Valve's Abrash Forecasts Consumer VR Hardware on PC by 2015
Steamworks Development
Consumer-priced virtual reality will arrive within two years because existing display and tracking technologies can cross the human perceptual threshold for presence without scientific breakthroughs.
AI-assisted summary of What VR Could, Should, and Almost Certainly Will Be within Two Years (Steam Dev Days 2014) — original source. Check the original for full context.
Source published:
Key ideas
- Valve forecasts that compelling, consumer-priced VR hardware capable of true psychological presence can be delivered by around 2015 using modifications of existing commodity components rather than speculative breakthroughs. (0:00)
- True presence requires meeting strict simultaneous hardware thresholds, including low-persistence displays (under 3 ms) at high refresh rates (~95 Hz), sub-20 ms motion-to-last-photon latency, and sub-millimeter 6DoF tracking. (16:06)
- The PC ecosystem remains the mandatory home for early VR evolution due to the immense compute power needed for 95 Hz stereoscopic rendering, unconstrained developer iteration, and open collaboration. (5:23)
- Software paradigms must be reinvented from scratch because traditional artificial FPS locomotion induces severe motion sickness, and conventional 3D art tricks like normal mapping break down under stereoscopic inspection. (23:35)
Detailed notes
## Background Published on YouTube by Steamworks Development on February 11, 2014, this video captures a presentation delivered by Valve's Michael Abrash at Steam Dev Days 2014. According to the video description, the talk covers what affordable consumer virtual reality hardware will be capable of achieving within two years, based on working prototypes built at Valve, and explains why developers should start building for the medium immediately.
## TL;DR Abrash outlines Valve's technical roadmap and thesis for bringing compelling, consumer-priced virtual reality to market by 2015. He defines the perceptual threshold of 'presence'—the subconscious conviction of being inside a virtual world—and details the strict hardware requirements needed to achieve it, including low-persistence displays, 95 Hz refresh rates, precise 6DoF tracking, and sub-20ms motion-to-photon latency. Highlighting Valve's open collaboration with Oculus, Abrash explains why high-performance PCs are the only viable home for early consumer VR and urges game developers to pioneer entirely new gameplay vocabularies rather than porting traditional 2D monitor titles.
## Atmosphere & style The presentation carries the measured, authoritative gravitas of a veteran systems architect delivering an industry prophecy rather than hype. Abrash speaks in a calm, analytical tone, blending rigorous perceptual neuroscience and display physics with personal anecdotes of working alongside John Carmack and standing frozen in visceral panic atop a simulated VR ledge. The room feels charged with quiet anticipation, as the talk methodically strips away science-fiction hand-waving to reveal an imminent, engineering-driven platform revolution.
## Bulletin - [00:00](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=0s) Valve forecasts that compelling, consumer-priced VR hardware capable of true psychological presence can be delivered by around 2015 using modifications of existing commodity components rather than speculative breakthroughs. - [16:06](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=966s) True presence requires meeting strict simultaneous hardware thresholds, including low-persistence displays (under 3 ms) at high refresh rates (~95 Hz), sub-20 ms motion-to-last-photon latency, and sub-millimeter 6DoF tracking. - [05:23](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=323s) The PC ecosystem remains the mandatory home for early VR evolution due to the immense compute power needed for 95 Hz stereoscopic rendering, unconstrained developer iteration, and open collaboration. - [23:35](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1415s) Software paradigms must be reinvented from scratch because traditional artificial FPS locomotion induces severe motion sickness, and conventional 3D art tricks like normal mapping break down under stereoscopic inspection.
