What is an Autofocus System? An In-Depth Look at Its Detection Methods and VCM Technology

The Question Every Embedded Vision Engineer Asks 

Here is a question that comes up in almost every embedded camera evaluation: "What is autofocus, and do I actually need it?" It seems simple. The answer is not. 

You may have seen a barcode scanner at a logistics terminal to miss every other package on a fast-moving conveyor. Not because the software failed. Because a fixed-focus lens, calibrated for one distance, was being asked to read targets at three distances. An autofocus camera solves that problem in under 300 milliseconds quietly, invisibly, every single time. 

What is autofocus at its core? It is a system that moves your camera lens to the one position that produces the sharpest possible image of your subject automatically, without human input, in real time. That is the short answer. 

The longer answer to the one that determines whether your embedded system performs in the field involves detection methods, actuators, hybrid architectures, and VCM physics. This guide covers all of it. 

What is Autofocus and Why Fixed-Focus Falls Short 

Fixed-focus cameras are genuinely excellent tools. When your subject distance is known, controlled, and constant, a fixed-focus lens correctly set at the factory delivers sharp images with zero latency and no moving parts to fail. 

The moment the subject distance varies, though, fixed focus becomes a liability. Medical endoscopes probe at changing depths. Retail kiosks serve customers of different heights. Industrial inspection cameras sweep between close-up surface detail and wider context shots. 

In each of these cases, auto focus is not a nice-to-have. It is a system requirement. Without it, you are shipping a product that works in a lab and struggles in the field. 

Modern autofocus technology achieves focus lock in under 300 milliseconds for phase-detection systems, and 500 to 800 milliseconds for contrast-detection, fast enough for real-time embedded applications when the right method is matched to the right task. 

How Does Autofocus Work? The Three Detection Methods 

How does autofocus work? Every camera autofocus system rests on one of three detection strategies. Each has its own physics, its own speed profile, and its own failure mode. Understanding all three is what separates an informed design decision from an expensive field of revision. 

Contrast Detection Autofocus (CDAF) The Hill Climber 

CDAF is the oldest and most widely deployed method. The principle is elegant: a focused image has sharp edges and high contrast. A blurred image does not. 

The algorithm moves the lens, measures contrast at each position, and searches for the peak. Engineers call it a "hill-climbing" pattern to start at one extreme, step toward the other, measure at every stop, then return to the peak once contrast starts falling. 

CDAF is reliable, hardware-simple, and cost-effective. It requires no specialized pixels on the sensor, which is why it appears across the widest range of modules. The cost is speeding; it must physically sweep the lens before knowing where the focus lies. For fast-moving subjects, the target may have shifted by the time the search finishes. 

Best For Document scanning, medical imaging, kiosk terminals, and any environment where the scene holds still long enough to let the algorithm complete its search. 

Phase-Detect Autofocus (PDAF): The Calculator 

Phase-detect autofocus was built specifically to solve the speed problem that limits CDAF. Instead of searching for focus by moving the lens and measuring what changes, PDAF calculates exactly where focus is before the lens moves at all. 

Here is how it works: specialized pixels on the sensor are masked so that each receives light from only the left or the right half of the lens aperture. In a focused image, both halves converge on the same point. In an out-of-focus image, they arrive at different points, phase-shifted. The autofocus processor reads the magnitude and direction of that shift, calculates the required lens travel, and sends the lens there in one deliberate move. 

One measurement. One correction. Done. Phase-detect autofocus achieves focus lock in 100 to 300 milliseconds, roughly 3 to 5 times faster than CDAF, making it the right choice for video, face tracking, and any application where the subject does not stay still. 

Best For Video autofocus, moving subjects, face and object tracking, and high-speed document capture, where a half-second focus search is not acceptable. 

Laser Autofocus: The Surveyor 

Laser autofocus takes a completely different approach. It emits an infrared laser pulse at the subject and measures the time it takes to return a time-of-flight (ToF) measurement. It reads distance directly, with no dependence on image texture or lighting. 

Laser AF thrives where both CDAF and PDAF struggle: low light, low-contrast surfaces, smooth glass, white walls. It does not need image detail because it is not analyzing the image at all. 

