Announcement Text
.png)
Contractors keep asking the same question: is iPhone LiDAR actually accurate enough for construction work?
The answer depends on what you mean by "iPhone LiDAR." The raw output from the sensor is not professional-grade. The output from properly processed iPhone LiDAR — with drift correction, motion filtering, and cloud-side alignment — is accurate to within half an inch on rooms up to 40 feet.
Here's what actually determines accuracy, what to expect on a real job site, and how Manifold's processing pipeline gets useful measurements out of consumer hardware.
The LiDAR Scanner first appeared on the iPad Pro in 2020 and the iPhone 12 Pro later that year. It's a time-of-flight sensor that fires invisible infrared pulses and measures how long they take to bounce back. Apple calls it a "direct time-of-flight" sensor, which is technically the same family as the multi-thousand-dollar terrestrial LiDAR scanners used in surveying — but with very different specs.
The phone's sensor:
A traditional terrestrial LiDAR scanner — a Leica BLK360 ($16,000) or a Faro Focus ($45,000+) — fires millions of pulses per scan, mounts on a tripod for stability, and produces sub-millimeter accuracy at 10+ meter range. The iPhone produces useful approximations of geometry, not survey-grade truth.
Independent testing is consistent: raw iPhone LiDAR output is accurate to roughly ±2-4 inches across a 20-foot room when used with the default Apple frameworks (RoomPlan, Object Capture). That's not professional accuracy. It's better than a tape measure for ballpark dimensions, but no commercial inspector or surveyor would sign off on it.
With proper processing — drift correction, motion filtering, and registration — that drops dramatically. Manifold's Orbit Measure pipeline produces output accurate to within ±0.5 inches across rooms up to 40 feet. That's good enough for renovation as-builts, restoration documentation, equipment surveys, and almost every contractor use case short of MEP coordination on a high-rise.
The gap between "raw" and "processed" is where most of this article lives. The raw sensor isn't the problem — it's what most apps do (or don't do) with the data.
Drift is the slow accumulation of tracking error as the phone moves through space. Every scan position is calculated relative to the previous position. If each measurement has a tiny error, those errors compound — by the end of a room, you can be inches off in absolute coordinate space.
This is the single biggest source of error in mobile scanning. Without correction, a 30-foot room might come out as 30 feet 4 inches at one wall and 29 feet 9 inches at the opposite wall. The numbers feel close but they're not consistent — and that breaks any downstream measurement work.
How Manifold corrects it: Cloud-side registration. Every scan is uploaded and aligned against itself using a globally consistent solver, the same general approach professional terrestrial LiDAR uses. Drift is eliminated rather than accepted.
LiDAR fires pulses at 60 Hz. If the phone is moving while a pulse is in flight, the depth calculation has a moving reference frame and the result is wrong. The faster the movement, the larger the error.
This is also why "scan a room in 10 seconds while walking quickly" apps produce noisy meshes. They're capturing motion-corrupted data.
How Manifold corrects it: Orbit Measure's capture flow asks the user to hold the phone still over each measurement target. The app actively waits for stillness before sampling. By separating motion from data capture, the per-sample error drops by an order of magnitude.
LiDAR depends on light bouncing back. On dark, glossy, or transparent surfaces, light scatters or passes through — and the sensor either gets no return or a wrong return. Glass walls, polished concrete, mirrors, dark cabinetry, and bright sunlit surfaces all produce holes or noise in the depth map.
How Manifold corrects it: Machine learning fills small gaps using surrounding geometry. Larger gaps (a full glass wall) require the user to capture the wall plane from a different angle, which the app prompts for during scanning.
Bright sunlight overwhelms the infrared return. This is why iPhone LiDAR struggles outdoors — the sensor can't see its own pulses against ambient solar IR. Shaded outdoor areas work; direct sun does not.