## Key takeaways - [00:00](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=0s) Compelling consumer-priced VR hardware capable of transforming entertainment is technically feasible within two years. - [07:33](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=453s) Presence is a distinct psychological phenomenon where automated, low-level perceptual subsystems trick the brain into believing it is truly in another physical location. - [02:54](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=174s) The exact same engineering thresholds that trigger presence also eliminate the sensory cue conflicts responsible for simulator sickness. - [02:54](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=174s) Valve actively collaborated with Oculus by sharing tracking, optics, and low-persistence research, directly influencing the Rift 'Crystal Cove' prototype. - [09:20](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=560s) Low-level human vision constructs depth and scale outside conscious awareness, meaning immersion cannot be achieved purely through intellectual reasoning. - [13:20](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=800s) A wide field of view between 80 and 110 degrees is necessary to engage peripheral vision cues essential for balance and spatial orientation. - [13:20](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=800s) To prevent debilitating retinal smearing caused by the vestibulo-ocular reflex (VOR), pixel persistence must be reduced to 3 milliseconds or less. - [16:06](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=966s) Reducing pixel persistence causes severe flicker at 60 Hz, requiring display panels to run at elevated refresh rates such as 95 Hz. - [16:06](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=966s) Global display illumination is necessary to avoid the shearing, bending, and tilting artifacts inherent to rolling shutter displays during head motion. - [16:06](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=966s) HMD optical assemblies face heavy physical trade-offs, relying on one or two compact lenses per eye to magnify smartphone-derived display panels. - [18:41](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1121s) Precise optical calibration is critical; uncorrected lens distortion produces visual swimming during head movement, instantly inducing motion sickness. - [18:41](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1121s) Robust 6DoF tracking requires positional accuracy within 1 millimeter and rotational accuracy within 0.25 degrees across at least a 1.5-meter cubed volume. - [20:21](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1221s) Motion-to-last-photon latency must be kept strictly at or below 20 milliseconds (with 25 ms as the absolute ceiling) to maintain physical presence. - [20:21](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1221s) Presence is an emergent property that fails completely if any single technical requirement across tracking, latency, optics, or display falls short. - [21:25](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1285s) Significant unsolved hardware problems remain for future headsets, including retail-ready inside-out tracking, optical IPD adjustment, eye tracking, and wireless tether elimination. - [23:35](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1415s) Traditional flat-screen design rules fail in VR; developers must invent a new design language focused on slower movement, closer scale, and stereoscopically valid art assets. - [25:21](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1521s) PC is uniquely positioned to drive the VR platform revolution over consoles and mobile due to maximum raw FLOPS, open development, and rapid hardware iterations.
## Detailed walkthrough ### [00:00](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=0s) The 59-Word Summary of VR's Near Future
Introduced on stage at Steam Dev Days 2014, Michael Abrash of Valve opens by addressing the crowd's visible enthusiasm for virtual reality. Before plunging into the meat of his technical presentation, he decides to encapsulate his entire thesis into an exact 59-word statement, knowing someone in the audience would likely count the words:
> “Compelling consumer-priced VR hardware is coming, probably within two years. It's for real this time. We've built prototypes, and it's pretty incredible. Our technology should work for consumer products. VR will be best and will evolve most rapidly on the PC. Steam will support it well, and we think it's possible that it could transform the entire entertainment industry.”
Abrash stresses that even if attendees remember nothing else from his talk, taking away those 59 words alone makes their time well spent.
```mermaid flowchart LR accTitle: Valve VR Thesis accDescr: Flow showing consumer VR timeline leading to PC platform evolution and entertainment transformation. A["R&D & Prototypes (Valve)"] --> B["Consumer-Priced VR Hardware (~2 Years)"] B --> C["PC Platform Leadership (Steam Support)"] C --> D["Transformation of Entertainment Industry"] ```
[01:15](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=75s) From Snow Crash and John Carmack to Modern VR
Abrash reflects on the winding 20-year path that led him to virtual reality. In the mid-1990s, while working at Microsoft, he understood networking and was already writing rudimentary 3D rendering packages. He consumed science fiction and grasped the abstract premise of virtual worlds, yet the concept still felt like pure fantasy rather than something practical to build. That mindset shifted when he read Neal Stephenson’s novel *Snow Crash*. Thinking through the requirements of building the "Metaverse," Abrash estimated he knew how to implement around 80% of it using the technology of that era—an estimate he admits was overly optimistic, quipping that he is “still waiting for my first VR sword fight.”
Shortly afterward, Abrash met with John Carmack, who spoke uninterrupted for two hours about his grand vision: generating persistent, networked 3D servers that users could connect together and build upon freely, eventually accreting into cyberspace. Inspired by this vivid technical future, Abrash joined Carmack at id Software to build *Quake*, laying the groundwork for millions of people who now inhabit networked 3D spaces every day. Abrash credits Stephenson and Carmack for setting that path in motion, noting that while he does not consider himself a visionary, he hopes this talk will similarly point developers toward the next missing chapter of the Metaverse: virtual reality.
