its pattern on a turn. A single line of proportions does not explain this.

This is the point from which the authors of the OctoNus project start. In their wording, the visual beauty of a diamond is composed of light, the stone’s optical properties and the conditions in which a person looks at it [1]. Cut grade is not cancelled by this. Grade and visual impression simply answer questions of a different scale.

Below is a review of open materials from the OctoNus diamond cut study. First their logic, then BLR, ETAS and the role of master stones. At the end we separate published claims from scenarios for the counter and e-com in 2026–2027. A related note on modelling fancy-color intensity: ML for fancy color.

Method and sources

This is desk research, with no BrillPulse experiment of our own. The corpus is four open OctoNus pages:

1. Determination of the cut quality of a diamond and visual evaluation of its beauty [1].
2. Diamond Cut Grading System Based on 3D Model. A Strategy for Development [2].
3. Problem of Grading Quality [3].
4. Study of Interdependence [4].

The pages name Sergey Sivovolenko (OctoNus Software), Yurii Shelementiev (Gemology Center of MSU) and Anton Vasiliev, linked to the optical part of the work; in some materials he is listed as “LAL”. English terms — light return, fire, Dynamic Contrast, BLR, ETAS — are left where a Russian equivalent would make the sense less precise. Quotations are from the published English texts.

The caveat here is one, and it is substantial: BrillPulse did not rerun DiamCalc on its own sample and did not run a blind comparison of stones. Where the text moves from the source to a possible market use, that is marked as a hypothesis.

Three groups of factors in the beauty of a cut

In [1] the authors do not try to derive beauty from a single characteristic. They split it into three groups of factors. The approach looks obvious until you look at an ordinary product card: it usually contains only the stone’s geometry and one shooting result.

1. Parameters of the light sources

For the authors, light is not background but part of the problem itself. What matter are the emission spectrum, the brightness and angular size of the source, its position relative to the stone, and the number of sources [1]. A small bright point and a large diffuse source will give a different pattern even if the stone does not move a millimetre.

So a single softbox in product photography and a window with several directional lamps cannot be treated as variants of the same test. In the first case the photographer controls the highlight; in the second the stone enters a more complex light field. The buyer will then see a third version — in an office, a restaurant or a daylight window.

2. Optical effects and properties of the stone

Into this group OctoNus puts light return — the share of incident light that returns to the eye — and fire, the splitting of white light into spectral flashes. Alongside sit colour, the pattern of facets and highlights, the symmetry of that pattern, Dynamic Contrast in motion, clarity, polish quality, facet flatness and surface smoothness [1].

The list is long, but it is a familiar experience. We do not perceive a stone as average brightness. The eye catches the alternation of light and dark zones, coloured flashes, the pace at which the pattern changes on a tilt. Lighting changes — and the “same” diamond becomes a different visual event.

3. Viewing conditions

The third group usually drops out of a technical description. Yet the authors account for the observer’s position, stereoscopic vision, pupil movement, colour background, the eye’s resolution and sensitivity, spatial and temporal contrast [1]. The last is especially important when the stone moves: a person sees not one frame but a whole sequence.

From this triad a fairly strict boundary follows. A model that has no light or observer describes part of the optics and geometry, but not the whole visual experience. That is not a defect of a parameter table; the table simply has a different job.

Links the authors state explicitly

Paper [1] has several especially valuable claims — precisely because they are stated without a marketing wrap. The first concerns colour and fire:

The radiation spectrum of the illumination source determines the perceived color of a diamond and affects its fire as well. The color depends on the product of the radiation spectrum of the source by the transmission spectrum of the gem. Both multipliers are equally important.OctoNus, Determination of the cut quality [1]

The authors separately tie fire to the angular size of the source:

The angular size of the illumination source strongly affects the fire of the diamond.ibid. [1]

And they warn against too literal a trust in photography:

A snapshot of a diamond is not adequate to its image visually perceived by an observer.ibid. [1]

Finally, they propose that cut be evaluated under lighting typical of the place where the stone is used:

The cut of a diamond should be evaluated under those illumination conditions being peculiar to the place where it is used.ibid. [1]

In the same row sits a constraint on mass:

It is inadmissible to improve the appearance of a diamond (as evaluated by an observer) by considerably decreasing its weight.ibid. [1]

From there it gets more interesting still. Size noticeably affects fire. Colouration, on the authors’ observation, usually weakens it and in extreme cases can almost remove it. The number and placement of sources change Dynamic Contrast and the highlight pattern. Poor polish and internal defects hit both fire and light return. Symmetry affects the overall impression. For colourless and weakly coloured stones the authors rank fire and Dynamic Contrast above light return; they also note the market’s wariness of unfamiliar proportions and facet patterns [1].

