Walk into a serious stone carver’s studio in 2026 and you’re as likely to see a camera rig and a colour calibration card as you are to see a sketchbook. The documentation craft that used to mean detailed drawings, plaster casts, and clay maquettes now includes photogrammetry, structured-light scanning, and software that turns hundreds of photographs into millimetre-accurate digital records of carved work.

This isn’t a story about technology replacing craft. The chisel work and the eye for stone are exactly what they were. The story is about the recording of carved work — for restoration, for archive, for client presentation, for insurance — getting much more accurate and much more accessible than it was a decade ago. Studios that have figured this out have a meaningful advantage on heritage commissions and conservation tenders.

What photogrammetry actually involves

The principle is older than the technology. Take many overlapping photographs of an object from many angles. Feed them to software that identifies common points across the images and reconstructs the three-dimensional surface. The output is a digital mesh that you can rotate, measure, archive, and use as a reference for restoration.

The practical setup at a working studio is unromantic. Decent camera (even a high-end mobile phone with a recent sensor will do for many applications), good lighting (overcast outdoor light works well; controlled studio light works better), patience, and good software. The commercial photogrammetry packages — RealityCapture, Metashape, plus open-source options like Meshroom — have all become genuinely capable.

Time involved: a thorough capture of a half-life-size figure might be 30-60 minutes of shooting and 2-4 hours of processing time on a modern workstation. For ornamental architectural elements it’s faster.

Where it’s adding real value

A few applications have become routine.

Pre-restoration baseline documentation. Before you start work on a damaged or weathered carving, capture it as it exists. The before-state record is invaluable when you’re explaining the work to the client, defending against criticism, or training another carver on what was done.

Restoration reference for missing elements. When a comparable element exists elsewhere on a building — another corbel in a series, another rosette in a frieze — capturing the intact one provides a reference for recreating the missing one. This used to require careful drawn measurement and judgement; the photogrammetry mesh gives you a starting point with much more dimensional fidelity.

Client presentation and approval. For commissions, especially heritage ones where the brief is “match what’s there” or “reproduce this element,” sharing a digital model of the existing work and the proposed addition gives the client something concrete to react to. Approval processes are faster and the conversations are better.

Insurance and archival. Carved work that’s outdoors and weathering, or that belongs to institutions concerned about future loss or damage, is increasingly being documented this way as part of routine maintenance. Several state heritage agencies in Australia now request photogrammetric records as part of conservation reporting.

Studio archive. Your own work, captured before it leaves the studio, becomes a permanent reference for the next time someone asks for “something like that piece you did for [the cathedral].”

Where it doesn’t replace anything

A few things photogrammetry doesn’t do that the old methods do.

It doesn’t capture material qualities. A mesh tells you the shape of the surface. It doesn’t tell you how the stone responds to light, the colour variation across the face, the way the carving sits in its setting. Photographs still matter, observation in person still matters, and notes about material qualities still matter.

It doesn’t substitute for tool-mark study. When you’re restoring a piece by an identifiable carver from a known period, the tool marks tell you a lot about how to match the work. A mesh smooths over the very details that matter to a sensitive restoration. Detailed close-up photography and direct observation still do what the mesh can’t.

It doesn’t replace measured drawings for some uses. Architects and engineers often still want orthographic drawings rather than meshes for setting out and integration. The mesh can produce the drawings, but the drawings remain the working document on most projects.

It doesn’t help with the carving itself. This should be obvious but is worth stating: nobody is carving better stone because they have a photogrammetric mesh. The mesh is documentation and reference; the carving is craft.

The AI-assisted layer

What’s newer in the last 18 months is AI assistance on top of the photogrammetry workflow. A few things that are working.

Automatic mesh cleaning. Removing noise, filling small gaps, smoothing artefacts from the capture. The current software handles much of this with less manual intervention than was required even two years ago.

Comparison and damage mapping. When you’ve captured a piece before and after a period of weathering, or comparing the same architectural element to a sister element nearby, software can flag the areas of difference automatically. This was tedious manual work until recently. Now it’s overnight processing.

Feature recognition for cataloguing. Studios with substantial archives of past work are finding it useful to have software that can identify patterns and elements across photographs and meshes, making the archive searchable in ways it wasn’t before. “Find me every ionic capital I’ve worked on in the last decade” is now a query that works.

A few Australian studios working at the larger end have built custom workflows around these tools, often in collaboration with software developers. The combination of stone-specific expertise and modern AI engineering is rare enough that the partnerships matter. Several consultancies, including Team400.ai’s custom development team and a few heritage-focused specialists, have done credible work in this space. The studios that get value typically have one person internally who understands both the craft side and what the tools can and can’t do. Without that internal anchor, the technology becomes shelfware.

Costs and barriers

The biggest barrier isn’t the cost of the equipment or software — both have come down meaningfully. The barrier is the time to learn the workflow and the discipline to use it consistently.

A studio that captures every significant piece, every significant restoration job, every weathered element they’re working from, builds a compounding archive that’s an asset. A studio that captures things sporadically gets the cost without the asset. The discipline matters more than the tooling.

For studios just starting, my practical advice would be: start with one project, do it thoroughly, learn the workflow on something you care about getting right, and then expand from there. Don’t buy a top-end scanner before you’ve done a serious project with a phone. The mesh quality is rarely the binding constraint.

The Australia ICOMOS heritage practice guidelines have started referencing photogrammetric documentation as appropriate practice for conservation projects. The major university heritage programmes at Canberra and Melbourne have published case studies that are worth reading.

The takeaway

Photogrammetry isn’t going to make anyone a better carver. It’s not the point. The point is that the documentation craft has caught up to a level where the work we do can be preserved, communicated, and referenced with much higher fidelity than was possible in the past. For conservation work in particular, where the relationship between past, present, and future hands on the stone matters, that’s a real benefit. Studios that have embraced this quietly are better positioned for heritage tenders than those that haven’t. The tooling is mature enough that the cost of staying in 1995 is now higher than the cost of moving forward.