While the video explains the reasons why we don't have as much FOV as many think we should have, and how expensive it is to implement, many people have left comments for potential solutions. 1. A number of people suggested led lights on the outer edges like TV backlights or light bars to simulate a wider FOV. 2. Many mentioned DFR which I mentioned in the video as something that would help a lot, but that has no momentum. 3. Some suggested NVidia should create some kind of DLSS type of compression for the outer edges so they wouldn't take as many distortion profile pixels to render. 4. A few mentioned curved displays. 5. There were even a couple suggesting going back to traditional or Fresnel lenses, but they don't hold up against higher resolution displays and have god rays and other artifacts. Fresnel lenses have physical ridges in them that cause higher ppd image quality to fall apart. Past that there were people who said they still wanted more resolution or the ability to see all the resolution they had. Many of them mentioned using non-displayport headsets like the Quest 3 or Pico 4. Others talked about how important FOV was to them and others said they had other things they wanted more like more resolution, higher frame rates or keeping headsets small and lightweight. In my video "The State of VR and the long game" the fragmentation of the VR market evident by comments here is covered, as well as the lack of profit incentive to add software support for DFR, which would also apply to Nvidia adding a special wide FOV DLSS. I'm not saying these things wouldn't work, but just that there isn't any motivation for them to be implemented. There are many things that are possible and that we would like to have, but unless it fits into a companies long term plans, or there is a profit incentive they won't happen. Much of the DFR support we have for OpenXR was the work of Matthieu Bucchianeri and is provided at no cost with source code on GitHub. github.com/mbucchia However, his work has mostly shown us what is possible in DCS( Digital Combat Systems) and created excitement over what is possible, but there hasn't been any major support picking up this torch and running with it.
It would help with the Compute needed for sure. The trick is figuring out how we can actually create the right incentive so VR titles actually add support for it.
@@markkovalcson7243 it will be supported in nearly everything retroactively on Quest 4 or whatever ends up shipping with eye tracking, but on PCVR? gabe newell is not interested in VR _AT ALL_, there is no incentive from anyone to anyone. you cant have a healthy VR ecosystem on PC without the biggest monopoly and biggest money printing factory in the world, STEAM, barring A LOT of the growing pain costs.
@@markkovalcson7243 The answer is we dont. You make it hardware/ steam integrated which is exactly what the next SteamVR headset will do. This takes the load off of devs while increasing performance and clarity for users.
The very wide field of view really pays off in Sim racing and flying, which helps provide you a more sense of speed when your peripheral view has objects passing by it. This is where a field of view is actually needed.
in a shooter it would be very good for seeing enemies on your side. It would also be better for just feeling comfortable in the headset since it feels less claustrophobic.
Its kinda crazy that my 4080 really struggles to run my Crystal Light, which is only 2880x2880 at only a (claimed) 115° HFOV. It does seem like the limiting factor on VR today isn't lenses or displays, but compute. I'm budgeting for a 5090, but even then, I'm under no expectation that it will take the lid off my PC's ability to run VR as I'd like. My last comment on one of your videos was to say that we're not even in the golden age of VR yet. We are still early adopters. We are buying Blackberrys while dreaming of iPhones. I look forward to watching this technology evolve, but in the mean time, we are just gonna keep throwing thousands of dollars at these companies while we wait for their tech to reach its potential.
The 2880 x 2880 is the panel resolution, once you add in the distortion profiles it's a lot more than that for the Crystal Light. In example, if you set Pimax Play to 1.0 resolution setting (100%) and steamVR or OpenXR to 100% then you get something like 4312 x 5102, at 0.9 in Pimax Play, you get 3880 x 4592 and at 0.8, 3448 x 4080 (all per eye). That's a lot of pixels to push, I use a 7800X3D with a 4090 and set my Crystal Light to 0.8 and leave it there, everything still looks amazing and my computer hardware can run it, at 1.0, my system is not enough to get decent frame rates.
As a flight sim enthusiast for 30 years, it never ends… I have a top of the line rig now, and I’ve resigned I’ll be chasing the dragon forever. I also think hedonic adaption is a factor. My current setup would have literally blown my mind 5 years ago, now I’m wondering what’s next
Higher FOV is better for dynamic VR gaming, higher resolution is better for sitting VR, menus, desktop mirroring, 2D gaming, movies, and avoids headache easier. The choice is obvious for the VR industry, but I still love higher FOV
@friendryan I've found it's better to see your dashboard or the road ahead of you over the higher FOV, racetracks hardly ever have incoming traffic to look out for
No, I don't think they will change the game when it comes to any specific detail but the complete package will take a big step forward in ways others can't. With software. That's where Half life, dynamic foveated rendering etc comes in. Looking at their history with the Index and Steam deck, they make smart decisions. They will increase the FOV and resolution enough for a great experience and find a balance for everything keeping the price and computer requirements somewhat low. The Deckard was one guy's fantasy and not the way Valve thinks. They will work towards the best experience for gamers right now at an affordable price.
@@UnimportantAcc That has to be a part of it. Half life is being teased for real now and that game is for innovating pushing software and hardware forward. No way, that's a flat screen game. A lot of people expecting that will be disappointed. Some of the HMD specs will also disappoint people, just like when the Index was released but that stood the test of time, not because of high specs but because it was just right for the time.
We really need eye tracking for foveated rendering in anything that's going to attempt a wider FOV. I think Meta made good decisions on the right tradeoffs with teh Quest 3. Aside from a few dumb things like port locations, flimsy headstrap attachments, and sub-optimal mic location. They definitely got the display system right.
I thought that FOV was essential for immersion, but after receiving my bigscreen beyond, my mind was completely changed. The beyond had the lowest FOV out of any headset I’ve ever tried, and it was also the most immersive. I have 2000+ hours in vr, but trying the beyond gave me the same wow as when I tried vr for the first time ever
It was a big disappointment to me. Not only small FOV, but I really found it too dim. I would like to see a lot brighter images from headsets as well as bigger FOV. Praying the wide version of the new Pimax Super will actually be usable.
Yup, same here. As a dev I've tried pretty much all HMDs again, but keep a VP1 at home because nothing really felt better than that or worth my money. The MeganeX 8k I tried a couple weeks back changed all that, and reinforces that point you brought up even further. I think now that we got to this PPD density, we can finally start slowly expanding FOV. But not before. Current PPD give us around 4K monitor density, and anything less is just not okay.
@@johnnyalbrecht41 The Varjo is great, absolute top tier if you ask me. But man, having an HMD with about the same fidelity (less FOV but better contrast)... at 1/5th the weight... REALLY makes a good argument (talking the MeganeX 8K here). It basically takes care of that barrier where you don't "feel like" going through the hassle of putting the VR HMD on.
One of the reasons that the Beyond feels more immersive for me is because it tracks with my face perfectly and doesn't move around on me. My Aero moved all over my face in my motion sim breaking immersion in a very noticeable way. The higher contrast and dark blacks also help convey some things in a more immersive believable way. Overall it's the best VR experience I've ever had, but people have different opinions depending on their priorities and sensitivities.
Thank you for the video and I’ve found your channel recently and been enjoying it. I hope FOV improves someday and was really happy to see you raise the topic as few do.
@@Anonnymouse53 They need to just re-release it with better panels and aspheric lenses and I think it would sell very well while the wait continues for the unicorn 12K
Yeah me too, I ordered a pimax crystal super. Hoping it's not too big a difference. I tried the 8kx in potato fov mode and it was fine to me. I think the super should be at least that big according to the reported numbers. The strange thing is the quest 2 fov annoyed me but the pimax 8kx in potato mode didn't when according to reported numbers there should only be like a 4 degree difference.
@markkovalcson7243 That was due to manufactures net yet having a large enough need to make a slightly larger OLED, right? But it could be done now with a traditional screen if they wanted to no? Years ago, there was a special headset that had two Curved screens. Havent seen it since. Seems like that would help fov/optics as well.
@@brettcameratraveler Remember that only 14% of light gets through pancakes, so it takes a very bright display to make that work. He said they haven't got the lenses or displays were he wants them yet. Also remember at at current resolutions, Rendering that wide would bring a 5090 to its knees. Your Demo didn't have tracking and simply played a video loop, so there was no rendering time considerations.
@markkovalcson7243 For sure. I was speaking to the display and lens combo already existing. Seemed plenty bright enough in the demo I tried. As far as rendering is concerned, that's going to be a challenge. Thankfully, the further from the center of our field of view the render is, the far less resolution the render needs to be. With these extreme fovs in mind, a more appropriate implementation of eye tracked fovated rendering is likely to shine and be a fraction of what the 5090 would have had to normally render. Anyway it was very cool to see such a wide fov lens and screen exists and works. Hopefully someone will combine it with some of the other existing solutions out there to achieve what we are looking for.
I share your view on the demand for FOV; I too crave more horizontal pixels! TBH, I'd settle for even 150° for now! Great breakdown on the 'why' though! You've earned my sub!
As a degree pilled FOVcel I agree. The main thing I’ve wanted since quest 1 was more FOV but it’s so hard to find on the market without ridiculous compromises. I play mostly Elite Dangerous and it’s hard to be “immersed” when I feel like I’m looking through binoculars.
On the quest 3 without the facial gasket, with the lenses almost touching my eyes, the field of view is just good enough to not distract me, probably about 110-120 degrees. I think if they could comfortably improve the FOV to 130-150, which seems reasonable to me, we will have something truly special. I don't think I need much more than that.
Fantastic analysis. I want to direct so many ignorant commenters to this video. How many times have we seen ranting and whining in the comments about FoV on every new headset announcement? I would add that a large amount of the onus is on Nvidia, there are so many different approaches that they could have taken to make PCVR a million times better at the GPU/driver level, but never will because the incentive is obviously just not there for them. If Zuckerberg hadn't gone all-in on standalone (which was the beginning of the end, in my opinion) I bet he could have strong-armed Nvidia into engineering some game-changing stuff. Native universal driver-level DFR alone could change everything, but the possibilities don't end there. Imagine a world where dual-GPU SLI systems came back, and each card was dedicated to one of the two panels, each card/panel with its own bespoke DP output, AND -- imagine this -- each card/panel with its own discrete DLSS 4 pipeline running. If that became feasible with native driver-level support to make the rendering pipeline universal, then we would, in 2025, be seeing the Cyberpunk VR mod running at 200+ FPS 4k per eye, fully ray-traced. Granted, with two 5090s in SLI, which I realize is an absurd thing to even say out loud, but that's obviously an extreme example of what could've been possible if the core of the industry (Nvidia) cared about PCVR and invested R&D to develop it.
