Canon EOS R5 & R6: Understanding Subject Detection
Understanding Subject Detection: Recognition and Tracking
Explore Canon EOS R5 and R6 subject recognition, Servo AF tracking, DIGIC processing, firmware developments and practical Birds in Flight applications.
Subject Recognition and Tracking in the Evolution of Intelligent Autofocus
Autofocus has undergone a significant transformation in the transition from conventional DSLR systems to Canon's mirrorless EOS R architecture.
Traditional autofocus was primarily concerned with establishing focus on a selected subject or area. The photographer determined where the camera should focus, while the autofocus system measured focus error and adjusted the lens accordingly.
Modern mirrorless autofocus introduces another dimension. The camera can analyse image information, recognise particular subject categories, identify relevant anatomical features and maintain focus on a moving subject.
Canon's development of Dual Pixel CMOS AF II, deep-learning-based subject recognition and increasingly sophisticated tracking algorithms represents an important part of this transition.
For photographers using the EOS R5 and R6 camera generations, understanding these developments is increasingly important. Subject recognition is no longer simply an additional autofocus convenience. It has become an integral part of how modern cameras acquire, prioritise and follow photographic subjects.
1. Understanding Dual Pixel CMOS AF II
The foundation of the original EOS R5 and R6 autofocus systems is Canon's Dual Pixel CMOS AF architecture.
Unlike conventional DSLR autofocus, which generally relies on a dedicated phase-detection module, Dual Pixel CMOS AF performs phase-detection autofocus directly on the imaging sensor.
Each effective pixel incorporates two photodiodes. These provide phase-detection information and contribute to image formation.
This arrangement allows the camera to evaluate focus direction and the amount of adjustment required. It also provides extensive autofocus coverage across the imaging area.
The original EOS R5 and R6 introduced this technology to a broader range of photographic applications, including moving subjects and Birds in Flight.
The processing architecture is equally important. DIGIC X handles substantial image-processing and autofocus-related computational workloads, including subject-recognition functions.
The sensor provides the focusing information, while the processing system interprets that information and executes autofocus decisions.
2. Subject Recognition: What Does the Camera Actually Recognise?
Subject recognition is the process through which the camera analyses incoming image information and identifies a subject belonging to a supported category.
Depending on the camera model and firmware, these categories include:
- People: face, head and eye detection.
- Animals: including birds, dogs, cats and horses on supported models.
- Vehicles: including cars, motorcycles, trains and aircraft on selected newer models.
An important development occurred with the original EOS R5 firmware 1.5.0, which introduced vehicle detection for automobiles and motorcycles. The update also improved head and eye detection performance.
Firmware 1.5.2 subsequently improved Eye Detection stability and autofocus tracking when photographing moving subjects.
These updates demonstrate that subject recognition can evolve after a camera's initial release.
Recognition should not, however, be confused with species identification. Selecting Animals instructs the camera to prioritise supported animal subjects. It does not mean that the camera can reliably distinguish between individual bird species.
For Birds in Flight photography, the ability to recognise a bird's head or body can be useful when its eye is temporarily difficult to detect.
3. Subject Detection versus Subject Tracking
Subject detection and subject tracking are closely connected, but they describe different operations.
Subject detection is the camera's ability to recognise and locate a subject within the image.
Subject tracking is the process of maintaining focus on that subject as its position changes.
Consider a bird entering the camera's field of view.
The camera recognises it as an animal, identifies an appropriate focusing feature and establishes an autofocus relationship. With Servo AF active, the system continues adjusting focus as the bird moves towards, away from or across the photographer's position.
Detection therefore provides the initial recognition, while tracking provides continuity.
A camera may recognise a bird successfully but still struggle to maintain tracking if the subject moves unpredictably, becomes obscured or leaves the effective tracking area.
This distinction is central to understanding intelligent autofocus.
4. AF Area Selection and Subject Acquisition
The photographer's choice of AF area influences how subject recognition operates.
Canon provides several autofocus methods, including:
- Face + Tracking.
- Spot AF.
- 1-point AF.
- Expand AF Area.
- Zone AF.