[02:54](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=174s) Achieving Presence and Valve's Collaboration with Oculus
While many developers viewed the original Oculus Rift DK1 as an interesting curiosity, Abrash insists VR will exert a major influence much faster than anticipated. Two years prior, Valve targeted VR as a transformative medium. Valve's R&D resulted in internal prototype head-mounted displays and bespoke room-scale demos that produce a profound psychological phenomenon known as "presence"—the unmistakable sensation of truly being inside another space, as if teleported.
Crucially, the exact same hardware capabilities required to induce presence also eliminate motion sickness. Abrash identifies Oculus as the leading candidate to deliver this level of consumer-grade hardware. While the DK1 lacked the hardware thresholds needed to induce presence, the newer Oculus Rift "Crystal Cove" prototype represents a major leap forward:
- Shipped over 50,000 Oculus Rift DK1 units to developers worldwide. - Built low-persistence prototype displays at Valve, demonstrating them to Oculus to convince them of low persistence's necessity in eliminating motion blur; this directly influenced the creation of the Crystal Cove prototype. - Shared Valve's internal tracking research, hardware architectures, optical calibration data, and lens designs directly with Oculus. - Crystal Cove integrates key presence enablers: positional translation tracking, lower latency, higher resolution panels, and low-persistence displays.
> “If Oculus executes well—and so far they seem to be—I think they could well deliver strong presence on the PC within the next two years.”
[05:23](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=323s) Why High-End PC Hardware Is Essential for Presence
Abrash argues that the advent of presence-capable VR will precipitate a sea change across entertainment, tipping the scales from passive, traditional media toward interactive computer technology. True presence cannot be achieved via streaming, films, or broadcast television because human perception demands extremely low motion-to-photon latency and massive rendering bandwidth. Those constraints mandate a head-mounted display wired directly to high-performance local hardware running heavy-duty real-time 3D rendering.
Because raw compute power is indispensable to satisfying these perceptual thresholds, the personal computer ecosystem—spanning Windows, Linux, and OS X—stands as the definitive home for VR:
> “A corollary is that the PC—Linux, Windows, and OS X—is going to be the best place for VR because that's where the most FLOPS live.”
Valve’s strategy to foster this ecosystem operates on three fronts: 1. **Steam Platform Integration:** Building native VR infrastructure into Steam (as detailed by Valve developer Joe Ludwig in another talk). 2. **Hardware R&D & Open Collaboration:** Continually advancing display, tracking, and optical research while collaborating openly with external partners like Oculus. 3. **Software Prototyping:** Experimenting internally to discover the native design language of VR and sharing findings with developers.
Abrash clarifies that while Valve has no current plans to manufacture or ship its own consumer VR hardware, that stance could change if necessary to ensure the PC platform thrives. He also extends an open invitation to any Steam partner visiting Seattle over the subsequent months to email him directly (`mabrash@valvesoftware.com`) to experience Valve's dedicated 30-minute room-scale VR demo firsthand.
[07:33](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=453s) The Neuroscience of Low-Level Perception
To explain why presence requires specific hardware specifications, Abrash delves into human perception. True immersion does not stem from conscious reasoning; rather, it requires convincing automated, low-level perceptual subsystems that function deep beneath conscious awareness.
To demonstrate how the subconscious brain manipulates perceived reality before conscious thought ever receives it, Abrash presents a classic visual illusion of an alleyway containing two spheres: - **Visual Cues Present:** Perspective convergence lines, surface textures, directional lighting, and cast shadows all signal significant spatial depth. - **The Task:** When participants are asked to adjust the nearer sphere until both circles project identical 2D sizes on the screen, they unfailingly resize the foreground circle so that it is physically larger than the background circle. - **The Reality:** In the slide demonstration, the sphere perceived as farther away looks significantly larger, yet the foreground circle is actually almost 20% larger in raw pixel dimensions.
Abrash points to functional MRI (fMRI) studies revealing that this calculation happens at the earliest visual processing layer of the brain (V1 in the visual cortex). Even before conscious awareness begins processing the image, the visual cortex allocates a larger surface area of brain tissue to the object perceived as distant, even when both objects project the exact same size onto the physical retina.
> “The visual system has already judged the sizes before the data reaches conscious awareness.”
