13. Immersive, Web, and Interactive Graphics

Introduction and Context

Computer graphics has moved beyond the production of images on a two-dimensional screen. Contemporary systems increasingly place users inside rendered environments, overlay digital information on the physical world, connect graphics to live data, and allow interaction through bodies, voices, gestures, controllers, and spatial movement. Games, web applications, industrial digital twins, virtual reality (VR), augmented reality (AR), and mixed-reality systems therefore require more than geometric modelling and rasterization. They require coordinated treatment of perception, interaction, tracking, computation, networking, accessibility, and human experience.

The central design problem is no longer simply how to draw an object. It is how to maintain a convincing, responsive, safe, and meaningful relationship between a user, a computational system, and a changing environment. A virtual training system must respond to head and hand movement with very low delay. A web-based visualisation must remain usable across devices and network conditions. An industrial digital twin must represent sensor data accurately enough to support decisions. An AR maintenance application must align virtual instructions with physical machinery without creating dangerous ambiguity.

This chapter examines the technical and design foundations of such systems. It connects rendering pipelines with human-computer interaction, spatial computing, stereoscopic perception, tracking, latency management, accessibility, and evaluation. The emphasis is on informed engineering decisions: selecting an interaction model, measuring performance, managing uncertainty, and balancing visual fidelity against responsiveness and usability.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Explain the distinguishing characteristics of immersive, web-based, and interactive graphics systems.

2. Analyse the roles of tracking, spatial mapping, stereoscopy, interaction, and rendering in VR and AR.

3. Compare interaction techniques and choose suitable methods for different users and tasks.

4. Design performance strategies for latency-sensitive and networked graphics applications.

5. Evaluate accessibility, safety, privacy, and ethical issues in immersive systems.

6. Apply relevant metrics and testing methods to games, digital twins, web graphics, and spatial applications.

7. Assess current developments such as WebXR, cloud rendering, artificial intelligence, and spatial computing.

From Display-Oriented Graphics to Spatial Computing

Traditional interactive graphics usually assumes a fixed viewpoint, a rectangular display, and input devices such as a mouse or keyboard. Immersive graphics changes these assumptions. The viewpoint may move continuously with the user’s head; the display may be stereoscopic; the environment may be mapped in three dimensions; and input may be distributed across hands, eyes, voice, and body position.

Interactive graphics can be understood as a closed loop:

sense or receive input → interpret user intent and world state → update the scene → render → display or present → observe the user’s response.

The quality of this loop depends on more than frame rate. A system can produce many frames per second yet feel unresponsive if input is sampled late, simulation is delayed, or presentation is poorly synchronized. Conversely, a visually simple application can feel highly responsive when its end-to-end delay is small and its interaction model is predictable.

Spatial computing extends this loop by treating physical space as an input and output context. A device may estimate its position, detect surfaces, identify spatial anchors, and place virtual objects relative to the real environment. The resulting graphics are not merely viewed; they are situated. This creates opportunities for natural interaction but also introduces uncertainty. Tracking errors, changing illumination, reflective surfaces, and incomplete scene understanding can cause virtual content to drift or appear incorrectly occluded.

Core Concepts and Terminology

An immersive system typically combines several coordinate spaces. Model space describes an object relative to its own origin. World space places objects in a common virtual or physical environment. View or camera space expresses geometry relative to the observer. Clip and screen spaces support projection and rasterization. In AR, an additional relationship must be maintained between the virtual world and the physical environment. This relationship is often represented through a tracked pose and one or more spatial anchors.

A pose consists of position and orientation. Position is commonly represented by a three-dimensional vector, while orientation may be represented by Euler angles, rotation matrices, or quaternions. Quaternions are widely used for smooth orientation interpolation and avoidance of certain singularities associated with Euler-angle representations.

Tracking is the estimation of a device, body part, or object’s pose over time. Inside-out tracking uses cameras and sensors mounted on the device; outside-in tracking uses external cameras or beacons. Six-degrees-of-freedom (6DoF) tracking estimates three translational and three rotational degrees of freedom. Three-degrees-of-freedom (3DoF) tracking generally estimates orientation but not full position.