For a seller this means a simple thing: an argument over which stone is “more beautiful” remains unfinished until light and viewing mode are named. Still, rotation and binocular vision are different conditions. Likewise one cannot discuss the “best proportion” without recalling mass: the authors themselves treat a radical loss of weight to improve the visual effect as inadmissible.

Why a parametric grade is not enough

In [2] OctoNus compares the parametric approach, direct optical measurements and work with a 3D model. The parametric path looks natural: start with table, crown and pavilion, then add more and more variables. But the problem, in the authors’ view, is the links among them. A small set of unconnected numbers gives too coarse a picture, and an attempt to account for all parameters and all dependencies quickly becomes an unmanageable construction.

Hence their hard conclusion: parametric systems cannot fully solve cut grading even for the round shape, and for fancy shapes the limit is more visible still [2]. Proportions are not declared useless. They are needed in cut planning. But on their own they work poorly as a universal engine of new designs and of a full visual assessment.

Direct instrument methods — Brilliancescope, Firescope, Hearts and Arrows Viewer, Idealscope and similar device classes — receive a more mixed assessment. A system may be sensitive to a particular type of stone; the eye still interprets the result; measuring light does not remove the question of subjective perception. Scintillation is especially hard to cover correctly: the stone changes tilt, light sources are many and small. There is also a simple physiological objection — the instrument often works from one viewpoint, while a person looks with two eyes [2].

Some stones that look good or bad to our eyes may be graded differently by direct methods.OctoNus, Diamond Cut Grading System [2]

This is not an argument against instruments. Rather, a reminder that one should measure the property one then wants to discuss with the buyer or the cutter.

BLR, ETAS and master stones

Basic Light Responses (BLR)

Instead of moving from one proportion to another, OctoNus proposes starting from an accurate 3D model and calculated Basic Light Responses. Among the BLR examples in [2] are return of light to the observer’s eye, leakage through the pavilion, “nail head”, and ETAS as part of the space seen through a moving stone.

Responses are divided into positive and negative. Among the former are scintillation, brilliance, fire and contrast; among the latter — fish-eye, nail head and light leakage [2]. The scheme is convenient for an engineer: instead of the general epithet “beautiful”, separate optical reactions appear that can be calculated and matched to observation.

Effective Total Angular Size (ETAS)

ETAS, Effective Total Angular Size, the authors describe as a coefficient independent of a particular lighting and viewing condition. It is meant to describe the optical potential of a faceted stone and to help predict its light responses. But the coefficient itself is not a final verdict: the BLR set still has to be calibrated against what a person actually sees [2].

From this information we wish to propose a new coefficient that is independent of lighting and viewing conditions; Effective Total Angular Size, or ETAS. This coefficient will be able to describe the optical potential of any polished and faceted diamond.OctoNus [2]

Master Stone Set

Here the Master Stones Set (MSS) appears. If there is no ready objective formula of beauty, the system has to be tied to a series of real stones and human assessments. MSS should show differences between grades, collect scores on each BLR, include successful and weak examples, ordinary and atypical solutions, round and fancy shapes. It is also needed to test different grading systems [2].

The authors allow a statistical matching of computer and expert clusters; the text mentions neural networks as one way of working with such arrays [2]. The target system, in their design, should not privilege one shape or set of proportions in advance. Similar stones should receive a similar score, and the market should depend less on the slogan of a single “Ideal”.

So BLR and ETAS are more reasonably read as a research engineering frame. This is not a call to replace GIA Excellent tomorrow with one new metric. Master stones, expert calibration and a check that the calculation matches perception are already built into the scheme.

There is a less obvious consequence. Master stones are needed not only to tune an algorithm, but for a conversation among the workshop, the laboratory and the sale. These participants have different habitual languages. The cutter sees tolerances and mass, the gemmologist a comparison system, the seller the buyer’s reaction at the window. A series of stones through which all these questions can be walked removes part of the argument over terms. It does not remove subjectivity; it makes it observable and repeatable.

Such calibration needs discipline in the details: light source, distance, background, viewing angles and the composition of the assessor group. Otherwise a new “standard” merely fixes the taste of one stand or one person. In the published strategy [2] this is not written out as a ready protocol for retail, yet the logic of MSS itself does not allow viewing context to be treated as secondary.

Fancy color: cut and light are inseparable

For a coloured stone the question becomes more sensitive still. In [1] visible colour is described as the product of the source spectrum and the stone’s transmission spectrum. One multiplier cannot be treated as secondary. Hence a familiar effect: saturation on a studio photograph can diverge from the impression under other light.