- Large Zone AF.
Their availability and interaction with subject detection vary between camera generations.
The original EOS R5 and R6 provide different approaches to subject acquisition. Face + Tracking offers extensive automatic selection, while Zone AF and Large Zone AF allow the photographer to restrict the initial search area.
This distinction becomes particularly relevant when photographing birds against complex backgrounds.
A larger AF area offers greater acquisition flexibility but can introduce competing subjects. A smaller area gives the photographer more control but demands greater framing accuracy.
The EOS R5 Mark II and R6 Mark III introduce further refinements to whole-area subject tracking.
On the R6 Mark III, Whole Area Tracking Servo AF can allow the camera to track a subject across the entire screen during Servo AF operation. When disabled, tracking remains restricted to the selected AF points.
This provides a useful balance between automatic subject acquisition and photographer-controlled focusing.
5. Servo AF and Tracking Intelligence
Servo AF is the operational foundation of continuous autofocus in the EOS R5 and R6 cameras.
Unlike One-Shot AF, which establishes focus and then stops adjusting, Servo AF continuously evaluates focus information and adjusts the lens as the subject moves.
When combined with subject recognition, the camera has two related responsibilities: maintaining focus on the selected subject and adjusting focus as its distance from the camera changes.
Canon provides additional control through the AF Configuration Tool.
Two important parameters are Tracking Sensitivity and Acceleration/Deceleration Tracking.
Tracking Sensitivity
Tracking Sensitivity controls how readily the camera changes focus when another subject or obstruction enters the active AF area.
Negative settings favour maintaining focus on the original subject.
Positive settings make the camera more responsive to subjects appearing at different distances.
This is particularly relevant when photographing birds against vegetation or other visually complex backgrounds.
Acceleration/Deceleration Tracking
This parameter controls how Servo AF responds when a subject changes speed suddenly.
Positive settings are useful when photographing subjects that accelerate, decelerate or stop unexpectedly.
Negative settings favour more consistent movement and can help stabilise tracking when minor movements or obstructions interfere with autofocus.
6. The DIGIC Accelerator and EOS R5 Mark II
The EOS R5 Mark II represents a significant development in Canon's autofocus processing architecture.
While the original EOS R5 and R6 generations use DIGIC X, the R5 Mark II combines DIGIC X with a dedicated DIGIC Accelerator.
Canon describes the Accelerator as handling substantial volumes of sensor information associated with autofocus detection and exposure metering.
The resulting architecture supports Canon's Dual Pixel Intelligent AF system.
Action Priority AF
One of the more distinctive developments is Action Priority AF.
Rather than simply recognising and tracking a selected person, this system analyses supported sporting situations and prioritises the person performing the relevant action.
The original supported sports include football, basketball and volleyball. Firmware 1.3.0 subsequently added American Football.
Action Priority AF is not a general-purpose predictive bird-tracking function. Its specific applications are defined around supported team sports.
Nevertheless, it demonstrates how Canon is moving beyond basic subject recognition towards interpreting more complex visual situations.
Registered People Priority
The R5 Mark II also provides Registered People Priority, allowing photographers to register up to ten faces for preferential recognition and tracking.
Firmware 1.3.0 improved detection and tracking in difficult situations, including profile views, blurred or partially obscured faces and smaller subjects.
This is an important example of how firmware can refine recognition performance without changing the camera's underlying sensor architecture.
7. Obstruction Handling and Tracking Continuity
One of the more important developments in modern autofocus is the ability to maintain subject tracking when the subject becomes partially obscured.
This is particularly relevant to nature photography, where subjects frequently move against complex backgrounds.
A bird may fly behind branches, pass another bird or briefly disappear against a visually distracting background.
The autofocus system must maintain a sufficiently reliable relationship with the original subject to avoid unnecessarily transferring focus elsewhere.
Recognition establishes that a subject is present. Tracking continuity concerns whether the camera can maintain its focus relationship with that subject when visual information becomes temporarily less reliable.
The EOS R5 Mark II introduces improved handling of briefly obscured subjects through its Dual Pixel Intelligent AF architecture.