Spatial mapping constructs a representation of surfaces or environmental geometry. A mapping may be a point cloud, depth image, polygonal mesh, occupancy representation, or semantic map. World understanding adds labels or categories such as floor, wall, table, or human body. These representations support occlusion, collision, navigation, and placement of virtual content.

Presence is the subjective sense of being located within a mediated environment. Immersion refers more broadly to the system’s objective capacity to surround or engage the user through visual, auditory, and interactive channels. The two are related but not identical: technically sophisticated hardware does not guarantee a strong sense of presence.

Latency is the delay between an event and its perceptible consequence. Motion-to-photon latency is the interval between a user’s movement and the corresponding updated image reaching the display. Jitter is variation in delay or timing. Both can undermine comfort and control. Frame rate describes the frequency of presented frames, whereas frame time is the duration available to produce one frame. A stable frame time is often more important than a high but irregular frame rate.

Rendering Architectures for Interactive and Immersive Systems

A conventional real-time rendering pipeline includes application logic, scene update, geometry processing, rasterization or ray-based rendering, shading, post-processing, and presentation. In an immersive application, the pipeline must usually render from two slightly different viewpoints for stereoscopic display. The views share much of the scene but differ in camera position and projection. Rendering systems may therefore use multiview techniques, instancing, foveated rendering, or single-pass approaches to reduce duplicated work.

Forward rendering computes lighting as objects are drawn and can be efficient for scenes with limited lights. Deferred rendering stores geometric and material information in intermediate buffers and computes lighting later, which is useful when many lights are present but increases memory and bandwidth demands. Physically based rendering improves consistency between material parameters and illumination, but visual realism should not be pursued at the expense of response time or legibility.

A practical immersive pipeline separates simulation rate from rendering rate when necessary. Physics, networking, and input processing may operate at different frequencies. Interpolation can smooth remote or simulated motion, while prediction can estimate the pose at display time. Prediction must be corrected carefully because visible correction can be more disturbing than modest visual simplification.

Level of detail (LOD) reduces geometric and shading complexity for distant or less important objects. Occlusion culling avoids rendering objects that cannot be seen. Dynamic resolution changes render resolution according to available processing capacity. Foveated rendering allocates greater detail near the user’s gaze and lower detail in peripheral regions. It can be fixed, based on an assumed viewing region, or gaze-contingent, using eye tracking. Gaze-contingent methods require accurate and timely eye tracking; otherwise, the user may notice a moving region of reduced quality.

For web graphics, rendering may use the 2D Canvas API, SVG, WebGL, or WebGPU. Canvas is appropriate for immediate-mode drawing and many dynamic visualisations. SVG provides a retained, structured, and accessible representation for scalable vector content, although very large scenes can become expensive in the document model. WebGL provides broad hardware-accelerated 3D access, while WebGPU exposes a more modern graphics and compute model. The choice should reflect scene complexity, compatibility, accessibility requirements, and development expertise rather than novelty alone.

Stereoscopy, Depth, and Perceptual Comfort

Stereoscopic systems present different images to the left and right eyes. The visual system interprets the disparity between corresponding features as depth. Important parameters include interpupillary distance, camera separation, convergence, focal distance, and the depth range of objects. Incorrect relationships can produce discomfort, visual fatigue, or difficulty focusing.

A major issue is the accommodation-convergence conflict. In natural vision, the eyes converge on an object and adjust focus to the same distance. In a head-mounted display, the eyes generally accommodate at the fixed display focal distance while convergence may vary with virtual depth. Excessive disparity or rapid depth changes can increase discomfort. Consequently, content should avoid unnecessary extreme depth, maintain stable visual references, and provide comfortable user-adjustable settings where possible.

