Virtual image
A virtual image is an optical image formed at a point from which light rays only appear to come, rather than a point where they actually meet. The image seen in a flat mirror is the standard example: it cannot be projected onto a screen, because the rays only appear to originate from a common point behind the mirror.[1] A real image, by contrast, can be caught on a screen because the rays physically pass through it.[1]
Conventional head-mounted displays depend on virtual images. The display panel or microdisplay sits a few centimeters from the eye, too close to focus on, and a lens or combiner turns it into a magnified virtual image that appears to float at a comfortable distance.[2][3] The IEC measurement standard for eyewear displays covers both VR goggles and AR glasses "using virtual image optics", and defines the eye box, eye relief and field of view of such devices in terms of the virtual image.[4] Because the distance at which that image appears is usually fixed, the virtual image is also central to the vergence-accommodation conflict.[5]
Definition
In geometric optics, a virtual image forms where diverging rays, traced backward, appear to meet. For a plane mirror the image lies as far behind the mirror as the object lies in front of it, so the object and image distances have opposite signs.[1] The difference from a real image is practical as well as geometric: a real image can be projected onto a screen and a virtual image cannot.[1]
A simple magnifier is the case closest to how headset optics work. When an object is placed within one focal length of a convex lens, its image is virtual, upright and larger than the object.[6] Placing the image far away gives slightly less magnification than placing it at the eye's near point, but OpenStax notes that it "provides for the most comfortable viewing conditions, because the eye is relaxed when viewing a distant object."[6]
Virtual images in VR headsets
The human eye cannot focus on a display held a few centimeters away. Lanman and Luebke of NVIDIA illustrated this in 2013 with a bare OLED microdisplay close to the eye, which is perceived as "a severely defocused image", and noted that conventional near-eye displays "require bulky magnifying optics to facilitate accommodation."[2]
In a typical VR headset the display panel is placed in front of the eye with imaging optics in between.[3] The gap between panel and lens is usually close to the focal length of the lens system, which is why conventional VR headsets usually have a considerably larger volume than eyeglasses, most of it empty space. Xiong and colleagues, in a 2021 review in Light: Science & Applications, give about 4 cm as the focal length of the transmissive lens in conventional VR headsets, and note that Fresnel lenses are thinner than conventional lenses but do not change the required lens-to-panel distance significantly.[3] Another way to reduce the overall length is to fold the optical path, as in headsets built on pancake optics.[3]
Because the distance between the panel and the optics is fixed in current VR display modules, "the VR imagery is displayed at a single depth", even though the content is rendered with stereoscopic parallax for two eyes.[3] The distance at which the eye must focus is therefore fixed as well; Microsoft's documentation calls it the focal distance of the display.[7] Reported values for fixed-focus headsets include:
| Device or class | Reported focal distance of the virtual image | Source |
|---|---|---|
| HoloLens (1st gen) and HoloLens 2 | About 2.0 m; Microsoft recommends placing holograms between 1.25 m and 5 m | Microsoft[7] |
| Windows Mixed Reality immersive headsets | Generally between 1.25 m and 2.5 m | Microsoft[7] |
| Most VR headsets (2018) | "Something around two meters" (Oculus researcher Douglas Lanman) | UploadVR[8] |
| Typical VR headsets (2022) | A single fixed focal plane at about two meters | MIXED[9] |
Such distances can also be given in diopters, the reciprocal of the distance in meters; Zhan and colleagues, for example, describe focal planes at 0.2 and 0.8 diopters as a viewing distance range of 1.25 m to 5 m.[5]
Vergence-accommodation conflict
In a stereoscopic headset the eyes converge on objects at whatever depth the rendered disparity suggests, but they must focus (accommodate) on the virtual image to keep it sharp. Microsoft's design guidance states that in most head-mounted displays users "will always accommodate to the focal distance of the display to get a sharp image, but converge to the distance of the object of interest to get a single image."[7] Zhan, Xiong, Zou and Wu describe the stimulus to accommodation as "fixed by the display panel and viewing optics", while the stimulus to vergence changes with the image content.[5] The resulting mismatch has been reported to cause visual fatigue, including eye strain, blurred vision and headache.[5]