Understanding these subconscious visual mechanisms is the foundational requirement for building hardware that can trigger genuine presence.
[09:20](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=560s) Low-Level Perception and the Cliff Demo
Michael Abrash presents an optical illusion to demonstrate how human visual processing constructs reality outside conscious awareness. Two visual targets show an "X" shape where one centerpiece appears gray and the other looks dark yellow. When the background is masked out without altering the centerpieces themselves, both patches are revealed to be the exact same shade. Because human vision evolved to maintain consistent perception across radically shifting illumination conditions, low-level visual processing relies heavily on context. Abrash emphasizes that this occurs entirely below conscious thought, just as the brain automatically warps the perceived size of shapes based on context. When virtual reality successfully stimulates these low-level sensory mechanisms, the user genuinely feels situated in a place; when it fails, they feel an indefinable disconnection.
```mermaid flowchart LR accTitle: Sensory Processing to Perception accDescr: Diagram showing sensory inputs passing through subconscious low-level processing before reaching conscious awareness.
subgraph SENSORY["Input Layer"] INP["Visual & Vestibular Cues"] end
subgraph SUBCONSCIOUS["Subconscious Processing"] LLP["Low-Level Context & VOR"] end
subgraph CONSCIOUS["Conscious Mind"] PERC["Sense of Presence & Reality"] end
INP --> LLP LLP --> PERC ```
To illustrate the visceral power of proper sensory stimulation, Abrash details a Valve VR demo room experience, issuing a humorous spoiler warning: “This is a bit of a spoiler for one of our demos so if you don't want to hear it cover your ears I'll wait a moment while you decide.” In the demo, the user stands on a diving-board-style stone ledge high above the floor of an enclosed box room textured with outdated web pages—including one announcing that Facebook is going public. Viewed on a conventional computer monitor, the flat or warped imagery appears completely unimpressive. Yet inside the head-mounted display, the physical response is uncontrollable:
> “Looking this at this on a screen even when it's not warped doesn't do anything for me but whenever I stand on that ledge in VR my knees lock up just like they did on top of the Empire State Building even though I know for certain that I'm in a demo room wearing a head mounted display looking at imagery of the inside of a badly textured box my body knows that I'm at the edge of a cliff.”
The subconscious reaction does not fade with time or repetition. Because the inputs satisfy the low-level perceptual system, the user's body remains convinced at an instinctive level that it stands over a lethal drop.
---
[11:25](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=685s) Defining Presence vs. Immersion
Abrash distinguishes between two terms often conflated in VR discussions: *immersion* and *presence*. Immersion simply means being surrounded by the imagery of a virtual world, whereas presence represents the profound sensation of actually existing inside that virtual space. Presence cannot be fully communicated in words; it can only be understood through firsthand experience.
Abrash describes the concrete physical manifestations that occur when presence takes hold: reaching out involuntarily to touch a rendered cube despite knowing the room is empty, ducking to avoid a hanging virtual pipe, experiencing visceral unease when a massive block suspends overhead, and feeling entirely unable to step off a virtual drop. Upon taking off the head-mounted display, users frequently suffer disorientation as their perceptual system readjusts to finding itself back in physical reality.
> “It's more than just looking at someplace interesting it's flipping the switch that makes you believe deep in your lizard brain that you are someplace interesting.”
Abrash explains that presence is among the most powerful non-real experiences imaginable precisely because it engages the exact same sensory channels as physical reality:
- **Visceral Engagement:** Presence operates on the "lizard brain" rather than intellectual belief, hooking deeper into the human perceptual system than any prior medium. - **Audience Variation:** Susceptibility to presence varies across individuals, but responses consistently strengthen as hardware systems refine and improve. - **The Core Thesis of Valve's VR Research:** Abrash summarizes their strategic direction: “Consequently we think that building Hardware that's capable of delivering a strong sense of presence is the key to VR success.”
---
[13:20](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=800s) Hardware Requirements: FOV, Resolution, and Low Persistence
Through iterative research and prototyping, Valve discovered that a strict baseline across multiple hardware specifications must be met simultaneously before presence can occur. Crucially, the exact same engineering factors required for presence also mitigate simulator sickness. Motion sickness typically stems from cue conflicts—such as tracking flaws that create a mismatch between visual input and the vestibular system in the inner ear.
