Subject tracking is the autofocus function that identifies a chosen subject and follows it across the frame, continuously moving the active focus area and adjusting focus distance as the subject moves. In photography it combines continuous autofocus with image analysis, so the camera rather than the photographer decides which focus point to use from moment to moment.
Tracking turns autofocus from a point-and-hold tool into a subject-centered system. It is essential for sport, wildlife, events and video, where a subject rarely stays under a single fixed point long enough for manual point selection.
How subject tracking works in focus and autofocus systems
Subject tracking runs as a loop of recognition, position estimation and focus drive. The camera first locks onto a target, either the subject under the initial focus point or one found by automatic detection. It then analyzes each new live view frame to locate the target again, using color, shape, contrast patterns and learned subject models.
Once the position is known, the camera measures distance at that spot, usually with phase detection AF pixels on the sensor, and commands the lens motor. A predictive algorithm extrapolates the subject’s speed and direction so that focus is set for where the subject will be at the moment of exposure, not where it was when measured.
| Stage | What the camera does | Main information used |
|---|---|---|
| Acquisition | Selects the initial target | Start point, detected subjects |
| Recognition | Finds the target in each new frame | Subject model, color, pattern |
| Distance measurement | Reads focus error at the target | Phase detection, sometimes contrast |
| Prediction | Estimates position at exposure time | Speed and acceleration history |
| Lens drive | Moves focus to the predicted distance | Lens motor and focus position data |
Detection types and tracking methods
Early systems tracked by color and luminance patterns read from a separate metering sensor. Current mirrorless cameras use the imaging sensor itself and machine-learned subject detection, which recognizes categories such as people, animals, birds and vehicles and can identify specific parts like eyes, heads or helmets.
| Approach | How it identifies the subject | Strength | Weakness |
|---|---|---|---|
| Pattern and color tracking | Follows the image area initially selected | Works on any object | Loses targets that change appearance |
| Face detection | Recognizes human faces | Reliable for people | Fails on turned heads |
| Learned subject detection | Recognizes trained categories and body parts | Holds subjects through pose changes | Limited to trained categories |
| 3D distance tracking | Uses depth data across the frame | Separates subject from background | Needs dense distance information |
Eye AF is a refinement of detection-based tracking: once the system has found a person or animal, it narrows the focus point to the nearest eye. Tracking keeps the subject in the frame identified, while eye detection decides which plane of that subject receives critical focus.
Settings that shape tracking behavior
Most cameras expose a small set of parameters that tune how aggressively the tracker reacts. The defaults suit general use, but matching them to the subject’s motion often decides whether a sequence is consistently sharp.
| Parameter | Typical options | Effect |
|---|---|---|
| Tracking sensitivity | Locked on to responsive, often five levels | How long the tracker ignores obstacles |
| Acceleration response | Steady to erratic | How quickly prediction adapts to speed changes |
| Starting area | Single point, zone or whole frame | Where the camera looks for the first target |
| Subject category | People, animals, birds, vehicles, auto | Which trained model the camera applies |
| Focus priority | Release priority or focus priority | Whether the shutter waits for confirmed focus |
Many photographers pair tracking with back-button focus, which separates focus activation from the shutter release. Holding the button keeps tracking active, and releasing it freezes the current distance, which is useful when a subject briefly stops behind an obstacle.
Subject tracking compared with other autofocus modes
Subject tracking differs from simpler continuous modes in who chooses the focus point. In single-point continuous AF the photographer keeps the point on the subject, and the camera only adjusts distance. In tracking, the camera also moves the point, which frees the photographer to compose but requires trust in the recognition.
| Mode | Point selection | Distance adjustment | Best use |
|---|---|---|---|
| Single-shot AF | Photographer | Once, then locked | Static subjects |
| Continuous AF, single point | Photographer | Continuous | Predictable motion, cluttered scenes |
| Zone AF | Camera within a chosen zone | Continuous | Erratic subjects in a known area |
| Subject tracking | Camera across the frame | Continuous with prediction | Moving subjects with free composition |
Limitations and failure modes
Tracking performance falls when the camera receives poor information. Low light reduces the signal for both detection and phase measurement, and low-contrast subjects against similar backgrounds can lead the tracker to jump to the background. Fast lateral movement close to the camera stresses both prediction and lens motor speed.
- Occlusion causes switching. A referee or branch passing in front of the subject can capture the tracker unless sensitivity is set to hold on.
- Multiple similar subjects confuse detection. In group sports the camera may move to another player with the same shape.
- Slow lenses limit tracking. An accurate tracker cannot help if the focus motor cannot follow the predicted distance.
- Readout and frame rate matter. Faster sensors provide more frequent position updates and therefore steadier tracking.
The idea in brief
Imagine pointing out a friend in a crowd and asking someone to keep watching them. Even when the friend walks to the other side of the room, the observer keeps their eyes on that person. Subject tracking does the same for a camera: once it knows what matters, it follows that subject and keeps it sharp.
This lets the photographer concentrate on timing and framing. The camera handles where to focus, while the person decides when to take the picture.
A practical scenario
A wildlife photographer photographs a heron taking off from a riverbank with a 500 mm lens. The camera is set to continuous AF with bird detection and whole-frame tracking, sensitivity one step toward locked on, and a burst rate of 20 frames per second at 1/3200 s.
As the bird lifts off and crosses in front of reeds, the tracker stays on its head instead of jumping to the vegetation, and prediction keeps up with the accelerating wingbeats. The photographer simply keeps the heron within the frame. Most frames of the sequence show a sharp eye, whereas a single fixed point would have required perfect manual following at a focal length where that is nearly impossible.