In practice, laser AF plays a supporting role, providing the initial distance estimate that positions the lens quickly, before PDAF or CDAF fine-tunes the result. Which leads naturally to the architecture that combines all three. 

Detection Method Comparison 

Property	CDAF	PDAF	Laser AF
Focus Speed	500-800 ms	100-300 ms	100 ms (distance)
Motion Handling	Best for static	Good	Excellent (ToF)
Low-Light	Degrades	Degrades	Excellent
Low-Contrast Surfaces	Struggles	Struggles	Excellent
Hardware Cost	Low	Moderate	Higher
Best Use Case	Kiosks, docs, medical	Video, tracking, retail	Low-light, clean surfaces

What is a Hybrid Autofocus System and When Do You Need One?

No single detection method is perfect across all operating conditions. That is exactly why the most capable imaging systems use a hybrid autofocus architecture combining multiple methods, so each covers the other's weaknesses. 

A typical hybrid autofocus system works in three stages: 

• Laser AF reads the approximate subject distance before the lens has moved a millimeter. 

• Phase-detect autofocus drives the lens to the correct zone in a single, calculated step. 

• CDAF validates final sharpness, making fine adjustments when conditions allow.  

Each method contributes to what it does best. The result is a hybrid autofocus system that is faster than CDAF alone, more reliable under poor lighting than PDAF alone, and far more adaptable than any single-method approach. 

The Actuator: What Actually Moves the Lens 

Detection tells the system where the focus is. But something still has to physically move the glass there. In the vast majority of autofocus camera modules from consumer smartphones to industrial embedded systems, that is a Voice Coil Motor, or VCM. 

How a VCM Works 

A VCM is an electromagnetic actuator built on the same principle as a loudspeaker. A coil of wire sits inside a permanent magnetic field. When current flows through the coil, it generates a force that moves the coil and the lens attached to it along the optical axis. Reduce the current, and a return spring pulls the lens back to its resting position. 

Position is directly proportional to current. Send more, the lens travels farther. Send less; it returns. This linear relationship is what makes VCMs fast, quiet, and precisely controllable with a simple DAC circuit. A dedicated driver IC, typically connected via I²C, receives a position code from the autofocus processor and converts it to the exact current output needed. 

Common driver ICs offer 10-bit position resolution: 1,024 discrete focus positions across the full travel range. That fine resolution is what enables smooth video auto focus; the lens moves in tiny, imperceptible increments rather than jumping between coarse steps. 

Actuator Comparison: VCM vs. Stepper Motor vs. Piezo 

Property	VCM	Stepper Motor	Piezo Actuator
Speed	Fast (< 200 ms)	Slow (300-600 ms)	Very Fast (< 50 ms)
Precision	High (10-bit)	Moderate	Very High
Power	Low (spring return)	Moderate	Very Low
Noise	Near silent	Audible clicking	Silent
Cost	Low	Low-Moderate	High
Best Use Case	Smartphones, kiosks, medical	Budget industrial AF	Precision optical systems

For most embedded camera autofocus applications, medical devices, retail kiosks, logistics scanners, and industrial inspection, the VCM hits the right balance of speed, precision, cost, and silence. Stepper motors work as a budget option when speed is secondary. Piezo actuators belong in precision optical instruments where performance outweighs cost. 

Vadzo Autofocus Camera Modules Match the Method to Your Application 

Vadzo offers a range of autofocus camera modules covering CDAF, PDAF, and VCM-driven actuation across USB, MIPI, and SerDes interfaces. Here is how to match the module to your application.  

Camera Model	Sensor	AF Method	Best For	Interface
Armor 1335-CRA-FPD3	onsemi AR1335 CMOS	CDAF (ROI AF)	Medical, Kiosk, Retail	FPD-Link III
Vadzo Falcon-1335 CRO	onsemi AR1335 CMOS	CDAF + OIS	Handheld, Mobile	USB 3.2 Gen 1 Type C Interface Backward Compatible to USB 2.0
Vadzo Falcon-258CRA	Sony IMX258 Exmor	PDAF	Fast subjects, Video	USB 3.0 Gen1 Type C Interface Backward Compatible to USB 2.0
Vadzo Falcon-234CGH	onsemi AR0234 CMOS	CDAF (GS)	Industrial, High-speed	USB 3.0 Gen1 Type C Interface Backward Compatible to USB 2.0
Vadzo Falcon-1335MRS	onsemi AR1335 CMOS	CDAF + VCM (M12)	Industrial, Interchangeable Lens	USB 3.0 Gen1 Type C Interface Backward Compatible to USB 2.0
Vadzo Falcon-5640CRA	OV5640 CMOS Sensor from Omnivision	CDAF	Compact / Cost-sensitive	USB 3.0 Gen1 Type C Interface Backward Compatible to USB 2.0