Depth perception is not produced by binocular disparity alone. Motion parallax, perspective, occlusion, shading, texture gradients, familiar size, and atmospheric effects also contribute. Good immersive design uses these cues consistently. Incorrect occlusion—for example, a virtual instruction appearing in front of a real machine part when it should be behind it—can be especially damaging in AR because it conflicts with physical perception.

Interaction and Input Interpretation

Interaction design begins with tasks rather than devices. A button, object-selection technique, or locomotion method should be chosen according to the user’s goals, environment, physical abilities, and error costs. Direct manipulation is effective when users need to grasp, rotate, or position objects. Ray-based selection is useful when objects are distant or direct reach would be tiring. Gaze selection can be efficient for pointing but may be inappropriate for confirmation because people look at many objects without intending to select them.

Common interaction techniques include controller input, hand tracking, eye gaze, voice commands, gesture recognition, haptic feedback, and tangible objects. Each has strengths and weaknesses. Hand tracking feels natural and reduces equipment, but it can be affected by occlusion and limited gesture vocabularies. Controllers provide reliable buttons and haptic events but may impose a learning burden. Voice input supports hands-busy tasks but raises privacy and recognition concerns.

The interpretation layer should distinguish raw sensing from intentional action. A tracked hand position is not automatically a command. Systems should use state machines, temporal filtering, confidence estimates, and contextual rules to reduce accidental activation. For safety-critical applications, confirmation, undo, guarded actions, and clear system status are essential.

Locomotion is a particularly important VR issue. Physical walking provides strong spatial correspondence but requires adequate space. Teleportation reduces motion sickness for many users but can weaken continuous spatial understanding. Smooth artificial movement offers greater freedom but may cause discomfort, particularly when visual acceleration lacks corresponding vestibular sensation. Snap turning, adjustable movement speed, seated modes, and rest points are useful alternatives rather than signs of incomplete design.

Haptics can communicate contact, selection, resistance, or error. Vibrotactile feedback is relatively accessible in consumer devices, while force feedback and mid-air haptics provide richer but more demanding implementations. Haptic signals should be distinguishable, brief, and meaningful; excessive vibration becomes noise.

Tracking, Registration, and Spatial Alignment

A tracking system estimates motion by combining sensor measurements. Cameras provide visual features, inertial measurement units provide acceleration and angular velocity, and depth sensors provide distance information. Sensor fusion algorithms, often based on variants of Bayesian estimation or nonlinear filtering, combine these imperfect observations. Visual-inertial odometry estimates motion from camera and inertial data; simultaneous localization and mapping (SLAM) estimates device motion while constructing an environmental map.

Registration is the alignment of virtual content with the real world. It requires a stable coordinate relationship, accurate calibration, and appropriate handling of drift. Marker-based tracking uses known visual patterns and can be reliable in controlled environments. Markerless methods are more flexible but depend on scene texture, lighting, and sensor quality. Anchors provide persistent reference points, but persistence across devices or sessions raises data-management and privacy issues.

Tracking quality should be evaluated through positional error, rotational error, drift over time, recovery after occlusion, update frequency, and confidence under varied environmental conditions. Developers should communicate uncertainty. A virtual measurement or instruction should not appear equally authoritative when the underlying registration is unstable.

Web Graphics, Games, and Digital Twins

Web graphics makes interactive visualisation broadly deployable. A browser-based application can combine responsive layout, data services, GPU rendering, and collaborative interaction. However, web deployment introduces variability in browsers, hardware, permissions, network bandwidth, display size, and input methods. Progressive enhancement is therefore valuable: provide a meaningful two-dimensional or low-fidelity experience when immersive or high-performance capabilities are unavailable.

Games integrate real-time graphics with rules, narrative, audio, networking, and player psychology. Their performance decisions are often made at the level of player experience. A small reduction in texture detail may be preferable to inconsistent frame pacing, while an apparently realistic effect may be removed if it obscures gameplay information. Multiplayer games additionally require client prediction, interpolation, lag compensation, authority models, and protection against cheating.

