Virtual Reality

Virtual reality (VR) is a computer-generated environment that a person perceives as a place and can act within. A typical system presents separate images to each eye, updates them in real time as the user moves their head or body, and accepts input through controllers, gestures, or other sensors. Its defining aims are immersion, the technical degree to which a system surrounds the senses, and presence, the psychological sense of being inside the simulated space. VR belongs to the broader field of computer science and draws on optics, neuroscience, and computer vision. It is usually placed alongside augmented and mixed reality on a continuum that runs from wholly physical environments to wholly synthetic ones.

The technology matters for several reasons. It has become a consumer entertainment medium and a platform for social interaction. It is also a tool for training, engineering design, and medical treatment, where simulated environments can be safer, cheaper, or more controllable than real ones. Its history, from mid-twentieth-century head-mounted displays to the consumer headsets of the 2010s and 2020s, shows repeated cycles of hype and retrenchment. Those cycles were driven by the balance between display quality, tracking accuracy, latency, cost, and comfort. As of 2026, market data show that dedicated headset shipments have been falling while smart glasses and other eyewear grow, which shows how the category is still changing.

This entry first defines VR and traces its origins from early stereoscopes and head-mounted displays through the research labs and commercial waves that followed. It then explains the perceptual principles, hardware, tracking, rendering, and software that make VR work, along with the main system types. The entry next examines human factors such as cybersickness, the major application areas in entertainment, training, medicine, and education, and the relation of VR to augmented and mixed reality. It closes with market dynamics, privacy and ethical questions, and open technical problems.

Definition and core concepts

Virtual reality (VR) is generally defined as a computer-generated environment that a user experiences as a place rather than as an image on a screen. The defining change is that the user is situated inside the scene, and the software updates what they see and hear as they move. A three-dimensional scene shown on a monitor and controlled with a keyboard and mouse is therefore not VR in this strict sense, even though it uses the same rendering techniques. A 360-degree video is also excluded by many practitioners: the viewer can turn their head to look around a recorded scene, but the viewpoint was captured by a camera and cannot respond to the viewer's movement through the space. Others accept 360-degree video as a lightweight, "non-interactive" form of VR, and this boundary remains a matter of convention.

Two terms are central to discussions of VR and are often confused. Immersion refers to the objective properties of the system: the field of view, the fidelity and latency of tracking, stereoscopic depth, audio, and how many sensory channels the system engages. Presence refers to the user's subjective, psychological sense of "being there" in the virtual environment. Researchers generally treat immersion as something a designer can measure and presence as something each person experiences differently. The two are linked but not equivalent: a well-engineered headset does not guarantee presence for every user, and compelling content can produce presence in a less technically capable system.

VR systems rely on a perception-action loop. The system senses the user's head orientation and, in more advanced configurations, position, hands, eyes, and body. It then renders the scene from that viewpoint and returns the result to the senses through displays, speakers, and sometimes haptic devices. The loop must close quickly, because delay between a movement and the corresponding change in the image undermines the sense of place and can produce discomfort. Because the user's own actions alter the sensory input, this closed loop is what separates VR from passive media such as film, although the boundary between interactive and passive experiences is itself debated.

There is no universally accepted definition of what counts as VR. One camp requires a head-mounted display with head tracking that blocks the physical surroundings. A broader camp accepts any system that surrounds the user with a stereoscopic, perspective-correct image in real time, which includes projection-based rooms such as CAVE systems and simulator domes. Proponents of the broader definition point to their wider sensory coverage and to decades of research on them. Critics argue that extending the term to cover non-head-mounted and even desktop configurations dilutes it and makes comparisons between studies difficult. Similar disputes arise over whether systems that overlay digital content on the physical world, which are usually classed as augmented reality, belong under the same heading, a question that the later sections of this entry address through the reality-virtuality continuum.

Precursors and early head-mounted displays, to 1968

The perceptual basis of VR predates computers by more than a century. In 1838 Charles Wheatstone showed that the brain fuses two slightly different flat pictures, one seen by each eye, into a single impression of depth. His mirror-based stereoscope made this effect demonstrable with drawings and later photographs. David Brewster's lenticular viewer of the late 1840s made stereoscopy a popular parlour pastime. The principle behind all of these devices, delivering a distinct image to each eye, is the one every later headset relies on.

Training simulators came next. In 1929 Edwin Link, a pilot and the son of an organ maker, built a fuselage-shaped trainer that rested on a universal joint and was tilted by pneumatic bellows, letting a student practise instrument flying on the ground regardless of weather. The device was commercialised that year and became a standard pilot-training aid during the Second World War. It had no visual display of a synthetic world, but it set the pattern later VR systems extended: an artificial environment that responds to the trainee's controls. The flight simulator lineage it began is the clearest institutional ancestor of simulation-based training.

Morton Heilig, a cinematographer, pursued a more overtly sensory goal. His Sensorama, developed from the late 1950s and patented in 1962, was an arcade-style cabinet combining stereoscopic 3D film, stereo sound, fans, scent generators and a vibrating seat. Its short films included a motorcycle ride through a city. It was mechanical rather than computational, and it failed commercially. In 1960 Heilig also patented the Telesphere Mask, usually regarded as the first head-mounted display. It used miniature television tubes to present stereoscopic images with wide peripheral vision and stereo sound. It had no motion tracking, so the picture did not change when the wearer turned their head. It was therefore a wide, immersive screen fixed to the face rather than a responsive environment.

Ivan Sutherland supplied the conceptual step in his 1965 essay "The Ultimate Display." He argued that a computer-controlled room could govern the behaviour of matter, so that a displayed object such as a bullet would act like a real one. In 1968 Sutherland and his student Bob Sproull built a head-mounted system at Harvard that became known as the Sword of Damocles. Because the headset was too heavy to wear comfortably, it hung from a jointed mechanical arm fixed to the ceiling. The arm also measured head position, allowing the computer to redraw the scene from the viewer's current perspective. Two small CRTs beside the user's ears fed images to half-silvered mirrors, which showed stereoscopic white wireframe graphics, such as a floating cube, superimposed on the real room. This dual character is why the system is often cited as an early precursor of both VR and AR. Its tracking was mechanical, and the fully integrated headset came only after Sutherland moved to the University of Utah later in the decade.

Research labs and early commercial attempts, 1970s to 1990s

Video-based responsive environments

Myron Krueger took a route that avoided head-mounted hardware. In 1969 he worked on Glowflow, a darkened room with sensors and phosphorescent tubes that lit up or changed sound when people moved through it. The sensors captured the movements of participants in the room and lit up the tubes or changed the sounds produced by a synthetizer. His 1970 Metaplay combined live video and computer graphic images and projected them in front of the viewer. VIDEOPLACE, the best-known result, put the participant's video silhouette inside a computer-generated scene. Krueger aimed for environments that worked without goggles or gloves, and he published his framework as Artificial Reality in 1983. The work done in the lab would form the basis of his much cited 1983 book Artificial Reality.

The origin of the term "virtual reality" is disputed. Krueger used "artificial reality," and that phrase remained popular in Japan for some time. At about this time the phrase "artificial reality" was popular in Japan. Jaron Lanier is generally credited with coining "virtual reality," and one account places the moment in a conversation about the VIEW lab at NASA Ames. He recalled a lunch in which Lanier described what Scott Fisher was doing at the NASA Ames Virtual Interactive Environment Workstation lab. Other sources date Lanier's usage to 1989. The term also appears earlier in Antonin Artaud's 1938 writing on theatre, a precedent that is not always mentioned.

Military and government work

Military programs ran in parallel. The US Air Force's Super Cockpit program, led by Thomas Furness in the 1960s and 1970s, explored helmet-mounted displays that would present sensor and flight data to pilots in three dimensions. The Air Force's helmet systems were expensive, which pushed NASA researchers toward cheaper personal simulators. Influenced by a range of prior projects that included Sensorama, flight simulators, and arcade rides, and surprised by the expense of the air force's Darth Vader helmets, Fisher's group focused on building low-cost, personal simulation environments.

At NASA Ames, the VIEW project was founded in 1985 under Scott Fisher. Its goal was telerobotics for space stations, but the group also considered entertainment and education. The VIEW workstation, called the Virtual Visual Environment Display when completed in 1985, established a standard suite of VR technology that included a stereoscopic head-coupled display, head tracker, speech recognition, computer-generated imagery, data glove, and 3-D audio technology.

VPL Research

VPL Research, founded by Jaron Lanier, turned these ideas into products. Its DataGlove used fibre-optic sensors to measure finger bending, with a magnetic tracker for hand position. VPL Research has designed a DataGlove which employs fiberoptics to monitor flexion and extension of the user's fingers, and a Polhemus magnetic tracking device to position the user's hand in a three dimensional matrix. Its EyePhone head-mounted display used two colour LCD screens behind an optical system, and each monitor is driven by a separate computing path. VPL's commercial visibility did much to spread the term "virtual reality" beyond research circles.

CAVE and projection-based rooms

In 1992 researchers at the Electronic Visualization Laboratory at the University of Illinois Chicago introduced the CAVE, a room in which stereoscopic images are projected onto several walls, and a tracked user sees the scene through shutter glasses. Unlike head-mounted systems, it allowed a small group to share the view and avoided heavy headsets. CAVE-style rooms were adopted for scientific visualization, automotive design and engineering review.

The mid-1990s setback: Virtual Boy

Consumer efforts stumbled in the mid-1990s. Nintendo's Virtual Boy was a tabletop stereoscopic console released in Japan on July 21, 1995, and in North America at a suggested retail price of US$179.95. Its display was a monochrome red-and-black screen, and its resolution was 384×224. Nintendo removed head tracking during development, citing concerns about motion sickness and the risk of eye problems in young children. This means the console's failure was not caused by poor tracking alone; the company had deliberately dropped motion tracking, which left a device that was no longer a true head-tracked VR system.

Sales were weak. Its failure has been cited as due to its high price, monochrome display, unimpressive 3D effect, lack of true portability, health concerns, and low quality games. Price cuts did not reverse this. Nintendo eventually had to lower it to $99 as sales stagnated. The console sold about 770,000 units, and twenty-two games were released commercially for the system before it was discontinued. The episode shows how cost, weak visuals, limited content and health concerns combined to undermine a product whose hardware was not yet mature.

Consumer resurgence, 2010s to 2020s

The modern consumer wave started with crowdfunding. Oculus, the company behind the Rift, opened a Kickstarter campaign in August 2012 with a goal of $250,000. The project passed that figure in two hours, and by the time the 30-day campaign closed, backers had pledged $2,437,429. Its first product was a development kit for developers, not a finished consumer device.

In March 2014 Facebook agreed to buy the company. The deal was reported at about $2 billion: $400 million in cash and 23.1 million Facebook common shares valued at $1.6 billion, plus an additional $300 million tied to milestones. Many early backers objected that the sale contradicted the independent ethos of crowdfunding, and game industry figures, including Minecraft developer Markus Persson, criticized the sale. The acquisition is the starting point of Meta's VR business, which later operated under the Reality Labs division.

The 2016 launches established the tethered model. The first consumer Rift, the CV1, was released on March 28, 2016, in 20 countries, at a starting price of US$599. HTC's Vive, developed with Valve, shipped the same spring and used room-scale external base stations. Sony's PlayStation VR followed in October 2016 for the PlayStation 4. These headsets depended on a powerful PC or console and, in the case of the Vive and Rift, on external tracking sensors.

Smartphone-based viewers offered a much cheaper route. Google's Cardboard, first unveiled at the Google I/O developer conference in June 2014, was designed as a low-cost, accessible entry point to virtual reality, letting users build their own viewer from cardboard and lenses. Google handed out 6,000 kits at the event and published a tutorial online, and homemade versions appeared within hours. Shipment figures conflict: a 2016 report cited 10 million units shipped, while a later reference lists 15 million units shipped. The official viewer was discontinued in March 2021. Samsung's Gear VR, launched in November 2015 and designed specifically for Samsung smartphones, showed that a more capable phone-powered headset could reach a mass audience. Phone-based systems have limits, however. Because they often lack positional tracking, the risk of motion sickness can be higher, and they usually have lower-resolution displays.

Standalone headsets removed the phone and the PC. They carry their own processors, displays, sensors and batteries, which makes them cheaper to buy and easier to set up, but they must work within tight limits of power and heat. Later models added color passthrough cameras; Meta's Quest 3, released in 2023, was positioned as an affordable mixed-reality headset with color passthrough.