How to Choose: Start with your subject's motion profile. Static or slow-moving? CDAF. Fast-moving or video? PDAF. Variable environment, low contrast, or demanding conditions? Hybrid. Then choose your interface based on your host platform and data bandwidth requirements. 

Frequently Asked Questions (FAQs)

What is autofocus, and how does it work on a camera?

Autofocus is a system that automatically moves your camera lens to the position that produces the sharpest image of the subject. A camera autofocus system detects whether the image is in focus using contrast or phase data, calculates the exact correction needed, and drives a motor typically a VCM to execute that correction in milliseconds. The entire loop completes faster than the human eye can register. 

What is the difference between CDAF and phase-detect autofocus?

Contrast Detection Autofocus (CDAF) physically sweeps the lens through a range of positions, measuring image contrast at each stop until it finds the sharpest point. Phase-detect autofocus (PDAF) calculates the exact lens position needed before moving by analyzing the phase shift between two sets of masked sensor pixels. PDAF is 3 to 5x faster than CDAF, making it the right choice when the subject moves. 

What does a hybrid autofocus system and when should I use one?

A hybrid autofocus system combines multiple detection methods typically laser AF for rapid distance estimation, PDAF for fast lens positioning, and CDAF for final sharpness verification. Use a hybrid autofocus approach when your application involves variable lighting, low-contrast surfaces, or unpredictable subject movement environments where no single detection method is reliable on its own. 

What is a Voice Coil Motor (VCM) in an autofocus camera module?

A VCM is the actuator that physically moves the lens inside most autofocus camera modules. It works on electromagnetic principles: current through a coil inside a magnetic field generates a force that moves the lens along the optical axis. Position is proportional to current, which makes VCMs fast (under 200 ms), near silent, and precisely controllable as the preferred actuator for medical, industrial, and retail embedded vision systems. 

How do I choose the right autofocus camera for my embedded system?

Match your detection method to your subject and environment. For static or slow-moving subjects: CDAF with a VCM is cost-effective and proven. For fast-moving subjects or continuous video: phase-detect autofocus delivers the speed you need. For challenging conditions, low light, low-contrast surfaces, variable distances, A hybrid autofocus system is the most robust architecture. Vadzo's lineup covers all three approaches across USB, MIPI, and SerDes interfaces. 

Smarter Autofocus 

What is autofocus when you strip it to its essence? It is the system that closes the gap between what a lens can theoretically see and what it actually delivers in the conditions you cannot always control. 

Autofocus technology is a layered decision by a detection method that finds the focus target, and an actuator that moves the lens there. The combination you choose shapes how fast your system responds, how well it handles motion, and how long it stays reliable in the field. 

Here are the three things to carry with you from this guide: 

• Match detection to motion. CDAF for static scenes. PDAF for moving subjects. Hybrid for unpredictable conditions. 

• The VCM is the default actuator for good reason. Fast, precise, quiet, and cost-effective fits the widest range of embedded applications. 

• Autofocus is a system, not a checkbox. The right combination of detection method and actuator, specified correctly, is the difference between a product that holds focus and one that does not.  

Vadzo's lineup of autofocus camera modules gives you that foundation across CDAF, PDAF, hybrid autofocus, and VCM actuation, on the interfaces your platform demands. 

Your Next Step: Browse Vadzo's autofocus camera module range at vadzoimaging.com filter by detection method, interface, and resolution to find the right module for your next embedded vision design. 

Which autofocus challenge are you solving right now and which detection method are you evaluating? Leave a comment or reach out to Vadzo's engineering team.