VR UI/UX design

VR UI/UX design is the design for user interfaces (UI) and user experiences (UX) in virtual reality. One of the goals of VR user interface design is to leverage the 3D environment of VR to create ergonomic applications that will increase the user's multitasking capabilities and overall productivity. Mike Alger set out one such approach in Visual Design Methods for Virtual Reality, a paper written in September 2015 for an MA in Moving Image at Ravensbourne in London. He also has 2 videos explaining his ideas.

The field overlaps with research on 3D user interfaces in human-computer interaction, which studies how people select and manipulate objects, move through virtual spaces and issue commands when wearing a head-mounted display (HMD). Platform holders publish their own guidance: Meta Horizon OS, Microsoft's HoloLens documentation, Apple's visionOS Human Interface Guidelines and Google's Android XR design guides each give recommended viewing distances, target sizes and comfort rules, and the numbers differ from platform to platform. Alger noted in 2015 that the VR community tends to treat VR and augmented reality as separate categories, but that many of the same interface principles apply to both, with the input method being the main thing that changes the design.

Reviewed 27 September 2026. Checked the new text against Alger's 2015 thesis and videos, Road to VR, the Meta, Microsoft, Apple (HIG and WWDC23) and Android XR design guidance, and the cited research papers (Jacob, Bowman and Hodges, Consumed Endurance, Beyond the HUD, Point and Teleport, Speicher et al., Fernandes and Feiner) plus DOI metadata. About review dates.

Scope

VR UI design covers the visible interface: panels, menus, buttons, text, cursors and heads-up elements, and where they are placed in the space around the user. Guidance on VR UX also covers the wider experience, including input, locomotion, comfort and sickness, and fatigue over a session. The textbook 3D User Interfaces: Theory and Practice by Joseph J. LaViola, Ernst Kruijff, Doug Bowman, Ryan P. McMahan and Ivan P. Poupyrev (second edition, Addison-Wesley, 2017) groups interaction techniques into selection and manipulation, navigation, and system control.

Alger argued that existing interfaces were designed for 2D screens rather than for the volume of VR: a user in a headset cannot see a keyboard to type, and moves their body in ways a mouse and keyboard never required. His research question was "How can the two dimensional paradigm of operating systems be most ergonomically redesigned for head mounted displays?" In an interview with Road to VR he said that "interaction design is all about looking at the properties of a given technology as well as the properties of humans".

History

Research foundations

In a 1990 CHI paper on eye movement-based interaction, Robert J. K. Jacob described the "Midas Touch" problem of using gaze as a pointer: "Everywhere you look, another command is activated; you cannot look anywhere without issuing a command." The Go-Go technique of Ivan Poupyrev, Mark Billinghurst, Suzanne Weghorst and Tadao Ichikawa, presented at UIST 1996, used a non-linear mapping that lets the user's virtual arm grow so that distant objects can be reached and manipulated directly.

In 1997 Doug Bowman and Larry Hodges compared six techniques for grabbing and manipulating remote objects in immersive virtual environments. They found that ray-casting, in which a virtual light ray extends from the user's hand, suffers from a "lever-arm" problem: an object attached to the end of the ray cannot easily be rotated in place, and the technique gives no direct control over its distance. Their study led to the HOMER (hand-centered object manipulation extending ray-casting) techniques, which treat grabbing and manipulation as separate tasks. Bowman, Kruijff, LaViola and Poupyrev published "An Introduction to 3-D User Interface Design" in the journal Presence in 2001, and the four, with Ryan P. McMahan, are the authors of the 2017 second edition of 3D User Interfaces: Theory and Practice.

Consumer VR era

Alger wrote in 2015 that the medium had "very few existing standards, protocols, and workflows", and that the community had long been undecided on input: Google Cardboard used a single button, Gear VR a trackpad and back button, and the development kits of the Oculus Rift shipped with no input device, while the Vive and the consumer Rift had been announced with tracked motion controllers. He cited a 2014 presentation by Alex Chu of Samsung Research, who used "greyboxing" (placing untextured primitive shapes in a scene and testing how their size and placement feel in the headset) when designing for Gear VR.

Hardware makers also published reusable interaction code. Leap Motion released its Interaction Engine as an early-access beta for Unity in August 2016; the company described it as "a layer that exists between the Unity game engine and real-world hand physics", intended to make grabbing virtual objects with bare hands work reliably. Google launched Daydream Elements at Google I/O in 2017 and detailed its first set of interactions that September; its designers described it as "a set of best practices and reusable code for some of the most fundamental things you do in VR". It included teleportation with an arc drawn to the destination, tunneling (the field of view gets smaller and the background is replaced with a grid during movement), a chase camera, and click and swipe menus for the Daydream controller. Google reported that a stable grid was the most effective tunneling treatment for reducing discomfort.