Apple's Vision Pro marked a shift in positioning. When the Vision Pro launched on Feb. 2, 2024, 600 apps were already available, and the device sold at a starting price of $3,500. Apple avoided the term "headset" and called the product a "spatial computer." Its hardware uses dual micro-OLED displays with roughly 23 million pixels combined, and its color passthrough uses external cameras to reconstruct the real world in real time. Apple itself described the device as more VR at the hardware level, but AR at the user experience level. Although Apple did not coin the term "spatial computing," its use of the term helped popularize it. Reviews praised the display and passthrough but also pointed to a battery pack and limited early software as drawbacks. Across the industry, the emphasis has moved from gaming toward wearable displays, sensing of the surrounding area, and interfaces based on hands, eyes and voice, with productivity and integrated apps and digital material in the user's real world as the selling points.

Perceptual basis: how the brain is convinced

Stereopsis and binocular disparity

The horizontal gap between the eyes, about 6.5 cm in an average adult, means each retina registers a slightly offset picture of the same scene. The visual cortex combines these two images into a single three-dimensional view. The brain reads the horizontal differences between them, known as binocular disparity, as distance: the greater the disparity, the closer the object is perceived to be. VR headsets reproduce this by drawing a separate image for each eye, offset by the amount a real object would produce at the simulated distance. A flat pair of displays can therefore produce volume, which is why stereoscopy essentially presents two-dimensional imagery on flat displays to trick the brain into perceiving a volumetric world.

Disparity is a trade-off rather than a free gain. Stereoscopic depth is most useful at close, arm's-length ranges, since stereopsis has been shown to matter mainly in close viewing. Experiments also show that a stronger disparity increases the perceived depth magnitude but tends to reduce the perceived realism, precision, and fusion of the image. Too much disparity makes the two images hard to fuse into one percept. Research has further found that disparity can clearly enhance subjectively experienced depth without producing matching gains in objective task performance, so a convincing sense of depth should not be equated with better behaviour.

Motion parallax, head tracking and vestibular agreement

Depth is also signalled by motion. When the head moves, nearer objects sweep across the visual field faster than distant ones; this motion parallax is a monocular cue that works with one eye. A head-tracked renderer recomputes the perspective of the scene from the user's current position, so the virtual world behaves like a window onto a stable space rather than a picture that slides with the head. Latency between head movement and image update weakens this effect, which is why tracking precision matters.

Balance adds another layer. The vestibular system, located in the inner ear, senses head rotation and linear acceleration, and its signals must agree with what the eyes report. When the display shows the body moving, for example during joystick-driven locomotion, while the inner ear reports that the body is still, the two streams disagree. This sensory conflict is among the leading explanations of cybersickness, discussed in the human-factors section.

Place illusion and plausibility illusion

Mel Slater's framework divides presence, the sense of being there, into two components. Place illusion (PI) is the qualia of having a sensation of being in a real place. Plausibility illusion (Psi) is the illusion that the scenario being depicted is actually occurring. Slater argues that the two are orthogonal, meaning they can vary independently: it is possible for PI to occur independently of Psi and vice versa. A user may feel located in a convincing room while the virtual characters in it fail to respond, leaving plausibility low. Both illusions are said to hold while the participant still knows that the experience is artificial. Measurement remains difficult; the traditional questionnaire methods have limitations, and researchers have turned to physiological signals such as EEG to assess these components. The framework is theoretical, and correlations between presence and outcomes such as learning or task performance have been inconsistent across studies.

Depth cues and the vergence-accommodation conflict

In ordinary vision, two processes are tied together. Vergence is the inward or outward rotation of the eyes to fixate an object, and accommodation is the change in the lens's focus. Normally they move in step, so the depth at which the eyes converge also matches the focal distance. Stereoscopic displays break this link. The vergence–accommodation conflict (VAC) is a visual phenomenon that occurs when the brain receives mismatching cues between vergence and accommodation of the eye. In a head-mounted display the eyes converge on virtual objects at many distances, but the focal plane of the optics stays fixed. The eyes rotate to fixate an object in the same way as with physical objects, but must focus on the fixed plane of the display in order to see the virtual object clearly. The optics are often designed so that the virtual image plane sits somewhere between about two metres and infinity, but the conflict persists at any fixed focal distance.

The practical consequence is discomfort. The conflict can be unpleasant and cause eye strain. A user experiencing only vergence or accommodation, and not both simultaneously, can experience some degree of fatigue or nausea. Designers respond by keeping frequently inspected content near the fixed focal distance, limiting extreme depth ranges, and developing optical solutions such as varifocal and light-field displays, which are treated in the displays section. Whether the conflict measurably harms performance on specific tasks is still an active research question.

Displays and optics

A VR headset places a small panel a few centimetres in front of each eye, then uses optics to stretch that panel across a wide part of the visual field at a comfortable apparent distance. The choice of panel, lens and pixel density together determines how sharp, bright and comfortable the image is.

Panel technologies

Affordable headsets mostly use LCD panels. An LCD depends on a backlight behind a liquid-crystal layer, polarizers and colour filters, so light leaks through pixels that are meant to be black. The result is a dark grey haze rather than true black, with contrast typically in the low thousands to one. LCDs are cheap, mature to manufacture and can reach high brightness. Their pixel transitions are slower, which can smear fast-moving content.

OLED panels are self-emissive: each pixel switches off completely, which gives true blacks and quick response. Many OLED panels, however, use a PenTile subpixel layout that shares subpixels between neighbouring pixels, which lowers effective resolution. Burn-in and uneven brightness across the panel (often called mura) are further concerns.

Micro-OLED (OLEDoS) fabricates OLED pixels directly on a silicon wafer, using processes close to those of chip manufacturing. The pixels are far smaller than those of conventional panels, so pixel density is very high. Apple Vision Pro and Samsung's Galaxy XR both use micro-OLED displays. Wafer-based production is difficult and costly, and panels of this kind remain expensive, which keeps them mainly in premium and enterprise products.

Persistence blur is a separate issue. In a conventional sample-and-hold display, each frame stays lit for its whole duration, and the eye's tracking of a moving object smears it across the image. Fast-switching OLED and micro-OLED panels reduce this, and some headsets additionally strobe the light to shorten the visible time of each frame.

Lens systems

A Fresnel lens is a thin element with concentric ridges that bends light with little material. It is cheap and can offer a wide field of view, but its ridges scatter light and produce "god rays" (bright streaks around high-contrast edges), along with colour fringing and pincushion distortion that software must correct. Sharpness also falls toward the edges, leaving a narrow central sweet spot. Fresnel designs need a relatively large gap between lens and screen, which makes headsets deep and heavy.

Pancake lenses fold the light path. Polarization-sensitive reflective layers and waveplates make light bounce between two or more elements, so the display can sit much closer to the lens. This allows thinner and lighter headsets, such as Meta Quest 3 and Apple Vision Pro, and gives sharper, less distorted images toward the edges. The cost is light loss at each reflection, which means the panel must be brighter and draws more power to produce a comparable image.

Lens design also governs eye relief, the distance between the lens and the eye. Adequate eye relief lets users wear glasses, while a larger field of view usually demands either more lens size or more compromise in edge quality. Designers therefore trade field of view, sharpness, eye relief and overall size against one another.

Key metrics

Field of view is usually quoted in degrees. Most consumer headsets fall between roughly 90° and 120° horizontally, and the Quest 3 is specified at 110° horizontally and 96° vertically. Because the same pixel count spread across a wider angle yields fewer pixels per degree, field of view and sharpness pull against each other. Pixels per degree (PPD) is calculated by dividing horizontal pixels by horizontal field of view, so a panel with 2,000 horizontal pixels across 100° gives 20 PPD. Human central vision is commonly estimated to resolve about 60 PPD. A 2020 analysis put commercial headsets at around 14 PPD, although modern headsets have improved considerably.

Low PPD makes the gaps between pixels visible, producing the screen-door effect. Reference thresholds vary; one display reference holds that it is hard to notice above roughly 25 PPD and essentially absent above about 50 PPD for most users. Refresh rate matters as well. Many current headsets run at 90 Hz or higher, which reduces judder and helps keep the image stable during head movement.

Interpupillary distance and focus

Headsets provide an adjustable interpupillary distance (IPD) so that each lens's optical centre lines up with the wearer's pupils. Adult IPD values broadly cluster between about 50 and 70 mm. Misalignment causes blur, distortion and eye strain, and some headsets offer only coarse steps or fixed settings. Fixed-focus optics place the virtual image at a single depth, which conflicts with the eyes' natural focusing behaviour and contributes to the vergence–accommodation mismatch discussed elsewhere in this article.

Research on varifocal displays aims to fix this by changing focal distance to match where the user is looking, typically using eye tracking to drive moving lenses or displays. Meta's 2018 Half Dome prototypes demonstrated mechanically varifocal optics. Light-field displays take a different approach, reproducing the directions of light rays so that the eye can focus naturally on virtual depths; they usually trade spatial resolution and require heavy computation. These approaches remain largely experimental in consumer products.

Tracking, input and haptics

Outside-in and inside-out tracking

Tracking estimates where the headset and controllers are, and it has two main architectures. Outside-in systems place sensors in the room and have the headset and controllers read them. The HTC Vive's Lighthouse system is the standard example: fixed base stations sweep infrared lasers across the play area, and photosensors on the devices compute their absolute position and orientation. This approach is widely regarded as the precision benchmark, but the base stations must be mounted, aligned and powered, which limits portability.

Inside-out systems put the cameras and sensors on the headset itself. They estimate motion by tracking features in the surrounding room, a method related to computer vision and simultaneous localization and mapping. Because nothing needs to be installed beforehand, inside-out tracking became the default for standalone headsets, and most recent headsets use it. Its weaknesses are sensitivity to low light, featureless walls, fast hand movement and occlusion of the cameras' view, and it is generally considered less precise than a well-calibrated outside-in setup for professional motion capture.

Inertial sensing, sensor fusion and degrees of freedom

Almost every headset contains an inertial measurement unit (IMU), which combines accelerometers, gyroscopes and often magnetometers. An IMU reports orientation and acceleration at very high rates, which keeps the view responsive between camera or base-station updates. However, integrating its signals drifts over time, so position estimates from an IMU alone degrade quickly. Sensor fusion algorithms, typically Kalman-filter or similar estimators, combine IMU data with optical or laser observations to correct the drift while keeping the low latency of the inertial data.

The distinction between three and six degrees of freedom (DoF) governs what a device can register. A three-DoF device tracks only rotation (pitch, yaw and roll), which is enough to look around in a seated experience such as a 360-degree video. Six-DoF tracking adds translation along three axes, so the user can lean, crouch or walk, and the view responds to real head displacement. This is what makes parallax and room-scale movement possible.

Controllers, hand, eye and body tracking

Handheld controllers were the first widely used input devices. They are tracked like the headset and usually include buttons, analog sticks, and vibration motors. Hand tracking uses cameras on the headset to reconstruct finger positions, removing the need for a controller, but it struggles with occluded fingers and gives less precise button-like input. Eye tracking uses small inward-facing cameras to measure gaze direction. It supports selection by looking, avatar expression and foveated rendering, which is discussed in the rendering section. Full-body tracking adds worn trackers on the limbs or torso, or infers posture from the headset and controllers, so that an avatar's legs and hips can move with the user. Each addition increases realism and expressive range, but also adds cost, calibration steps and occlusion problems.

Haptic feedback

Haptic feedback can be grouped by the physical effect it produces. The most common form is vibrotactile, in which small motors in controllers or wearables create buzzes that signal impacts, button presses or textures. It is cheap and widely deployed but conveys little about the size or stiffness of an object. Kinesthetic, or force, feedback resists the user's movement. Force-feedback gloves typically attach a small actuator to each finger, sometimes through a tendon system, so that a virtual object can block the fingers from closing. Exoskeletal arms provide larger forces and whole-limb resistance, but they are bulky, restrict movement and are mainly found in laboratories or specialized settings. Research prototypes explore lighter alternatives, such as pneumatic gloves that aim to be untethered, and skin-deformation devices that press against the skin to simulate contact and shape.

Locomotion

Moving through a virtual world that is larger than the physical room is a central design problem. Joystick locomotion is simple but is associated with more motion sickness than the alternatives. Teleportation, in which the user points to a destination and jumps there, is popular because it avoids much of the sensory conflict that causes discomfort, although it reduces the amount of natural walking. Redirected walking subtly rotates or translates the virtual scene so that a user who believes they are walking straight actually follows a curve within a small physical space. In a room-scale comparison, redirected walking supported the best unconscious spatial learning, while participants preferred both it and teleportation to joysticks. Treadmills and omnidirectional platforms offer another route to natural walking, though they require substantial hardware. Many applications combine methods, using redirected walking for nearby targets and teleportation for distant ones.

