Reality Capture in the Studio

The first and most important studio skill is not operating a scanner - it is routing the job correctly. For any given capture task there are really three destinations: do it yourself in-house, commission a scanning specialist, or hand it to a licensed surveyor. Choosing well saves money, avoids embarrassment, and keeps you on the right side of the professional line. Choosing by habit - "we always scan it ourselves" or "we always outsource" - is how studios either take on work they cannot deliver or pay for work they could have done in an afternoon.

A clean way to decide is to ask three questions in order. First: is the result binding, survey-grade or georeferenced? If the deliverable must stand up legally or structurally - a boundary or cadastral position, a setting-out grid, deformation or structural monitoring, a survey tied accurately to real-world coordinates - then it belongs to a licensed surveyor or geospatial professional, full stop. No amount of in-house enthusiasm changes that, and knowing it is part of being a professional rather than a hobbyist. Second, if it is not binding: is the job large, complex, or does it demand high accuracy? A whole multi-storey building, a site with difficult access, an intricate heritage facade, or any job where a few millimetres of error would be costly is usually work for a scanning specialist with professional terrestrial equipment, proper control and the experience to deliver a checked, registered dataset. Third, if it is none of those: is it a small space, an as-built record, or a visual reference at modest accuracy? A single room, a quick fit-out verification, a shell to design joinery against, a space to visualise - these are exactly where in-house self-capture with a phone, a handheld device and photogrammetry earns its keep.

Two rules keep the routing honest. When in doubt, commission - the cost of an outside capture is almost always less than the cost of designing against a dataset you could not really trust. And whatever the route, verify the deliverable against an independent check: a few tape or laser-distance measurements of known lengths, compared against the same features in the data, will catch a gross error before it reaches the drawings. The point of the decision is to match the method to what the job actually demands - its accuracy, its access and its risk - rather than to whatever equipment happens to be in the cupboard.

For the self-capture tier, you do not need a lakhs-worth laser scanner to start being useful. A genuinely lightweight capability, built around accessible tools and honest about its limits, covers a surprising amount of everyday studio work. The anchor is usually one of two things: a phone or tablet with a LiDAR sensor and a good scanning app, or a photogrammetry workflow that turns ordinary overlapping photographs into 3D data. Many studios run both, using the phone for instant rough scans and walk-throughs and photogrammetry when they want better detail or texture on a space or an object. A modest handheld SLAM scanner is a sensible next step up for firms doing a lot of interior or small-building work, but it is an upgrade, not a prerequisite.

The capability is more than the capture device, though, and this is where lightweight setups succeed or fail. You need somewhere to process - a reasonably capable computer and the photogrammetry or scan-processing software, which is increasingly available at accessible or even free entry tiers. You need simple, repeatable capture discipline: enough coverage to avoid gaps, overlap between photos or scan positions, a scale reference or known dimension in the scene, and a note of the conditions. And you need at least one person who owns the workflow - who knows how to capture a space well, process it, spot when it has gone wrong, and model or measure from it. That human skill is the scarcest and most valuable part of the whole capability; the hardware is the cheap bit by comparison.

Start deliberately small. Pick one recurring job your studio does - a room survey before a fit-out, an as-built of a space you are renovating, a record of an existing facade - and make in-house capture the standard way you do that one thing. Build the routine on real work until it is reliable, then widen it. Resist the urge to buy the most impressive device first; buy the cheapest tool that answers the questions your jobs actually ask, learn its limits cold, and scale the method to the work. A phone used confidently on the right jobs delivers far more value than an expensive scanner nobody has learned to trust. And keep the line clear: this capability handles modest, non-binding work, and the moment a job outgrows it - in size, in accuracy, or in legal weight - you commission rather than push the little tool past what it can honestly do.

A capture that never leaves the laptop of the person who made it has changed nothing. The value appears only when the data reaches the people making decisions, on a basis they can use. So the real integration question is: how does a point cloud or a mesh become part of how your team already works? For most studios the answer runs through BIM. The captured existing conditions become the reference the model is built against - the subject of this whole course's scan-to-BIM half - so that the design model sits on top of an accurate record of what is really there, and the out-of-plumb wall, the sagging floor and the beam nobody drew are in the data from the start rather than discovered on site.

Getting this to work in practice turns on a few unglamorous disciplines. Agree a coordinate basis so the capture, the model and any site survey share one frame and line up rather than floating in separate spaces - even a simple, consistent local origin is far better than none. Decide the level of detail and accuracy the model actually needs before anyone starts modelling, so you capture and build to the job rather than lavishing survey-grade effort on a visual reference or, worse, trusting a rough scan where precision matters. Keep the point cloud itself available alongside the model, not just the tidied geometry, because the cloud is the measured evidence you return to when a dimension is questioned. And manage the data: point clouds are large, so agree where they live, how they are named, and who can reach them, or the capability will drown in unshareable files (Module 5's theme).

Just as important is the human integration. Capture is only routine if the team knows it exists and trusts it. That means the person who captures and the people who design and coordinate talk to each other: the designer says what the model needs to answer; the capturer says what the data can and cannot support and where the gaps and occlusions are. A captured dataset handed over with a plain note of its accuracy, its coverage and its blind spots is worth far more than a silent, beautiful file that the next person over-trusts. The aim is a shared, measurable basis that the whole team works from - one version of what is really there - rather than each person carrying their own assumptions about the existing building into the project.

Routine is the whole game. A capability used once is a novelty; a capability used on every suitable job is a genuine advantage, and the difference is a handful of durable habits rather than any single clever trick. The backbone is a repeatable loop that the studio runs the same way every time: plan the capture to the accuracy the job needs, capture on site with good coverage, process and clean the data, register and check it against an independent measurement, bring it into the model, and share it with the team on the agreed coordinate basis. Running the same loop each time is what turns capture from an anxious improvisation into a normal first step, and it is what lets the skill compound: every job makes the next one faster and more trustworthy.

A few specific habits carry most of the weight. Plan before you capture - even five minutes deciding what accuracy the job needs, where to stand or shoot, and what could be occluded saves an expensive return visit. Always take a check measurement - a known length or two, recorded independently, so you can confirm the data is right rather than merely detailed; of the six steps in the loop, this is the one people skip and the one that keeps you honest. Record the context - date, conditions, what was and was not captured - because a scan without that note loses half its value the moment memory fades. Name and store the data consistently so it can be found and reused rather than lost. And debrief briefly after jobs that went wrong, so the mistake becomes a fixed step in the workflow rather than a lesson relearned the hard way.

There is also a habit of restraint, and it matters as much as any technique. A studio that has invested in capture feels a pull to use it everywhere, including on jobs that are too big, too precise or too binding for its tools. Resisting that pull - knowing the decision rule and honouring it, commissioning a specialist or a surveyor when the job demands it - is what separates a capable practice from an overreaching one. Reality capture becomes a true studio asset not when you own the cleverest device, but when capture is a normal, well-judged, checked part of how you start work, and when everyone knows both what it gives you and where it stops.