A digital twin is a computational representation of a physical asset, process, or system that is connected to data about its real counterpart. Graphics provides the spatial and visual interface, but a digital twin is not merely a 3D model. It needs identity, state, time, provenance, update mechanisms, and a defined relationship between simulation and observation. A useful twin distinguishes measured data from inferred or predicted data. Visual dashboards can expose anomalies, maintenance status, energy consumption, or production flow, but they should not suggest a level of precision unsupported by the sensors.

Latency, Performance, and Quality of Experience

Performance engineering should begin with a latency budget. The total delay may include sensor sampling, operating-system scheduling, application update, rendering, compositor processing, transmission, and display persistence. Profiling should measure these components rather than relying only on average frames per second. Spikes, dropped frames, thermal throttling, and garbage collection pauses can be more harmful than a slightly lower but stable frame rate.

Useful performance practices include asynchronous resource loading, asset streaming, memory pooling, batching, shader simplification, occlusion culling, LOD, fixed time-step simulation where appropriate, and careful synchronization. Networked applications should separate authoritative state from presentation state. A client may interpolate received data, predict short-term motion, and reconcile with authoritative corrections. The visual design should make uncertainty and connection loss understandable.

Quality of experience includes comfort, task completion, learnability, presence, confidence, perceived responsiveness, and error recovery. A technically efficient system can still fail if users cannot understand where to look, how to select, or why an action did not occur. Evaluation should therefore combine instrumentation with user studies. Metrics may include completion time, error rate, workload, simulator-sickness reports, gaze or movement patterns, retention, and subjective comfort. Physiological data can be informative in research but requires strong consent and careful interpretation.

Accessibility, Safety, and Responsible Design

Accessibility must be designed into the interaction model. Users may have limited vision, hearing, mobility, reach, balance, or ability to perform precise gestures. Systems should support adjustable text and contrast, captions, spatial audio alternatives, seated and standing modes, dominant-hand selection, one-handed operation, remapping, dwell or switch input, reduced motion, and non-immersive alternatives. Colour should not be the sole carrier of meaning. Visual alerts should have audio or haptic equivalents when appropriate.

Immersive systems can cause cybersickness, eyestrain, fatigue, falls, collisions, and disorientation. Safety design includes boundary warnings, passthrough or environmental awareness where available, safe default speeds, session breaks, guardian systems, and clear exit procedures. Designers should avoid placing critical information where it is hidden by overlays or requiring users to move their bodies in unsafe ways.

Privacy is particularly significant because immersive devices may collect room geometry, body movement, eye gaze, voice, facial expressions, and social interaction data. Such data can reveal health conditions, interests, emotional states, or household information. Data minimisation, purpose limitation, informed consent, access control, encryption, retention limits, and transparent user controls are professional responsibilities. Biometric and behavioural data should not be treated as harmless telemetry.

Ethical design also concerns manipulation and social power. Persuasive environments can exploit attention or induce emotional responses. Avatars and tracking systems may reproduce bias through inaccurate body representation, speech recognition, or gesture interpretation. Digital twins used for organisational decisions should preserve auditability and distinguish recommendations from automated decisions. Accessibility, privacy, and accountability are not optional additions after implementation; they influence architecture and product requirements from the beginning.

An Evaluation Framework for Immersive and Interactive Graphics

A useful evaluation can be organised around five questions. First, is the system technically stable? This includes frame timing, tracking accuracy, crash rate, thermal behaviour, and network reliability. Second, is the interaction understandable and controllable? This concerns discoverability, feedback, consistency, and recoverability. Third, is the experience comfortable and accessible? Measures include fatigue, cybersickness, posture, visual legibility, and success across user groups. Fourth, does the system support the intended task or organisational outcome? A beautiful environment that does not improve training, maintenance, learning, or play has limited value. Fifth, are risks and uncertainties communicated honestly?

