PCVR has been stagnating for years. Why?
- Price.
Affordable devices ($200~$300) are not there. The cheapest PCVR headset is about ~800$, like a high-end smartphone, and for that money only a couple of quite heavy (~0.8kg) devices available, with rigid display port cables and questionable optical solutions. Modern & lightweight high-end PCVR devices did exist, but their high price (around 2000$ with all requirements covered) isolates them in a tiny niche.
For the absolute majority of PCVR users’ only option is to start toxic relationships with standalone VR headsets from Meta - Quest 3S & 3 ($400-$600). Those can be connected to the PC via cable or even WiFi (with a legendary amount of friction and compression artifacts). Quests are heavy and clumsy, but at least they’re mass produced so the price is lower and they’re, while weighing ~0.55kg, still lighter than those PCVR-first entry models from Pimax or HTC.
- Friction, friction, friction.
Prohibitive prices is not the only problem in the PCVR domain. You simply can’t buy something that is not a Meta Quest in a local hardware store. The only option is to pre-order a PCVR headset and wait - maybe for a few days, maybe for a month or even half of a year.
Default headstraps on Quest’s are poor so you need to replace them. And there are no adjustable diopters - half of the users need to pay for a visit to an optometrist for a prescription, then order lenses, then wait for delivery and, finally, install those.
All devices in the mass market lack face/eye/brows tracking, so the majority of the PCVR users have paralyzed faces in social VR, but this can be softened with extra effort - crafty default animations and audio plug-ins to animate mouth by sounds from your microphone. With extra $100 and extra wait there are options to have at least mouth tracking, bulky and tricky to mount on the headset. Face tracking is not a standard feature even in most expensive devices - high-end PCVR hardware brands don’t care much about foveated rendering or social VR needs and only recently things started to improve - basic eye tracking, reluctantly, started to appear here and there, glitchy and barely usable in social VR.
Hardcore PCVR enthusiasts usually own more than one headset - one for high-end gaming and one for social VR with facial features tracking, but with poor specs and often deprecated by its manufacturer (Quest Pro, Pico 4 Enterprise, HTC Vive focus vision).
- Dated paradigm.
Contemporary PCVR hardware brands are still focused on the old approach back from 2014 - “VR headsets are only good for VR games”. In the beginning those were low resolution devices (~1080x1440 per eye at best) with a door screen effect, so it was logical to focus only on gaming, they were useless for productivity or anything serious. But now this decade-old paradigm does more harm than good, locking the whole VR industry in a vicious circle “no AAA PCVR games market → no point investing into mass production of the affordable PCVR headsets → no AAA PCVR games market”.
In the current state PCVR can’t grow, it’s only capable of providing value for a few millions of VR enthusiasts, who are ready to deal with all kinds of friction and high cost - through their creativity, time, money and community knowledge; a way to fulfill emptiness of human existence.
What can make PCVR much, much more popular?
Let’s start with the obvious: increase display resolution to overreach threshold, that makes it usable in productivity workloads. Then make those high-res headsets lightweight and comfortable for long sessions. Add a decent face+eyebrows+eyes tracking module/option for normal social VR experience for those who need that. Now we’re almost ready to break away from that dated gaming paradigm, just produce those headsets at scale with affordable prices… but something is missing here, this plan still feels hollow.
Maybe that missing thing is that old goal, buried in the sands of time at the dawn of the modern PCVR era, and it’s the right time to dig it out - to replace hundreds of millions of flat screens with PCVR headsets. This is only possible if PCVR can offer real, substantial advantages as a better, affordable alternative for PC desktop monitors, laptop screens and TV panels:
- Inside VR it costs a mouse click to add extra monitors or TVs, of any size and curvature. If a VR headset is using 10 bit HDR OLED panels - all those extra virtual monitors/TVs will be 10 bit HDR OLED, wall-sized if you wish them to.
- Highly personalized VR space for working and relaxing. Holograms, whole locations and all kinds of 3D spatial media, it’s not only about having more virtual flat screens.
- Energy consumption for VR headset is around 5-10W while monitors and TVs are 30W-600W, each. Lower energy bills, less stress on the environment, longer battery life for the laptops.
- Foveated tech stack performance benefits.
That last one, foveated tech stack, can be one of the key features in PCVR. If there is information where eyes are focused now and where is your head located in the virtual space - it opens up many possibilities to cut off a lot of all kinds of redundant work for 3D hardware and raise up quality only where it can be perceived in this specific moment. For example:
- Foveated streaming for textures & meshes, so only 3D objects that are in your focus right now will have top LODs and texture MIPs, so VRAM usage in 3D can drop, making 8GB video cards capable of handling games that previously need 16GB. Lighting fast streaming hardware is needed, though.
- Foveated resolution adjustment. In real life it’s common to look at a 4K TV at a distance, where your eyes physically can’t notice the difference between 1080p and 4K, wasting most of GPU performance on rendering details that cannot be perceived. When playing 3D games on a flat virtual screen inside a VR environment it’s possible to dynamically calculate optimal rendering resolution for that virtual screen, even taking into account angle of looking on this particular part of this virtual screen at a given moment and amount of real pixels from PCVR headset in this area, so no single extra pixel will be rendered.
- Foveated anti-aliasing - apply “godmode” super-sampling AA only to the tiny spot in the focus of the user’s eyes, with a fraction of the performance cost.
Right now the foveated tech stack feels undervalued and undeveloped, providing up to 30% of performance only in VR-native games and doing superb job in foveated streaming, but it’s not used to boost non-VR games on the flat virtual screen in VR. But with all of the above combined, it has the potential to turn user’s RTX 5060 8GB GPU into RTX5080 16GB, or RX9600XT 8GB into RX9700XT 16GB; for free. I have a sneaking suspicion that Valve was smart enough to implement something like that, so the combo of a quite weak Steam Machine with the Steam Frame VR headset will be punching surprisingly high above its weight, when playing 3D games on a flat screen in a virtual space. If this is true it will be interesting to see how far Steam Machine’s weak GPU (equivalent of Radeon RX7600 8GB) can be leveraged by a foveated tech stack. However, knowing Valve’s pricing it can be not worth it even if 50-70% performance gain will be achieved…
PCVR has quite a potential to break free from its stagnation, if it will stop trying to be a VR-gaming-only device and focus on being better alternative for the hundreds of millions users of the flat screens. Creating PCVR headsets is a tough art because of how many technical, financial and biological issues needed to be addressed for those devices to feel good, comfortable and affordable.
[Here is my previous article about situation in standalone VR]