Rendering, computing and audio

Rendering load and stereo pipelines

A headset must produce two slightly offset images at once, and both must be redrawn at a high rate, so the rendering cost of a VR scene is roughly doubled before any other work is added. The problem has grown with display resolution. Per-eye pixel counts rose from about 2.3 million on a 2019 Oculus Rift to about 14 million on the 2024 Varjo XR-4. Stereo pipelines therefore try to share work between the eyes, for example by culling geometry once and drawing both views in a single pass, rather than running the full scene twice. The frame budget is tight: GPU rendering time depends on scene complexity and GPU power, and for a 90 Hz target it must be under 11.1 ms. This is why VR content is usually tuned around the GPU and parallel throughput of the host, and why mobile headsets must work within a far smaller power envelope.

Motion-to-photon latency

Motion-to-photon latency is the interval from a physical movement, usually of the head, to the moment the display emits light reflecting that movement. It is also called end-to-end latency, and it typically includes sampling the head-tracking sensors, combining sensor information, rendering, and scanning out the image. The budget is spread across several stages. Inertial sensors sample at roughly 1000 Hz, while cameras run more slowly, at about 30–90 Hz. Without mitigation, the total can easily exceed 50–100 ms. Ideally, one paper argues, the delay should be only a few milliseconds, and it should be 20 milliseconds or less to avoid problems.

Asynchronous timewarp and reprojection

The main software countermeasure is timewarp. Pose prediction estimates future head position and orientation from motion-sensor data and kinematic models such as Kalman filters or neural networks, and it serves as a foundation for latency reduction. Shortly before display, the compositor retrieves fresh orientation data and calculates a transformation matrix that moves each eye texture from its render-time pose to its display-time pose. The same warp step corrects for optical distortion of the lenses while reducing motion-to-photon delay, so one pass handles two jobs. Asynchronous timewarp runs separately from the stereoscopic renderer, so it can increase the perceived frame rate and smooth out inconsistent frame rates. Meta's documentation now marks its native mobile ATW interface as deprecated in favour of OpenXR, which shows how these techniques are increasingly managed by runtimes rather than individual applications.

Foveated rendering

Foveated rendering exploits the fact that acuity is highest at the fovea. It uses eye tracking to render high detail only in the user's central gaze area while using lower quality in peripheral vision, which substantially reduces computational load. Early versions were static: the first version of one vendor's technique increased effective resolution at the lens centre while decreasing quality elsewhere. Dynamic versions move the supersampled area to wherever the eye is looking. The idea dates to 2014, when a VR startup called Fove showed foveated rendering on a headset that was also the first VR headset to feature eye tracking. Latency is the critical constraint: one platform's eye-tracked pipeline runs its cameras at the same frame rate as the display and is synchronised with the rendering pipeline to minimise the delay between eye movement and display time. Cost remains a barrier, since hardware costs add hundreds of dollars to headset prices, and some systems instead use gaze prediction, a lower-cost alternative for devices without eye-tracking hardware.

Spatial audio and head-related transfer functions

Spatial audio supports presence by giving sound a believable origin. Its central tool is the head-related transfer function (HRTF). The geometry of the head and ears changes sound depending on its direction, and the HRTF captures these directional effects to localise sound. Formally, an HRTF is a set of filters describing the acoustic transformation of a sound as it travels from a specific location in space to a listener's ear canals. In practice, a computed HRTF convolved with a single audio signal yields left and right channels that carry a spatial component. Because the result must track the user, the sound's placement has to be recalculated as the listener turns or moves. Accuracy is limited by individual differences: because of high inter-individual variability in anatomy, HRTF measurement, modelling and simulation are critical for high-fidelity binaural audio. Studies also find that general, non-individualised HRTFs limit three-dimensional audio perception in VR.

Computing architectures: tethered, streamed and standalone

The computing location determines the trade-offs. A tethered PC headset draws on a desktop GPU with the most power and the widest content range, but the cable limits movement. Wireless PC streaming removes the cable, yet it adds video encoding, decoding and network transit to the motion-to-photon chain. With cloud-based or streaming VR, the speed and stability of the internet can have a negative impact on latency. Cloud rendering moves the heaviest computation to remote servers, which can be useful for cloud computing economics but is most sensitive to network jitter. Standalone headsets avoid these links by running on their own processors and are constrained by battery capacity and thermal limits, so they typically use mobile-class chips and lower resolution or simpler shading than tethered systems. Each approach therefore trades visual fidelity, latency and mobility against one another.

Software, content and standards

VR content is usually built in a small number of real-time engines. Game engines such as Unity and Unreal Engine supply the pieces that most VR projects need: stereo rendering, scene graphs, physics, audio, animation, and input handling. Both engines also include VR templates and device plugins, so a studio can start with a working locomotion and interaction system instead of writing low-level drivers. Engines have become the main point where hardware differences are hidden from developers. Content that is not created in a game engine, such as scanned environments or recorded performances, is often imported into one. photogrammetry, which reconstructs 3D geometry and textures from overlapping photographs, is widely used for real places and objects. Volumetric video goes further by capturing a moving person or scene as a sequence of 3D frames that a viewer can look at from different positions. Both techniques produce very large, heavy assets, which makes polygon budgets, texture compression, and streaming important in headsets with limited memory and power.

Cross-platform access is provided by standards. OpenXR is a royalty-free, open API maintained by the Khronos Group that exposes common functions for application lifecycle, rendering, tracking, frame timing, and input. Its purpose is to let one application run on several headsets without a separate port for each vendor's SDK. Without such a layer, each device could run only the software written for its own platform, which raises development costs and limits market growth. OpenXR 1.1, released in April 2024, moved several widely used extensions into the core specification, including standard interaction profiles for controllers and hand-related inputs. The Khronos announcement also lists Unity, Unreal Engine, and Godot among the major engines that support the standard. Conformance testing and a registry of certified products are meant to make behavior consistent across vendors, although differences in runtime quality, supported features, and performance still exist in practice.

On the web, the web platform offers a separate route through the WebXR Device API. It is a W3C standard developed by the Immersive Web community and working groups, and it gives web pages access to head tracking, controllers, hand input, and display output through standard APIs, usually together with WebGL. Its main advantage is distribution: a user can enter an immersive session by opening a link, without an app store submission or a separate installation. WebXR replaced the earlier experimental WebVR effort and covers both VR and AR sessions under a single API. Browser support has improved since Chrome 79 in 2019, but it is still uneven across devices, and feature support for hand tracking, anchors, and other modules depends on the browser and the hardware. Web experiences are therefore usually simpler and lighter than native titles, and developers often treat them as prototypes, demos, or marketing tools.

Distribution through app stores is a separate issue from technical standards, and it strongly affects developer economics. The Meta Horizon Store, which began in 2013 as Oculus Share, is the largest storefront for standalone headsets, and Meta describes it as the main game and app store for Meta Quest devices. Developers pay a revenue share defined in their agreement with Meta; the developer dashboard reports that share per transaction, and the portion remaining after it is the developer's payout. Analysts' estimates of the typical split vary, with one 2026 analysis assuming that developers usually keep about 70 percent after taxes. Meta reported that the store's gross revenue rose only slightly in 2025 compared with 2024, while more than 100 titles earned over one million dollars in gross revenue that year. Because a single dominant store controls review, discoverability, and payment, platform policies on content, pricing, and in-app purchases shape which projects are viable. Smaller developers often rely on sideloading, early-access channels, or alternative stores where these are permitted.

Interoperability problems do not end with the runtime. Assets are stored in different formats, input schemes are mapped differently, and performance targets differ between a tethered PC headset and a mobile-chip standalone device. A scene built for a high-end PC may need reduced geometry, lower-resolution textures, simplified lighting, and different shader code before it runs on a standalone headset. Shared interchange formats, such as the glTF standard for 3D models, reduce the cost of moving assets between tools, but they do not solve differences in material models, physics behavior, or input conventions. Porting therefore usually means rebuilding parts of an application rather than copying it, and many experiences are released on one platform first. Persistent user data, multiplayer services, and account systems are also platform-specific, so moving a game can mean losing progress or social connections unless the developer builds an export and migration path.

System types and device classes

Virtual reality systems are commonly grouped by how much of the user's visual field and body they enclose, and by where the computing happens. These groupings are practical rather than formal, and writers draw the boundaries differently.

Non-immersive, semi-immersive and fully immersive systems

Desktop VR shows a three-dimensional scene on an ordinary monitor, and the user navigates with a keyboard, mouse or game controller. The physical room stays fully visible and the screen frame bounds the view, so this is the least immersive form.

Semi-immersive systems place the image on large surfaces around the user. The CAVE (Cave Automatic Virtual Environment) is the best-known example. It was first built at the Electronic Visualization Laboratory of the University of Illinois at Chicago and demonstrated at SIGGRAPH in 1992. A CAVE is an immersive virtual reality environment where projectors are directed to three, four, five or six of the walls of a room-sized cube. The user typically wears stereoscopic glasses to see 3D images projected onto the walls, floor and sometimes the ceiling of the room. By tracking the user's head and adjusting the images projected in real-time based on their perspective and position, the CAVE creates the illusion of being inside the scene. Because the body stays in a real room, several participants can share one installation, though only the tracked user receives the correct perspective. Simulator domes, which project onto large curved screens for flight and planetarium-style training, belong to the same family.

Fully immersive systems use a head-mounted display (see head-mounted displays). Most VR devices are closed-style headsets or helmet-mounted devices (HMD) that block ambient light and allow the user to fully immerse themselves in the virtual world.

Headset classes

Tethered PC-VR headsets connect by cable to a computer, which performs the rendering. They make use of the potentially limitless processing power of the computer and, therefore offer the most immersive VR experience. The trade-offs are cost, since a capable computer is needed, and reduced freedom of movement. They come with the limitation of being tethered, restricting movement and potentially causing cable management issues during use.

Standalone headsets carry their own processors, displays, sensors and batteries. They offer increased mobility and comfort. The cost is performance: this comfort comes at the cost of reduced power and consequently lower resolution and refresh rates. Standalone devices suit simpler, more casual applications, like social VR or fitness, which require less processing power.

Console headsets, such as PlayStation VR, attach to a game console rather than a PC and share the tethered design's cable and placement constraints while depending on the console's hardware.

Smartphone-based viewers use the phone's own screen, sensors and processor. Smartphone headsets or holders enclose a compatible smartphone and use the phone's display, sensors, Wi-Fi support, and speakers. Designs range from folded cardboard to more elaborate holders, and it offers lower-quality graphics and processing power than tethered and standalone headsets. Because the hardware cost is largely the phone, this is often the least expensive option.

Hybrid headsets can run standalone and also connect to a PC or console. They attempt to combine the best aspects of standalone and tethered headsets.

Enterprise and industrial devices are a separate procurement category. These are often standalone or tethered units sold with fleet-management software, support contracts and ruggedized or clinically certified hardware. Enterprise-grade standalone devices that offer robust software management and support are typical of this market.

Video passthrough and optical see-through

Headsets that show the real world have two main designs, and they blur the line between VR and augmented reality (augmented reality). With video pass-through augmented reality headsets, users are looking at a screen that is displaying a continuous live feed of the world captured by one or more cameras on the front of the headset. Virtual imagery is superimposed on the video feed. With optical see-through augmented reality headsets, users are looking through a transparent lens or other optical component directly at the real world. Passthrough is the more common approach, because it is easier to design and is traditionally significantly less expensive.

The designs differ in several trade-offs. Passthrough can render opaque content: optical see-through devices can only display light, not black or opaque content. Passthrough also lets graphics and camera video share one pipeline, so it's much easier to match the video latency with the computer graphics latency. Its cost is peripheral vision, since video passthrough wearable devices provide a field of view (FoV) between 90 and 120 degrees, thereby limiting the user's peripheral vision. Optical see-through leaves the real view largely intact, and it keeps working in a failure: an optical see-through solution will allow you to see something in that extreme situation if power fails. Current VR headsets with color cameras are often classed as mixed reality devices, since VR HMDs utilizing passthrough can also be classified as VST HMDs or mixed reality (MR) HMDs.

Choosing a system type

Cost, comfort, mobility and computing power push buyers in different directions. A high-end tethered PC is the most capable but the most expensive, and it adds cables and a room layout that limits movement. Standalone headsets remove the computer and the cable at the price of graphical fidelity. Smartphone viewers are cheapest but offer the weakest graphics. CAVEs and simulator domes provide large-scale shared experiences, but they need dedicated rooms, projectors and maintenance, so they are used mainly by research labs, training centers and industry. Passthrough adds room awareness and mixed-reality use to a headset at the cost of optical clarity and peripheral view, which is why the same hardware decision can depend heavily on the intended task.