On the Apple Vision Pro, people perform most actions by looking at a virtual object and making an indirect gesture, such as a tap, to activate it. At WWDC23 Apple's designers explained that in visionOS "combined with eyes for targeting, hand gestures are the primary way to interact across the system", with a pinch of the fingers acting as the equivalent of a tap on a phone screen. The combination had been studied earlier in research such as the 2017 "Gaze + Pinch" paper by Ken Pfeuffer, Benedikt Mayer, Diako Mardanbegi and Hans Gellersen, which used eye gaze to select targets and indirect freehand gestures to manipulate them, whether the target is near or far.

Mike Alger's design methods

Alger's 2015 paper presented pre-visualization design methods for "volumetric user interfaces and experiences within the larger scope of a virtual reality operating system". He published an initial manifesto video in January 2015, presented at Ravensbourne and described as a declaration of intent for "the research and design of user interfaces for ergonomic multitasking in virtual reality, focused primarily on the interaction design of operating systems and browsers", and a final summary video, VR Interface Design Pre-Visualisation Methods, in October 2015. His workflow used video and motion graphics tools to mock up animated interfaces and test content zones before development.

In the Road to VR interview, Alger said that projecting UI onto a sphere around the user feels more comfortable because every element is the same distance from the eyes, which helps avoid the vergence-accommodation conflict. He also suggested turning the depth cues that 2D interfaces imitate, such as drop shadows, into real depth in VR, and designing interfaces that do not force repetitive, fatiguing motions. By January 2016 he had taken a job in Google's VR design department.

Load video

YouTube

YouTube might collect personal data. Privacy Policy

ContinueDismiss

VR Interface Design Manifesto by Mike Alger

Load video

YouTube

YouTube might collect personal data. Privacy Policy

ContinueDismiss

VR Interface Design Pre-Visualisation Methods by Mike Alger

Content placement and ergonomics

Alger's content zones

Alger defined guideline zones for a seated user in a non-rotating chair wearing the Oculus Rift DK2, whose horizontal field of view he took as 94.2 degrees from Oculus' Unity camera settings. Citing Chu, he gave comfortable horizontal head rotation as 30 degrees from center and the maximum as 55 degrees. Combined with half the field of view, this gave a comfortable content area reaching 77 degrees to each side (94/2 + 30) and a peripheral area, visible with strain, reaching 102 degrees (94/2 + 55). Content behind that was placed in a "curiosity zone" that users would see only by turning their bodies.

In depth, he treated a sphere of about 0.5 meters around the head as a "no-no zone" for persistent interface elements, because the eyes strain to focus on close objects. For the far limit he calculated the depth at which stereo disparity falls below one pixel. On the DK2 each pixel covers roughly 0.1 degrees, and with an interpupillary distance of 63 mm this gives a maximum perceivable depth of 20.34 meters (14.9 to 22.3 meters across IPDs of 52 to 78 mm). He wrote that other headsets of the time produced nearly the same 20 meter figure, so content beyond that distance loses the benefit of depth perception. Vertically, again citing Chu, he gave upward head movement as 20 degrees comfortable and 60 degrees maximum, and downward movement as 12 degrees comfortable and 40 degrees maximum. Drawing on office ergonomics, he placed content meant for long working periods between 15 and 50 degrees below eye level and beyond 0.6 meters, with text for long reading at the focal distance of the lenses (1.3 meters for the DK2).

For touch interfaces he defined a zone within two-thirds of the user's arm extension, minus the no-no zone, to limit fatigue. He noted that this touch zone overlaps part of the main workspace, so buttons can get in the way of the content being viewed, and that adding a virtual floor (which helps users feel balanced) cuts off much of the workspace below the horizon. He stressed that the zones apply to a non-rotating seated setup on the DK2 and would change for swivel chairs, room-scale experiences, social applications or higher-resolution displays.

Platform guidance

Current platform guidelines give similar but not identical figures. Meta's Horizon OS documentation says that objects a user fixates on for an extended time, such as a menu, should be rendered at least 0.5 meters away, and that "many have found that 1 meter is a comfortable distance for menus and GUIs". Microsoft's comfort guidance for HoloLens gives the optimal zone for hologram placement as 1.25 m to 5 m, recommends fading content out at 40 cm with a clipping plane at 30 cm, and says the optimal region for content is 0 to 35 degrees below the horizon. Apple's visionOS guidelines recommend placing content people will read or engage with over time at least one meter away. Google's Android XR guidance keeps spatial panels between 0.75 and 5 meters from the user, spawns a panel's center 1.75 meters away with its vertical center 5 degrees below eye level, and recommends keeping content within the central 41 degrees of the field of view.