After other discussions I wish that I emphasized a bit heavier that there was an issue with headsets being based off of smartphone displays. Smartphone displays had no reason to improve in the ways we wanted them to for VR. So a source of cheap large displays along with less expensive Fresnel lenses that could work well with that size display were not viable and were dropped. Meanwhile miniLED displays are getting larger and are at a much higher resolution, so Fresnel's are no longer viable at least for the center of the displays in order to utilize that resolution well. So we had the issue of usable inexpensive large displays and the fact Fresnel lenses are not as viable as the resolutions get higher, fighting against them being sustainable.
I see it the same way, for me the obstacle lies 50% with Nvidia, because the problem of computing power with 2 graphics cards has been overcome and if the VR glasses had 2 video cables I wouldn't even need SLI. At the end of the '90s and after, that wasn't an issue. This is a pure sales strategy that I hold against Nvidia. But the greed of the user not to be satisfied if you don't get everything immediately and then to talk badly prevents the manufacturers from taking smaller steps. EVERYTHING, and possibly even yesterday, is the topic everywhere! As mentioned in the video, if the time is not yet right, let them say this. Then you concentrate more on the color display, then come more pixels and then the optics. Why all three steps at once? An 8kx with the color quality of the Crystal at first would be a good improvement for me. For me, the VR glasses primarily take me away from my desk. And the FOV takes my diving goggles off so that I can see everything through my frameless glasses as if I were in real life. One step at a time would be a wise approach.
To add to Nvidia's onus, their next line of graphics cards seems to be hyper-focused on AI upscaling and frame generation, which might work really well for flat gamers, but I'm highly skeptical it'll be a good fit for VR. Any sort of artifacting or mismatches between each eye's images will be a glaring issue in a VR headset. In terms of pure rasterization it doesn't seem like the 5000 series will be a substantial upgrade from the 4000 series from what I've seen and heard. I hope I'm wrong.
Why wide FOV hasn't been adopted boils down to how some games and game engines haven't been prepared for non-parallel displays. Valve has canted displays in the index, however they factor in that most games will not properly support non-parallel headsets, and this is where parallel projection comes in, which makes games behave properly at the cost of increasing the number of pixels being rendered to the headset. iirc Nvidia and AMD made solutions to this with adding Quad View support, which renders 4 sets (2 parallel and 2 non-parallel) of images in one pass. Only way to make the industry adapt is if we see companies like Meta, Valve, HTC or any member of OpenXR create a guideline for how to make non-parallel rendering behave with no issues
"some games and game engines haven't been prepared for non-parallel displays">that's simply not how that works, that's all handled by the headset "Valve has canted displays in the index" >incorrect honestly the rest of this i more of the first statement having to do with "non parallel displays" and "parallel projection" which is none of that is handled by the game engine but rather openvr,openxr,steamvr wide fov means more pixels and a wider sweet spot in optics required because people have different faces, aswell as wider fov causes motion sickness easier we're just now adopting pancake lenses, and going for micro-oled which all has a higher cost of manufacturing and higher demands of total pixel count, and theirs a drive more mobility in vr for bigger adoption, so methods are being made for boosting stability while adding clarity, and you can't just slap frame generation and upscaling on vr and call it a day
@quadstrike You seem to not understand how wider FOVs work relative to how they're rendered. There are methods you could do to ensure things don't feel extremely distorted or wonky without requiring parallel projection. Again, most game engines are designed to assume VR is parallel, which means the wider FOV with parallel projection disabled will make particles look weird, cause random head flipping issues, etc. Games can implement ways to mitigate these problems, but the purpose of PP is to compensate older VR titles or titles that do not factor non-parallel displays. You cannot have a wider FOV by rendering to a flat display, things will look stretched hence why you should look up things like Nvidia Quad View and the issues Pimax users have had years ago.
@Smash_ter i didn't say distorted, i said focal sweet spot, parallel projection has nothing to do with the reason people dont have wider fov, pimax demonstates this steam vr renders a concave surface thus no distortion if the display is set right. not everyone has their eyes in the same area in their eye sockets, and may even be off canter. but they telling people they can't use wide fov headsets because their biological bone structure, also adapting for perception is even harder on high fov, all the while we're chatting under a video where this guy's opinion is that the index has a wide field of view when the quest 3 is a clear improvement, the distance from the lens that your eyes are the better your FOV. a lot of the chatter I'm seeing makes me scrutinize if half of the people here have used anything recent or spend that much time in headset. probably the types that are sitting and waiting for something to get released that is specifically catered to them, The jump between quest 2 and 3 is very huge, as for high FOV there's several headsets in the works just for this, That's why I know the main complaint is the focal sweet spot.
I think part of the reason is there's a significant chunk of devs live day to day with restricted FOV: They wear glasses 🤓. As a childhood glasses wearer, I hardly ever notice the restricted veiw in VR. It wasn't until last year, when I finally bought a real wide set of glasses instead of the narrow format I prefered as a young adult, that I started to occasionally feel the FOV was in the way. Even then, it was the vertical angles that I noticed, not the width. In terms of immersion, I'd much rather get the weight off my face. The pressure on my face, especially when I'm moving my mouth and eyes, is very distracting and fatiguing.
Love how you broke everything down Mark! It's a really much needed discussion in the VR space that more people need education on -- that VR is literally one of the most difficult technological problems, and the folks working on it are still doing a great job. The mere fact we have something like the Quest 3 at less than 500 dollars at all at this point is astounding to me, and I try to appreciate the slow and long journey to us getting lighter and wider FOV headsets. I feel like we're still a few GPU/ tech generations away from really having headsets that will be a no brainer for "normal" folks outside of enthusiasts to grab. It has to be light, really ease to use, have incredible optics and passthrough and be dummy-proof for it to be consumer grade tech. I also think that having a reliable low-latency Wifi6/7 connection to the headset is essential as well for those looking for high end wireless experiences. All of those things will take for it to become normal for the regular consumer to have.
There have been plenry of momemts in vr where the only missing peice was the FOV, I could easily imagine being IN the location ratger than looking through a scuba mask , and the feeling was palpable, so it is frustraring. We dont have even 130 deg now as standard, or at the very least more news out of Meta on their plans for increasing FOV, I figured this was a given, getting rid of the scuba mask experience, i am willing to have a bigger headset provided it uses a halo style band. The situational awareness gained by wider FOV is huge.
BTW this video isn't suggesting that we don't want a wider FOV. It simply explains the real reasons that we don't have it now and there are multiple reasons. There is currently no way Meta can deliver a stand alone wide FOV headset right now for the reasons their CTO specified.
Yup I agree. I bought a new facemask, amvr if anyones curioius, tbh the nose flap on it itches me weird but maybe I have it on wrong. Anyway, the new facemask lets me get a higher fov because the contour of my face vs the old facemask just wasnt right. I could see light leaking through below my nose before. Now, its a soft plush material that my face sinks into and it forms a dark seal. I can noticeably tell that the fov is higher from switching the facemask personally.
Ive had this thought for a while that headset engineers really need to start thinking outside the box with FOV. Whether it be a mix of slightly better real fov + an ambi-light system or something else idk but Im sure there are some effective creative solutions out there
Minor thing, but Quest 3 and Pico 4 do not have MiniLED backlights, they're just normal LCD panels. Quest Pro does have an LCD with a quantum dot layer and MiniLED backlight though.
I truly wonder if something like Samsung's MicroLED tech like in their "The Wall" TVs will be the ticket to small, customized wide FOV displays. They just showed the new versions at CES and their new seam tech makes it nearly invisible. Shrink that down and you can make a custom panel (or two if a specific shape is necessary) to get the desired FOV
The smallest Wall is 109 inches 4k. That's due to density limits, they would be selling TVs by now if they could. They're nowhere _near_ shrinking it down to the ~5 inches needed for VR. What's wrong with OLED?
I think a major influence is that it's easier to market PPD than FOV. I think you're right that more experienced VR users want FOV, but it's not an attribute that new customers really understand. Especially because there's been so much outright lying from manufacturers about their claimed FOVs (I'm sorry, but individual variances do not cover how large these discrepancies are. Manufacturers routinely claim FOVs that are wider than their drivers even render). FOV is a big confusing mess for most people. But PPD they understand. And especially since FOV can be traded away for more PPD... well, I think that's fundamentally what's happened.
Resolution is easy to sell. We saw this with digital cameras for many years even past the point where smaller pixel sizes were hurting dynamic range, and making the sensors noisier. It took a little while for the market to wisen up. With VR displays once we pass retina ppd, there is no gain. So we are almost at the point where the only point in having more pixels will be to extend the FOV or maybe to hide moire' artifacts.
@@markkovalcson7243 We're actually farther from retina PPD than marketing departments would like us to believe. Especially Pimax's 57 PPD claim for the Crystal Super which is much higher than other manufacturers using the same panels. Even assuming it's only peak PPD, the claim is dubious when you look at all the numbers. But customers don't have a way to measure it themselves, so manufacturers can claim whatever peak PPD they want. A real retina PPD display would need ~60 PPD _average_ across the entire display, not just peak in the very center. And that we can determine just by simple math dividing the resolution by the FOV. And with the new batch of 3840x3552 VR headsets coming out this year, that massive resolution still only amounts to ~35 PPD. So there's still a long way to go there. But I think your comment still stands because of how much it's diminishing returns as we get closer to real retina displays. So redirecting additional resolution to FOV is coming back into vogue.
What are you smoking. No one in the general public understands PPD. Go watch any regular review of VR or AR. They will state the resolution of the displays and maybe the FOV. And they'll just spit out marketing gobblygook like "Up to a 200 inch display." All of that stuff means nothing. PPD and FOV are completely interrelated mathematically. Same panels, larger FOV = lower PPD. People just do not understand that speaking about VR like stationary TV or monitor resolution is not the same.