Human factors and health effects

Cybersickness and its explanations

Cybersickness is the umbrella term for discomfort during or after immersion. Reported symptoms include nausea, dizziness, headache, eye strain, disorientation and unsteadiness, and a motion sickness-like profile is common. Designers and hardware makers have identified several contributors: wide fields of view, low frame rates, display latency, artificial locomotion methods, and content with abrupt acceleration or strong self-motion illusions.

The most widely cited explanation is sensory conflict theory. Symptoms are often linked to a clash between the visual motion the eyes report and the signals from the vestibular and proprioceptive systems. Because the inner ear senses no acceleration while the image shows movement, the central nervous system receives contradictory input. A second account, postural instability theory, places the problem in balance control. Under this view, the body's usual balance strategies work less well in an unfamiliar virtual setting, and sickness follows a sustained failure to keep posture stable. Other proposals, including eye-movement (oculomotor) theory and the older poison theory, have received less support. Causality remains disputed: studies find that losses of postural stability coincide with the onset of cybersickness, but it is unclear whether they are the cause or a consequence.

Susceptibility varies widely. Individual differences track age, prior history of motion sickness, and how efficiently a person integrates sensory information. Some studies also report that people with impaired vestibular function are especially vulnerable, which supports the sensory conflict account. Newer headsets generally produce lower symptom scores than early systems, but substantial symptoms persist, driven mainly by sensory mismatch and perceived self-motion.

Visual fatigue and focus-convergence mismatch

Eye strain has a distinct optical cause. In a vergence-accommodation conflict, the brain receives mismatched cues between where the eyes converge and where they focus. In a headset, the screen sits at a fixed focal distance, while stereoscopic images make the eyes converge on objects at other depths. The resulting decoupling places physiological stress on the oculomotor system. Researchers have linked this mismatch to visual fatigue and binocular instability, and it is considered a significant contributor to cybersickness as well.

Interpupillary distance (IPD) adds a second source of strain. If the lens centers are not aligned with the user's pupils, images are distorted and the eye muscles work harder to fuse them. Many headsets offer adjustment over a limited range; one consumer guide reports that most modern headsets allow adjustment between 58 and 72 mm. Users with unusual IPD values may find that no setting fits well. Common symptoms of prolonged use include tired eyes, dryness and temporary difficulty refocusing after removing the headset. Long-term effects have not been established.

Physical safety

Users wearing a headset cannot see their surroundings, so collisions with walls, furniture, pets, people and sharp edges are a recognized hazard. Tethered headsets add tripping risks from cables. Most room-scale systems therefore include guardian or chaperone boundary features, which define a play area during setup and display a visible grid or warning when the user approaches its edge. These systems depend on accurate mapping at setup and do not reliably detect objects moved afterward, so designers generally recommend clearing the space and using seated or stationary modes where possible.

Age advisories and research gaps

Manufacturers set minimum ages on the basis of precaution and product liability. One guide to current ratings reports that Meta sets a minimum age of 10 for Quest headsets, with parent-managed accounts required for children aged 10 to 12 and standard accounts available from 13. Sony recommends the PSVR2 for ages 12 and up, and Apple recommends the Vision Pro for 13 and up. Such ratings are primarily based on product liability and a precautionary principle rather than on extensive long-term longitudinal studies. Some professional guidance is stricter; one summary of pediatric advice reports that children under 6 should not use immersive headsets at all, since their visual systems and sense of balance are still developing.

The main research gap is the lack of long-term and developmental data. Most studies involve short sessions with adults, and few track children over years. It is therefore unknown whether repeated exposure during visual development alters depth perception or binocular function. Similarly, the effects of extended daily use on balance, sleep and attention have received limited study. Consumer-facing summaries often present these questions as open rather than answered.

Entertainment, games and social platforms

Entertainment is the largest consumer use of virtual reality. One market-research estimate put gaming at about 63% of VR end-user revenue in 2025. The Meta Quest store is the largest content marketplace, with cumulative revenue of roughly $3 billion by March 2025, and over 100 Quest games each passed $1 million in gross revenue in 2025. Genres map onto what VR does well. Physics-driven combat sandboxes such as Blade & Sorcery: Nomad depend on the player's hands and body, rhythm games such as Beat Saber use motion controllers to make movement the core mechanic, and simulation and sports titles rely on a convincing sense of place. Horror draws on the same sense of presence, since the threat appears close to the body rather than on a screen.

Free-to-play titles now dominate how people actually spend time in VR. Meta's own disclosures attribute about 70% of time spent on Quest headsets to free-to-play apps. Its top-revenue lists for 2025 include several free titles built around multiplayer play and social activity, such as Gorilla Tag and VRChat, alongside paid games. This pattern pushes studios toward games with long-lived communities and in-app purchases, and it has made engagement, not only unit sales, a key measure of the market.

Non-game entertainment spans home viewing and public venues. Cinematic VR presents linear films and narrative pieces in which the viewer looks around a scene, a format discussed in film and narrative. Live events such as concerts, sports and theatre are also staged for headsets, and industry analysts describe demand for special interactive experiences at these events as a driver of the location-based market. Location-based entertainment (LBE) is the most commercially visible branch. Operators run arcades, free-roam arenas, escape rooms and theme-park attractions. One market report estimated that traditional VR arcades generated about 45% of 2025 application revenue, while free-roam arenas were forecast to grow faster, at about 31.5% a year over 2026–2031. Group operators such as Sandbox VR use motion sensors to map players' movement inside a physical arena, and sell sessions for groups of two to six players. Dreamscape Immersive adds motorized floors, wind and heat to cinematic scripts.

Social VR platforms center on avatars and shared spaces rather than on a single game. VRChat, founded in 2014, lets users wear custom Unity-built avatars, supports full-body tracking and eye tracking on compatible hardware, and hosts user-built worlds. Its own published data show a peak of about 101,000 concurrent users between July 2023 and July 2024, and nearly 90,000 people joined a single New Year's Eve event in 2022. Rec Room, launched in 2016, takes a more curated approach, with in-game creation tools that give worlds a consistent look. Its avatars long lacked full-body inverse kinematics, which made arms and legs appear absent, and it has drawn a younger audience. Designing presence among many people raises problems that single-player games do not face. Avatars that track only the head and hands convey body language poorly. Eye tracking and lip sync are available only on some hardware, so users see different levels of social expression. Crowds also strain networking, audio and rendering. Moderation and personal safety, covered in the privacy and ethics section, shape these spaces directly. Platform demographics are hard to measure because the main companies are private and do not publish full user data.

Business models combine several revenue streams. Hardware is often sold at a loss or subsidized by the platform owner, and the owner recovers the cost through store sales and revenue shares. Meta's Reality Labs segment reported an operating loss of about $4.6 billion in Q2 2026, which illustrates how large these hardware and platform costs can be. Stores typically take a commission on app sales. Subscriptions add recurring income, and Meta reported more than one million Horizon+ subscribers in March 2026. Social platforms add creator economies: Rec Room lets premium users sell avatar accessories and rooms for real money, while VRChat's creators sell avatars and assets. These mechanisms give developers a strong incentive to build for the largest platform, which in turn reinforces the market concentration discussed elsewhere in this article.

Training, simulation and industrial design

Training is one of the oldest and most heavily funded uses of VR. A simulated environment can reproduce situations that are dangerous, expensive or impossible to stage, and the case for it usually rests on three arguments: errors made in simulation cause no injury, the consumables and equipment of real operations (aircraft hours, fuel, ammunition, vehicle wear) are not used up, and a scenario can be repeated, varied and recorded at will.

Flight, vehicle and military training

Flight simulation is the most regulated application. In 2024 the US Federal Aviation Administration qualified its first VR-based flight training device, a Loft Dynamics simulator for the Airbus H125 helicopter that uses a Varjo headset, so that US helicopter pilots can now use this simulator to train and receive credit toward official pilot ratings. The same product had been EASA certified since 2021. The manufacturer also claims the device is both cheaper and smaller than traditional FAA-qualified flight simulators, which use a complex array of huge screens or projectors. Qualification also shapes design. The cockpit in this device is not mixed reality, because the passthrough view on the manufacturer's headset isn't depth and scale correct and thus wouldn't meet the standards of regulators. A different system, TRU Simulation's Veris, was qualified at Level 7 and adds a dual-use element: using real avionics and aircraft components, some training can be conducted without a VR headset, which keeps the headset for the flight portion of training.

Military programs follow a similar logic. The US Army's Synthetic Training Environment blends merging live, virtual, constructive [LVC] and gaming platforms into an interoperable training experience, and the blending reduces wear on tactical equipment. Trials at Fort Cavazos in Texas tested VR replicas of vehicle crew stations, including the Abrams, Bradley and Stryker simulators that replicated training on the M1 Abrams main battle tank (MBT), the M2 Bradley Infantry Fighting Vehicle, and the STRYKER armored vehicle. Terrain generation is a further element: the Army's approach can rapidly generate these Synthetic Training Environments from satellite imagery and other public geographic data, and AI is used to analyze soldier performance to make the trainings more effective, and introduce variability into the simulations. Haptic devices that provide tactile feedback for weapon recoil, impacts, and environmental effects are being added to these systems. Emergency-response training follows the same pattern, with simulated casualties on which medical staff can practise first aid and emergency procedures in a safe, controlled environment.

Industrial prototyping, digital twins and maintenance

In manufacturing, VR is used for design review at full scale, before physical mock-ups are built. Engineers can check reach, clearances, sightlines and assembly sequences, and changes are cheaper to make in the virtual model than in tooling. Assembly and maintenance training uses the same models: technicians rehearse procedures on a virtual machine, which avoids taking production equipment out of service. The link between a virtual model and a physical asset is often described as a digital twin, a model updated with real sensor or operational data, which VR can make walk-through and inspectable. These uses are generally reported by vendors and manufacturers, and independent measurement of their savings is sparse.

Architecture, urban planning and automotive prototyping

Architects and planners use room-scale walkthroughs so that clients and the public can judge spatial proportions, daylight and circulation before construction. Urban planning applications add the experience of street-level scenes, such as crowd movement or the visibility of a crossing, to models of proposed districts. In the automotive sector, VR supports early evaluation of interior ergonomics, control placement, mirror and pillar visibility, and styling, which can reduce the number of physical clay models and prototypes needed before tooling. Fidelity is a key constraint in all these settings: a model that looks correct may still mislead a reviewer if scale, lighting or interaction differs from reality.

Evaluating transfer to real tasks

The central scientific question is whether skill acquired in VR is retained and transfers to real performance, as opposed to a skill that exists only inside the simulator. Researchers distinguish near transfer, where the real task closely resembles the training, from far transfer, where it differs, and they test retention after delays. Controlled comparisons are the standard method. In one study, forty-seven healthy adults were randomly assigned to train either in a virtual replica of the aisle or directly in the physical setting. Those who practised in VR later completed the real task faster and with shorter travel distances than people attempting it for the first time, but direct real-world practice produced larger gains in completion time, a result the authors relate to encoding specificity. Another study compared a virtual copy of a sensorimotor workstation with non-interactive text and video instruction; the VR group had significantly higher accuracy and produced fewer timeouts in the workstation task.

Evidence remains uneven. A systematic review of 64 studies on creativity, critical thinking, collaboration and communication found that only two studies focused on skill transfer. Motor-learning researchers also argue that simple laboratory tasks may not shed light on performance of more complex real-world skills. Designers therefore need to specify which features of the real task, such as perception, timing or force, must be preserved, and to test whether training effects last after the simulator is no longer used. Both the transfer of training literature and the cognitive-science study of skill learning provide the frameworks for these questions.

Medicine, psychology and rehabilitation

Exposure therapy for anxiety and phobias

Virtual reality exposure therapy (VRET) applies the logic of exposure-based treatment inside a generated scene. A patient is repeatedly brought into contact with a feared object or situation, such as a height, an animal or a crowd, at a dose the clinician can adjust. The approach is most established for specific phobias. A 2008 meta-analysis of 13 studies (n=397) found VRET highly effective against control conditions and slightly more effective than real-world (in vivo) exposure, though its authors cautioned that many studies reported their methods poorly. A 2019 meta-analysis restricted to randomized trials with matched exposure time found that VR and in vivo exposure both produced large effects, with no significant difference between them. A 2025 review of social anxiety and specific phobia reached a similar conclusion, reporting comparable moderate effects.

