Fitting a facet design to a piece of rough

[RoughCreations]

I have been doing a proof of concept for fitting a chosen faceting design to a piece of gem rough using 3D software and an automatic algorithm. Yes, I probably have too much time on my hands, but I do a lot of 3D modelling in my paid work, so it doesn't hurt to have a play. The aim is to get the best yield possible from a design-rough geometry perspective.

The fairly random facet design I used for testing purposes is "Watch" by Dan Starr



The gem rough is a chunk of smoky quartz. I painted it in ochre to reduce reflections which allows a better 3D model to be created by the software.



34 images of the rough were taken at different angles. A patterned "Zebra" base was used which allowed the 3D processing software to better understand the orientation.
The set of images were first processed into a point cloud, and then a mesh object.





The gem design and the rough were imported into 3D software and then enmeshed together.



The challenge is to see if doing the best-fit step can be automated. This is the key point in the whole process, it can be done manually with the software as I have done in the picture above, but full automation is the aim.

RC

[Manficat]

This is way beyond me  this would probably be quite helpful for faceting. what kind of things do you 3d model in your paid work?

[RoughCreations]

It may be useful for the computer-keen person with a piece of high value rough.
I 3D-model mineral deposits in my paid work as a geologist and then calculate a resource (tonnage and grade).
Incidentally, if anyone reading this (or someone they know) needs a mineral resource calculated, let me know - special price for Forum members, lol.

RC

[bush]

What cad are you using?
I'll see if I can create a script or plugin to calculate the best fit

[RoughCreations]

What cad are you using?
I'll see if I can create a script or plugin to calculate the best fit

Meshroom for the photogrammetry, and Blender for the best fit. I might also try MeshLab for the best fit as it is easier to use. They're all free and open-source software.

[Alaskan]

I was asking andrewbrown about the small stones we see selling from around the world and not having any clue as to how to choose a design...

... or the expectation of having a finished product of what... 50% of the size of the rough being optimistic?

What kind of product to expect from a 2.5 carat sapphire or garnet... will wait for your experience... thaks.

[RoughCreations]

I was asking andrewbrown about the small stones we see selling from around the world and not having any clue as to how to choose a design...

... or the expectation of having a finished product of what... 50% of the size of the rough being optimistic?

50% yield from a specific piece of rough is generally very optimistic, generally only achievable if your rough has a freakily-good shape or has been pre-formed by the seller (rare) and would therefore command an extra premium in the asking price.
I think a 25 - 33% yield calculated over many cut stones seems to be closer to the average that normal facetors achieve. Carefully choosing a design to match the shape of your rough would get you closer to 33% average, with nice prismatic rough such as tourmaline crystals etc definitely helping in this regard. Highly irregularly-shaped rough with shallow projections usually means closer to the 25% yield, with unexpected inclusions discovered and requiring removal during the cutting process also affecting the final yield. Other factors such as human error while cutting can also combine to reduce the yield for a given stone, and affect your average too.

What kind of product to expect from a 2.5 carat sapphire or garnet... will wait for your experience... thaks.

So, bearing in mind all that, the answer to your question is unfortunately: That depends....
If I was buying a 2.5ct piece of sapphire rough, I would be focussing on nice clean regularly-shaped offerings only, with the aim (hope?) of getting an 0.8ct finished stone.

RC