Oblique photography at scale: where GenAnywhere fits a review
When one site becomes twelve, the problem stops being the capture and becomes the pack: naming, consistency and the cost of opening the file.
A single capture reviewed on its own is a manageable object. A programme of twelve sites, each with a flight, a model and a note, is a different problem. The evidence is the same kind; the difficulty is keeping it comparable, findable and light enough to use.

1 What scale changes
1.1 Why does scale break a method?
Because the small decisions that nobody records become expensive when repeated. A file named by habit is fine once and chaotic twelve times.
Because consistency replaces detail as the thing reviewers notice. A pack where one site is photographic and the next is a grey block invites the room to compare the wrong attributes.
Because access becomes the constraint. Files that open quickly on the author's machine may be unusable on a reviewer's laptop, and a pack nobody can open is not evidence. The pack becomes a small information system, and information systems fail at their edges: names, versions and access.
1.2 What stays the same?
The questions do. Every site in a programme still needs a date, a coverage boundary and a statement of what the record does not cover. Those three items are the minimum for any record, whatever the number of sites.
The separation does. Oblique photography remains surface evidence, and a base remains arrangement, whatever the number of sites.
The discipline does. Scale multiplies the value of a naming convention and the cost of skipping one. A programme also makes the exceptions visible: the site with no capture, the one with an old base, the one nobody has visited. Standards bodies such as the Open Geospatial Consortium publish conventions for describing geospatial datasets that are worth borrowing when a pack has to be indexed.
2 Set the unit of review
2.1 One base per site or one corridor?
Per site when the decisions are separate. A programme of individual planning applications is easier to manage as separate scenes with a shared convention.
Per corridor when the decisions interact. A transport route or a river frontage often needs one continuous base, because the argument is about relationship rather than any single plot.
Choose before the work starts. A pack that mixes the two units is hard to compare and harder to hand over. Write the choice at the top of the index, so the unit is not renegotiated at every meeting.
2.2 Naming and versions
Use a pattern that carries the site, the date and the version, and apply it to every file in the programme.
Keep the capture and the base distinguishable in the name. A reviewer should know which file is evidence and which is a model before opening either. That distinction is harder to maintain across twelve sites than across one, which is why the name matters more than the folder.
Retire old versions to a dated folder rather than overwriting. A pack with three files called "final" has no final. A name that carries the site and the date is easier to search than a folder tree that its author alone understands.
3 Build a consistent base
3.1 Draw and edit
Draw each area with the same margin beyond the site boundary. Consistency at the edges makes the scenes comparable when they are shown side by side. The margin is also a reminder that every site has neighbours.
Apply the same layer discipline everywhere. Buildings, terrain, water and vegetation separate in every scene, so a reviewer who learns one pack can read the next.
Record the edits in the same place in every note. A replacement or a height change should be findable without hunting. Consistency of edge treatment matters as much as consistency of detail.
3.2 Layers across sites
GenAnywhere builds each site model from map and satellite data, with buildings, terrain, water and vegetation on separate layers, and exports a general GLB. Applied across a programme, that gives every scene the same structure.
The same structure is what allows a reviewer to compare sites without relearning the interface each time. A reviewer who knows where the terrain layer sits can check twelve sites in an afternoon. The model keeps buildings, terrain, water and vegetation apart, and the GLB export preserves that structure for the receiving tool.

It also allows a programme lead to check one site quickly, because the layer that matters is always in the same place. Where a site departs from the standard, note the departure rather than hiding it in the scene.
4 Compare across sites
4.1 Same viewpoint
Capture each site from a comparable direction and height where the argument allows. A pack where every site is shown from the south makes comparison possible. Where the terrain makes that impossible, note the exception in the index.
For map-derived scenes, fix a standard camera for the programme. A consistent view is a quiet form of fairness.
Where a site needs a special view, add it rather than replacing the standard one. The standard view keeps the pack coherent. Note the camera settings in the index so the view can be reproduced.
4.2 Same level of detail
Decide the level once. If the programme compares massing, every scene should show massing; a single photographic scene will dominate attention.
Keep texture for the sites where surface condition is part of the decision, and say so in the note. A single photographic scene in a programme of block models will draw the eye and distort the comparison.
Resist the temptation to improve one scene for a meeting. A pack that is uneven is read as advocacy even when it is not. An uneven pack also makes the improved site look better than it is.

5 Manage the pack
5.1 Index
Keep a single index listing the site, the capture date, the base date, the version and the note. Twelve sites is the point where memory stops working.
Link each entry to its files rather than duplicating them into folders. Duplicates drift.
Update the index when a scene is replaced. An index that lags is worse than none, because people trust it. The index should be readable without opening any scene.
5.2 Access
Check the file sizes against the machines that will open them. A programme reviewed on site laptops needs scenes that load there.
Keep the note with the scene in the same package. A scene separated from its note loses its context within a week. A package that opens without a call is the sign of a clean handover.
Decide who can edit. A shared scene edited by everyone becomes a scene nobody can explain. A pack that opens on the reviewer's machine is worth more than one that opens on the author's.
6 Keep it usable
6.1 When to split
Split a scene when it covers more ground than the question needs, or when the file size stops people opening it.
Split when two audiences need different levels of detail. A committee view and a design team view rarely belong in one file. Splitting is not a failure of the method; it is how the method survives a large programme.
Split when the sites are genuinely separate decisions. Combining them saves file count and costs clarity. A scene that covers three sites and answers none of them is worse than three small scenes.
6.2 When to drop detail
Drop texture where the decision is about arrangement. A programme comparing heights does not need façades.
Drop vegetation where it hides the ground and the ground is the subject.
Drop the far context where it adds weight without answering a question. The level of detail should follow the decision, not the available data.
6.3 When to stop
Stop when the pack can be read without the author present. That is the practical test of a handover at scale. If a colleague needs the author to narrate it, the index or the notes are still doing too little.
Stop when the index is accurate and the versions are retired. Further tidying has diminishing returns. A programme lead should be able to walk a new reviewer through the pack in ten minutes.
Draw each area in the GenAnywhere app and keep one convention across the programme.
7 FAQ
FAQ-1 Can GenAnywhere handle a programme of sites?
It builds one site model per drawn area, and each exports a general GLB. A programme is managed by applying the same drawing, layer and naming conventions across areas rather than by a single project container. That convention is a human agreement, and it is worth writing down. It is the part that makes a programme feel like one project instead of twelve.
FAQ-2 How many sites can a review pack hold?
As many as the audience can compare in one sitting. Beyond about a dozen, an index and a consistent camera matter more than any additional scene, and splitting into stages usually helps. A short index and a standard camera do more for comparability than another dozen scenes. A stage boundary is also a natural moment to retire old versions.
FAQ-3 Do I need a capture for every site?
No. Sites where the decision is about arrangement can run on the map-derived base. Capture the sites where surface condition or recent change is part of the argument. Where a site changes quickly, a capture still earns its place. Where the base is old, a visit is the way to confirm the change before the comparison is made.
FAQ-4 How do I keep file sizes manageable?
Review at the level of detail the decision needs, split scenes by audience, and export just what the receiving tool will use. A scene that opens quickly is used more often than one that is richer and slower.
See how a map area becomes an editable site model in the GenAnywhere workflow.
Content and technical notes
This article was prepared by the GenAnywhere team with AI-assisted drafting and human review.
Product features, formats, and availability change as the platform evolves; refer to the official GenAnywhere website for current details.
For technical questions or corrections, contact support@genanywhere.com.
Last reviewed: 2026-11-06.