For HoloLens, Microsoft advises keeping content that users converge on as close to the display's 2.0 m optical distance as possible, recommends against showing holograms closer than 40 cm, and suggests fading content out at 40 cm with a clipping plane at 30 cm.[7]
Virtual images in AR displays
Optical see-through AR displays overlay a virtual image on the user's direct view of the world, using a combiner such as a beam splitter, birdbath, freeform prism or waveguide.[3] Xiong and colleagues note that the conflict between vergence and accommodation "may be more serious in AR than VR", because the virtual image is superimposed directly on real objects that carry correct depth cues.[3] Waveguide combiners that use grating out-couplers produce an image with a fixed depth at infinity; the authors describe generating multiple depths with a varifocal or multifocal scheme as the most practical fix, at the cost of stacking extra waveguide layers.[3]
The Magic Leap One used a multiple-plane design. According to UploadVR's 2018 description, it used color-specific waveguides to present images at two distinct focal planes, each made of red, green and blue layers.[8]
Head-up displays
Head-up displays (HUDs) in aircraft and cars also present virtual images through a windshield or combiner.[10] In aviation HUDs the light is collimated so that the rays "appear parallel out to infinity"; as a result, a pilot does not need to refocus when looking between the symbology and the outside scene.[11] In automotive HUDs the windshield or a separate combiner shows an enlarged virtual image of driving information.[10] Radiant Vision Systems, a display measurement company, states that a conventional automotive HUD projects information at a fixed distance of 2 to 4 m, while augmented reality HUDs could place it 10 or 20 m away.[12] A 2021 research prototype from the Beijing Institute of Technology used three holographic optical elements to place red, green and blue images at 150 cm, 500 cm and 1000 cm.[10]
Research on variable virtual image distance
Two families of displays try to move or multiply the virtual image so that accommodation matches vergence. Varifocal displays track where the eye is focusing and place one image at that depth; multifocal displays show several virtual image planes at once without tracking focus.[5] Zhan and colleagues describe multifocal displays as generating "multiple 2D virtual images spanning the desired range of spatial focal depths", and noted in 2020 that the technology was rarely used in commercial products.[5]
Early prototypes moved optical parts mechanically. Their review describes a 1996 display by Shiwa and colleagues whose relay lens moved 4 mm in 0.3 s to sweep the virtual image from 20 cm to 10 m, following the viewer's gaze. A design by Shibata and colleagues moved a 6-inch LCD panel instead, using a telecentric optical system that kept the size of the virtual image unchanged over a range of 30 cm to 200 cm.[5]
In consumer VR research, the Oculus Half Dome prototype shown at Facebook's F8 conference in May 2018 combined a 140-degree field of view with varifocal displays that "physically move back and forth" to shift the focus of the optics.[13] Meta's Reality Labs Butterscotch Varifocal prototype, shown in 2023, drew on the mechanical varifocal mechanism of Half Dome 1 and 2.[14] Other approaches to the same problem include light field displays, holographic displays and Maxwellian-view retinal projection, which removes the accommodation cue rather than supplying it.[3][2]
See also
- Near-eye display
- Optical combiner
- Eye box
- Exit pupil
- Vergence-accommodation conflict
- Varifocal display
- Focal surface display
- Head-up display
References
- ↑ 1.0 1.1 1.2 1.3 Samuel J. Ling, Jeff Sanny, William Moebs. "2.1 Images Formed by Plane Mirrors". University Physics Volume 3. OpenStax. https://openstax.org/books/university-physics-volume-3/pages/2-1-images-formed-by-plane-mirrors. Retrieved 2026-09-27.
- ↑ 2.0 2.1 2.2 Douglas Lanman, David Luebke (2013-07). "Near-Eye Light Field Displays". ACM SIGGRAPH 2013 Emerging Technologies. https://history.siggraph.org/wp-content/uploads/2022/03/2013-11-Lanman_Near-EyeDisplays.pdf. Retrieved 2026-09-27.