Presence appears to matter for the mechanism. One meta-analysis found that the correlation between self-reported presence and anxiety was strong for fear of animals but close to zero for social anxiety, and that six-degree-of-freedom tracking and wider fields of view were associated with larger effects. Evidence for more disabling conditions is thinner. Critics have argued that most earlier reviews concentrated on specific phobias, which are generally less severe, and that some included non-randomized studies. Treating agoraphobia or social anxiety may require varied public settings and responsive virtual people, which makes the scenarios harder to build and to validate. Work on post-traumatic stress disorder (PTSD) is less mature than work on phobias, and the research is still developing.

Pain distraction is a separate application. Immersive scenes are used to divert attention during procedures and painful episodes, based on the idea that attention competes with pain signals. The available studies are generally small and short, and they focus on acute and procedural pain, not chronic pain. Their outcomes are therefore harder to generalize.

Surgical planning, anatomy and training

Patient imaging such as CT or MRI can be converted into three-dimensional models for rehearsal before an operation, and VR offers a way to explore these models at scale. Anatomy teaching uses similar interactive three-dimensional models, though the educational gains depend on the design of the lesson.

Surgical simulators are the area with the largest evidence base, but the picture is mixed. A 2025 scoping review identified 235 studies from more than 30 countries, about half of them randomized trials, and most reported improvements in performance, time, errors and confidence. The same review noted a lack of evidence across all surgical specialties. Evidence on robotic surgery is less consistent. A 2020 meta-analysis found no compelling evidence that skills gained on robotic VR simulators transferred to the operating room. A later systematic review reported nine positive studies, three with no effect and one negative, and noted that much of the evidence came from non-human models. Simulator limitations also affect transfer. Haptic feedback is often poor, and some urology simulators render bleeding inadequately, while cadavers offer superior anatomical fidelity. Orthopedic training reviews identify latency and difficulties with depth perception as continuing problems.

Motor and balance rehabilitation after stroke

Most of the evidence on VR in rehabilitation after stroke concerns motor function. A meta-review of systematic reviews concluded that VR benefits upper-limb, lower-limb, gait and balance recovery, particularly when added to conventional therapy. A meta-analysis of immersive VR published in 2025 included 23 studies with 395 patients in the systematic review, and 13 studies in the quantitative synthesis. Fully immersive systems showed the largest gains in gross motor function, while non-immersive systems were more helpful for fine dexterity. Programs longer than six weeks and those started within six months of stroke were associated with better results. Not all reviews agree. An earlier review found no statistically significant advantage over conventional therapy for upper-limb function, although its meta-analysis favored game-based VR. A meta-review of balance studies also found that after overlapping trials were removed, only four studies remained to support the claim. These differences reflect variable study quality, heterogeneous protocols and small samples.

Vestibular assessment and rehabilitation involve balance and dizziness rather than limb movement. VR can, in principle, present controlled visual scenes and motion to test how a patient responds to conflicting sensory information. The clinical value of such tools is less well established than the evidence for limb rehabilitation.

Evidence strength and clinical adoption

Evidence is strongest for specific phobias and for VR as an adjunct to standard stroke therapy, and weaker for severe psychiatric conditions, pain, and transfer of surgical skills to real operations. Clinical adoption depends on randomized trials that compare VR with established treatments, on regulatory decisions about medical devices, and on whether health systems and insurers will pay for the equipment, software and staff time involved.

Education, science and cultural heritage

Classrooms and virtual field trips

In schools, VR is mostly used to take students to places they cannot easily reach, to let them rehearse procedures, and to make abstract material visible. Smartphone-based systems made the first use case cheap. Google Expeditions, for example, lets a teacher run a shared field trip for a whole class from a companion app while students use phone-based headsets. Virtual trips have also been used to give rural pupils access to career exploration that would otherwise be unavailable. Evidence of learning benefit is promising but uneven. A controlled study of a VR lesson on porcelain production in Jingdezhen found that students using the interactive version reported stronger engagement and scored higher on concentration and knowledge tests than those who watched conventional video, though the study was small and tied to one topic. A bibliometric review of 97 peer-reviewed articles from 2018 to 2024 concluded that the field still has notable research gaps. Teacher preparation is a separate challenge: in one exploratory course, pre-service secondary teachers learned a VR geometry tool and then had to design a teaching sequence around it.

Science and molecular visualization

Molecules are three-dimensional, and many of their interactions are hard to convey on a flat screen. John Stone of the University of Illinois notes that VMD, a molecular viewer, began as a CAVE application in 1993, and that scientists' use of immersive tools remained limited by cost, technical complexity and the lack of local facilities. He argues that inexpensive head-mounted displays could remove much of that barrier, and that most research tasks in VR concern analysing and moving atomic structures. Stereoscopic viewing is often cited as the key advantage, because it can reveal structural features that stay hidden on a 2D display, although some perspectives in VR can make certain structural motifs harder to recognise. Researchers also describe the experience of entering a molecule and walking around it as a qualitatively different way of inspecting it.

Interactive molecular dynamics (iMD-VR) goes further by letting users push and pull simulated molecules with hand movements while the calculation runs in real time. The open-source Narupa framework supports several researchers sharing one simulated space, and web-based tools such as ProteinVR aim to run across desktop, mobile and headset platforms. Network-based approaches address a different problem: a protein interaction map can contain roughly 18,000 nodes and 300,000 connections, which quickly overwhelms conventional screen-based software. Earlier headsets were tethered to powerful computers and required complicated setups, which slowed adoption; these obstacles have since eased, though not entirely.

Museums and cultural heritage

Digital documentation has become a tool for recording heritage threatened by nature, conflict, terrorism, climate change and development. After the damage to Notre Dame, and to sites such as Aleppo and Palmyra, 3D records were used to guide physical reconstruction, and public-facing virtual reconstructions are often credited with raising awareness of neglected sites. Early digital heritage work emphasised faithful representation of objects and sites, while later work recognised that reconstructions are especially useful where context and meaning are not obvious from the remains alone. A five-year fieldwork project at Kandovan in Iran, a village of cone-shaped dwellings, combined data collection with a VR presentation that was shown at an exhibition in Australia, with the aim of showing how recent development has altered the settlement. At Hadrian's Villa, digital reconstructions of inaccessible structures have been shared online for virtual visits. Museums also use 3D scanning to offer tours of sites that are remote, fragile or endangered. Reconstructions of this kind involve interpretive choices, and AI-assisted restorations, such as recent work on the Great Mosque of Aleppo, add further questions about how far a reconstruction should go beyond surviving evidence.

Equity and access in schools

Cost is the most frequently raised barrier. In a survey of more than 1,400 US high school teachers, most agreed that virtual learning experiences provide quality information, yet more than half believed the costs of these technologies could widen existing equity gaps. Low-cost headsets have been presented as a way to give institutions that lack specialised equipment access to simulated practice, but hardware still needs charging, updating, storage and replacement, and these recurring costs fall on schools that often have limited technical staff. Training is equally important: teachers need time to learn the software and to fit it into lesson plans, and that preparation is not yet standard in teacher education. Access also varies by location and by student need, so schools with stable connectivity and funding are better placed to benefit than under-resourced ones.

Relation to augmented, mixed and extended reality

Milgram and Kishino's 1994 framework treats the continuous scale between the completely virtual and the completely real as a single axis, and categorized existing mixed reality hardware within the continuum based on a three dimensional framework based on extent of world knowledge, reproduction fidelity and extent of presence. The axis was first labelled "virtuality-reality," and its name was changed to the Reality-Virtuality Continuum the following year. On this scheme, VR is the virtual endpoint, and everything between the two endpoints is mixed reality. Within that span, the real-leaning side is mixed reality systems that are more based in reality, known as augmented reality (AR), and the virtual-leaning side is augmented virtuality (AV). Because the continuum is a spectrum rather than a set of boxes, the labels describe where a system sits, not what hardware it uses. That is why a consistent definition of AR does not exist across the literature, even though the continuum is the most common reference point.

"Extended reality" (XR) is an industry umbrella term for AR, mixed reality and VR taken together. It is used loosely, and it overlaps heavily with "mixed reality" in product marketing, so the two labels are not interchangeable in academic writing. The boundary between AV and VR is also blurry. Writers who favour AV as a category point to cases like a game in which sensors in the controllers provide data from the physical world to the virtual environment. A VR title that tracks the player's real hands therefore already contains a small amount of physical world, which is why some taxonomists argue that the original scheme needs extension. One group notes that most of the mixed-reality applications found in the literature simply juxtapose real and virtual objects through the projection of visual artefacts.

The two main ways of building an AR headset differ in how the user sees the real world. In an optical see-through design, the wearer looks through a transparent lens or other optical component directly at the real world, and virtual imagery is combined into that view by optics. In a video passthrough design, the wearer sees a screen that is displaying a continuous live feed of the world captured by one or more cameras on the front of the headset. Passthrough designs are the more common of the two, and they are easier to design and are traditionally significantly less expensive. Each approach has known trade-offs. Optical systems show the physical world with no camera delay, but they can only add light, so dark or opaque virtual content looks hazy; one vendor argues that those images are hazy and ghostlike, because optical see-through devices can only display light, not black or opaque content. Waveguide optics, a common form of optical combiner, have a related limit: light signals can't render the color black. Passthrough systems give full control over every pixel, but they depend on camera-to-display latency, and the latency of video frame transformations may need to be as low as possible to keep head movement from looking delayed. Passthrough also has an alignment advantage: AR objects mapped to real world objects are accurate because the processing happens on the same pipeline.

VR, AR and smart glasses share most of their technical base. All three need to know where the head is and how it moves, which requires computer vision and inertial sensors; all three need displays and optics; and all three need some model of the surrounding space. The requirements differ in one important respect. In immersive VR the environment is coherent because it is entirely virtual, so the main requirement is that virtual objects behave consistently with the user's actions. In augmented environments, visual alignment issues can arise between virtual and real parts of the scene, even for small spatial errors. AR therefore depends more heavily on spatial mapping and registration than VR does, and this is one reason that improvements in tracking and room mapping have benefited both fields.

Smart glasses are the newest part of this landscape. Omdia reported that global shipments of AI glasses reached 8.7 million units in 2025, up 322 percent year over year, and Meta dominated this category in 2025, accounting for 85 percent of global AI glasses shipments. Many of these products have cameras, microphones and speakers but no display, and they connect to a cloud-based AI assistant that answers questions about what the wearer is seeing. A smaller group adds a display in the lens, which allows text and notifications to appear in the wearer's view. The shift has also changed how companies allocate their resources. Omdia notes that major technology companies including Meta, Apple, Samsung and Huawei have adjusted their product roadmaps, delaying or scaling back VR headset launches in favor of lightweight AI smart glasses and AR glasses.

Market data on these categories disagree, partly because the categories themselves are defined differently. IDC reported that VR / MR headset shipments fell 23.3% to 548,000 units in the second quarter of 2026, while Counterpoint put the VR decline at 18% year-over-year for the same quarter and reported that global shipments of AR devices surged 138% YoY. A third firm, Smart Analytics Global, gave a much steeper figure, a 79% drop in shipments for VR headsets. IDC also separates display-less smart glasses from display-equipped eyewear, so growth rates that appear to describe "smart glasses" may refer to different product sets. Analysts tend to attribute the AR and AI-glasses growth to lightweight designs and AI features, but they also note that these categories remain small relative to the wider smartphone and wearables markets.

Market development and economics

Headset and glasses shipments

Headset shipments have fallen while the wider XR market has grown, and the difference is driven by smart glasses. IDC reported that global XR device shipments grew 44.4% in 2025, reaching 14.5 million units. Smart glasses drove nearly all of that growth. Headsets moved the other way. In the first half of 2025, standalone VR and mixed reality headsets declined 14%, and Meta Quest shipments specifically fell 42.3% compared to the previous year. Part of that drop reflects a high base, since 2024 was the Quest 3 launch year, which inflated the comparison base.

The decline continued into 2026. Counterpoint Research recorded VR headset shipments falling 17% year over year and 39% quarter over quarter in Q1 2026. Over the same period, IDC put display-less smart glasses at a Q1 2026 gain of 167%, reaching about 2.25 million units in one quarter. Counterpoint's H1 2026 data show AI glasses up 263% year over year, while AR and AI glasses shipments surged by 449 percent YoY, from a much smaller base.

Concentration of the headset market

The headset segment is highly concentrated. Meta has been the dominant vendor, and Meta captured 72.2% of global XR device shipments in 2025. In Counterpoint's Q1 2026 ranking of AR/VR OEM shipments, Meta held 58% of global AR/VR OEM shipments in Q1 2026, down from 66% in Q4 2025, and Sony ranked second at 7%, followed by RayNeo at 6%. Those figures blend headsets and glasses, so they show the share of the broader category rather than of headsets alone.