How Manifold corrects it: It doesn't, fully — this is a hardware limit. Orbit Measure works best for interior spaces and shaded exteriors. For sunlit outdoor scanning, photogrammetry-only modes are more reliable.
The honest summary: a Leica BLK360 is a better instrument. It's also a $16,000 instrument that requires training, takes longer per room, and for 90% of contractor field documentation work produces accuracy that nobody actually needs. The iPhone with proper processing is good enough for the work most field teams actually do.
Where traditional LiDAR wins: sub-millimeter accuracy at 10m+ range, perfect handling of sunlight and reflective surfaces, full-site capture in single coordinate space. Where it loses: $16,000–$45,000 hardware cost, $1,500–$5,000/year software, requires trained surveyor, 5–15 minutes per setup.
Where iPhone LiDAR wins: $999 hardware, anyone can use it, 60 seconds per room, fits in a pocket. Where it loses: ±2–4 inch raw accuracy, struggles with glass and direct sun, limited to ~5m range.
Where Manifold-processed iPhone LiDAR sits: ±0.5 inch accuracy, $24/user/month, 60 seconds per room, anyone can use it. Same limits on sun and glass as raw, but cloud processing closes the accuracy gap.
Renovation as-builts: ✅ iPhone LiDAR is sufficient. Half-inch accuracy is more than enough for renovation work. You're documenting what's there before you tear it out, and the dimensions you need are within consumer-grade tolerance.
Restoration damage documentation: ✅ iPhone LiDAR is sufficient. You're capturing existing conditions for insurance, scope, and dispute protection. Photo-tagged 3D records are the deliverable, not survey-grade coordinates.
HVAC equipment surveys and commissioning: ✅ iPhone LiDAR is sufficient. Equipment positions, room dimensions, duct paths — all within tolerance. Photo evidence linked to a 3D context is the win.
Construction handover and punch lists: ✅ iPhone LiDAR is sufficient. You're documenting completion against scope, not certifying as-built coordinates for a structural engineer.
MEP coordination on a high-rise: ❌ Use professional LiDAR. Sub-millimeter clash detection across vertical stacks is what BLK360-class instruments are built for.
Land surveying or boundary work: ❌ Use professional surveying equipment. Different problem entirely. iPhone LiDAR can't see far enough or accurately enough for survey-grade exterior work.
High-precision factory layout: ❌ Use professional LiDAR. Tolerances tighter than half an inch matter here.
The LiDAR sensor is identical on iPhone Pro and iPad Pro. iPad's larger form factor makes it easier to hold steady, which marginally reduces motion error during capture. Beyond that, no meaningful accuracy difference.
For most field crews, the iPhone wins on practicality — it's already in everyone's pocket. iPad is better for office processing and review, not job-site capture.
There are dozens of consumer LiDAR scanning apps in the App Store. Most produce noisy, drift-affected output that's fine for hobby capture but not for documentation work. The reason is what they do (and don't do) after capture:
Manifold's pipeline runs on the cloud after upload, which is what makes the half-inch tolerance possible. It's also why Orbit Measure works on phones that don't have LiDAR at all — the same processing approach handles photogrammetry-derived geometry.
For 90% of field contractor use cases, yes — when the data is properly processed. Raw LiDAR is not professional-grade. Cloud-processed LiDAR with drift correction, motion filtering, and ML gap-fill produces output that's reliable for as-built documentation, renovation records, restoration evidence, equipment surveys, and punch list documentation.
For the remaining 10% — survey-grade work, BIM coordination, sub-half-inch tolerance applications — you need a real LiDAR scanner. Don't try to make consumer hardware do work it wasn't built for.
Manifold's Orbit Measure handles the LiDAR processing pipeline described above, plus the photogrammetry version for non-LiDAR phones. Floor Plan Scan generates structured 2D floor plans from iPhone 12 Pro and newer.
Start a free trial — no credit card required — or book a 15-minute demo.
Compare Manifold
Real pricing, honest comparisons. No spin.