Evaluation methods should match development stage. Expert inspection can identify interaction and accessibility problems early. Instrumented prototypes reveal performance bottlenecks. Controlled experiments can compare locomotion or selection techniques. Field studies expose environmental variation and organisational constraints. A/B testing may be useful for web interfaces but should not replace safety and accessibility assessment. For research claims, investigators should report hardware, rendering settings, participant characteristics, task design, and limitations sufficiently for interpretation and replication.

Illustrative Case Study: An AR Maintenance Assistant

Background

A manufacturing company wants to reduce errors during maintenance of complex industrial equipment. It proposes an AR application that displays component names, torque values, inspection sequences, and warnings while a technician works near the machine.

Situation and Problem

The initial prototype uses fixed 3D labels and a marker attached to the machine. In laboratory conditions the labels align well, but in the factory the marker is partly obscured, lighting changes during shifts, and technicians report that labels block their view. Some users wear gloves and cannot perform precise hand gestures. Others work in restricted spaces where a head-mounted display is uncomfortable for long periods.

Analysis

The problem is not simply insufficient graphical realism. It involves registration uncertainty, task interruption, ergonomics, accessibility, and the consequences of incorrect information. A label that drifts by a few centimetres may be tolerable for general identification but unsafe when indicating a small valve. The application also lacks a clear distinction between live sensor data, standard operating procedure, and technician-entered observations.

Approach and Solution

The design team adopts a hybrid architecture. Visual-inertial tracking is supplemented by multiple robust reference features, and the system displays a confidence indicator when registration quality decreases. Critical instructions are anchored to stable machine regions rather than floating in the user’s field of view. Labels collapse when not needed, and a gaze-and-button method is offered alongside hand gestures. A voice command can repeat an instruction, while captions and a handheld display provide alternatives.

The rendering system uses simplified geometry and aggressive occlusion handling to preserve frame timing. It presents warnings in a consistent screen region and uses colour, shape, text, and audio rather than colour alone. Sensor values are timestamped and labelled as measured, estimated, or manually confirmed. Before deployment, technicians test the system under different illumination, glove types, body heights, and network conditions.

Outcome

The revised application is less visually elaborate than the original prototype but is easier to read, more stable, and safer to use. Testing shows that users complete routine sequences with fewer view-obstructing overlays and recover more effectively when tracking is interrupted. The company retains a non-immersive tablet workflow for tasks where wearing a headset is impractical.

Learning Points

The case illustrates five principles. Immersive content must be designed around work rather than novelty. Registration confidence should be visible when accuracy matters. Alternative input and output methods improve resilience and inclusion. Stable frame timing and legibility may be more valuable than visual detail. Finally, a digital representation must communicate the status and provenance of its data.

Current Developments and Emerging Directions

Spatial computing increasingly combines real-time rendering, scene understanding, language interfaces, and machine learning. Neural methods can assist with depth estimation, segmentation, denoising, super-resolution, avatar animation, and predictive tracking. Their use must be evaluated for failure modes, bias, computational cost, and explainability. A learned depth estimate that fails on transparent or reflective surfaces can produce serious AR errors.

Cloud and edge rendering may allow demanding graphics to run on remote hardware, but transmission adds latency and creates dependence on connectivity. Hybrid architectures can keep tracking, safety functions, and immediate interaction local while outsourcing selected rendering or data processing. Digital twins are becoming more connected to Internet of Things platforms, simulation engines, and collaborative workspaces. Open standards and interoperable scene descriptions can reduce platform lock-in, although conversion between representations remains difficult.

Head-mounted displays are becoming lighter, and advances in eye tracking, passthrough video, spatial audio, and display optics are expanding interaction possibilities. At the same time, the field is moving toward more modest forms of spatial interaction embedded in phones, vehicles, workstations, and public environments. The durable lesson is that immersion is not a goal in itself. Its value depends on whether spatial representation and interaction improve understanding, performance, communication, or experience without imposing unacceptable cost and risk.