Field of view (FOV) is the first critical requirement. Beyond creating visual immersion, a wide FOV supplies peripheral vision cues essential for motion sensing, human balance, and situational awareness. Presence begins to emerge around an 80-degree FOV and scales up substantially through at least 110 degrees, the maximum field Valve had tested.
Resolution presents a difficult hurdle because stretching an image across a wide FOV severely magnifies the display's individual pixels. In an HMD with a 110-degree FOV and a 1K × 1K per-eye resolution, the angular pixel density is merely one-seventh that of a standard big-screen television and one-tenth the resolving limit of the human eye—falling below the pixel density of the original *Quake* running at 320 × 200. Valve determined that a 1080p display represents the minimum viable resolution for presence, with 1440p and 2160p anticipated as massive visual upgrades once panels of appropriate physical dimensions become manufactured.
Pixel persistence—the duration each pixel remains illuminated per frame—is uniquely critical for head-mounted VR. On traditional stationary monitors, televisions, or mobile phones, full persistence goes largely unnoticed. In VR, however, the vestibulo-ocular reflex (VOR) keeps the human eye fixed on targets in the physical world while the head moves rapidly. If pixels stay lit throughout the entire refresh frame on a full-persistence display, the image drags across the moving retina:
> “The image on the left is with the head held still and the image on the right is a simulation of what happens with a leisurely 120° per second head turn rate on a 60 HZ full persistence display this results of in two degrees of smearing across the retina per frame which as you can see reduces detail considerably.”
To eliminate this debilitating retinal blur, Valve determined that pixel persistence must be cut to 3 milliseconds or less at 1K × 1K and 110 degrees FOV, with even shorter illumination windows required as pixel densities rise.
---
[16:06](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=966s) Refresh Rates, Global Display, and Optical Lens Design
Cutting pixel persistence to 3 milliseconds introduces a severe side effect: visible display flicker at conventional 60 Hz refresh rates. To eliminate flicker across a wide FOV, display refresh rates must be increased. Valve built an experimental low-persistence headset capable of running at 95 Hz, which successfully eliminated all perceptible flicker. Abrash notes that while an intermediate refresh rate might suffice, delivering consistent 95 Hz stereoscopic rendering will present an immense computational challenge for game engines as panel resolutions scale upward.
```mermaid flowchart TD accTitle: Trade-offs in VR Display Design accDescr: Flowchart illustrating the chain reaction from low persistence to high refresh rate demands and optical trade-offs.
A["Low Persistence (<3ms)"] -->|"Prevents retinal smear"| B["Flicker at 60 Hz"] B -->|"Requires higher refresh"| C["95 Hz Display Rate"] C -->|"Demands high performance"| D["Intense GPU Stereo Rendering Load"]
E["Cell Phone Panels"] -->|"Magnified by 1-2 lenses"| F["Severe Optical Aberrations"] F -->|"Corrected by trade-offs"| G["Subjective In-HMD Optimization"] ```
Display illumination timing also dictates image stability. Valve's prototype relies on *global display*, where every pixel in the panel lights up simultaneously. This eliminates the severe spatial distortions inherent to standard *rolling displays*, in which scan lines illuminate progressively over the frame, causing the virtual image to shear, compress, stretch, or tilt during head motion:
- **Correcting Rolling Displays:** While software reprojection could theoretically compensate for rolling shutter artifacts by updating the frame buffer mid-scan based on eye movement, that technique remains unproven. - **Eye-Tracking Bottlenecks:** Head movement cannot serve as a reliable proxy for eye motion across all edge cases, and low-latency head-mounted eye tracking remains an unsolved engineering hurdle. Consequently, global display remains the only proven method to preserve image geometry.
Turning to optics, Abrash demystifies the assumption that designing an HMD optical path is trivial. Practical consumer constraints require leveraging mass-manufactured smartphone panels magnified by lenses, as microdisplays and optical waveguides cannot provide a sufficiently wide FOV. Weight, form factor, and industrial design restrict the optics to only one or two lenses per eye.