Apple's premium Vision Pro has had a much smaller footprint. According to IDC data reported in one industry analysis, Apple shipped only 45,000 units in the 2025 holiday quarter, despite a chip upgrade in October 2025. Valve's forthcoming standalone hardware is expected to keep the category alive for gaming and enterprise use. HTC and ByteDance's Pico, which are long-standing headset makers, appear as smaller players in the sources consulted, and this section does not give precise shares for them.

Component costs and enterprise demand

Pricing pressure has come mainly from component costs. Memory shortages raised prices: one analysis notes that the DRAM shortage pushed Quest prices up by $50 to $100 in April, suppressing demand exactly when spring buying typically picks up. Higher prices hit consumer demand hardest, because headsets are discretionary purchases that compete with games and other entertainment.

Enterprise demand has been more stable. IDC expects headset sales to rebound in 2026 with new launches, but notes that their appeal will remain largely limited to gamers and businesses focused on design, training, and collaboration. That profile suggests the professional segments, which buy for specific training or design workflows, are less exposed to consumer slowdowns than the entertainment market. IDC's forecast for 2026 is 8.8 million units, rising to 21.0 million units shipped by 2029.

Why market-size estimates diverge

Estimates of the VR market differ widely, and the gaps are largely definitional. Fortune Business Insights values the market at USD 20.83 billion in 2025, while Mordor Intelligence uses a narrower scope and projects USD 12.92 billion in 2025. Straits Research's 2025 base is USD 13.23 billion. Growth rates diverge as well: Technavio forecasts a CAGR of 38% for 2024–2029, while Fortune projects a CAGR of 26.20% through 2034.

Several factors drive these differences. Mordor notes that published XR estimates often differ because firms choose dissimilar device mixes, price assumptions or update cadences. In its benchmark comparison, one firm bundles XR content, services and accessories and another counts headset hardware only; omits software and integration spend. Some estimates also rely on vendor self-reported sales. Forecasts can also change within a single year: IDC's earlier 2025 projections were later superseded by revised figures, which shows how sensitive shipment-based estimates are to new data.

Privacy, ethics and social impact

Data collection and privacy

A headset observes its wearer more closely than most screens do. Inward-facing cameras, inertial sensors and optical tracking can capture gaze direction, pupil behaviour, head and hand movement, facial expression, and the layout of the room around the user. One analysis of VR privacy notes that a headset may record posture, gaze, gestures, facial expression and interpersonal distance up to ninety times per second. Eye-tracking data is especially sensitive. Research on eye-movement biometrics has shown that gaze patterns can support continuous authentication in VR, and a recent study using a VR video-watching task reported identifying around 97 per cent of more than four hundred participants from eye movements alone. Gaze data can also reveal inferences about attention, stress or physical limitations, so the risk extends beyond identity.

Because biometric traits cannot be reset the way a password can, their exposure is hard to reverse. This is the main reason privacy researchers treat headset sensing as a distinct problem. Legal analyses argue that eye-tracking data may count as biometric data under regimes such as the EU's GDPR, which generally requires explicit consent for such processing, and under some US state laws. Consent is complicated, however, when the same signal serves several purposes. Foveated rendering, a performance technique, needs gaze data in real time, yet the same data could be retained for advertising measurement or sold to partners. Privacy advocates therefore call for separate consent options for each purpose, clear on-screen indicators when sensors are active, on-device processing where feasible, short retention periods and deletion rights. Room mapping raises a related issue: it can capture bystanders, such as family members or colleagues, who never agreed to be recorded.

Harassment, safety and moderation

Shared virtual spaces create forms of abuse that have no close equivalent in text-based forums. Because users appear as embodied avatars and talk in real time, the interaction can feel physically immediate. A systematic review of 85 studies proposed a typology that includes spatial intrusion, sexual harassment, verbal abuse, identity-based harassment, group-based aggression and systemic harm, and it reported disproportionate effects on marginalized groups. Researchers have also found that embodiment and presence make harassment feel more intense, and that the ephemeral nature of real-time interaction makes it hard to escape or report.

Moderation faces specific obstacles. Many platforms do not log voice chat or record spatial movement, which leaves victims with little evidence and perpetrators with little accountability. Interview studies report that blocking tools are hard to use in crowded spaces and do not remove a harasser's presence from the scene. Norms for acceptable behaviour are still forming, and some attackers regard harassment as "just a game" in which offline norms do not apply. Proposed remedies include AI-based moderation of voice and behaviour, proactive safety designs and community norms, but these remain under study. One controlled laboratory experiment with 52 participants found that a harassing avatar was rated as significantly more negative and disturbing, while well-being measures did not change significantly in that scenario, which runs against some earlier reports.

Accessibility and the digital divide

The question of who can use VR is only partly answered by the technology. Cost remains a barrier: capable headsets, powerful computers or subscriptions are expensive, and this limits access in schools, public institutions and lower-income households, which is part of the wider digital divide. Physical requirements also exclude some people. Room-scale experiences assume space and the ability to stand and move, and many interaction schemes assume two working hands. Visual conditions can complicate use, since fixed focus, lens fit and interpupillary distance settings may not suit people who wear corrective lenses or have binocular vision differences. Audio-centred cues and social voice features can exclude deaf or hard-of-hearing users when captions or visual alternatives are absent. For people with disabilities, the design of comfortable seated modes, alternative controllers and captioning is therefore a precondition of access rather than an optional feature. The research base on these issues is thinner than the base on hardware, and assessments of accessibility remain uneven across platforms.

Psychological effects of violent and realistic content

Whether realistic simulated experiences change behaviour is a long-running debate, and VR sharpens it because immersion is thought to strengthen emotional and cognitive effects. Some researchers argue that the sense of being present could make violent acts feel more real and therefore more consequential. Evidence is mixed. A three-wave longitudinal study of 1,340 Chinese adolescents found that exposure to violent video games, which was not limited to VR, was associated with higher aggression, desensitization and lower empathy. By contrast, two pre-registered experiments found no evidence that immersive VR, alone or combined with violent content, increased aggression compared with playing the same kind of game on a flat screen. Another study found that difficulty raised hostility in VR but violent content did not, and a further study reported that a competitive context mattered more than the violence itself. VR horror games, meanwhile, are reported to produce intense fear responses, which raises separate questions about distress and consent.

Clinical studies complicate the picture further. A randomized trial of a VR game for aggression regulation among forensic psychiatric outpatients (N = 30) found no difference from a control game, although participants reported more insight into their behaviour. Other trials of VR-assisted aggression prevention training showed short-term improvements that did not all persist at follow-up. Open questions include how long exposure effects last, how they vary by age and sex, whether realism matters more than interactivity, and how far laboratory measures predict real-world conduct. These questions are unresolved, and the findings should be read as provisional.

Open problems and future directions

Resolution, field of view and focus

Visual fidelity is the most obvious gap. One 2025 review of wireless VR gives about 720 pixels per degree as the density needed for full visual fidelity, while the wireless headsets it surveys reach only about 20 to 30 due to bandwidth limitations. Field of view and resolution pull against each other: wider FOV demands higher resolution, increasing the strain on the wireless link. Pixel density also has a physical cost. Higher-resolution displays add the amount of data that would need to be transmitted to a headset, and the physical size and weight of the hardware. Current premium devices show the trade-off. Apple's Vision Pro is around 34 pixels per degree, while a lightweight PC headset such as the Pimax Dream Air is listed at 170 g with a 110-degree horizontal field of view. Focus cues are a separate problem. Most consumer optics still place the image at a fixed focal distance, and varifocal and light-field approaches, discussed elsewhere in this entry, remain largely experimental in shipping products.

Latency, wireless streaming, battery and heat

Latency is closely tied to comfort. A consumer buying guide argues that motion sickness stems primarily from latency, inconsistent frame timing, and vestibular-visual mismatch, and it suggests that under 15ms is ideal for nausea-free locomotion. Wireless operation makes the problem harder. Current wireless links below 6 GHz cannot carry 8K video at more than 90 frames per second, so the research literature argues that the success of wireless VR hinges on bringing enough computing power to the HMD via dedicated ASICS or to the cloud or fog within a latency budget. Foveated rendering, which uses eye tracking to lower detail outside the gaze point, is one way to ease this load; the newest PC-linked headsets advertise built-in eye tracking for that purpose. Battery life and heat remain limits for untethered devices. Vision Pro, for example, is described as lasting about 2 hours in intensive use. Packing a GPU-class workload, a battery and chips into a head-worn enclosure leaves little room for heat dissipation, which in turn limits sustained performance.

Haptics, locomotion and multi-user interaction

Haptic feedback is still the least mature input channel. Researchers note that one key criterion is extremely low latency, as low as 1 ms, and that latency and robustness across hand sizes still need work. Wearable gloves face power and complexity limits: system complexity and power consumption remain challenging for prolonged wearable use, and a fully wireless glove validated in a realistic VR setting has yet to be demonstrated. Adding haptics does not automatically improve the experience. In a 162-participant study of a virtual museum, haptic gloves significantly decreased usability and increased cognitive workload, and the authors conclude that advanced interaction technologies do not necessarily enhance immersion. Locomotion is similarly unsettled, since empirical evidence on the combined effects of advanced interaction technologies on key user-experience metrics remains limited. Multi-user systems add further demands, because an interconnected VR service must handle the resource distribution, quality of experience (QoE) requirements, and the interaction between multiple users. Longer-term research also explores brain-computer interfaces as input or feedback channels, but this work is at an early, exploratory stage and raises unresolved questions about signal reliability, safety and neural privacy.

Generative AI and the future of content

Generative AI is widely expected to change how VR worlds are made. Some tools can generate entire 3D scenes from text prompts, and a 2026 review in Science China Information Sciences notes that such methods offer a broader array of options for constructing virtual characters/agents and realistic virtual scenes. Market forecasts are also optimistic, though estimates differ widely; one industry press release projects growth from $5.08 billion in 2025 to $6.73 billion in 2026. Open questions remain. Synthetic avatars and characters raise concerns regarding consent of individuals whose voices or likenesses were used as training data, which connects to data protection, intellectual property and copyright disputes over who owns generated assets. It is also unclear whether AI-generated worlds that adapt to user behavior will strengthen presence or simply produce more generic content at scale. Answers will depend on how developers, platforms and ethicists set norms for transparency, authorship and the use of personal data, and on whether generated content can be verified against the physical and perceptual demands that make VR convincing in the first place.