Platform	Placement distance	Target size and spacing	Other placement rules
Meta Horizon OS	Fixated content at least 0.5 m, about 1 m for menus; direct-touch UI at 42 to 46 cm; indirect (ray or gaze) UI at roughly 0.8 to 3 m; avoid the 0.5 to 0.8 m middle distance	Minimum hit target 48dp at panel scale (roughly 22 mm at touch distance); minimum angular collider size roughly 2.5 to 3 degrees; 12 mm between interactables	Place frequently used elements in the bottom half of touch panels; do not anchor menus to the wrist
Microsoft HoloLens	Optimal 1.25 to 5 m; encourage interaction at 1 m or farther; fade at 40 cm, clip at 30 cm	Not specified in the comfort guidance	Content 0 to 35 degrees below the horizon; avoid gaze more than 10 degrees above or 60 degrees below the horizon and neck rotation more than 45 degrees off center
Apple visionOS	At least 1 m for content viewed over time	Minimum 60 points for an eye target; item centers at least 60 points apart or a 16 point margin	Keep main content in the center of the field of view; avoid anchoring content to the head
Android XR	Panels 0.75 to 5 m; spawn at 1.75 m	Recommended target 56dp or larger, visual affordance 48dp or larger	Panel center 5 degrees below eye level; content in the central 41 degrees; text at 14dp or larger

Microsoft recommends a gaze angle at rest of between 10 and 20 degrees below the horizon, because the head tends to tilt down slightly during activities, and advises limiting most interactions to either horizontal or vertical head movement rather than both. Apple notes that looking at the edges of the field of view can tire the eyes, so the edges suit secondary content that is not needed all the time.

Depth, text and the vergence-accommodation conflict

Meta's documentation explains that in fully immersive headsets the stereoscopic images are always presented at the same optical distance, so the eyes' focus (accommodation) stays fixed while their convergence (vergence) changes with the depth of the virtual object; its minimum distance for fixated content is meant to prevent the resulting eye strain. Microsoft states that HoloLens displays sit at an optical distance of about 2.0 m and that the focal distances of immersive headsets are generally between 1.25 m and 2.5 m. It adds that content moving in depth, or tasks that force rapid switching between near and far focus, are more likely to cause discomfort, and suggests a "depth budget" for time spent viewing near content that moves in depth, for example no more than 25 percent of the time.

Both Apple and Meta advise against adding depth to text. Apple writes that "text that appears to hover above its background is difficult to read", while Meta recommends using depth for large, important elements like UI panels but "avoid using depth on text components as it can make them harder to read". Alger made a related point in 2015: extruding text into 3D does not help readers recognize letter shapes, so text stays an essentially 2D element even in VR.

Head-locked, body-locked and world-locked UI

UI can be attached to the user's head (head-locked), follow the body, or stay fixed in the world. Microsoft warns that heads-up displays rigidly locked to head orientation are likely to cause discomfort, and recommends body locking instead: the display translates with the user immediately but only reorients to the field of view after the head has turned past a threshold. Apple's guidelines say that anchoring content so it stays statically in front of someone "can make them feel stuck, confined, and uncomfortable", and recommend anchoring content in the person's space instead. Meta gives the same advice for mixed reality, noting that content fixed in front of the camera reduces awareness of the surroundings when it covers passthrough.

Meta also advises against menus that follow the user's hand or wrist, because the menu shifts while the other hand tries to poke it and the wrist moves during gameplay. Spawning a menu from the wrist, for example when the user looks at their palm, is acceptable as long as the menu then stays static in world space.

Diegetic and non-diegetic interfaces

Game UI can be classified by whether elements belong to the virtual world. In a 2009 master's thesis at Chalmers University of Technology, written with EA DICE about first-person shooters, Erik Fagerholt and Magnus Lorentzon mapped UI along two dimensions: whether an element is shown in the 3D game space, and whether it exists in the game's fiction. Non-diegetic elements are overlays outside both; geometric (spatial) elements sit in the 3D world without being part of the fiction; meta-representations belong to the fiction but are shown as an overlay; diegetic elements are both spatial and fictional.

In VR research, Paola Salomoni and colleagues designed and tested three diegetic interfaces on the Oculus Rift for two common GUI components, a shell and a global control interface, arguing that many GUI components should be redesigned with a diegetic approach because simple adaptations of them can seem unnatural, awkward or out of place.