Best-Practice Guidelines

A professional development process should establish user tasks, environment, safety conditions, accessibility needs, and performance targets before selecting hardware or a rendering technique. Teams should prototype interaction early, measure end-to-end latency, test tracking in realistic conditions, and provide graceful degradation when capabilities are unavailable. Visual complexity should be allocated according to perceptual and task importance. Every important action should have feedback and a recovery path.

Teams should maintain a clear data model for spatial anchors, sensor values, timestamps, confidence, and ownership. Privacy and retention requirements should be documented before collecting gaze, voice, room scans, or body data. Evaluation should include diverse users and actual operating environments. Finally, systems should be monitored after deployment because hardware updates, browser changes, environmental conditions, and user practices can alter performance and risk.

Chapter Summary

Immersive, web, and interactive graphics integrate rendering with sensing, interaction, networking, and human perception. VR and AR depend on accurate pose estimation, spatial alignment, stereoscopic presentation, and careful management of latency and comfort. Web graphics must address platform variation and progressive enhancement, while games balance visual fidelity with responsiveness and playability. Digital twins require more than 3D models: they connect geometry to live, historical, and inferred data with explicit provenance.

Effective systems use suitable interaction techniques, confidence-aware tracking, stable frame timing, accessible alternatives, and meaningful feedback. Evaluation should combine technical metrics with user, task, safety, and organisational outcomes. Privacy, consent, bias, data governance, and environmental safety are core engineering concerns. Emerging technologies such as WebGPU, WebXR, machine learning, cloud rendering, and spatial computing broaden the design space, but they do not replace the need for disciplined requirements, measurement, and human-centred judgement.

Key Terms

Augmented reality: A system that overlays computer-generated content on a user’s perception of the physical environment.

Digital twin: A data-connected computational representation of a physical asset, process, or system.

Foveated rendering: Rendering that allocates greater detail to the user’s gaze region and less detail elsewhere.

Frame time: The time available to produce and present one rendered frame.

Immersion: The objective extent to which a system surrounds or engages users through sensory and interactive channels.

Inside-out tracking: Pose estimation using sensors mounted on the tracked device.

Latency: Delay between an event and its observable system response.

Motion-to-photon latency: Delay from user movement to the corresponding displayed image.

Pose: The position and orientation of an object or device.

Presence: The subjective feeling of being located within a mediated environment.

Registration: Alignment of virtual content with a physical or shared reference frame.

Six-degrees-of-freedom tracking: Tracking of three translations and three rotations.

SLAM: Simultaneous localization and mapping, in which motion and environmental structure are estimated together.

Spatial mapping: Construction of a computational representation of surrounding surfaces or geometry.

Stereoscopy: Presentation of separate views to the two eyes to create depth perception.

WebGPU: A web graphics and compute interface designed for modern GPU capabilities.

WebXR: Web technologies that support access to immersive and augmented-reality devices.

World understanding: Interpretation of an environment in terms of surfaces, objects, or semantic categories.

Review Questions

Short Answer Questions

1. Distinguish immersion from presence in an interactive graphics system.

2. What is motion-to-photon latency, and why is it important in VR?

3. Explain the purpose of spatial anchors in AR.

4. State two differences between a digital twin and a conventional 3D model.

5. Why is progressive enhancement useful in web-based graphics?

Long Answer Questions

1. Explain the complete interaction loop in an immersive graphics system and identify major sources of delay.

2. Discuss stereoscopic depth, accommodation-convergence conflict, and design measures for visual comfort.

3. Compare controller, hand, gaze, and voice interaction with respect to precision, accessibility, and environmental constraints.

4. Describe how tracking, SLAM, spatial mapping, and registration work together in an AR application.

5. Discuss a comprehensive framework for evaluating the technical, experiential, organisational, and ethical quality of immersive graphics.

Analytical / Application Questions

1. A VR training application has a high average frame rate but frequent frame-time spikes and users report discomfort. Diagnose likely causes and propose an evaluation and optimisation plan.