The R6 Mark III also provides improved subject recognition, including detection of partially obscured subjects.
The technology assists the photographer, but its effectiveness remains dependent on available visual information.
8. Comparing the EOS R5 and R6 Generations
The five cameras illustrate different stages in the evolution of Canon's intelligent autofocus.
| Camera | Contribution to intelligent autofocus |
|---|---|
| EOS R5 | Established the original deep-learning subject-recognition foundation, with subsequent firmware improvements and additional vehicle detection. |
| EOS R5 Mark II | Introduced DIGIC Accelerator, Dual Pixel Intelligent AF, Eye Control AF, Action Priority AF and Registered People Priority. |
| EOS R6 | Established the original R6 implementation of Dual Pixel CMOS AF II and animal recognition. |
| EOS R6 Mark II | Expanded recognition capabilities, including animals and vehicles, while retaining DIGIC X processing. |
| EOS R6 Mark III | Further developed recognition and tracking, with Whole Area Tracking Servo AF and expanded subject-recognition capabilities. |
The R5 Mark II and R6 Mark III should not simply be ranked according to their release generation.
The R5 Mark II has the more advanced processing architecture, while the R6 Mark III offers a substantially developed recognition and tracking system within the DIGIC X platform.
For nature photographers, both cameras provide significant autofocus capabilities, although their operational differences remain important.
9. Practical Applications in Birds in Flight Photography
Birds in Flight photography provides an excellent practical environment in which to examine the relationship between subject recognition, autofocus acquisition and tracking continuity.
Birds do not follow predictable trajectories. Their movement can be irregular, sudden and influenced by environmental conditions.
A bird may approach, turn, dive, climb or change direction while the photographer attempts to maintain focus.
Subject acquisition
The first challenge is getting the subject into the active AF area.
With longer telephoto lenses, this can become particularly difficult. A bird may occupy only a small portion of the viewfinder.
Subject recognition cannot compensate entirely for poor initial framing.
The photographer still needs to locate the subject and maintain sufficient space around it.
Eye Detection and Body Recognition
Eye Detection is desirable when the bird's eye is sufficiently visible and large enough for reliable recognition.
However, the eye is not always the most practical initial focusing reference.
A bird flying towards the photographer may present a clear head and body but a relatively small eye. In other circumstances, the bird may turn away, making eye detection temporarily unavailable.
Broader animal recognition provides another focusing reference.
The photographer should therefore avoid treating Eye Detection as the only measure of successful animal autofocus.
Background complexity
Birds flying across trees, buildings or water can present considerable tracking challenges.
Tracking Sensitivity becomes particularly important when competing objects enter the active AF area.
A more locked-on setting may help maintain the original bird, while a more responsive setting may be appropriate when the photographer needs to acquire a new subject quickly.
Focal length and AF coverage
Lens selection also influences autofocus performance.
Canon specifies that AF coverage can vary according to the lens used. This is particularly relevant when using slower super-telephoto lenses such as the RF 600mm F11 IS STM and RF 800mm F11 IS STM.
The RF 800mm provides considerable reach, but its narrow field of view places additional demands on initial subject acquisition.
A more flexible focal length can make it easier to locate and maintain a bird in flight.
This is an important practical consideration when comparing different telephoto configurations. Effective autofocus is not determined by the camera body alone.
Practical configuration principles
For Birds in Flight photography, the following general principles remain useful:
- Use Servo AF for moving subjects.
- Select Animals as the subject-detection category where available.
- Enable Eye Detection when appropriate.
- Choose an AF area that provides sufficient acquisition flexibility.
- Adjust Tracking Sensitivity according to background complexity.
- Consider Acceleration/Deceleration Tracking when photographing unpredictable movement.
- Test configurations with the actual camera and lens combination.
10. Future Developments in Intelligent Autofocus
Canon's recent developments raise an interesting question: where might Subject Recognition and Tracking go next?
Several areas offer potential for further development.
Improved subject recognition: Future systems could improve recognition of smaller subjects, unusual orientations and partially obscured anatomical features.