References

1. Presence and Immersion in Virtual Realitylink.springer.com
2. Immersion (virtual reality)en.wikipedia.org
3. A brief chronology of Virtual Realityarxiv.org
4. Immersion Metrics for Virtual Realityarxiv.org
5. Morton L. Heilig: The inventor of virtual reality, decades ahead of his timebtw.media
6. A Brief History of Virtual Reality, in Patentsfinnegan.com
7. History Of Virtual Reality - Virtual Reality Societyvrs.org.uk
8. Sensorama - Virtual reality - Historydrafthistorydraft.com
9. USC HMH Foundation Moving Image Archive » MORTON HEILIG : INVENTOR VRhmharchive.com
10. Forgotten genius: the man who made a working VR machine in 1957techradar.com
11. Morton Heiligen.wikipedia.org
12. Ivan Sutherland%27s head-mounted 3D displayen.wikipedia.org
13. Ivan Sutherland's head-mounted 3D displayen.wikipedia.org
14. History of Virtual Environmentseducation.siggraph.hosting.acm.org
15. A Brief History of Augmented Realityinformit.com
16. Ivan Sutherlanden.wikipedia.org
17. Head-mounted displayen.wikipedia.org
18. Sword of Damocles - VR & AR Wiki - Virtual Reality & Augmented Reality Wikivrarwiki.com
19. Link Traineren.wikipedia.org
20. Link Trainersandiegoairandspace.org
21. Edwin Link Builds the First Flight Simulator : History of Informationhistoryofinformation.com
22. Link Aviation Devicesen.wikipedia.org
23. Edwin Albert Link - National Center For Simulationsimulationinformation.com
24. Myron Krueger — VIDEOPLACE, 1975 - DIGITAL ART MUSEUMdam.org
25. Myron W. Kruegerdada.compart-bremen.de
26. Videoplaceen.wikipedia.org
27. Virtual Reality: A Definition History - A Personal Essayarxiv.org
28. Directing Avatars in Live Performances -- an Autonomy Simulacrum of Virtual Entitiesarxiv.org
29. Smallest Federated Wikivw.fed.wiki.org
30. Methods and apparatus for generating and processing synthetic and absolute real time environmentsimage-ppubs.uspto.gov
31. Virtual Boynintendo.fandom.com
32. The Real Reason The Nintendo Virtual Boy Flopped - SlashGearslashgear.com
33. Virtual Boyen.wikipedia.org
34. Virtual Boyultimatepopculture.fandom.com
35. One of the Nintendo's biggest failures: Virtual Boy - InspireIPinspireip.com
36. Artificial Realityen.wikipedia.org
37. Myron W. Kruegeren.wikipedia.org
38. Oculus Rift’s $2 Billion Purchasetime.com
39. Reality Labsen.wikipedia.org
40. Google Cardboard - VR & AR Wiki - Virtual Reality & Augmented Reality Wikivrarwiki.com
41. Smartphone VR Viewer Rounduphackaday.com
42. Google Shipped More Than 10 million Cardboard VR Viewers Since 2014wccftech.com
43. Google Cardboarden.wikipedia.org
44. Evolution Of Mobile VRwalkovr.com
45. Inside Apple Vision Pro: How Spatial Computing Is Reshaping the Mixed-Reality Platform Warscurrenttrends.news
46. Apple Vision Pro could kick off spatial computing revolutionmastercard.com
47. Apple Vision Pro Launches: 'The Era of Spatial Computing Has Arrived' - MacRumorsmacrumors.com
48. Apple Vision Pro Launch: A Retrospective and Look Ahead for Spatial Computeabiresearch.com
49. New on Yahootech.yahoo.com
50. Oculus Rift backers upset over Facebook acquisition - CBS Newscbsnews.com
51. Reality Labsen.wikipedia.org
52. Oculus VRencyclopedia.pub
53. Apple Vision Pro Review: The New Era of Spatial Computing — UC Todayuctoday.com
54. Google Cardboard - Virtual Reality Wiki - Fandomvr.fandom.com
55. Understanding Stereopsis: The Magic of Depth Perception - Burman & Zuckerbrod Ophthalmology Associates2020detroit.com
56. Stereoscopy and Depth Perception in XR — Utilizing Natural Binocular Disparity — Petra Palusovapetrapalusova.com
57. The Active Side of Stereopsis: Fixation Strategy and Adaptation to Natural Environmentsnature.com
58. [PDF] Place illusion and plausibility can lead to realistic behaviour in immersive virtual environmentssemanticscholar.org
59. (PDF) A Separate Reality: An Update on Place Illusion and Plausibility in Virtual Realityresearchgate.net
60. Vergence%E2%80%93accommodation conflicten.wikipedia.org
61. Vergence-Accomodation Conflictdeveloper-docs.magicleap.cloud
62. Focus adjusting headsetimage-ppubs.uspto.gov
63. Magnitude, precision, and realism of depth perception in stereoscopic vision - Cognitive Research: Principles and Implicationslink.springer.com
64. Binocular disparity can augment the capacity of vision without affecting subjective experience of depthnature.com
65. Towards an objective measurement of presence, place illusion, and plausibility illusion in virtual reality using electroencephalographylink.springer.com
66. Beyond Being Real: A Sensorimotor Control Perspective on Interactions in Virtual Realityarxiv.org
67. TCL's High-Density OLED Could Spawn The Midrange Headsets VR Needsuploadvr.com
68. VR LCD Screen Guide: How to Choose the Right Displayelectronics.alibaba.com
69. Micro OLED vs Traditional OLED in VR: Which Display Technology Delivers Better Performance?panoxdisplay.com
70. Why choose Micro OLED for VR headsetsdisplaymodule.com
71. VR Headset Display Comparison: A Deep Dive into Resolution, Optics, aninairspace.com
72. Pancake vs Fresnel VR Lenseskomete-xr.com
73. Pancake lenses - VR & AR Wiki - Virtual Reality & Augmented Reality Wikivrarwiki.com
74. Technologies for efficient head-mounted display with pancake lensesimage-ppubs.uspto.gov
75. Fresnel vs pancake lenses: what's the difference? - VRX by VR Expertvrx.vr-expert.com
76. Screen Door Effect in VR: What It Is and How to Avoid Itvrvisiongroup.com
77. What is PPD (Pixels Per Degree) & How High PPD Improves VR Image Qualistore.pimax.com
78. Cellular-Connected Wireless Virtual Reality: Requirements, Challenges, and Solutionsarxiv.org
79. How Does Pixel Response Time Affect the Visibility of Screen Door Effect?us.ktcplay.com
80. VR Display Screen: The Gateway to Immersive Virtual Realitiesinairspace.com
81. Pose Tracking Methods: Outside-in VS Inside-out Tracking in VRstore.pimax.com
82. Choosing the Right VR Tracking System: Lighthouse and Inside-Outstore.pimax.com
83. Co-Located VR with Hybrid SLAM-based HMD Tracking and Motion Capture Synchronizationarxiv.org
84. What types of tracking systems are used in VR (e.g., inside-out vs. outside-in)? - Zilliz Vector Databasezilliz.com
85. Exoskeleton Robotics Glove with A Dual Force Feedback Haptic Device for Immersive Virtual Reality Interactionieeexplore.ieee.org
86. HaptGlove—Untethered Pneumatic Glove for Multimode Haptic Feedback in Reality–Virtuality Continuumncbi.nlm.nih.gov
87. Creating a haptic experience with VR-glovessenseglove.com
88. Electrostatic brake-based haptic deviceimage-ppubs.uspto.gov
89. Evaluation of Locomotion Techniques for Room-Scale VRdl.acm.org
90. Teleportation With Reorientation for Redirected Walking in VR - PubMedpubmed.ncbi.nlm.nih.gov
91. What is VR locomotion (virtual reality locomotion)?techtarget.com
92. Measuring eye-tracking accuracy and its impact on usability in apple vision proarxiv.org
93. (PDF) The asynchronous time warp for virtual reality on consumer hardwareresearchgate.net
94. Motion-to-photon latency - VR & AR Wiki - Virtual Reality & Augmented Reality Wikivrarwiki.com
95. The Asynchronous Time Warp for Virtual Reality on Consumer Hardwaredl.acm.org
96. Predictability-Aware Motion Prediction for Edge XR via High-Order Error-State Kalman Filteringarxiv.org
97. Asynchronous TimeWarp (ATW) (Deprecated)developers.meta.com
98. Efficient VR rendering: Survey on foveated, stereo, cloud, and low-power rendering techniques - ScienceDirectsciencedirect.com
99. NVIDIA Adds Eye-tracking Support to VRSS Foveated Rendering Techroadtovr.com
100. What is foveated rendering? VR tech tipped for PS5 explainedgearbrain.com
101. Save GPU with Eye Tracked Foveated Renderingdevelopers.meta.com
102. GazeProphet: Software-Only Gaze Prediction for VR Foveated Renderingarxiv.org
103. Overview of VR Spatial Audio Technologiesallpcb.com
104. Efficient personalization of head-related transfer functions for improved virtual spatial audioimage-ppubs.uspto.gov
105. Spatial sound conference system and apparatusimage-ppubs.uspto.gov
106. Spatial Audio in VR and AR Streaming - Fora Softforasoft.com
107. HRTF & Mesh Grading for Spatial Audioemergentmind.com
108. JMIR Serious Games - Usability of Individualized Head-Related Transfer Functions in Virtual Reality: Empirical Study With Perceptual Attributes in Sagittal Plane Sound Localizationgames.jmir.org
109. Motion-to-Photon Latency definition and descriptionthespatialstudio.de
110. Individual Differences in Training Naive Listeners to Localize Spatial Audio in Virtual Realitybiorxiv.org
111. Khronos Releases OpenXR 1.1 to Further Streamline Cross-Platform XR Developmentkhronos.org
112. News Archiveskhronos.org
113. Khronos announces OpenXR cross-platform Virtual Reality standardvideocardz.com
114. Khronos Group announces release of OpenXR 0.90 API standard for cross-platform AR and VR compatibilityauganix.org
115. Khronos Releases OpenXR 1.1 For Cross-Platform AR/VR Development - Phoronixphoronix.com
116. WebXR Device API Explainedimmersive-web.github.io
117. WebXR - Immersive Webimmersiveweb.dev
118. Virtual reality comes to the webweb.dev
119. Fundamentals of WebXR - Web APIsdeveloper.mozilla.org
120. WebXRen.wikipedia.org
121. Meta Horizon Storeen.wikipedia.org
122. The State of VR at GDC 2026: Building a Sustainable Futuredevelopers.meta.com
123. Get financial reportsdevelopers.meta.com
124. How Many XR Devices Did Meta Sell in Q1? - AR Insiderarinsider.co
125. Meta's Latest Quest Store Revenue Figure Signals a Steady but Stagnant Marketplaceroadtovr.com
126. Cave Automatic Virtual Environmentmixedreality.fandom.com
127. A Review of Prototyping in XR: Linking Extended Reality to Digital Fabricationarxiv.org
128. Transparent XR displays from near-eye to head-updispelix.com
129. Types of VR Headsets: Tethered vs Standalone vs Smartphonesimulatorhardware.com
130. VR Headset Types (Tethered, Standalone, Mixed Reality) - Geek Modegeekmode.blog
131. What are the differences between tethered and standalone VR headsets? - Zilliz Vector Databasezilliz.com
132. VR Headsets Buyer's Guidesvrt.org
133. What are VR Headsets? — updated 2026ixdf.org
134. Different Kinds of VR Headsets: Your Ultimate Guide to Virtual Realityinairspace.com
135. Reflective lens headsetimage-ppubs.uspto.gov
136. Video Pass-Through Mixed Reality Changes Reality As You Know Itvarjo.com
137. Optical See Through vs Video See Through on ThirdEye’s X1 Smart Glassesmedium.com
138. Passthrough vs. See-through: Impacts of Display Technology on Peripheral Vision - AREAthearea.org
139. The Perceptual Cost of Passthrough: How Video See-Through HMDs Degrade Human Visual Perception of Acuity, Contrast, and Colorarxiv.org
140. Immersive Virtual Reality CAVE Systemsmechdyne.com
141. Passthrough - VR & AR Wiki - Virtual Reality & Augmented Reality Wikivrarwiki.com
142. Difference between Tethered and Standalone VR headsets?progamerblog.com
143. Portable CAVE automatic virtual environment systemimage-ppubs.uspto.gov
144. Cybersickness in Virtual Environmentsnature.com
145. Effects of virtual reality exposure on postural control: a comparative analysis before and after exposure in individuals with binocular dysfunction and normal vision - PMCpmc.ncbi.nlm.nih.gov
146. The relationship between postural stability and cybersickness: It's complicatedsciencedirect.com
147. Vergence-accommodation conflicten.wikipedia.org
148. Vergence-Accommodation Conflictemergentmind.com
149. VR Effects on Eyes: Risks & Protection Explainedkomete-xr.com
150. Best VR Headset for Kids 2026: Safe, Age-Appropriate Picksvr.org
151. How Old Do You Have to Be to Have a VR Headset - A Parent's Guide to Virtual Reality Age Limitsinairspace.com
152. Is VR Bad For Kids? Age Limits For Quest, PSVR2, Vision Proscienceabc.com
153. How Early Can You Tell? Early Eye Gaze Dynamics and Cybersickness Progression in Virtual Realityarxiv.org
154. Multisensory integration and the experience of scene instability, presence and cybersickness in virtual environments - ScienceDirectsciencedirect.com
155. 50+ Virtual Reality Statistics: Meta Holds 72% of the 2026 XR Marketg2.com
156. Top 5 Meta Quest Developers in 2026treeview.studio
157. VR Gaming Statistics 2026: Users, Sales & Market Sizeshanethegamer.com
158. VR Gaming Headset Statistics 2026 (Ownership & Sales)rec0ded88.com
159. Meta Quest Statistics 2026: Steam Share, Sales & QLL Dataaxis-intelligence.com
160. 'Rec Room' Adds More Flexibility to Avatar System with New Full-body, Animated Costumesroadtovr.com
161. Social VR Hits 'Rec Room' and 'VRChat' Coming to Oculus Questroadtovr.com