2. Design an interaction strategy for an AR maintenance system used by technicians wearing gloves in noisy, poorly lit environments. Justify your input, feedback, accessibility, and safety choices.

3. An industrial digital twin displays predicted equipment failure as a red warning without indicating uncertainty. Analyse the technical, ethical, and organisational risks and redesign the presentation.

22. Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes a digital twin from a conventional 3D model?

A. A digital twin must always use stereoscopic rendering.

B. A digital twin is connected to data about a physical asset or process and represents its changing state.

C. A digital twin contains only photographic textures.

D. A digital twin cannot be used through a web browser.

Correct Answer: B. A digital twin is connected to data about a physical asset or process and represents its changing state.

Question 2

A VR application produces 90 frames per second but still feels uncomfortable because input is processed late and frame timing varies. Which conclusion is most appropriate?

A. Average frame rate alone is insufficient; end-to-end latency and frame-time stability must be measured.

B. The application should increase polygon count to improve realism.

C. Stereoscopic rendering should be removed from every VR application.

D. Network bandwidth cannot affect perceived responsiveness.

Correct Answer: A. Average frame rate alone is insufficient; end-to-end latency and frame-time stability must be measured.

Question 3

What is the main purpose of registration in augmented reality?

A. To compress textures before transmission.

B. To align virtual content with a physical or shared reference frame.

C. To convert raster images into SVG graphics.

D. To increase the number of audio channels.

Correct Answer: B. To align virtual content with a physical or shared reference frame.

Question 4

Which technique allocates greater rendering detail near the user’s gaze?

A. Occlusion culling

B. Deferred shading

C. Foveated rendering

D. Texture baking

Correct Answer: C. Foveated rendering

Question 5

A technician’s hand gestures frequently fail because gloves and machinery occlude the hands. Which design response is most robust?

A. Require more complex gestures.

B. Remove all feedback from the application.

C. Offer alternative input such as buttons or voice, with clear confirmation.

D. Increase visual texture resolution.

Correct Answer: C. Offer alternative input such as buttons or voice, with clear confirmation.

Question 6

Why can excessive virtual depth variation cause discomfort in a stereoscopic display?

A. It eliminates all motion parallax.

B. It can intensify the accommodation-convergence conflict and visual fatigue.

C. It prevents the GPU from using shaders.

D. It guarantees that tracking will drift.

Correct Answer: B. It can intensify the accommodation-convergence conflict and visual fatigue.

Question 7

Which statement reflects responsible handling of immersive-device data?

A. Room scans and gaze data are harmless and need no consent.

B. All collected data should be retained indefinitely for future analysis.

C. Data collection should follow minimisation, informed consent, access control, and retention limits.

D. Privacy applies only to voice data, not body movement.

Correct Answer: C. Data collection should follow minimisation, informed consent, access control, and retention limits.

Question 8

A web visualisation must work on powerful desktops and basic mobile devices. Which approach best follows progressive enhancement?

A. Require immersive hardware before showing any content.

B. Provide a meaningful baseline experience and add advanced graphics when capabilities permit.

C. Disable accessibility features to reduce processing.

D. Use the highest resolution regardless of device capability.

Correct Answer: B. Provide a meaningful baseline experience and add advanced graphics when capabilities permit.

Question 9

What is a key advantage of separating measured, estimated, and manually confirmed values in a digital twin?

A. It hides uncertainty from users.

B. It makes all values appear equally authoritative.

C. It communicates data provenance and supports better decisions.

D. It removes the need for sensors.

Correct Answer: C. It communicates data provenance and supports better decisions.

Question 10

Which evaluation plan is most appropriate for a safety-sensitive AR maintenance application?

A. Evaluate only visual realism using screenshots.

B. Measure tracking accuracy, latency, task errors, accessibility, environmental robustness, and user safety.

C. Ask one developer whether the interface feels intuitive.

D. Evaluate only average frame rate in a laboratory.

Correct Answer: B. Measure tracking accuracy, latency, task errors, accessibility, environmental robustness, and user safety.