← Back to Blog
GuideJun 26, 202610 min readBy Lidar Scanner Team

USDZ vs OBJ vs RoomPlan JSON: 3D Scan File Formats Explained

You finish a LiDAR scan, tap export, and a list of formats appears: USDZ, OBJ, RoomPlan JSON. If you've ever paused on that screen wondering which one you actually need, this guide is for you. Picking the wrong format usually isn't fatal — but it can mean an extra conversion step, a model that won't open where you wanted it, or losing the structured data that made the scan useful in the first place.

The short version: these three formats represent three fundamentally different ways of describing the same room. USDZ is a finished 3D scene built for viewing and AR. A textured OBJ is a raw mesh built for editing. RoomPlan JSON is the underlying data — the parametric description of the space itself. Understanding that distinction makes the choice obvious. Let's walk through each one.

The Big Picture: Scene, Mesh, and Data

Before the format-by-format breakdown, it helps to understand what an iPhone LiDAR scan actually produces under the hood.

When you scan a room with Lidar Scanner, ARKit's RoomPlan engine doesn't just capture a cloud of points. It understands the room — it recognizes walls, doors, windows, openings, and furniture as distinct objects with dimensions and positions. That semantic understanding is the special sauce, and it's why the three export formats differ so much:

  • USDZ packages the result as a tidy, viewable 3D scene — geometry plus materials, ready to drop onto your desk in AR or open in design software.
  • OBJ (textured mesh) gives you the raw surface geometry as triangles, with the captured texture, for when you want to edit, sculpt, or remesh.
  • RoomPlan JSON hands you the parametric data itself — the list of walls and objects with their measurements — so the scan can be re-imported, analyzed, or rebuilt programmatically.

Same scan, three lenses. Here's when to reach for each.

USDZ — The AR and Presentation Format

USDZ is Apple's 3D format, built on Pixar's Universal Scene Description (USD) framework. If you've ever tapped a 3D model link on an iPhone and watched it appear on your floor through the camera, that was USDZ and AR Quick Look working together. It's a single compressed archive that bundles geometry, materials, and textures into one tidy file.

For LiDAR room scans, Lidar Scanner generates a parametric USDZ — a clean scene where the walls, openings, and furniture are organized as recognizable elements rather than an undifferentiated blob of triangles. That structure is what makes it pleasant to work with downstream.

Where USDZ shines

USDZ's killer feature is native AR and viewing support across Apple's ecosystem, plus surprisingly broad acceptance in professional 3D and CAD tools:

  • AR Quick Look: Tap a USDZ in Safari, Messages, Mail, or the Files app and it opens directly in AR — place your scanned room or object in real space, no extra app required. Perfect for showing a client a captured space on their own iPhone.
  • Reality Composer: Build AR scenes and experiences using your USDZ scan as a component.
  • SketchUp, Rhino, and Revit: These import USDZ (some natively, some via readily available plugins or USD support), so a scanned room can become the base context for architectural and design work.
  • macOS Quick Look: Press space on a USDZ in Finder and preview the full 3D scene instantly.

Because the USDZ from a room scan is parametric and well-organized, it lands in these tools as something you can actually reason about — a room with walls and openings — rather than a mesh you have to interpret.

USDZ trade-offs

The main limitation is ecosystem fit. USDZ is happiest in Apple-and-design contexts; it's not the universal mesh-editing format that OBJ is, and it's not the structured data that JSON is. If your next step is heavy mesh surgery in a sculpting tool, or feeding measurements into a script, USDZ isn't the most direct route. But for viewing, presenting, and importing into design and CAD software, it's the first format to reach for.

RoomPlan JSON — The Data Format

This is the format most people overlook, and it's arguably the most powerful of the three for the right job. RoomPlan JSON isn't a 3D model you look at — it's the parametric description of the room as structured data: the walls, doors, windows, openings, and detected objects, each with their dimensions, positions, and relationships, written out in plain JSON.

Why parametric data is a big deal

A mesh tells you where the surfaces are. RoomPlan JSON tells you what the room is. That difference unlocks things a mesh can't:

  • Re-import and rebuild: Because the JSON describes the room parametrically, the scan can be brought back in and reconstructed — walls as walls, not as frozen triangles. Lidar Scanner can import RoomPlan JSON back into your scan library, which means your captures are portable data, not one-way exports.
  • Data portability and pipelines: JSON is universal. Any language, any tool, any database can read it. If you're building a workflow — feeding room dimensions into a space-planning system, populating a property database, or running your own analysis — the JSON is the clean, machine-readable source.
  • Future-proofing: A mesh is a snapshot of one moment's geometry. The parametric data is the model. Keeping the JSON means you keep the option to regenerate plans, models, or measurements later.

RoomPlan JSON trade-offs

The flip side is that JSON is not a viewable model. You can't AirDrop it to a client and expect them to see their living room — they'll see code. It's the format for systems and workflows, not for presentation. Think of it as the master record you keep, from which the viewable formats are derived. RoomPlan JSON export and import are part of the Pro feature set; see pricing for the details.