@@carlosdominguez3108 A TV is PPI just like a display is. PPD is the other side of the lens. You are probably right about the average quest owner. But in the enthusiast ranks PPD is discussed a lot and more people than ever are starting to understand what retina PPD means in terms of a VR headset. A lot of talk about PPD came up with the Varjo Aero, then Crystal, and Somnium VR1, and Super etc.. etc.. PPD is becoming much better known among the people buying at the high end. They may not be main stream, but they are dominant when headset costs are getting that high.
@@markkovalcson7243 Small problem...the demographic you speak of is miniscule in the scheme of VR, and not the direction the industry as a whole is moving. The average random dude or girl who buys a Quest 3 has never heard PPD. But they do understand what stands for. Giving only a resolution of the hardware displays when marketing VR tech means absolutely nothing. That's how Pimax can continually lie and say their headsets are 8K.
It's fascinating that in VR, FOV as experienced can be different from what the numbers suggest: There are other headsets with very similar stats, but the Index feels way more "open", because it feels like one's eyes are sharing the same space, rather than being in individual holes in a mask.
The reason flight simmers wanted more resolution is because earlier VR headsets didn't have the resolution needed to read the instruments in our VR cockpits. Increased FOV comes with increased graphics rendering workload and is hampered by optical and physical limitations of lenses and headsets. I live for multiplayer dogfights in DCS, and I would love to see a wide FOV display that lets me use my peripheral vision, but I would need a much beefier PC/GPU, and the headset would be incredibly expensive to boot. I ordered a VR1 with increased FOV, and at the same time I'm buying a 50xx NVidia card to handle the increased graphics rendering workload. I don't know how much better it will be than my Pimax Crystal, but I'll find out around February or March.
Good Luck! Odds are that you will see substantial improvements!
4 дні тому+3
Right on the money. I have the quest 1,2,3S. Odyssey 2. I had the vive. They all basically have the same FOV but are essentially a decade of devices. Absurd. I’d happily take a headset with the 2,3S resolution and a huge FOV. I don’t care about resolution at all. I want FOV and OLED blacks.
DFR is the biggest no show that would make the biggest difference to what is possible. But there is no way to force software developers to add meaningful support.
Great analysis. It's important to remind ourselves that consumer VR has really only been around for at best a decade. It's still early days. There's still going to be growing pains in the VR industry. Technologies like mico-oled and pancake lenses are absolutely worth it for VR headsets, but we can't expect them to be as cheaply or easily produced as the older fresnel or amoled designs of the past. Personally for me, FOV is very important, but only up to a point. Just like resolution, there's diminishing returns. I think in my perfect dream HMD I'd be perfectly happy with 140-160 HFOV. Right now, that's not feasible though. I'm not willing to trade-off weight, size, and comfort for it. I'm happy enough now with my Pico 4, 104 FOV both vertically (Most headsets suffer with VFOV which I dislike a lot) and horizontal, great binocular overlap, decent resolution, less than 600 grams. I'll definitely be excited to see what comes out of hypervision in the future.
we need both, it's just resolution scales better at the same time, we want headsets to be lesser in size, and more fov currently requires more space for it
Interesting video! I wonder if the solution to ultra-wide FOV isn't a traditional lens/screen solution, but something more like retinal projection combined with eye-tracking and deflecting mirrors? It might be possible to deflect the projected image through a series of mirrors to hit each eye fairly straight on regardless of where the user is looking. I'm sure that would come with a ton of its own issues/complications, but it could also bypass a lot of the issues with the current paradigm of screens/lenses.
I wonder if some sort of bias lighting could be a middleground. People have modded that into headsets, but it might cause extra glare from the light bouncing on your face into the lenses?
really interesting video, thanks for posting this. I find myself really distracted by the looking through a periscope feel that I get from quest 3 and others and looking for a happy medium still. Still really undecided with the options currently out there.
I know a few people who love their 8KX's and a number of people who couldn't make them work. Similarly I loved my Index and know many people who still do, but others didn't like it. It's great that we have so many options to choose from.
Ever since rtx20series we have VRS which can be used to shade outer pixels at reduced resolution. 4x reduction is always available and 16x reduction was already available back then, and there's a bunch of steps in between which can be useful for anisotropic type lens distortions. Not enough GPU bandwidth is not really an excuse, more like lacking sufficient software foundation to make use of it. But focusing ever larger FOV on an eyeball is the REALLY hard part.
The peripheral view doesn't have to be rendered at full resolution. It can probably be rendered at 1/16 maybe even 1/24 resolution. Your peripheral view cannot even see color so that data can be removed as well. All you need is some blurry motion to make it convincing.
There is a Hypervision prototype with two panels per eye with lower resolution panels on the outside. Both panels have smaller distortion panels and with the right software driving it would be a viable solution. Unfortunately there are two issues. If you had a lot less pixels going to the outside of the lens, they would look darker. Unfortunately, the way single lens distortion profiles tend to work requires rending more and more pixels to reach outwards.
I doo miss my pimax 5k plus was running it with the 140 fov and loved it the distorsions was not to bad only on the very side/end of the fov, sure it had alot of issues like housing cracking and really enoying setup everytime you put it on adjusting the software to get a nice sweetspot, but when it was dailed in it was sweet feeling off you having a skii mask on your head instead of benoculars was amazing.
@markkovalcson7243 but a headset having a 100% Binocular overlap doesn't mean it covers 100% of natural human vision overlap right? I mean my understanding is that humans have a 120° overlap so for a VR headset to cover that it needs to have more than 120° of fov, no?
@NowoNvr what the 100% overlap gives you is very comfortable viewing, but you are right that our eyes see more degrees of overlap than any VR headset delivers.
@@markkovalcson7243 thanks for the clarification, I hear it also affects the 3D effect... I'd settle for a comfortable overlap alongside taller vertical fov if we never manage to break out from the current ~ 110° degree that's becoming a standard.
Would really like a little more vertical field of view as well esp for games where you are managing mags and shit on a rig means you don’t have to look down as much
I want more resolution and refresh rate before FOV. It’s more useful and immersive when the image and motion is clear. I even used G2 over my Index due to clarity. Now that I have Beyond I found the perfect headset for me. All I want right now is another beyond with more resolution and slightly more refresh rate. I’m happy to wait for FOV. I want to keep the form factor.
I love my Beyond too which I've had for just over a year now, and it has been my best VR experience to date. New technology is coming and the SuperLight 8K has more resolution and appears to have no issues with persistence. At 90fps it should offer a sharper image and without persistence it should appear a bit faster. It's still not released, and there are still changes being made, but hopefully that will be one step further in the direction you are looking for.
Personally, refresh rate has been most important for me. Minimum 120hz, yet all these companies keep pushing 90hz and insane resolution. Will say, using the Quest 3 with the BoboVR head-strap (*without face gasket so you can push it as close to your eyes) has been the absolute game changer for me. LCD sucks, but hopefully one day we get an OLED solution with pancake lenses!
Curious what your thoughts are on an Ambilight style solution. Using RGB lights at the edge of the display to fill in that black space with at least matching light at the edges. The data is already there and very low cost. There have been some DIY versions of this. Would be similar to Varjo's headset with a super resolution center display.
I think you are at least the 4th person to mention this exact concept in the comments of this video. There is at least one DIY solution for the Valve Index. My feelings are mixed. Maybe I would like it, I think I'd have to see it.
I've been asking this for more than 5 YEARS now! Even the PiMax wasn't a great option from what I've heard.. though I could never get my hands on one to try out.
Probably should wait till I've watched the entire video.. but all the early points you've made are of struggles that designs with large FOV's had on cutting edge hardware from 4-5 years ago. As with all tech like this if the industry kept developing and refining the tech by now would have reached a useable and affordable point. The 'prohibitive' expense will always come in the early days. So long as no one strives for it we will never get it in a consume grade format.
As you could see there are many prototypes by Hypervision, Meta has one and Vrgineers has an ultrawide FOV headset in production. So people are still striving for this. The fact Pimax announced the 12K before it was a feasible headset shows they still hope to do this some day.
Don’t need a larger fov over 110 degrees if they can somehow put a light on the borders that match what you see in game. Kind of like what you see on TV when they blur the edges when playing a cellphone portrait video.
I tried the MeganeX 8K. They have better FOV than the BS. Its around 100-ish horizontal, and slightly more vertical. I'm totally fine butchering FOV in favor of PPD. FOV to me is nice have, but I care MOST about PPD and panel tech before I care about FOV. Also, doing numbers for that same HMD... It is using pretty much 100% of the DP1.4a bandwidth, so like you said, 2.0 is very needed and welcome for future HMDs! Happy to see finally someone that understands that FOV isn't just a free thing that is easy to do. Most people online cry about the FOV without understanding how incredibly hard and flatout not possible to do right with current tech is. Funnily enough, it tends to be the same people that want 120 or 144hz on their HMDs as well, ridiculous.
I've heard that it can be the most optically comfortable headset yet and some people seem to really love it. I've also heard some people estimate it at over 100 H+V. I haven't seen WimFOV numbers to support that yet. We need to get feedback from real customers once it is released. One issue seems to be that the halo strap adjustments required to get the clarity of this headset dialed in are more than some people want or took time to deal with at CES, which is a shame. I hope it does well! I will give MaganeX credit for not releasing fictious FOV, ppd, etc.. numbers. But by not releasing any numbers other than the display resolution and binocular overlap which is impressive at 100%, it has left people wondering.
@@markkovalcson7243 Yeah, there are no WimFOV numbers since we basically get to try it with VRChat in the events they do around here in Tokyo hahah I am very used to my Vive Pro 1 though, and its quite similar to that one (A bit better even, especially the vertical). Once it gets to me in the next few weeks I will be able to A-B test them properly though.
I’m the complete opposite. I’d much rather sacrifice PPD in favor of extreme FOV. I don’t care about resolution nearly as much as I care about FOV. I would love to have a 72-90 Hz refresh rate, at least 150-160° FOV, and I don’t care much about cost as long as the FOV is extremely high. Hell, I’d take $3K for a 160° FOV headset. FOV is the single most important factor for immersion for me
@@anakinlowground5515 After enjoying 5K+ large FOV for almost 6 years it's so hard to go back to toilet roll FOV my PCL has. I'm considering selling it and getting used 8KX, or see if Super 50 PPD version will deliver.