162. Across the Metaverse: My trip though VR social platformsspacebar.news
163. VRChat or Rec Room - VR.Spacevr.space
164. Social VR Face-Off: VRChat vs. Rec Room - VRInsightsvrinsights.io
165. Location-Based Entertainment Market Report 2024- 2029, By Applications, Geo, Techmarketsandmarkets.com
166. Location Based VR Explained - Sandbox VRsandboxvr.com
167. From Zero To Immersive: A Look At 3 Top-Ranked Location-Based VR Venuesarpost.co
168. Location Based VR Market - Industry Growth & Companiesmordorintelligence.com
169. What Is LBVR? Location-Based VR Guide for Operatorssynthesisvr.com
170. VR Flight Simulator FAA Qualified For The First Timeuploadvr.com
171. TRU Simulation’s Veris virtual reality flight simulator achieves FAA qualification - Vertical Magverticalmag.com
172. REALITY CHECK - USAASC - Army.milasc.army.mil
173. “Virtual Reality” Training Could Be a Godsend for the US Army - The National Interestnationalinterest.org
174. Soldiers are training in virtual environments generated from real citiesfuturism.com
175. Reality Check: Haptics improvements to Army simulation training makes virtual environments feel more realistic.lineofdeparture.army.mil
176. Virtual Reality in the Military: Simulating Combat Training - ArborXRarborxr.com
177. From virtual to reality: A systematic review of the impact of virtual reality on skill acquisition, retention, and transfer - ScienceDirectsciencedirect.com
178. Transference of spatial and visual memory from virtual environments to the real world. Implications for clinical and health interventions - PMCpmc.ncbi.nlm.nih.gov
179. Frontiersfrontiersin.org
180. Learning and transfer of complex motor skills in virtual reality: a perspective reviewncbi.nlm.nih.gov
181. Virtual reality exposure therapy for anxiety disorders: a meta-analysis - Database of Abstracts of Reviews of Effects (DARE): Quality-assessed Reviews - NCBI Bookshelfncbi.nlm.nih.gov
182. Inferiority or Even Superiority of Virtual Reality Exposure Therapy in Phobias?—A Systematic Review and Quantitative Meta-Analysis on Randomized Controlled Trials Specifically Comparing the Efficacy of Virtual Reality Exposure to Gold Standard in vivo Exposure in Agoraphobia, Specific Phobia, and Social Phobia - PMCpmc.ncbi.nlm.nih.gov
183. Examining the comparative effectiveness of virtual reality and in-vivo exposure therapy on social anxiety and specific phobia: A systematic review & meta-analysis - ScienceDirectsciencedirect.com
184. A Meta-Analysis on the Relationship between Self-Reported Presence and Anxiety in Virtual Reality Exposure Therapy for Anxiety Disordersncbi.nlm.nih.gov
185. The Effectiveness of Virtual Reality Exposure–Based Cognitive Behavioral Therapy for Severe Anxiety Disorders, Obsessive-Compulsive Disorder, and Posttraumatic Stress Disorder: Meta-analysisncbi.nlm.nih.gov
186. Meta-analysis of virtual reality exposure therapy for social anxiety disorderncbi.nlm.nih.gov
187. A Scoping Review on the Evidence for Virtual Reality Simulation in Surgical Training - PubMedpubmed.ncbi.nlm.nih.gov
188. Virtual reality simulation in robot-assisted surgery: meta-analysis of skill transfer and predictability of skillacademic.oup.com
189. Does training on robotic virtual reality simulators improve the post-training robotic surgical skills of surgeons? A systematic review and meta-analysis - PMCpmc.ncbi.nlm.nih.gov
190. Virtual reality vs. physical models in surgical skills training. An update of the evidence - PMCpmc.ncbi.nlm.nih.gov
191. Virtual Reality Simulation for Orthopedic Surgical Training: A Narrative Review of Current Evidence and Educational Impactncbi.nlm.nih.gov
192. Virtual reality in stroke recovery: a meta-review of systematic reviewsncbi.nlm.nih.gov
193. Immersive Virtual Reality in Stroke Rehabilitation: A Systematic Review and Meta-Analysis of Its Efficacy in Upper Limb Recoveryncbi.nlm.nih.gov
194. The efficacy of virtual reality for upper limb rehabilitation in stroke patients: a systematic review and meta-analysisncbi.nlm.nih.gov
195. Journal of Medical Internet Research - Effects of Virtual Reality Training on Upper Limb Function and Balance in Stroke Patients: Systematic Review and Meta-Meta-Analysisjmir.org
196. A systematic review of techniques and clinical evidence to adopt virtual reality in post-stroke upper limb rehabilitationlink.springer.com
197. Virtual reality in primary educationen.wikipedia.org
198. Research on the impact of immersive virtual reality classroom on student experience and concentrationlink.springer.com
199. Virtual reality in education: global trends, challenges, and impacts—game changer or passing trend?link.springer.com
200. Frontiersfrontiersin.org
201. Virtual Reality in Education: Benefits, Uses & Examplesbuiltin.com
202. Development of a Virtual Reality Application for Oculomotor Examination Education Based on Student-Centered Pedagogyarxiv.org
203. Molecular Visualization and Simulation in VRks.uiuc.edu
204. VR visualization supports research on molecular networksphys.org
205. State of the Art of Molecular Visualization in Immersive Virtual Environments - Kuťák - 2023 - Computer Graphics Forum - Wiley Online Libraryonlinelibrary.wiley.com
206. The MolecularWeb Universe: Web-Based, Immersive, Multiuser Molecular Graphics And Modeling, for Education and Work in Chemistry, Structural Biology, and Materials Sciencesarxiv.org
207. A Perspective on AI-Guided Molecular Simulations in VR: Exploring Strategies for Imitation Learning in Hyperdimensional Molecular Systemsarxiv.org
208. Measuring the limit of perception of bond stiffness of interactive molecules in VR via a gamified psychophysics experimentarxiv.org
209. Interactive molecular dynamics in virtual reality from quantum chemistry to drug binding: An open-source multi-person frameworkpubs.aip.org
210. Virtual reality-based digitisation for endangered heritage sites: Theoretical framework and application - ScienceDirectsciencedirect.com
211. The Virtual Living Museumwisdomlib.org
212. Digital preservation and development of architectural heritage from a virtual perspective: a systematic reviewnature.com
213. Reconstruction of Cultural Heritage in Virtual Space Following Disastersdoi.org
214. A Comprehensive Review of Multimodal XR Applications, Risks, and Ethical Challenges in the Metaversearxiv.org
215. 3D reconstruction of cultural heritage sites; A case study of Patan Durbar Square - ScienceDirectsciencedirect.com
216. Reality%E2%80%93virtuality continuumen.wikipedia.org
217. Capstone Experiences in Developing Augmented Reality Tables for Community Organizationsarxiv.org
218. Milgram Reality-Virtuality Continuumresearchgate.net
219. Staging Reality: The AR, VR & MR Continuum - AR Insiderarinsider.co
220. Notes on Pervasive Virtualityarxiv.org
221. XR Talks: Will Pass-through Video Outshine See-through Optics? - AR Insiderarinsider.co
222. System and method for parallax correction for video see-through augmented realityimage-ppubs.uspto.gov
223. Engineering Reliable Interactions in the Reality-Artificiality Continuumarxiv.org
224. AR and VR Enter a New Era: How AI is Reshaping the Display Ecosystem - Dash - 2026 - Information Display - Wiley Online Librarysid.onlinelibrary.wiley.com
225. Analysis: VR market under pressure as AI smart glasses take-offadvanced-television.com
226. Global smart eyewear shipments grow 35.3% in Q2 as China posts first decline · TechNodetechnode.com
227. Global XR (AR & VR Headsets) Market Share: Quarterlycounterpointresearch.com
228. AI Smart Glasses Shipments Are Up 776% While VR Headsets Just Collapsedandroidheadlines.com
229. IDC - Augmented and Virtual Reality Headsets Market Insightsidc.com
230. AI Smart Glasses Market Growth 2026: Why Shipments Climb << Virtual Reality News :: Next Realityvirtual.reality.news
231. Smart Glasses Are Now Outselling VR Headsets 3 to 1. The Numbers Tell the Whole Story.vr.org
232. Virtual Reality Market Statistics 2026quantumrun.com
233. Rokid Leads Global AR and AI Glasses Shipments as AI Glasses Market Surges in H1 2026prnewswire.com
234. The global VR headset market is undergoing a transformation: An economic assessmentxpert.digital
235. Global XR Shipments Rebound Behind Glasses-First Momentum, IDC Reportsmy.idc.com
236. Worldwide Augmented, Extended, Mixed, and Virtual Reality Headset Forecast Update, 2025–2029: CY 1Q26my.idc.com
237. Virtual Reality (VR) Market Size, Growth, Sharefortunebusinessinsights.com
238. Virtual Reality (VR) Market Size, Report, Share & Growth Trends 2025-2031mordorintelligence.com
239. Virtual Reality (VR) Market Size to Grow by USD 133.17 Billion from 2024 to 2029technavio.com
240. Virtual, Augmented & Mixed Reality (VR/AR) Market Size 2031mordorintelligence.com
241. IDC gloomy on AR/VR device shipments - Mobile World Livemobileworldlive.com
242. Security and Privacy in Virtual Reality: A Literature Surveyarxiv.org
243. Through the looking glass: eye tracking biometrics and the loss of anonymity in extended realitylink.springer.com
244. Privacy in Virtual and Augmented Realitysjipl.mainelaw.maine.edu
245. VR/AR Biometric Compliance: Immersive Tech Consentsecureprivacy.ai
246. Biometric Privacy in VR Shoppinginfluencers-time.com
247. Differential privacy for eye tracking with temporal correlationsncbi.nlm.nih.gov
248. Harassment in Virtual Reality: A Systematic Reviewdoi.org
249. How Harassment Shapes Self-Perception and Well-Being in Social VR: Evidence from a Controlled Lab Studycomputer.org
250. Disturbing the Peace: Experiencing and Mitigating Emerging Harassment in Social Virtual Realityresearchgate.net
251. Harassment in Social Virtual Reality: Challenges for Platform Governanceresearchgate.net
252. VR Safety: Harassment Controls Studyemergentmind.com
253. Towards Leveraging AI-based Moderation to Address Emergent Harassment in Social Virtual Realitydl.acm.org
254. Violent and non-violent virtual reality video games: Influences on affect, aggressive cognition, and aggressive behavior. Two pre-registered experiments - ScienceDirectsciencedirect.com
255. Violent and non-violent virtual reality video games: Influences on affect, aggressive cognition, and aggressive behavior. Two pre-registered experimentsresearchgate.net
256. Aggression and VR Violent Gaming: The Power of Moral Disengagement and Immersionresearchgate.net
257. Video games, frustration, violence, and virtual reality: Two studies - PubMedpubmed.ncbi.nlm.nih.gov
258. Testing the Effects of a Virtual Reality Game for Aggressive Impulse Management: A Preliminary Randomized Controlled Trial among Forensic Psychiatric Outpatientsncbi.nlm.nih.gov
259. A pilot study on treatment content in virtual reality-assisted aggression therapy at a maximum-security forensic psychiatric clinicncbi.nlm.nih.gov
260. Violent Video Games in Virtual Reality: Re-Evaluating the Impact and Rating of Interactive Experiencesresearchgate.net
261. AI Enabled 6G for Semantic Metaverse: Prospects, Challenges and Solutions for Future Wireless VRarxiv.org
262. High-density energy directing devices for two-dimensional, stereoscopic, light field and holographic head-mountedimage-ppubs.uspto.gov
263. Highest Resolution VR Headset 2026: Sharpest Displays Rankedvr.org
264. Best Gaming VR Headsets 2026: PC VR vs Standalonesimracingcockpit.gg
265. Best VR Headset Resolution Guide: How to Choose Wisely in 2026electronics.alibaba.com
266. Highest Resolution VR Headset Guide: How to Choose in 2026electronics.alibaba.com
267. Towards Low-Latency and Ultra-Reliable Virtual Realityarxiv.org
268. Low-Latency Haptic Open Glove for Immersive Virtual Reality Interaction - PMCpmc.ncbi.nlm.nih.gov
269. Untethered wearable haptic feedback glove integrated with soft thin-film motors for immersive VR interaction - ScienceDirectsciencedirect.com
270. More immersion, better experience? The trade-offs of haptic gloves and treadmill locomotion in virtual museumslink.springer.com
271. The Role of AI in Shaping the Future of Virtual Reality Developmentframesixty.com
272. Artificial intelligence for virtual reality: a reviewcad.zju.edu.cn
273. Generative Artificial Intelligence (AI) In Virtual Reality Market Report Examines Leading Companies And Growth Opportunitiesopenpr.com
274. Generative AI pornographyen.wikipedia.org
275. Best VR Headsets 2026: Steam Frame, Crystal Super & PCVR Choicesstore.pimax.com
276. Low-Latency Haptic Open Glove for Immersive Virtual Reality Interactionncbi.nlm.nih.gov
277. Top VR Technologies And Trends To Watch In 2026 - Vagonvagon.io