OBJ and the Textured Mesh — The Editing Format

OBJ is one of the oldest 3D formats still in heavy use, created by Wavefront in the 1980s and supported by essentially every 3D application on the planet. Its enduring appeal is simplicity and universality: if a tool touches 3D geometry, it reads OBJ.

Lidar Scanner can produce a textured mesh of your scan — the raw captured surface as a dense triangle mesh with the photographic texture mapped onto it. Where the parametric USDZ gives you a clean, idealized room, the textured mesh gives you the actual scanned surface, bumps and all, which is exactly what you want for certain jobs.

Where the textured mesh shines

  • Mesh editing and cleanup: Import into Blender (free and open source), Maya, 3ds Max, ZBrush, or Cinema 4D to retopologize, decimate, sculpt, or repair the geometry. OBJ imports seamlessly into all of them.
  • Capturing reality, warts and all: When you need the true surface — a textured record of an existing condition, an as-found object, a heritage piece — the mesh preserves what was really there rather than an idealized reconstruction.
  • Object Capture workflows: Lidar Scanner also includes Object Capture, a guided turntable photo workflow that produces photogrammetry-style textured 3D models of individual objects. Those models live naturally in the mesh world, where a textured OBJ is the standard currency for editing and finishing.

Textured mesh trade-offs

A mesh is "dumb" in the helpful sense and the limiting sense: it knows where the surface is but not that the surface is a wall. You lose the parametric structure, so you can't easily re-derive a floor plan or pull clean dimensions from it the way you can from the JSON. OBJ also typically travels as multiple files (the .obj geometry plus a material file and texture images), so keep them together. Use the mesh when you want to work the geometry; use USDZ or JSON when you want structure.

Side-by-Side Comparison

USDZ (parametric scene)RoomPlan JSON (data)OBJ (textured mesh)
What it isViewable 3D sceneParametric room dataRaw surface mesh
Primary useAR, presentation, design importRe-import, pipelines, analysisMesh editing, sculpting
AR / Quick LookYes, nativeNoNo
Editable as meshLimitedNoYes, fully
Re-importable to Lidar ScannerYes
Machine-readable dataStructured sceneFully (plain JSON)Geometry only
SoftwareAR Quick Look, Reality Composer, SketchUp, Rhino, RevitAny language/tool that reads JSONBlender, Maya, 3ds Max, ZBrush, Cinema 4D
Best whenSomeone will view itA system will consume itYou will edit the geometry

Which One Should You Pick?

Map the format to your next action, not to a vague sense of which is "best."

You want to show someone the space in AR, or import a clean room into SketchUp / Rhino / Revit. Choose USDZ. It opens in AR Quick Look with a tap and lands in design and CAD tools as an organized scene.

You're building a workflow, populating a database, or want to keep the master record so you can re-import or regenerate later. Choose RoomPlan JSON. It's the structured, portable, machine-readable source of truth — and Lidar Scanner can import it back in.

You need to edit, remesh, repair, or sculpt the geometry — or you captured an object with Object Capture. Choose the textured mesh (OBJ). It's the universal currency of 3D editing tools.

You're not sure and you want options. Export more than one. The scan is the same; the formats are just different views of it. A common pattern: keep the RoomPlan JSON as your archival master, export USDZ for the client-facing AR preview, and pull a textured mesh only when a specific edit demands it. If your scans are headed for 2D drawings instead of 3D, see how to export iPhone floor plans to CAD for the DXF, PDF, and SVG side of the export menu.

How to Export from Lidar Scanner

The export itself is the same simple flow regardless of which format you land on:

  1. Scan your room or object. Use live RoomPlan capture for a room, following the on-screen guidance, or the guided Object Capture turntable flow for a single object. If you're new to it, how to scan a room in 3D with iPhone LiDAR covers the capture basics, and the scanning tips guide covers getting clean, accurate results.
  2. Open the scan from your library. Every capture is saved locally and can be browsed, renamed, and re-exported anytime.
  3. Choose export and select your format — USDZ, RoomPlan JSON, or textured mesh.
  4. Share via the native iOS share sheet. AirDrop to your Mac, save to Files or iCloud Drive, email it, or open it in another app.

Everything is processed on-device — your scans are private and the export works offline, with no upload or cloud round-trip. High-Fidelity capture, textured mesh, RoomPlan JSON, and CAD exports are part of the Pro tier; the full breakdown lives on the pricing page, and the how it works walkthrough shows the capture-to-export pipeline end to end.

The Short Version

Three formats, three jobs. USDZ is the finished scene you show people and import into design and CAD tools. RoomPlan JSON is the parametric data you keep, re-import, and feed into systems. The textured mesh (OBJ) is the raw geometry you edit. They're not competitors — they're three ways of carrying the same scan to three different destinations, and the right pick is simply wherever your scan is headed next. Browse the full feature set to see how capture, the floor-plan editor, and these exports fit together.

Lidar Scanner is available on the App Store. If you want to scan a room once and walk away with an AR-ready USDZ, an editable mesh, and the underlying RoomPlan data all from a single capture, download it from the App Store.

Ready to scan your space?

Download Lidar Scanner and turn any room or object into a precise 3D model — all on-device.

Download on the App Store