Eliminating all optical aberrations and geometric distortions with perfection would require a massive, multi-element optical assembly. Abrash points out a diagram of an ideal optical stack where the front two lenses measure over a foot in diameter and the complete optical housing exceeds five pounds—an absurd impossibility for wearable headgear. With a hard limit of one or two compact lenses per eye, compensating for one aberration inevitably exacerbates another. Because focal lengths, eye relief, and optical distortions cannot be evaluated purely through automated mechanical metrics, every lens design requires subjective, time-consuming evaluation directly inside VR headsets across diverse users. Valve has produced lenses capable of sustaining presence, but the overall design space remains a delicate web of trade-offs.
[18:41](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1121s) Optical Calibration and Tracking Requirements for Presence
Optical calibration may initially sound like a secondary engineering detail, but Michael Abrash emphasizes that it is both hugely challenging and utterly critical because the human visual system is exquisitely sensitive to slight geometric and lens errors. A virtual scene might look immaculate when inspected statically, but the moment a user turns their head from side to side, miscalibrated optics cause straight lines and surfaces to warp and ripple horribly. This visual swimming instantly shatters the illusion of presence and triggers acute motion sickness almost immediately. Getting calibration dialed in is also a prerequisite for debugging the rest of the VR pipeline; Valve discovered that until their optical calibration was dead-on, they could not even detect minor head-tracking glitches because the optical instability masked them entirely. As Abrash notes, “The bottom line is that a highly accurate process for characterizing the lenses and correcting the rendered images is absolutely essential.”
Once visual clarity and calibration are achieved, the virtual image must be tracked and positioned so that the brain’s perceptual system accepts it as an unyielding physical reality. This necessitates rock-solid 6-degree-of-freedom (6DoF) head tracking that reports translation along the X, Y, and Z axes alongside angular orientation. Valve’s internal testing determined that sustaining presence requires tracking accuracy of within a single millimeter of translational position and a quarter of a degree in orientation, maintained continuously across a tracking volume of no less than 1.5 meters on each side. Delivering that precision is considerably harder than it sounds, and no existing consumer tracking solution at the time came close. Showing the room-scale tracking setup used in Valve's Dev Days demos, Abrash jokes that while their fiducial-marker rig works, “it isn't exactly consumer friendly” because it effectively requires “wallpapering your house.” However, Valve is developing alternative tracking architectures that dispense with wall markers, giving them high confidence that consumer-grade tracking is viable in the near term.
- Tracking precision targets for presence: ≤ 1 mm position error, ≤ 0.25° orientation error. - Minimum tracking volume required: at least 1.5 meters cubed. - Full 6DoF tracking (translation in X, Y, Z plus rotational orientation) is mandatory; rotational tracking alone fails to convince the human perceptual system.
[20:21](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1221s) Latency Budgets and Presence as an Emergent Phenomenon
Low latency is the next non-negotiable pillar: the virtual image must appear in the exact right place at the exact right moment. Valve measures latency strictly from the physical instant head motion occurs to the precise moment the final photon generated by that motion strikes the eye. Testing reveals that a motion-to-last-photon latency of 20 milliseconds, coupled with robust sensor prediction, reliably induces presence, while 25 milliseconds represents the absolute upper ceiling. Exceeding that threshold causes the virtual environment to feel detached rather than nailed down in space, causing the human perceptual system to reject the experience as reality.
Abrash underscores that presence is not an isolated attribute that can be achieved piecemeal through a single standout component; it is an emergent state that arises only when every technical element clears its required threshold simultaneously. If the lenses are improperly calibrated, scenes warp during head rotations; if tracking drifts or latency creeps above 20 to 25 milliseconds, the perceptual illusion collapses. As Abrash summarizes the challenge, “I want to emphasize that presence is not a property of any one of the elements I've discussed; it's a property that emerges when all the elements are good enough.”
```mermaid flowchart TD accTitle: Pipeline of Key Requirements for Inducing Presence accDescr: Interdependent hardware and tracking elements that together produce the emergent sense of VR presence. subgraph S1["Visual Presentation"] A["Optical Calibration<br/>Characterize lenses & correct rendering"] B["Low-Persistence Display<br/>Global illumination eliminating blur"] end subgraph S2["Spatial Tracking & Speed"] C["6DoF Tracking<br/>≤ 1 mm position, ≤ 0.25° orientation"] D["Low Latency<br/>≤ 20 ms motion-to-last-photon"] end subgraph S3["Emergent Outcome"] E["Perceptual Presence<br/>Brain accepts virtual world as real"] end A --> E B --> E C --> E D --> E ```
- Motion-to-photon latency limit: ≤ 20 ms (with 25 ms being the borderline limit). - Latency metric: measured from physical head motion onset to the emission of the last corresponding photon. - Emergent nature of presence: “Presence can't be induced if even one of the key elements is subpar.”