Input and interaction

Direct and indirect interaction

Platform guidelines distinguish direct interaction, where the user touches a control with a hand or controller, from indirect interaction at a distance through a ray, gaze or both. Meta describes direct touch as the most intuitive option for UI within arm's reach with a small number of buttons, and indirect interaction as the most comfortable, since it needs less arm motion; a hybrid model switches from ray to touch as the hand approaches a panel. Apple likewise states that indirect gestures "are comfortable to perform at any distance" and recommends them for UI and common components, while direct gestures are "best for infrequent use" because people may find it tiring to keep their arms raised for long periods.

Alger's 2015 paper surveyed earlier conventions: a ray cast like a laser pointer as a cursor, sometimes widened into a cone to make distant targets easier to hit; radial menus around the hand, which can be reached quickly with muscle memory; and the "World in Miniature", a small copy of the environment whose changes apply to the full-size world. Bowman and Hodges concluded that grabbing and manipulation should be treated as separate problems, which their hybrid HOMER techniques do.

Gaze and hand tracking

Eye-tracked interfaces have to deal with Jacob's Midas Touch problem, in which looking at something is mistaken for choosing it. visionOS highlights an element when the user looks at it and activates it only on a pinch. Apple states that the hover effect runs outside the app's process, so an app only learns which element was focused when the user performs a gesture on it. Meta's guidance for hand tracking says gaze handles distance targeting where eye tracking is available, with ray casting as the fallback, and warns against using both at once because they compete for intent.

Hand tracking has no haptic feedback, so Meta recommends clear hover and pressed states and audio cues for every action, a default button travel depth of 7 mm, "touch limiting" so the virtual hand cannot pass through a panel, and reserving direct touch for the index finger to avoid accidental activations. It also names the "Midas Touch effect" of accidental selections as a risk of densely packed hand-tracked UI.

Text entry

In 2015 Alger listed text input, along with locomotion, among the "less solved problems" of VR interaction. In a 2018 CHI study, Marco Speicher, Anna Maria Feit, Pascal Ziegler and Antonio Krüger implemented six methods for selecting characters on a virtual keyboard and found that pointing with tracked hand-held controllers outperformed all the others, while head pointing could be a viable alternative depending on the available hardware.

Arm fatigue

Mid-air interaction is prone to fatigue and a feeling of heaviness in the arms known as the "gorilla-arm effect". Juan David Hincapié-Ramos and colleagues at the University of Manitoba proposed a metric called Consumed Endurance at CHI 2014 to quantify it; in their study, interaction with the arm bent at the center plane consumed the least endurance, while working at shoulder height or with the arm extended consumed more. Microsoft recommends that experiences avoid constant, repeated gesture input, for example by adding short breaks or mixing gesture and speech input. Alger reached a similar conclusion for his VR operating system concept, moving from a standing design with raised arms to a primarily seated one with minimal wrist movement.

Comfort and locomotion

Meta notes that motion sickness can occur when there is a mismatch between what the user sees and how they move. Meta's design guidance favors teleportation and snap turning, customizable comfort settings, and avoiding sudden acceleration, rapid direction changes and unnatural camera movements. The "Point & Teleport" technique, in which users point where they want to be and are moved there instantly, was compared with walk-in-place and joystick locomotion by Evren Bozgeyikli and colleagues at CHI PLAY 2016; the authors expected it not to cause motion sickness because it "does not involve any visible translational motion", and their 16 participants found it fun and user friendly, while adding a direction component degraded the experience. Microsoft describes a "blink" approach for immersive headsets that moves the user rapidly (at 100 m/s) toward the chosen spot while fading the image out and back in.

Restricting the field of view during movement is another technique. Ajoy Fernandes and Steven Feiner of Columbia University presented dynamic field-of-view restrictors at IEEE 3DUI 2016, where the paper won the Best Paper Award. In their study of 30 participants, those using the restrictors felt more comfortable and stayed in the virtual environment longer, without a reduced sense of presence, and most participants were not aware of the intervention. Google's Daydream Elements included a similar tunneling technique.

Frame rate also affects comfort. Alger cited a minimum of about 75 frames per second recommended by Oculus in 2015 and the 90 frames per second Valve required for the Vive. Microsoft recommends rendering at a minimum of 60 FPS on its mixed reality devices, and at higher rates on devices that support them, partly because lower frame rates add latency.

Accessibility

Apple lists VoiceOver, Switch Control, Dwell Control, Guided Access and Head Pointer among the accessibility technologies supported on Apple Vision Pro, and advises giving people multiple ways to interact with an app. Meta's comfort guidance asks developers to make locomotion work for standing, seated and assistive-device users.

See also

• Human-computer interaction
• Locomotion
• Virtual reality sickness
• Vergence-accommodation conflict
• Hand tracking
• Eye tracking
• Heads-up display