I have Pico 4. For me fov was never on the first place, except games where you compite with other people. What I really want is to have every game rendered in true display resolution 2000x2000 and not 1600x1600
I don’t even need much resolution at the outer edge, just give me something like those Philips ambilight TVs where color LEDs light up the wall. Just so there’s not that black outline like swim goggles.
I hate that companies DON'T use a cord to move as much hardware off the headset as possible. If I could get 2x the performance and quality by carrying a puck I'm all for it.
I’ve been using VR for sim racing for 4 years now and wouldn’t have entered the sim racing arena if it hadn’t been for VR. Reality is that nobody wants a helmet on their head. Costs good be overcome, and newer technologies could help reduce them. You could even have amazingly powerful VR, runs smooth, 144hz refresh, etc etc. people aren’t going to wear helmets. Geeky and uncomfortable and very isolating. VR for the masses will be some sort of glasses that you wear everywhere with limited HUD’s. VR for the geeks like myself will be the big domes.
Personally I definitely think resolution and clarity is the most important thing in vr. Went from a wide fov but lower ppd headset to a high ppd but average fov headset and started to actually enjoy VR! Wide fov is useless if it looks soft and makes you sick. Also manual ipd adjustments can be hard to get right. Add to that the added weight many have. I can't imagine buying a hmd without automatic ipd anymore after getting used to it.
I've seen many people go from wide FOV to sharper headsets with less FOV over the years. It's funny how something someone will scream is a dealbreaker becomes no big deal after they've got a new headset. As far as autoIPD goes, the feature is flawed. Part of the reason is that most people do not have symmetrical eyes. My left eye is a bit closer to my nose and my right eye a bit further away. For some people the difference can be a few mm difference. Auto IPD adjusts both sides the same so for most people it can't align the eyeballs properly. Also most headsets with this feature have a very large eye box which means you don't need the IPD aligned that closely. I could run my Index or Aero which had auto IPD from about 60 - 68mm and the display was still clear. The main difference would be FOV and binocular overlap. But both displays always looked sharp to me.
Funny how different peoples eyes are. My ipd is very wide but almost symmetrical within 0.5mm so auto ipd works for me. I definitely notice when it's off even by a fraction as I need to see the gauges clearly in sims. The previous hmd I had I could not get clear what ever I tried to adjust it to manually. I also use prescription lenses in the hmd. Anyway with a clear view I can use vr for hours when before it was very tiring to even do a short race due to the low ppd.
@@OldNTired68 The differences in people's faces, eyes, and sensitivities can be pretty extreme. I think this is a reason that there is so many highly emotional love/hate comments about almost every headset out there.
Where's your FOV? You can't "Moore's Law" your way out of an optics challenge. The broader consumer market will continually improve display resolution but optics is pretty much a VR only problem.
Clarity > FOV. I like being able to win the eye test game in VRChat. Anyway, FOV is hard, in order to keep clarity you have to push a LOT of pixels, which has a HUGE VRAM and bus bandwidth overhead, it goes up in exponents. Width * height * color depth in bits / 8 bytes. (Pimax Dream Air requires 273 MB JUST for the frame buffer, for one frame. Bus has to 'MOVE' all that data in 1/90th a second too. Double/triple that if your buffer is set to 2 or 3 frames. Doesn't include primary monitors which also impact frame buffer size). To get around that with brute force, you need a good GPU, which reduces the number of potential customers. Offset R&D and manufacturing by raising prices, since you can't scale up due to too few people having the supporting hardware. (This is one of the reasons GPU upscalers and frame generation are becoming so popular, but those technologies aren't really ready for VR, their flaws very apparent on displays 1 inch from the eye) Panels get too big and it makes it VERY hard to make an optical stack, & headset gets clunky. And clunky headsets only really work in sim-pits. It gets hard to 'whip' your head around in fast paced shooters. In order to reduce optical distortion panels HAVE to be smaller, which drives up cost a LOT. Lens tech never really trickled down (makes sence considering cinimatography lens tech never really trickled either) Patent hoarding is stifling innovation. I'm damn glad we're going to pancake and aspheric stacks. Trying to go back to an Index or any Fresnel lens stack is like wiping Vaseline over my eyes. Just a mess of blur. (Also, Quest 3 / 3s is LCD with LED backlight, not true MiniLED. Not too far from Quest Pro's displays, just no FALD or Quantum Dots)
DFR must be added to all PCVR applications, and DLSS and FSR must be better implemented so VR can forgo the 6 to 40 grand for a professional GPU. I don't see the brute-force hardware approach to VR problems ever being affordable for consumers.
Many of us would LOVE it if DFR was a priority, but there is no profit incentive for any of the VR development shops to add this. Where it matters most in complex racing and flight sims, the MSFS development staff have shown no interest and the iRacing development staff have shown no interest. If the larger more profitable titles don't add support the smaller shops will never consider it. With no incentive, there is no momentum.
I'm on the fence. The Quest 3 is the first headset I owned that I feel is visually good enough, but I can still see pixels, and of course blacklevels suck. I can definitely understand we want resolution at least to the point of where the Apple Vision Pro is. After that, I think FOV needs to be the main priority.
My Pimax 8KX has the widest FOV but the clarity is lacking. I just bought the Pimax Crystal Light and the clarity is nice but the FOV is not as wide as the Pimax 8 KX.
The 3S also has a lower resolution display. The HTC Focus Vision also has Fresnel lenses, but they are just releasing earlier tech. Since Fresnel rely on physical grooves in the lens, it's very hard to make them work well with high ppd. That's why aspheric and pancake lenses are being used.
I don't need "ultra wide" 180°+ FoV, 160 sounds mind blowing, and I think if any manufacturer can hit 140° it would likely be the point of diminishing returns/ "Good enough"/ accepted by the majority level... I just want something that improves on the Index without going backwards in other areas, even if just a modest bump in FoV, resolution, lens quality, and a slight decrease in size. Apparently that's too much to ask for from anyone besides PiMax, so I'll just keep waiting. That Sominum VR1 sounds kinda promising.
tbf your human skull doesnt allow for much fov because you can see the ingress of your eyes, the outer bit of your eye socket bone. but the peripheral vision effect is most welcome if they get to it tho
It will be interesting to see what finally comes to market based all their R&D and prototyping! But it still may be many years before we see these in enthusiast headsets.
Great video, very informative, thanks! The low FOV has killed VR for me. I can't buy the new headsets because they've low FOV. I am very sensitive to small FOV.
For hi resolution there was nice test what Varjo did at first prototypes but was unpractical. "flying mirror" for hires display that tracked your eye movement. If I remember right next proto's was one fixed hires display that overlapped normal vive resolution display for "not sweetspot" display. I think they achieved human resolution (pixels / deg) or close to it in sweet spot. That also needed 2 optical links. I think their current "consumer" headset use one display / eye for cheaper optics (No need to combine two displays for one eye) and don't need yearly license. Maybe when we get displays with pix/inc regions we can lower calculating need from gpu and can give better resolution for optics sweet spot.
If UA-cam recommends 30 Mbps to watch 2160p 60FPS videos, why do we need such high bitrate for VR? I don't notice much difference in my quest 3 between 400 Mbps and 80 Mpbs. Genuinely curious.
UA-cam streams HIGHLY compressed video, which means you only get a full frame periodically and then it is mostly only sending the pixels that change. An uncompressed 4K/60Hz signal is 18Gbps. Increase 60H to 90Hz = 27 Gbps Now double that for 2 eyes = 54 Gbps. a 4K signal is only 3240x2160 or about 7Mp However the rendering resolution for most headsets is far above that. For the VR1 it could be about 4 X more pixels. 5600x5600 = 31.4Mp 31.4/7 = 4.48 x 54 Gbps = 242 Gbps But most VR is only 8 bit rather than 12 bit color. so drop that down by 2/3 x 242 Gbps = 161.2 Gbps Now add DSC ( display stream compression ) 3:1 visually lossless compression 161.2 /3 and you are back down to 54 Gbps at 90fps Display Port 1.4 has a maximum data rate of 25.92 Gbps. Display Port 2.0 has a maximum data rate of 77.4 Gbps. I may have missed something but those should be the main bits.
Issue is the degredation of brightness at higher fov. PIMAX tackled this with their 8kx by using comparatively MASSIVE display panels. If you want a compact package and decent fov im not sure the tech is there yet. If it were we would have it.
We don't need 200 degree FOV in a headset. I'd take a true 140 degree and be comfortable at calling it a day there as a compromise between res and FOV. PSVR2 has a slightly wider FOV than the Index, and is also higher res. It's really weird how that headset basically gets no credit (its even ommited from this video) or generally gets shit on, when you will not find a better headset for the money. Quest 3 is great and all, but you have the issue of compression of the image with that. If a headset had 3k per eye and the 140 degrees with native display and a small form factor, I think we're in a area where you've got a fair acceptable standard to go forward from there.
Yes I don't get the focus on crazy wide FOV right now when we struggle so hard with tradeoffs at only 110-120, all I need is 140-150 with additional pixels to fill that increase, and eye tracking / foviated rendering, with maybe 20% increase in pixel density, especially in the first 90 deg - if that is even possible
I'm guessing because I haven't made any of those "yet", but typically a VR video stream would bounce around with the players head and I think many people use some kind of image stabilization which might blur the image. I know some people have made these types of videos sharp, so there is a way to do it properly. Another reason might be that the video stream being shared my be rendered at a lower priority than what the player is seeing. Once again only guesses. Now is when someone who knows the answer can reply with a correction to my erroneous information :)
I have had the star vr one,pimax 8kx,varjo xr3 and many more... i realy hope there comes a good micro oled headset with a great binoculair overlap and great brightness with a decent FOV
Honestly, I think it’s just an expensive aspect to develop for such a small audience that actually cares about it. As an enthusiast, I appreciate a higher FOV, but I would assume most people are just fine with the FOV the Quest 3 offers
Cows have wide pupils because they are prey animals and need to see predators along a narrow stretch on the horizon in focus. Predators have vertical pupils so they can see their prey directly ahead of them clearly at different distances in focus. Humans have round pupils that are balanced in terms of what is in focus in both directions. The key is that this isn't field of view related. This is focal plane related. A good analogy to this would be a Tilt and Shift camera lens that allows you to get an in focus view along a single axis. The lens still sees the exact same area, but things are in focus along a specific section of it.