[21:25](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1285s) The 2015 Consumer Headset Specification and Hardware Roadblocks
Synthesizing Valve’s internal research and hardware prototypes, Abrash outlines what a viable consumer head-mounted display (HMD) could look like by 2015. Valve's working prototypes already achieve genuine presence using modified commodity parts, proving that cost-effective manufacturing at scale is within reach without waiting for speculative breakthroughs. While future headsets will eventually demand custom silicon and massive research investments, delivering an introductory consumer product requires only focused execution. Explaining this optimism, Abrash states, “In contrast, we believe everything on this slide is doable with relatively minor tweaks of existing technology—no breakthroughs or miracles are needed, just solid engineering.”
At the same time, Abrash is careful to distinguish between an internal laboratory prototype and a consumer-ready retail product. Turning an R&D demonstration into an affordable, manufacturable device entails significant commercial hurdles. In an ideal world, presence would benefit from vastly superior specs: displays could easily leverage up to 100 times as many pixels, wider fields of view, and optical assemblies that accommodate per-user lens positioning and interpupillary adjustments. Eye tracking would provide huge gains for rendering and presence but remains far from solved. Eliminating tether cables through wireless transmission represents another major ergonomic hurdle, while securing low-persistence, global-update, ultra-low-latency panels will require deep partnerships with display manufacturers.
- 2015 hardware forecast: an affordable consumer HMD capable of inducing true presence can be delivered within two years using modified commodity parts. - Display headroom: VR displays could realistically use up to 100× the pixel counts of current prototypes alongside wider fields of view. - Unsolved hardware engineering challenges: retail-ready 6DoF tracking, per-user optical and IPD calibration, robust eye tracking, tetherless wireless video transmission, and custom low-persistence panel manufacturing.
[23:35](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1415s) Unexplored Frontiers: Input, Locomotion, and a New Game Design Vocabulary
Beyond the visual headset, the non-visual facets of virtual reality represent even larger, more stubborn technical hurdles. Spatial 3D audio, haptic feedback, body tracking, and dedicated VR controllers are essential to long-term immersion, yet their interaction with core game mechanics is almost entirely uncharted territory. Abrash warns that figuring out these sensory domains will require years of collective experimentation across the game industry. In particular, software and game design must be reinvented from scratch because traditional 2D monitor design paradigms break down inside an immersive head-mounted display.
Locomotion exemplifies this design crisis: conventional first-person shooter (FPS) artificial movement instantly induces severe motion sickness in VR. Developers must likely shift toward slower character movement, confined scale, and intimate, close-up object manipulation, and then discover how to make those mechanics genuinely fun. Multiplayer VR promises to be far more compelling and socially profound than monitor-based gaming, evoking the groundbreaking early eras of deathmatch, player-created mods, and persistent dedicated servers. Visual aesthetics also require total re-evaluation. Normal mapping falls apart under stereoscopic inspection, high-frequency screen textures can look like cheap theatrical stage props, while surprisingly simple geometric scenes can evoke breathtaking realism. As Abrash points out, “In the end, a whole new VR gameplay vocabulary will need to be developed, just as was the case with FPSs.”
- Traditional FPS artificial locomotion causes severe motion sickness; VR demands new movement systems, likely focusing on slower speeds and close-up interactions. - Traditional art pipelines fail: normal maps look synthetic and flat in stereo 3D, and photorealistic monitor assets often look like cheesy stage sets. - Multiplayer VR carries the potential to become the most powerful online social medium yet devised. - Input and haptics remain harder and less explored than display hardware.
[25:21](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1521s) The Next Major Platform Shift and Why VR Belongs on PC
Abrash places the emergence of VR in the lineage of seminal industry platform shifts that enabled entirely new classes of software, comparing its trajectory to the arrivals of *Myst*, *Quake*, *Wii Sports*, and *Angry Birds*. Straightforward ports of existing flatscreen titles can offer interesting visual immersion, but true greatness demands experiences built natively from the ground up for virtual reality. Drawing a parallel to the dawn of real-time 3D gaming, Abrash notes the nostalgic cultural weight carried by id Software’s classic shooter on its recent twentieth milestone: “The first game that nails a VR experience is going to be remembered fondly for a long, long time.”