More sophisticated tracking continuity: Further refinement could improve the transition between initial recognition, temporary obstruction and reacquisition.
Greater subject-priority control: Registered People Priority demonstrates how photographers can designate subjects for preferential tracking. Similar concepts could potentially be extended to other photographic applications.
More advanced processing: The DIGIC Accelerator demonstrates Canon's interest in dedicated processing architecture for handling increasingly complex autofocus information.
Firmware-based refinement: The original R5 and R5 Mark II firmware updates demonstrate that recognition and tracking can improve after a camera's release.
These are possible development directions, not announced Canon product commitments. Firmware also has physical and architectural limitations. Sensor design, processing capacity, lens communication and optical characteristics remain important constraints.
11. The Photographer Remains Part of the Autofocus System
There is a tendency to evaluate modern autofocus according to its ability to operate automatically.
The more sophisticated the recognition system becomes, the easier it is to assume that the photographer's involvement is gradually becoming less important.
I would argue the opposite.
Intelligent autofocus increases the importance of understanding how the camera interprets a photographic situation.
A photographer who understands subject recognition, AF area selection, Servo AF characteristics and tracking sensitivity is better equipped to use the technology effectively.
The camera can recognise a bird, identify its eye and maintain focus as it moves through the frame. But the photographer still needs to anticipate movement, establish initial acquisition, maintain appropriate framing and select suitable exposure parameters.
Autofocus is therefore not an isolated camera function. It is part of a larger photographic process involving observation, anticipation, technical knowledge and decision-making.
Conclusion: The Evolution of Intelligent Autofocus
Canon's EOS R5 and R6 camera generations demonstrate how autofocus has evolved from conventional focus-point selection towards increasingly sophisticated subject recognition and tracking.
The original EOS R5 and R6 established the deep-learning recognition foundation. The R5 Mark II introduced a more advanced processing architecture through DIGIC Accelerator and Dual Pixel Intelligent AF. The R6 Mark II expanded the capabilities of the original R6 platform, while the R6 Mark III further developed subject recognition and Whole Area Tracking Servo AF.
These developments represent more than incremental improvements in autofocus speed.
They reflect a broader transition towards cameras that can interpret visual information, recognise photographic subjects and maintain focus relationships under changing conditions.
Yet intelligent autofocus remains dependent on the photographer's ability to understand and control the system.
The technology provides increasingly sophisticated assistance, but successful photography still requires observation, anticipation and informed decision-making.
For the modern EOS R photographer, understanding Subject Recognition and Tracking is therefore not simply about learning camera settings. It is about understanding how the camera and photographer work together to achieve consistent photographic results.
References
Canon. (n.d.-a). EOS R5 firmware update, version 2.2.0. Canon Asia. https://asia.canon/en/support/0401059502
Canon. (n.d.-b). EOS R5 Mark II: Product specifications. Canon USA. https://www.usa.canon.com/support/p/eos-r5-mark-ii
Canon. (n.d.-c). EOS R6 Mark II: AF menu settings. Canon Camera Museum and Product Manuals. https://cam.start.canon/en/C012/manual/html/UG-05_AF-Drive_0020.html
Canon. (n.d.-d). EOS R6 Mark III: Product specifications. Canon USA. https://www.usa.canon.com/support/p/eos-r6-mark-iii
Canon. (n.d.-e). EOS R6 Mark III: Selecting the AF area. Canon Camera Museum and Product Manuals. https://cam.start.canon/en/C022/manual/html/UG-05_AF-Drive_0060.html
Canon. (n.d.-f). EOS R6: Servo AF characteristics. Canon Camera Museum and Product Manuals. https://cam.start.canon/en/C004/manual/html/UG-04_AF-Drive_0100.html
Canon. (2021, December 1). EOS R5 firmware update, version 1.5.0. https://sg.canon/en/support/0400709602
Canon. (2026, April 27). EOS R5 Mark II firmware update, version 1.3.0. https://asia.canon/en/support/0401146102
Canon Photography Training Milnerton Woodbridge Island | Kirstenbosch Cape Town | Intaka Island
© Vernon Chalmers Photography. All rights reserved.