In the OctoNus/DiamCalc logic, a good cut should not win only in one specially assembled lighting scenario. For a fancy stone this is especially visible, because cut, light distribution and colour work together. BrillPulse’s empirical contour on fancy yellow — scanner features and CatBoost — is treated in fancy-color-octonus-gia.

Synthesis: what this means for the certificate, the photo and the window

A certification cut grade gives the market comparability inside an accepted scale. The OctoNus approach tries to describe the stone’s behaviour in a light environment and in motion. These descriptions overlap but do not coincide. The phrase “Excellent, so it will play on any photo” attributes to the certificate a promise that sources [1]–[2] do not make.

A product photograph is a snapshot. The authors say plainly: a snapshot is not equal to the visual image [1]. So a short video, a 3D model, a turn of the stone and several lighting modes can be useful not as decorative excess, but as a way to show the original conditions more honestly. Even good video will not replace a personal viewing: the buyer will still have their own light, distance and background.

Another conclusion concerns cut planning. The wish to “add sparkle at any cost” is limited in the OctoNus papers by weight: a noticeable loss of mass cannot be a normal way to improve appearance [1]. That is a technical and a commercial constraint at once.

It does not follow that every stone in the window needs a complex lighting installation. Rather, it is useful for the seller to know which effect their light amplifies and which it hides. In e-com the task is more down-to-earth still: the buyer should not see a single retouched highlight. Several frames under understandable conditions will often say more than one flawless frame without context.

Scenarios and hypotheses

Below are not published market facts, but testable scenarios that follow from the OctoNus frame.

Limitations

The article does not measure stones and does not assign a cut grade. Paraphrase and quotations rest on the available OctoNus pages; completeness of the authors’ full archive is not claimed. BLR and ETAS are treated as parts of a system-development strategy, not as a public mandatory laboratory standard. Scenarios for e-com and tools are not a forecast of market share.

Conclusions

1. OctoNus describes the beauty of a cut through light, the stone’s optics and the observer [1].
2. Spectrum and angular size of the source affect colour and fire; a snapshot is not equal to visual perception [1].
3. The authors treat parametric cut grading as principally limited, especially for fancy shapes [2].
4. Direct optical instruments are useful but remain dependent on the measurement scheme and on a single viewpoint [2].
5. A 3D model, BLR, ETAS and master stones form a scheme in which calculation is checked against human perception [2].
6. The authors treat a considerable reduction of mass to improve appearance as inadmissible [1].
7. Certificate, photograph and the impression of a stone in motion describe different sides of one object.
8. Dual light, an extra language of metrics and differentiation of cuts remain BrillPulse hypotheses that need testing.

FAQ

Is a grade of Excellent enough to assess the beauty of a cut? No, if the subject is the full visual impression. Grade is needed for comparability, and OctoNus separately treats lighting, optical effects and viewing conditions [1][2].

Why does a stone in a photo differ from its look in the store? A snapshot is static and usually made in one lighting mode. In the store, spectrum, size and number of sources, the viewer’s position and dynamic contrast are at work [1].

Does a deeper cut make a stone better? There is no universal rule. The authors separately object to a noticeable reduction of mass for the sake of appearance [1].

Does fancy color depend on cut? Yes. Visible colour involves cut, light and the stone’s transmission spectrum together [1].

What is BLR? Basic Light Responses — basic light responses calculated on a 3D model and then calibrated to observed perception [2].

What is ETAS? Effective Total Angular Size — an OctoNus-proposed coefficient of optical potential, independent of a particular combination of light and observer [2].

Does the OctoNus method replace a GIA grade? No. It is a research and engineering frame for analysing visual quality, not a replacement for a laboratory cut grade.

Bibliography

1. OctoNus. Determination of the cut quality of a diamond and visual evaluation of its beauty. https://www.octonus.com/projects/diamond-cut-study/determination-of-the-cut-quality (accessed 27.07.2026).
2. OctoNus. Diamond Cut Grading System Based on 3D Model. A Strategy for Development. https://www.octonus.com/projects/diamond-cut-study/diamond-cut-grading-system (accessed 27.07.2026).
3. OctoNus. Problem of Grading Quality. https://www.octonus.com/projects/diamond-cut-study/problem-of-grading-quality (accessed 27.07.2026).
4. OctoNus. Study of Interdependence. https://www.octonus.com/projects/study-of-interdependence (accessed 27.07.2026).
5. BrillPulse. Two ML versions for fancy color. https://brillpulse.com/research/fancy-color-octonus-gia.html

BrillPulse records the content of open OctoNus papers and separates them from its own scenario assumptions. The material does not value a specific stone and does not promise a result for a given cut. Contact: info@brillpulse.com