When that breakthrough software appears, Abrash argues it will almost certainly debut on the PC rather than on consoles or mobile. First, PC hardware iterates rapidly and openly, as demonstrated by the rapid evolution of the Oculus Rift, whereas game consoles remain locked into fixed hardware cycles for years. Second, the open PC ecosystem allows a far broader, unconstrained community of software and hardware developers to freely experiment. Third, VR demands staggering computing performance—rendering independent stereoscopic viewpoints at 95 Hz with ultra-low latency requires top-tier graphical power that consoles cannot match, while mobile chips remain strictly bounded by heat and battery limits. With Valve baking native VR support into Steam and actively sharing its R&D discoveries, Abrash concludes with an emphatic message to developers: “Today's key takeaway is simple, but it has powerful implications: a great VR system at a consumer price in 2015 is more than just possible, it's sitting there waiting to happen, and it will happen, if not in 2015, then soon after.”
- Historical parallels: VR represents a foundational platform shift comparable to CD-ROM (*Myst*), true 3D engines (*Quake*, *Doom*), motion controllers (*Wii Sports*), and touch interfaces (*Angry Birds*). - Performance baseline: VR requires dual-buffer stereo rendering running at 95 Hz minimum refresh rates. - PC advantages over consoles and mobile: - Rapid, continuous hardware revision cycles instead of static multi-year console lifespans. - Open experimentation for independent developers without platform gatekeepers. - Raw compute capacity uninhibited by mobile thermal envelopes or console power constraints. - Active ecosystem integration and driver investment through Steam and Valve R&D.
[27:35](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1655s) Building the Future of VR Together
Michael Abrash brings his presentation to a close by framing the advancement of virtual reality not as a solitary effort by Valve, but as an open, collaborative journey with the entire developer community. He emphasizes the necessity of an active dialogue, calling for “a two-way exchange so together we can make VR one of the great entertainment revolutions.” Acknowledging that the technical challenges, creative opportunities, and hardware breakthroughs discussed during the session only scratch the surface of what lies ahead, Abrash invites attendees to continue the discussion directly during the upcoming Q&A session as well as through his blog. With that, he thanks the audience and concludes his talk to enthusiastic applause.
- **Collaborative development**: Abrash stresses that realizing the promise of consumer VR requires open communication and shared problem-solving between hardware pioneers and software developers. - **Continuing the dialogue**: Attendees are invited to follow up with Abrash in the subsequent Q&A session and follow along with his ongoing writings and technical deep dives on his blog.
## Notable quotes > [25:21](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1521s) The first game that nails a VR experience is going to be remembered fondly for a long, long time. — Michael Abrash
> [20:21](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1221s) Presence can't be induced if even one of the key elements is subpar. — Michael Abrash
> [23:35](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=1415s) In the end, a whole new VR gameplay vocabulary will need to be developed, just as was the case with FPSs. — Michael Abrash
> [02:54](https://www.youtube.com/watch?v=G-2dQoeqVVo&t=174s) If Oculus executes well—and so far they seem to be—I think they could well deliver strong presence on the PC within the next two years. — Michael Abrash
## Mentioned Michael Abrash, Valve, Steam, Steam Dev Days, Neal Stephenson, Snow Crash, John Carmack, id Software, Quake, Doom, Oculus, Oculus Rift DK1, Crystal Cove, Joe Ludwig, Facebook, Empire State Building, Myst, Wii Sports, Angry Birds
## Video description We've figured out what affordable VR hardware will be capable of within a couple of years, and assembled a prototype that reveals that that level of VR hardware is capable of stunning VR experiences. That hardware is almost certainly going to appear in that timeframe, and it will be worth starting to develop for it now. This talk will discuss what that hardware is, and what it makes possible.
Speaker: Michael Abrash (Valve) Slides (PDF): http://media.steampowered.com/apps/steamdevdays/slides/vrshouldbe.pdf Slides (PowerPoint): http://media.steampowered.com/apps/steamdevdays/slides/vrshouldbe.ppt
[Source](https://www.youtube.com/watch?v=G-2dQoeqVVo)