FOV is difficult to get right optically. Pimax shows you how much of an issue it is. This is because we need to use curved screens on this scale. In addition to this, other market contenders that don't push FOV on their products but instead of absolute res, pushes other companies looking to innovate in FOV to drop that endeavor and compete with the race to the highest res instead of working on FOV development.
much better would to just add microLED’s just outside the peripheral vision that match the color of the edge of the screen.. simply because our FOV isn’t that wide to begin with (would say 120° H and 188°V.. so why have screens that reach 200° + FoV when the display moves with your head anyways.)
Are pancake lenses superior to aspheric ones refarding required render resolution. Do aspheric lenses require much bigger resolutions (for the same panel resolution) due to distorsion correction?
@@davidzx692 I don't have that answer. This is something that Meta and Hypervision should know, but we haven't had any wider FOV pancake lenses in the wild to look at yet.
The advantage of pancake lenses is that they use reflection rather than refraction as their means of bending light. This allows for more aggressive angles and generally higher lens performance. And it produces much less chromatic aberration.
This is the reason I keep my pimax 8xk, nobody else even to this day offers large FOV which besides wireless is probably the most important thing to feeling detached from reality. I can easily get higher pixel density, the market has no more modern solution for FOV to this day.
As far as FOV goes, it’s always puzzled me why no one is looking at wrap around curved panels. Well I say no one, I mean public ally. There’s prob some really interesting prototypes in labs around the world.
You are right. The panels wouldn't be a big problem. That's a good idea. However, no company in the world has the tooling to build the curved lenses these curved panels would have to be paired with. That is a huge problem. Somebody must invest in the high-tech machines to create such lenses first. I'm not sure the consumer VR headset market can absorb the cost for that.
why cant they just have an extra low-resolution panel do the extra fov, like glasses, you have your high detail in the middle and you can see around the edge but it is blurry/low rez
I Want Wide FOV Too! The Pimax 8KX, 5K+, and XTAL 5K and 8K units all gave me eye strain. So Im on the hunt. Im gonna take a chance on the VR1. No way Im going back to Pimax.
It was hit or miss on the Pimax wide FOV headsets. Some had eye strain and headaches, and I know some people who were OK with the optics and loved those headsets. From what I hear the VR1 is better optically and has better binocular overlap than the Pimax wide FOV lenses do. If you can afford it and can pair it with a 5090, you should be very happy.
@@markkovalcson7243 Yes! it was hit or miss for sure, some had eye strain and some didnt. I loved the wide FOV with the Pimax and when it worked, the immersion was fantastic. The 8KX also had an issue where if you had low IPD, the image was blurry. Pimax wasnt addressing the issue and touted it was use error. It was not. I had to heavily modify the 8KX to get rid of the blurriness. When I say modify, I had to disassemble the headset and carve out plastic to allow the lenses to come closer together, not because my IPD is very low, but to allow the lenses to line up properly with my eyes, long story on that but it worked!! . I had a huge write up on the mod at the Pimax forum back in the day and provided pictures and a small tutorial. It was a daunting mod and wouldnt recommend trying it unless your a seasoned electronics or small device repair tech. My mod fixed the slight blurriness experienced with lower IPD users, however the eye strain continued to be and issue, some believe its the canted lens design that XTAL and Pimax implement. Im really hoping that Somnium didnt take that design from the XTAL and Pimax playbook. Its hard to tell from any of the VR1 pics if they have canted displays/lenses. Somnium has a 14 day return window so if it has canted lenses at least I can return it.
Correct. The PS VR2 is a headset adapted for PC use because they were not selling well. Also the 2000x2040 resolution is less than the Reverb G2 that used Fresnel lenses. So as a trajectory, it is near the end of the line. What I was saying was that to handle new higher resolution displays Fresnel doesn't hold up.
@@markkovalcson7243 This is innacurate. Sony always had designs of making the PSVR2 for the PC, and the smoking gun on that is the USBC Displayport. When they were developing the hardware, USBC Virtual Link was thought to be the industry standard going forward, so Sony left it in, and built the port into the PS5. As we all know, Nvidia had other ideas and stopped using it after the 20 series cards. It has nothing to do with the headset "not selling well" as we don't know any information on that because Sony has been a bit tight lipped about everything recently. We know one thing, though: the headset flys off the shelves every time they have a sale on it. Also, fresnel has no bearing on if a display holds up well or not. The optics are all about how light passes through and reaches your eye, and if you have a HDR display, you need to allow more light to pass through and you can't get better than fresnels at doing this, unless you produce a really bright OLED to mitigate the offset of light if you want to use HDR with something like pancake lenses. That would drive the cost up to enterprise levels like the Xtal.
What I don't understand is why headset manufacturers don't at least put some leds around the lenses (in the facial interface) that match the colour of whatever is at that part of the screen, so at least it's not completely black outside the lenses. I think that would already improve the perceived FOV significantly!
While the video explains the reasons why we don't have as much FOV as many think we should have, and how expensive it is to implement, many people have left comments for potential solutions.
1. A number of people suggested led lights on the outer edges like TV backlights or light bars to simulate a wider FOV.
2. Many mentioned DFR which I mentioned in the video as something that would help a lot, but that has no momentum.
3. Some suggested NVidia should create some kind of DLSS type of compression for the outer edges so they wouldn't take as many distortion profile pixels to render.
4. A few mentioned curved displays.
5. There were even a couple suggesting going back to traditional or Fresnel lenses, but they don't hold up against higher resolution displays and have god rays and other artifacts. Fresnel lenses have physical ridges in them that cause higher ppd image quality to fall apart.
Past that there were people who said they still wanted more resolution or the ability to see all the resolution they had. Many of them mentioned using non-displayport headsets like the Quest 3 or Pico 4. Others talked about how important FOV was to them and others said they had other things they wanted more like more resolution, higher frame rates or keeping headsets small and lightweight.
In my video "The State of VR and the long game" the fragmentation of the VR market evident by comments here is covered, as well as the lack of profit incentive to add software support for DFR, which would also apply to Nvidia adding a special wide FOV DLSS. I'm not saying these things wouldn't work, but just that there isn't any motivation for them to be implemented. There are many things that are possible and that we would like to have, but unless it fits into a companies long term plans, or there is a profit incentive they won't happen. Much of the DFR support we have for OpenXR was the work of Matthieu Bucchianeri and is provided at no cost with source code on GitHub. github.com/mbucchia However, his work has mostly shown us what is possible in DCS( Digital Combat Systems) and created excitement over what is possible, but there hasn't been any major support picking up this torch and running with it.
Eye tracking may be the solution to foveated rendering of ultrawide fov VRs.
It would help with the Compute needed for sure. The trick is figuring out how we can actually create the right incentive so VR titles actually add support for it.
@@markkovalcson7243 it will be supported in nearly everything retroactively on Quest 4 or whatever ends up shipping with eye tracking, but on PCVR? gabe newell is not interested in VR _AT ALL_, there is no incentive from anyone to anyone. you cant have a healthy VR ecosystem on PC without the biggest monopoly and biggest money printing factory in the world, STEAM, barring A LOT of the growing pain costs.
@@markkovalcson7243 The answer is we dont. You make it hardware/ steam integrated which is exactly what the next SteamVR headset will do. This takes the load off of devs while increasing performance and clarity for users.
@@markkovalcson7243Sony could have kick-started the incentive, but they decided to not support it on pc for psvr2. Thanks a lot Sony.
I've been waiting for eye-tracked foveated rendering since I first got into VR. The fact that there have been like 2 headsets with it drives me nuts.
The very wide field of view really pays off in Sim racing and flying, which helps provide you a more sense of speed when your peripheral view has objects passing by it. This is where a field of view is actually needed.
in a shooter it would be very good for seeing enemies on your side. It would also be better for just feeling comfortable in the headset since it feels less claustrophobic.
You'd be surprised how immersive resolution is. You can read street signs and see details from other cars on the road.
@@JasonNosajasoNosaj Right, but the resolution is basically already here. You can read the g1000 in MSFS in VR. We're there, we need the FOV now.
Its kinda crazy that my 4080 really struggles to run my Crystal Light, which is only 2880x2880 at only a (claimed) 115° HFOV. It does seem like the limiting factor on VR today isn't lenses or displays, but compute. I'm budgeting for a 5090, but even then, I'm under no expectation that it will take the lid off my PC's ability to run VR as I'd like.
My last comment on one of your videos was to say that we're not even in the golden age of VR yet. We are still early adopters. We are buying Blackberrys while dreaming of iPhones. I look forward to watching this technology evolve, but in the mean time, we are just gonna keep throwing thousands of dollars at these companies while we wait for their tech to reach its potential.
The 2880 x 2880 is the panel resolution, once you add in the distortion profiles it's a lot more than that for the Crystal Light. In example, if you set Pimax Play to 1.0 resolution setting (100%) and steamVR or OpenXR to 100% then you get something like 4312 x 5102, at 0.9 in Pimax Play, you get 3880 x 4592 and at 0.8, 3448 x 4080 (all per eye).
That's a lot of pixels to push, I use a 7800X3D with a 4090 and set my Crystal Light to 0.8 and leave it there, everything still looks amazing and my computer hardware can run it, at 1.0, my system is not enough to get decent frame rates.
Pimax isn't really a native steamvr headset, I'm sure it's their software stack adding extra compute.
thats because you dont know what youre doing
I think the solution here is foveated rendering and eye tracking.
As a flight sim enthusiast for 30 years, it never ends… I have a top of the line rig now, and I’ve resigned I’ll be chasing the dragon forever. I also think hedonic adaption is a factor.
My current setup would have literally blown my mind 5 years ago, now I’m wondering what’s next
Higher FOV is better for dynamic VR gaming, higher resolution is better for sitting VR, menus, desktop mirroring, 2D gaming, movies, and avoids headache easier. The choice is obvious for the VR industry, but I still love higher FOV
No Higher FOV is what is needed. Doesn't matter if you are sitting. I don't want to feel like IM in a tunnel or cage. Just sucks.
@friendryan I've found it's better to see your dashboard or the road ahead of you over the higher FOV, racetracks hardly ever have incoming traffic to look out for
@@Blazzes1001If you cant see incoming traffics with even last gen's resolutions, you need glasses or to turn off dynamic resolution
Doesn't look like Valve's next HMD will be a game-changer when it comes to FOV either
It's literally not possible at this stage
No, I don't think they will change the game when it comes to any specific detail but the complete package will take a big step forward in ways others can't. With software. That's where Half life, dynamic foveated rendering etc comes in.
