Picture a dig, somewhere — far from any data centre. On the table lie the finds of the week: a sherd with incised decoration, a fibula, a small figure. The archaeologist on site has no graphics card, no 3D training and no time to make up for either. He has a camera. And that is enough: he photographs each piece as a clean series — by the rules you now know — hands over the images, and a thousand kilometres away a museum computes the digital twins from them and puts them online with the Orbit module. Between find site and worldwide availability lie one to three days.
This pattern — capture on site, compute centrally, show globally — is more than an anecdote. It is the division of labour that first brings this technology into the wide world: the rare resource (computing power, 3D experience) is decoupled from the common one (a person with a camera in the right place). Just as photojournalism invented the correspondent, the 3D correspondent arises here. And for cultural heritage one circumstance comes on top that makes this path not only convenient but necessary: the artefact often may not travel at all — protective regulations, fragility, export bans. The photos always may. The twin goes where the original would never come, and the original in the process does not leave its country by a single metre.
AI image
For small objects the circuit from Part 3.2 applies, with a field adaptation: the piece lies on a texture-rich, matte underlay (7.3 — a printed cloth, at a pinch a newspaper), the light is diffuse (an overcast sky is field equipment at zero cost; avoid or shade bright sun). Then two moves that turn the twin from a matter of appearance into a scientific document:
And over it all hovers the one condition that sets this scenario apart from all others: there is often no second attempt. The dig moves on, the piece goes into storage, the return flight is booked. That is why in the field one checks before packing — the checklist for it:
Now the images travel — and on this journey a banal, devastating mistake lurks: the messenger. Whoever sends photos over chat services sends compressed copies — shrunk, softened, often without the capture data that Motion Guard and the camera computation need. The rule is absolute: original files, unaltered — via cloud folder, storage medium or file transfer, one folder per object, never mix series.
Why this strictness? Because in this whole chain — capturing, transmitting, computing, publishing — exactly one link is not curable from afar: the capture. The museum can recompute, export differently, present more beautifully. What it cannot do: sharpen a shaken series or invent a missing back (Part 5.8). The person at the find site holds the key — that is no burden, that is the dignity of this work. And it is the reason why this entire work addresses itself to that person.
On the receiving side runs the program you know from the tool manuals of this library: import, automatic image check, reconstruction path according to image count (Part 4), training, finished model. One detail rounds off the division of labour: the Motion Guard finding comes minutes after the handover — not days later. If it reports problems while the dig is still running, the second attempt is suddenly possible after all: a phone call, a follow-up series the next morning. The check at the start of the chain turns the worst case (unusable, and everyone has departed) into an email.
The finished model is only half the fulfilment of the scenario — the other half is publishing without a line of code. In the Orbit module the twin becomes a narrated exhibition: hotspots mark the incision, the repair spot, the stamp; the gallery shows detail and find photos; the audio guide speaks the history of the piece; all of it multilingual. And beneath the visible surface lies the invisible one: DFX SplatCore writes the object information machine-readably into the page — search engines and AI assistants understand what this object is, where it comes from, who shows it. The sherd is not only looked at; it is found.
The final point of the scenario carries further than any cloud: the Orbit export is fully self-contained. It runs on the exhibition computer in the museum, on a trade-fair device without internet, on a tiny computer at the display case — and it does not cease to exist when a provider changes its prices or discontinues its service, because there is no provider in between. For cultural heritage this is not a convenience feature but the condition: a digital memory that depends on a subscription is none. Meydenbauer's glass plates survived a hundred years and then made reconstructions possible — the claim on their digital successors may be no lesser.