- ↑ 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 Jianghao Xiong, En-Lin Hsiang, Ziqian He, Tao Zhan, Shin-Tson Wu (2021). "Augmented reality and virtual reality displays: emerging technologies and future perspectives". Light: Science & Applications, vol. 10, article 216. doi:10.1038/s41377-021-00658-8. https://pmc.ncbi.nlm.nih.gov/articles/PMC8546092/. Retrieved 2026-09-27.
- ↑ "IEC 63145-20-10:2019 Eyewear display - Part 20-10: Fundamental measurement methods - Optical properties (Edition 1.0, preview sample)". iTeh Standards. International Electrotechnical Commission. 2019-08. https://cdn.standards.iteh.ai/samples/101828/8ac7b0875e974aa085e18a1d6cb73b49/IEC-63145-20-10-2019.pdf. Retrieved 2026-09-27.
- ↑ 5.0 5.1 5.2 5.3 5.4 5.5 5.6 Tao Zhan, Jianghao Xiong, Junyu Zou, Shin-Tson Wu (2020-03-30). "Multifocal displays: review and prospect". PhotoniX, vol. 1, article 10. https://doi.org/10.1186/s43074-020-00010-0. Retrieved 2026-09-27.
- ↑ 6.0 6.1 Samuel J. Ling, Jeff Sanny, William Moebs. "2.7 The Simple Magnifier". University Physics Volume 3. OpenStax. https://openstax.org/books/university-physics-volume-3/pages/2-7-the-simple-magnifier. Retrieved 2026-09-27.
- ↑ 7.0 7.1 7.2 7.3 7.4 "Comfort". Microsoft Learn (Mixed Reality documentation). Microsoft. https://learn.microsoft.com/en-us/windows/mixed-reality/design/comfort. Retrieved 2026-09-27.
- ↑ 8.0 8.1 Ian Hamilton (2018-08-27). "What Magic Leap One And Facebook's Half Dome Have In Common". UploadVR. https://www.uploadvr.com/magic-leap-one-half-dome-common-differences/. Retrieved 2026-09-27.
- ↑ Tomislav Bezmalinovic (2022-07-24). "Varifocal displays simulate natural vision in VR - watch these demos". MIXED. https://mixed-news.com/en/varifcoal-vr-simulates-natural-vision-in-virtual-reality-watch-these-demos/. Retrieved 2026-09-27.
- ↑ 10.0 10.1 10.2 Zhenlv Lv, Juan Liu, Liangfa Xu (2021). "A multi-plane augmented reality head-up display system based on volume holographic optical elements with large area". arXiv preprint 2104.14315. https://arxiv.org/abs/2104.14315. Retrieved 2026-09-27.
- ↑ Radiant Vision Systems (2021-01-13). "Quality Considerations for Aviation Head-up Displays (HUDs)". Aviation Today. https://www.aviationtoday.com/2021/01/13/quality-considerations-aviation-head-displays-huds/. Retrieved 2026-09-27.
- ↑ Matt Scholz (2020-11-23). "Head-Up Displays: System Benefits from 2D to AR". AZoOptics. Radiant Vision Systems. https://www.azooptics.com/Article.aspx?ArticleID=1878. Retrieved 2026-09-27.
- ↑ Ben Lang (2018-05-02). "Oculus Reveals 140 Degree VR Headset Prototype with Varifocal Displays". Road to VR. https://www.roadtovr.com/facebook-oculus-half-dome-prototype-vr-headset-140-degree-varifocal-f8/. Retrieved 2026-09-27.
- ↑ "Demo or Die: How Reality Labs' Display Systems Research Team Is Pushing the VR Industry Toward the Future". Meta Quest Blog. Meta. 2023-07-31. https://www.meta.com/blog/reality-labs-research-display-systems-siggraph-2023-butterscotch-varifocal-flamera/. Retrieved 2026-09-27.