Looking at their history with the Index and Steam deck, they make smart decisions. They will increase the FOV and resolution enough for a great experience and find a balance for everything keeping the price and computer requirements somewhat low. The Deckard was one guy's fantasy and not the way Valve thinks. They will work towards the best experience for gamers right now at an affordable price.
@@dangoran yes I believe so too - I really hope they do bring foveated rendering (and encoding) further forward
@@UnimportantAcc That has to be a part of it. Half life is being teased for real now and that game is for innovating pushing software and hardware forward. No way, that's a flat screen game. A lot of people expecting that will be disappointed. Some of the HMD specs will also disappoint people, just like when the Index was released but that stood the test of time, not because of high specs but because it was just right for the time.
We really need eye tracking for foveated rendering in anything that's going to attempt a wider FOV. I think Meta made good decisions on the right tradeoffs with teh Quest 3. Aside from a few dumb things like port locations, flimsy headstrap attachments, and sub-optimal mic location. They definitely got the display system right.
Same, I wouldnt take a drop in graphics or resolution over FOV on the Meta 3.
I thought that FOV was essential for immersion, but after receiving my bigscreen beyond, my mind was completely changed. The beyond had the lowest FOV out of any headset I’ve ever tried, and it was also the most immersive. I have 2000+ hours in vr, but trying the beyond gave me the same wow as when I tried vr for the first time ever
It was a big disappointment to me. Not only small FOV, but I really found it too dim. I would like to see a lot brighter images from headsets as well as bigger FOV. Praying the wide version of the new Pimax Super will actually be usable.
Yup, same here. As a dev I've tried pretty much all HMDs again, but keep a VP1 at home because nothing really felt better than that or worth my money.
The MeganeX 8k I tried a couple weeks back changed all that, and reinforces that point you brought up even further.
I think now that we got to this PPD density, we can finally start slowly expanding FOV. But not before. Current PPD give us around 4K monitor density, and anything less is just not okay.
Just happy with my varjo x4 steam edition for now
@@johnnyalbrecht41 The Varjo is great, absolute top tier if you ask me.
But man, having an HMD with about the same fidelity (less FOV but better contrast)... at 1/5th the weight... REALLY makes a good argument (talking the MeganeX 8K here).
It basically takes care of that barrier where you don't "feel like" going through the hassle of putting the VR HMD on.
One of the reasons that the Beyond feels more immersive for me is because it tracks with my face perfectly and doesn't move around on me. My Aero moved all over my face in my motion sim breaking immersion in a very noticeable way. The higher contrast and dark blacks also help convey some things in a more immersive believable way.
Overall it's the best VR experience I've ever had, but people have different opinions depending on their priorities and sensitivities.
Thank you for the video and I’ve found your channel recently and been enjoying it. I hope FOV improves someday and was really happy to see you raise the topic as few do.
Still using Pimax 8KX (160 HFOV)
Enjoy!
Same! As nice as OLED screens are, it's difficult to go back to
Same. It's not perfect. But the fov makes it all worth it
@@Anonnymouse53 They need to just re-release it with better panels and aspheric lenses and I think it would sell very well while the wait continues for the unicorn 12K
Yeah me too, I ordered a pimax crystal super. Hoping it's not too big a difference. I tried the 8kx in potato fov mode and it was fine to me. I think the super should be at least that big according to the reported numbers.
The strange thing is the quest 2 fov annoyed me but the pimax 8kx in potato mode didn't when according to reported numbers there should only be like a 4 degree difference.
I demoed a very compact and thin prototype 240 degree fov Hypervision headset at AWE last year. It's possible.
It's possible, but Artur said it was still 4+ years away. He left a comment under this video about this.
@markkovalcson7243 That was due to manufactures net yet having a large enough need to make a slightly larger OLED, right? But it could be done now with a traditional screen if they wanted to no?
Years ago, there was a special headset that had two Curved screens. Havent seen it since. Seems like that would help fov/optics as well.
@@brettcameratraveler Remember that only 14% of light gets through pancakes, so it takes a very bright display to make that work. He said they haven't got the lenses or displays were he wants them yet.
Also remember at at current resolutions, Rendering that wide would bring a 5090 to its knees.
Your Demo didn't have tracking and simply played a video loop, so there was no rendering time considerations.
@markkovalcson7243 For sure. I was speaking to the display and lens combo already existing. Seemed plenty bright enough in the demo I tried. As far as rendering is concerned, that's going to be a challenge. Thankfully, the further from the center of our field of view the render is, the far less resolution the render needs to be. With these extreme fovs in mind, a more appropriate implementation of eye tracked fovated rendering is likely to shine and be a fraction of what the 5090 would have had to normally render. Anyway it was very cool to see such a wide fov lens and screen exists and works. Hopefully someone will combine it with some of the other existing solutions out there to achieve what we are looking for.
I share your view on the demand for FOV; I too crave more horizontal pixels!
TBH, I'd settle for even 150° for now!
Great breakdown on the 'why' though! You've earned my sub!
As a degree pilled FOVcel I agree. The main thing I’ve wanted since quest 1 was more FOV but it’s so hard to find on the market without ridiculous compromises. I play mostly Elite Dangerous and it’s hard to be “immersed” when I feel like I’m looking through binoculars.
On the quest 3 without the facial gasket, with the lenses almost touching my eyes, the field of view is just good enough to not distract me, probably about 110-120 degrees. I think if they could comfortably improve the FOV to 130-150, which seems reasonable to me, we will have something truly special. I don't think I need much more than that.
Fantastic analysis. I want to direct so many ignorant commenters to this video. How many times have we seen ranting and whining in the comments about FoV on every new headset announcement? I would add that a large amount of the onus is on Nvidia, there are so many different approaches that they could have taken to make PCVR a million times better at the GPU/driver level, but never will because the incentive is obviously just not there for them. If Zuckerberg hadn't gone all-in on standalone (which was the beginning of the end, in my opinion) I bet he could have strong-armed Nvidia into engineering some game-changing stuff. Native universal driver-level DFR alone could change everything, but the possibilities don't end there. Imagine a world where dual-GPU SLI systems came back, and each card was dedicated to one of the two panels, each card/panel with its own bespoke DP output, AND -- imagine this -- each card/panel with its own discrete DLSS 4 pipeline running. If that became feasible with native driver-level support to make the rendering pipeline universal, then we would, in 2025, be seeing the Cyberpunk VR mod running at 200+ FPS 4k per eye, fully ray-traced. Granted, with two 5090s in SLI, which I realize is an absurd thing to even say out loud, but that's obviously an extreme example of what could've been possible if the core of the industry (Nvidia) cared about PCVR and invested R&D to develop it.
After other discussions I wish that I emphasized a bit heavier that there was an issue with headsets being based off of smartphone displays. Smartphone displays had no reason to improve in the ways we wanted them to for VR. So a source of cheap large displays along with less expensive Fresnel lenses that could work well with that size display were not viable and were dropped. Meanwhile miniLED displays are getting larger and are at a much higher resolution, so Fresnel's are no longer viable at least for the center of the displays in order to utilize that resolution well. So we had the issue of usable inexpensive large displays and the fact Fresnel lenses are not as viable as the resolutions get higher, fighting against them being sustainable.
I see it the same way, for me the obstacle lies 50% with Nvidia, because the problem of computing power with 2 graphics cards has been overcome and if the VR glasses had 2 video cables I wouldn't even need SLI. At the end of the '90s and after, that wasn't an issue. This is a pure sales strategy that I hold against Nvidia.
But the greed of the user not to be satisfied if you don't get everything immediately and then to talk badly prevents the manufacturers from taking smaller steps.
EVERYTHING, and possibly even yesterday, is the topic everywhere! As mentioned in the video, if the time is not yet right, let them say this.
Then you concentrate more on the color display, then come more pixels and then the optics. Why all three steps at once?
An 8kx with the color quality of the Crystal at first would be a good improvement for me.
For me, the VR glasses primarily take me away from my desk. And the FOV takes my diving goggles off so that I can see everything through my frameless glasses as if I were in real life. One step at a time would be a wise approach.
To add to Nvidia's onus, their next line of graphics cards seems to be hyper-focused on AI upscaling and frame generation, which might work really well for flat gamers, but I'm highly skeptical it'll be a good fit for VR. Any sort of artifacting or mismatches between each eye's images will be a glaring issue in a VR headset. In terms of pure rasterization it doesn't seem like the 5000 series will be a substantial upgrade from the 4000 series from what I've seen and heard.
I hope I'm wrong.
Why wide FOV hasn't been adopted boils down to how some games and game engines haven't been prepared for non-parallel displays. Valve has canted displays in the index, however they factor in that most games will not properly support non-parallel headsets, and this is where parallel projection comes in, which makes games behave properly at the cost of increasing the number of pixels being rendered to the headset. iirc Nvidia and AMD made solutions to this with adding Quad View support, which renders 4 sets (2 parallel and 2 non-parallel) of images in one pass. Only way to make the industry adapt is if we see companies like Meta, Valve, HTC or any member of OpenXR create a guideline for how to make non-parallel rendering behave with no issues
"some games and game engines haven't been prepared for non-parallel displays">that's simply not how that works, that's all handled by the headset
"Valve has canted displays in the index" >incorrect
honestly the rest of this i more of the first statement having to do with "non parallel displays" and "parallel projection" which is none of that is handled by the game engine but rather openvr,openxr,steamvr
wide fov means more pixels and a wider sweet spot in optics required because people have different faces, aswell as wider fov causes motion sickness easier
we're just now adopting pancake lenses, and going for micro-oled which all has a higher cost of manufacturing and higher demands of total pixel count, and theirs a drive more mobility in vr for bigger adoption, so methods are being made for boosting stability while adding clarity, and you can't just slap frame generation and upscaling on vr and call it a day
@quadstrike You seem to not understand how wider FOVs work relative to how they're rendered. There are methods you could do to ensure things don't feel extremely distorted or wonky without requiring parallel projection. Again, most game engines are designed to assume VR is parallel, which means the wider FOV with parallel projection disabled will make particles look weird, cause random head flipping issues, etc. Games can implement ways to mitigate these problems, but the purpose of PP is to compensate older VR titles or titles that do not factor non-parallel displays. You cannot have a wider FOV by rendering to a flat display, things will look stretched hence why you should look up things like Nvidia Quad View and the issues Pimax users have had years ago.
