Static vs Dynamic ADAS Calibration

Static calibration puts printed targets at measured positions in front of a stationary car so the camera can compare what it sees against geometry the shop controls. Dynamic calibration sends the car down a real road so the camera can work its own aim out from lane markings and motion over several minutes. You do not get to pick. The manufacturer's service procedure for your VIN names the family, the shop follows it, and a growing number of vehicles require both in sequence.

What each procedure is actually doing to the camera

Both procedures write the same thing into the module: a small set of numbers describing exactly how the camera is mounted. How high the lens sits above the road, how far behind the front axle it is, and the three angles it points at (pitch, yaw and roll). The software needs those numbers to turn a pixel into a position. Everything downstream, lane centering, adaptive cruise following distance, emergency braking timing, is arithmetic performed on top of them. Calibration is the act of measuring those numbers again and storing the corrected values. The rest of the mechanism is covered in why moving the glass moves the camera.

Static gives the camera a known answer to check itself against. A target board is placed at a position the shop has measured relative to the vehicle, and the software already knows where the pattern should land on the imaging sensor if the aim were perfect. It compares that expectation with reality and stores the difference as a correction.

Dynamic gives the camera a road and asks it to solve for itself. There is no known object. Instead the module leans on structure that is statistically reliable over enough miles: two lane lines that are meant to be parallel and should therefore converge at a single vanishing point, a horizon that should sit at a predictable row of pixels, and the way stationary objects flow outward through the frame as the car moves forward. Over a qualifying stretch of road it converges on the aim that makes all of those things agree, then commits it.

One sentence holds the difference: static measures aim against something the shop knows, dynamic infers aim from something the road supplies.

Why a fraction of a degree becomes feet of error

This is the reason the procedure is fussy about floors and target stands rather than being a five minute reset. Aiming error is angular, and angular error opens out with distance. An error of half a degree is about eleven inches of sideways error at 100 feet and roughly two and a half feet at 300 feet, which is the range band where adaptive cruise picks a lead vehicle and where emergency braking decides whether the object ahead is in your lane or the next one.

Now look at what half a degree costs at the glass. Across a camera mount roughly four inches long, half a degree of tilt is about three hundredths of an inch of height difference between one end and the other, near enough the thickness of a few sheets of paper. Nothing about that is visible to you, and nothing about it feels wrong when you drive away. That mismatch, invisible at the bracket and large down the road, is the whole safety argument for calibrating every time. The bonding tolerances behind it are covered in what the camera bracket has to hold.

The two axes fail differently. Yaw (pointing left or right) throws lateral position, so lane centering settles off center and sign recognition can read the road next to yours. Pitch (pointing up or down) throws distance, because a single camera infers range partly from where the base of a vehicle meets the road surface in the image. Tilt the camera down a little and the assumed road plane rises, so everything ahead reads as nearer than it is. Tilt it up and traffic reads farther away, which is the direction that delays a braking intervention.

The two families side by side

Where both are specified the bill lands higher, commonly $400 to $800, because you are buying both cost structures back to back. The full pricing picture, including who performs it and whether a claim pays, is in the calibration cost guide, and you can model your own job in the cost estimator.

Which systems tend toward which family

Treat this as a way to set your expectations before you call, not as a specification. Procedures change between model years and between trims of the same car, and the service information tied to your VIN is the only authority.

• Stereo camera systems such as Subaru EyeSight lean heavily on target based work in a controlled bay, for reasons explained in the next section.
• Mono camera systems from Toyota and Honda are commonly dual procedure across recent generations, with a bench setup followed by a confirming drive. See Toyota Safety Sense calibration and Honda Sensing calibration.
• Several domestic and Korean platforms publish a drive based routine for the forward camera on many models while still requiring targets on others in the same showroom. The brand pages for Ford Co-Pilot360 and Hyundai and Kia SmartSense cover how to pin yours down.
• Camera dense vehicles that use several cameras cooperatively, including Tesla vehicles after a windshield, often replace the target board with a guided setup drive that the car itself scores and reports on.

The ADAS calibration section carries a page for each of the major systems. Two vehicles parked next to each other, same badge and same year, can need different procedures because one has a higher assistance package. This is why phone quotes built from year, make and model alone go wrong so often.

Why a road drive replaces targets on some cars and not others

There are two separate reasons, one physical and one commercial, and it helps to keep them apart.