@Smash_ter i didn't say distorted, i said focal sweet spot, parallel projection has nothing to do with the reason people dont have wider fov, pimax demonstates this
steam vr renders a concave surface thus no distortion if the display is set right. not everyone has their eyes in the same area in their eye sockets, and may even be off canter. but they telling people they can't use wide fov headsets because their biological bone structure, also adapting for perception is even harder on high fov, all the while we're chatting under a video where this guy's opinion is that the index has a wide field of view when the quest 3 is a clear improvement, the distance from the lens that your eyes are the better your FOV.
a lot of the chatter I'm seeing makes me scrutinize if half of the people here have used anything recent or spend that much time in headset. probably the types that are sitting and waiting for something to get released that is specifically catered to them, The jump between quest 2 and 3 is very huge, as for high FOV there's several headsets in the works just for this, That's why I know the main complaint is the focal sweet spot.
I think part of the reason is there's a significant chunk of devs live day to day with restricted FOV: They wear glasses 🤓. As a childhood glasses wearer, I hardly ever notice the restricted veiw in VR.
It wasn't until last year, when I finally bought a real wide set of glasses instead of the narrow format I prefered as a young adult, that I started to occasionally feel the FOV was in the way. Even then, it was the vertical angles that I noticed, not the width.
In terms of immersion, I'd much rather get the weight off my face. The pressure on my face, especially when I'm moving my mouth and eyes, is very distracting and fatiguing.
Love how you broke everything down Mark!
It's a really much needed discussion in the VR space that more people need education on -- that VR is literally one of the most difficult technological problems, and the folks working on it are still doing a great job.
The mere fact we have something like the Quest 3 at less than 500 dollars at all at this point is astounding to me, and I try to appreciate the slow and long journey to us getting lighter and wider FOV headsets.
I feel like we're still a few GPU/ tech generations away from really having headsets that will be a no brainer for "normal" folks outside of enthusiasts to grab.
It has to be light, really ease to use, have incredible optics and passthrough and be dummy-proof for it to be consumer grade tech.
I also think that having a reliable low-latency Wifi6/7 connection to the headset is essential as well for those looking for high end wireless experiences.
All of those things will take for it to become normal for the regular consumer to have.
There have been plenry of momemts in vr where the only missing peice was the FOV, I could easily imagine being IN the location ratger than looking through a scuba mask , and the feeling was palpable, so it is frustraring. We dont have even 130 deg now as standard, or at the very least more news out of Meta on their plans for increasing FOV, I figured this was a given, getting rid of the scuba mask experience, i am willing to have a bigger headset provided it uses a halo style band. The situational awareness gained by wider FOV is huge.
BTW this video isn't suggesting that we don't want a wider FOV. It simply explains the real reasons that we don't have it now and there are multiple reasons. There is currently no way Meta can deliver a stand alone wide FOV headset right now for the reasons their CTO specified.
Yup I agree. I bought a new facemask, amvr if anyones curioius, tbh the nose flap on it itches me weird but maybe I have it on wrong. Anyway, the new facemask lets me get a higher fov because the contour of my face vs the old facemask just wasnt right. I could see light leaking through below my nose before. Now, its a soft plush material that my face sinks into and it forms a dark seal. I can noticeably tell that the fov is higher from switching the facemask personally.
Ive had this thought for a while that headset engineers really need to start thinking outside the box with FOV. Whether it be a mix of slightly better real fov + an ambi-light system or something else idk but Im sure there are some effective creative solutions out there
There was someone who attempted an ambilight system and it seemed to work great, but don’t know if there’s happened anything on that front recently
i still use my index and want to update, but just never see anything with a higher fov so I don't upgrade.
You aren't alone in that.
Minor thing, but Quest 3 and Pico 4 do not have MiniLED backlights, they're just normal LCD panels.
Quest Pro does have an LCD with a quantum dot layer and MiniLED backlight though.
Thanks! I'll make a note of that.
I truly wonder if something like Samsung's MicroLED tech like in their "The Wall" TVs will be the ticket to small, customized wide FOV displays. They just showed the new versions at CES and their new seam tech makes it nearly invisible. Shrink that down and you can make a custom panel (or two if a specific shape is necessary) to get the desired FOV
It's apparently harder to use MicroLEDs for something like VR, which is why MicroOLED is being used in the industry
@@Smash_ter Apparently based on what? When you say stuff like that you should link to evidence.
The smallest Wall is 109 inches 4k. That's due to density limits, they would be selling TVs by now if they could. They're nowhere _near_ shrinking it down to the ~5 inches needed for VR. What's wrong with OLED?
@@carlosdominguez3108 you know posting links puts my replies on a spam filter. But look it up on Google with relation to the AVP
I agreed , FOV really adds the the immersion and its what long term VR users need to feel like were moving forwards
I think a major influence is that it's easier to market PPD than FOV. I think you're right that more experienced VR users want FOV, but it's not an attribute that new customers really understand. Especially because there's been so much outright lying from manufacturers about their claimed FOVs (I'm sorry, but individual variances do not cover how large these discrepancies are. Manufacturers routinely claim FOVs that are wider than their drivers even render).
FOV is a big confusing mess for most people. But PPD they understand.
And especially since FOV can be traded away for more PPD... well, I think that's fundamentally what's happened.
Resolution is easy to sell. We saw this with digital cameras for many years even past the point where smaller pixel sizes were hurting dynamic range, and making the sensors noisier. It took a little while for the market to wisen up.
With VR displays once we pass retina ppd, there is no gain. So we are almost at the point where the only point in having more pixels will be to extend the FOV or maybe to hide moire' artifacts.
@@markkovalcson7243 We're actually farther from retina PPD than marketing departments would like us to believe. Especially Pimax's 57 PPD claim for the Crystal Super which is much higher than other manufacturers using the same panels. Even assuming it's only peak PPD, the claim is dubious when you look at all the numbers. But customers don't have a way to measure it themselves, so manufacturers can claim whatever peak PPD they want.
A real retina PPD display would need ~60 PPD _average_ across the entire display, not just peak in the very center. And that we can determine just by simple math dividing the resolution by the FOV. And with the new batch of 3840x3552 VR headsets coming out this year, that massive resolution still only amounts to ~35 PPD.
So there's still a long way to go there. But I think your comment still stands because of how much it's diminishing returns as we get closer to real retina displays. So redirecting additional resolution to FOV is coming back into vogue.
What are you smoking. No one in the general public understands PPD. Go watch any regular review of VR or AR. They will state the resolution of the displays and maybe the FOV. And they'll just spit out marketing gobblygook like "Up to a 200 inch display." All of that stuff means nothing. PPD and FOV are completely interrelated mathematically. Same panels, larger FOV = lower PPD.
People just do not understand that speaking about VR like stationary TV or monitor resolution is not the same.
@@carlosdominguez3108 A TV is PPI just like a display is. PPD is the other side of the lens. You are probably right about the average quest owner. But in the enthusiast ranks PPD is discussed a lot and more people than ever are starting to understand what retina PPD means in terms of a VR headset. A lot of talk about PPD came up with the Varjo Aero, then Crystal, and Somnium VR1, and Super etc.. etc.. PPD is becoming much better known among the people buying at the high end. They may not be main stream, but they are dominant when headset costs are getting that high.
@@markkovalcson7243 Small problem...the demographic you speak of is miniscule in the scheme of VR, and not the direction the industry as a whole is moving. The average random dude or girl who buys a Quest 3 has never heard PPD. But they do understand what stands for. Giving only a resolution of the hardware displays when marketing VR tech means absolutely nothing. That's how Pimax can continually lie and say their headsets are 8K.
Mark, awesome and really interesting video!
Thank you!
Have you not tried the PIMAX 8K or 8KX? A lot wider than valve index.
I have not and you are absolutely correct.
Very thought provoking video, thanks for publishing.
It's fascinating that in VR, FOV as experienced can be different from what the numbers suggest: There are other headsets with very similar stats, but the Index feels way more "open", because it feels like one's eyes are sharing the same space, rather than being in individual holes in a mask.
The reason flight simmers wanted more resolution is because earlier VR headsets didn't have the resolution needed to read the instruments in our VR cockpits. Increased FOV comes with increased graphics rendering workload and is hampered by optical and physical limitations of lenses and headsets. I live for multiplayer dogfights in DCS, and I would love to see a wide FOV display that lets me use my peripheral vision, but I would need a much beefier PC/GPU, and the headset would be incredibly expensive to boot. I ordered a VR1 with increased FOV, and at the same time I'm buying a 50xx NVidia card to handle the increased graphics rendering workload. I don't know how much better it will be than my Pimax Crystal, but I'll find out around February or March.
Good Luck! Odds are that you will see substantial improvements!
Right on the money. I have the quest 1,2,3S. Odyssey 2. I had the vive. They all basically have the same FOV but are essentially a decade of devices. Absurd.
I’d happily take a headset with the 2,3S resolution and a huge FOV. I don’t care about resolution at all. I want FOV and OLED blacks.
Maybe fovaeted rendering could help the rendering issue.
DFR is the biggest no show that would make the biggest difference to what is possible. But there is no way to force software developers to add meaningful support.
Great analysis. It's important to remind ourselves that consumer VR has really only been around for at best a decade. It's still early days. There's still going to be growing pains in the VR industry. Technologies like mico-oled and pancake lenses are absolutely worth it for VR headsets, but we can't expect them to be as cheaply or easily produced as the older fresnel or amoled designs of the past.
Personally for me, FOV is very important, but only up to a point. Just like resolution, there's diminishing returns. I think in my perfect dream HMD I'd be perfectly happy with 140-160 HFOV. Right now, that's not feasible though. I'm not willing to trade-off weight, size, and comfort for it. I'm happy enough now with my Pico 4, 104 FOV both vertically (Most headsets suffer with VFOV which I dislike a lot) and horizontal, great binocular overlap, decent resolution, less than 600 grams. I'll definitely be excited to see what comes out of hypervision in the future.
Headset size and comfort is the problem companies just wanna play it safe
You have to admit if a headset isn't comfortable, no one will want to wear it.