The physical reason is what the system needs to recover. A single camera that models the road as a flat plane with lanes of a fairly standard width has enough built in assumptions to solve its own aim from ordinary driving. The scene supplies the constraints. A stereo pair is different: it measures depth from the tiny disparity between two lenses held at a fixed spacing, so it needs its geometry re-established against something at a known distance. Roads rarely hand a car an object of known size at a known range, which is why stereo systems tend to want a target. Systems that fuse camera data with a radar have an independent range reference already on board, and that extra source is part of why some of them can be trusted to learn on the move.

The commercial reason is whether the routine exists at all. A manufacturer has to write, validate and publish a learn procedure, and expose it through the diagnostic interface. Where that work was not done, targets are the only sanctioned path even on a vehicle that could in principle self-solve. That is why "my friend's car just needed a drive" tells you nothing about yours.

What a static procedure needs from your shop

• A floor that is genuinely flat and level. Not "close enough". Most bay floors slope toward a drain, and a slope tilts the whole car, which tilts the camera against a target that is not tilted with it.
• Clear length and width around the car. The target stand sits ahead of the vehicle at a measured distance, and other vehicles, parts racks and roll up doors have to be out of the working area.
• Even lighting and a plain background. Direct sun, a window behind the target or a shiny toolbox in frame can wash out the pattern the camera is looking for.
• Setup against the thrust line, not the body. The target is squared to the direction the car actually travels, which is set by the rear axle. A car with a rear alignment problem can be geometrically incapable of passing until the alignment is corrected.
• Ride height at spec. Tire pressures set, fuel level as specified, cargo and roof loads out, nobody sitting in the car. All of it changes the angle at which the camera looks down the road.
• Targets in good condition. A creased, faded or coffee stained board is a bad reference, and the equipment needed to keep a bay honest is set out in what a shop needs to calibrate your car.

What a dynamic procedure needs from the road

• Painted lines the camera can actually see. Faded, patched or snow covered markings starve the routine of the structure it is solving from.
• A speed the technician can hold steadily, inside whatever window the manufacturer sets for your model, for as long as the routine asks.
• Dry conditions and usable light. Spray, glare and dusk all degrade the image the module is learning from.
• Traffic that cooperates. Stop and go driving resets progress on many systems, and some routines want a vehicle ahead to track for part of the drive.
• A technician who knows a qualifying route. Shops that do this daily have one, and that is a fair thing to ask about.

A dynamic calibration that could not complete is not a partial success. If conditions defeat it, the correct outcome is a second attempt in better conditions, not a car handed back with the routine still running. What to do when either family will not finish is covered in why calibrations fail.

Working out what your car needs, before you book

1. Confirm you have a camera at all. Look behind the rearview mirror for a housing with one or two forward facing lenses. The full trigger list, including work other than glass, is in when calibration is required.
2. Ask the shop to look up the procedure by VIN, not by model, and to tell you the family: static, dynamic or both.
3. Ask where the car goes. A specific bay on site, or a named calibration partner, are both fine answers. "It relearns as you drive" is not.
4. Ask what happens if it fails, including who pays for a second attempt or different glass. That conversation belongs before the work, not after.
5. Ask for the paperwork by name. Pre-scan, post-scan and a certificate naming the procedure performed.

Once you know which family applies, two pages take you further: the step by step walkthrough of a calibration so you can follow along on the day, and how to choose a glass shop so the calibration question is settled before the glass is ordered rather than after.

Frequently asked questions

Is dynamic calibration less accurate than static?

Neither is better in general. Each is the correct method for the systems it was designed around. What matters is doing the one your vehicle's service procedure specifies. A dynamic drive substituted for a required static procedure is not a cheaper option, it is an incomplete job, and the reverse can also fail to satisfy the routine.

Why do some cars need both procedures?

Because the two do different work. The static stage sets the aim against known geometry in controlled conditions, and the drive confirms that the result holds against real lane markings and moving traffic. Manufacturers that specify both treat the road stage as verification rather than as the calibration itself.

Can static calibration be done in my driveway?

Usually not. Driveways slope for drainage, and a slope tilts the car relative to a target that is not tilted with it. Mobile static calibration is real but only where the surface is genuinely flat and level, with room ahead of the car and no direct sun on the target. Ask the operator how they verify the surface.

How do I find out which procedure my car needs?

Ask the shop to pull the procedure by VIN rather than by year, make and model. Trim level changes the answer, since two identical looking cars can carry different assistance packages. A dealer service department can also confirm what the factory information specifies for your build.

The shop says my car recalibrates itself as I drive. Is that true?

It is true only if the manufacturer publishes a drive based learn routine for your vehicle, and even then a technician normally has to start that routine with a scan tool. No mainstream system silently re-aims itself after glass work with no procedure at all. Ask for the procedure name in writing.