I have really good peripheral vision, and it makes the limited FOV even that much more of an issue. I have to be incredibly engaged to not notice it.
we need both, it's just resolution scales better
at the same time, we want headsets to be lesser in size, and more fov currently requires more space for it
Interesting video!
I wonder if the solution to ultra-wide FOV isn't a traditional lens/screen solution, but something more like retinal projection combined with eye-tracking and deflecting mirrors? It might be possible to deflect the projected image through a series of mirrors to hit each eye fairly straight on regardless of where the user is looking. I'm sure that would come with a ton of its own issues/complications, but it could also bypass a lot of the issues with the current paradigm of screens/lenses.
I always thoroughly enoy your videos. Thank you so much.
I wonder if some sort of bias lighting could be a middleground.
People have modded that into headsets, but it might cause extra glare from the light bouncing on your face into the lenses?
I think I need a running counter for this suggestion.
really interesting video, thanks for posting this. I find myself really distracted by the looking through a periscope feel that I get from quest 3 and others and looking for a happy medium still. Still really undecided with the options currently out there.
I had the 8k x and the valve index didn't even come close. There have been wide FOV for more than 6 years. It does add it's challenges
I know a few people who love their 8KX's and a number of people who couldn't make them work. Similarly I loved my Index and know many people who still do, but others didn't like it. It's great that we have so many options to choose from.
@markkovalcson7243 agree, they all fill a different niche. I love my Crystal OG but miss the FOV
Ever since rtx20series we have VRS which can be used to shade outer pixels at reduced resolution. 4x reduction is always available and 16x reduction was already available back then, and there's a bunch of steps in between which can be useful for anisotropic type lens distortions. Not enough GPU bandwidth is not really an excuse, more like lacking sufficient software foundation to make use of it.
But focusing ever larger FOV on an eyeball is the REALLY hard part.
The peripheral view doesn't have to be rendered at full resolution. It can probably be rendered at 1/16 maybe even 1/24 resolution. Your peripheral view cannot even see color so that data can be removed as well. All you need is some blurry motion to make it convincing.
There is a Hypervision prototype with two panels per eye with lower resolution panels on the outside. Both panels have smaller distortion panels and with the right software driving it would be a viable solution.
Unfortunately there are two issues. If you had a lot less pixels going to the outside of the lens, they would look darker. Unfortunately, the way single lens distortion profiles tend to work requires rending more and more pixels to reach outwards.
The issue is that you can turn your eyes in your head.
I doo miss my pimax 5k plus was running it with the 140 fov and loved it the distorsions was not to bad only on the very side/end of the fov, sure it had alot of issues like housing cracking and really enoying setup everytime you put it on adjusting the software to get a nice sweetspot, but when it was dailed in it was sweet feeling off you having a skii mask on your head instead of benoculars was amazing.
Justice for Binocular fov!
The MaganeX SuperLight 8K is currently the only headset with 100% binocular overlap, but the FOV isn't massive.
@markkovalcson7243 but a headset having a 100% Binocular overlap doesn't mean it covers 100% of natural human vision overlap right? I mean my understanding is that humans have a 120° overlap so for a VR headset to cover that it needs to have more than 120° of fov, no?
@NowoNvr what the 100% overlap gives you is very comfortable viewing, but you are right that our eyes see more degrees of overlap than any VR headset delivers.
@@markkovalcson7243 thanks for the clarification, I hear it also affects the 3D effect... I'd settle for a comfortable overlap alongside taller vertical fov if we never manage to break out from the current ~ 110° degree that's becoming a standard.
Would really like a little more vertical field of view as well esp for games where you are managing mags and shit on a rig means you don’t have to look down as much
Absolutely! I'd love to have 120V FOV.
I want more resolution and refresh rate before FOV. It’s more useful and immersive when the image and motion is clear. I even used G2 over my Index due to clarity. Now that I have Beyond I found the perfect headset for me. All I want right now is another beyond with more resolution and slightly more refresh rate. I’m happy to wait for FOV. I want to keep the form factor.
I love my Beyond too which I've had for just over a year now, and it has been my best VR experience to date. New technology is coming and the SuperLight 8K has more resolution and appears to have no issues with persistence. At 90fps it should offer a sharper image and without persistence it should appear a bit faster. It's still not released, and there are still changes being made, but hopefully that will be one step further in the direction you are looking for.
Personally, refresh rate has been most important for me. Minimum 120hz, yet all these companies keep pushing 90hz and insane resolution.
Will say, using the Quest 3 with the BoboVR head-strap (*without face gasket so you can push it as close to your eyes) has been the absolute game changer for me. LCD sucks, but hopefully one day we get an OLED solution with pancake lenses!
Curious what your thoughts are on an Ambilight style solution. Using RGB lights at the edge of the display to fill in that black space with at least matching light at the edges. The data is already there and very low cost. There have been some DIY versions of this.
Would be similar to Varjo's headset with a super resolution center display.
I think you are at least the 4th person to mention this exact concept in the comments of this video.
There is at least one DIY solution for the Valve Index.
My feelings are mixed. Maybe I would like it, I think I'd have to see it.
I've been asking this for more than 5 YEARS now! Even the PiMax wasn't a great option from what I've heard.. though I could never get my hands on one to try out.
Probably should wait till I've watched the entire video.. but all the early points you've made are of struggles that designs with large FOV's had on cutting edge hardware from 4-5 years ago. As with all tech like this if the industry kept developing and refining the tech by now would have reached a useable and affordable point. The 'prohibitive' expense will always come in the early days. So long as no one strives for it we will never get it in a consume grade format.
As you could see there are many prototypes by Hypervision, Meta has one and Vrgineers has an ultrawide FOV headset in production. So people are still striving for this. The fact Pimax announced the 12K before it was a feasible headset shows they still hope to do this some day.
Don’t need a larger fov over 110 degrees if they can somehow put a light on the borders that match what you see in game. Kind of like what you see on TV when they blur the edges when playing a cellphone portrait video.
Definitely need a running counter for this suggestion.
another great video as always
I tried the MeganeX 8K. They have better FOV than the BS. Its around 100-ish horizontal, and slightly more vertical.
I'm totally fine butchering FOV in favor of PPD. FOV to me is nice have, but I care MOST about PPD and panel tech before I care about FOV.
Also, doing numbers for that same HMD... It is using pretty much 100% of the DP1.4a bandwidth, so like you said, 2.0 is very needed and welcome for future HMDs!
Happy to see finally someone that understands that FOV isn't just a free thing that is easy to do. Most people online cry about the FOV without understanding how incredibly hard and flatout not possible to do right with current tech is. Funnily enough, it tends to be the same people that want 120 or 144hz on their HMDs as well, ridiculous.
I've heard that it can be the most optically comfortable headset yet and some people seem to really love it. I've also heard some people estimate it at over 100 H+V. I haven't seen WimFOV numbers to support that yet. We need to get feedback from real customers once it is released. One issue seems to be that the halo strap adjustments required to get the clarity of this headset dialed in are more than some people want or took time to deal with at CES, which is a shame. I hope it does well! I will give MaganeX credit for not releasing fictious FOV, ppd, etc.. numbers. But by not releasing any numbers other than the display resolution and binocular overlap which is impressive at 100%, it has left people wondering.
@@markkovalcson7243 Yeah, there are no WimFOV numbers since we basically get to try it with VRChat in the events they do around here in Tokyo hahah
I am very used to my Vive Pro 1 though, and its quite similar to that one (A bit better even, especially the vertical).
Once it gets to me in the next few weeks I will be able to A-B test them properly though.
I’m the complete opposite. I’d much rather sacrifice PPD in favor of extreme FOV. I don’t care about resolution nearly as much as I care about FOV. I would love to have a 72-90 Hz refresh rate, at least 150-160° FOV, and I don’t care much about cost as long as the FOV is extremely high. Hell, I’d take $3K for a 160° FOV headset. FOV is the single most important factor for immersion for me
@@anakinlowground5515 After enjoying 5K+ large FOV for almost 6 years it's so hard to go back to toilet roll FOV my PCL has. I'm considering selling it and getting used 8KX, or see if Super 50 PPD version will deliver.
I have Pico 4. For me fov was never on the first place, except games where you compite with other people. What I really want is to have every game rendered in true display resolution 2000x2000 and not 1600x1600
Wonderful video. I love this commentary style
I don’t even need much resolution at the outer edge, just give me something like those Philips ambilight TVs where color LEDs light up the wall.
Just so there’s not that black outline like swim goggles.
5th or 6th time this has been suggested in the comments.
Would it be possible to implement technology like dlss to vr games for better fps and potentially being a solution for rendering ultra wide for vr?
Extremely informative video!!
I hate that companies DON'T use a cord to move as much hardware off the headset as possible. If I could get 2x the performance and quality by carrying a puck I'm all for it.
I’ve been using VR for sim racing for 4 years now and wouldn’t have entered the sim racing arena if it hadn’t been for VR. Reality is that nobody wants a helmet on their head. Costs good be overcome, and newer technologies could help reduce them. You could even have amazingly powerful VR, runs smooth, 144hz refresh, etc etc. people aren’t going to wear helmets. Geeky and uncomfortable and very isolating. VR for the masses will be some sort of glasses that you wear everywhere with limited HUD’s. VR for the geeks like myself will be the big domes.
Personally I definitely think resolution and clarity is the most important thing in vr. Went from a wide fov but lower ppd headset to a high ppd but average fov headset and started to actually enjoy VR! Wide fov is useless if it looks soft and makes you sick. Also manual ipd adjustments can be hard to get right. Add to that the added weight many have.
I can't imagine buying a hmd without automatic ipd anymore after getting used to it.
I've seen many people go from wide FOV to sharper headsets with less FOV over the years. It's funny how something someone will scream is a dealbreaker becomes no big deal after they've got a new headset.
As far as autoIPD goes, the feature is flawed. Part of the reason is that most people do not have symmetrical eyes. My left eye is a bit closer to my nose and my right eye a bit further away. For some people the difference can be a few mm difference. Auto IPD adjusts both sides the same so for most people it can't align the eyeballs properly. Also most headsets with this feature have a very large eye box which means you don't need the IPD aligned that closely. I could run my Index or Aero which had auto IPD from about 60 - 68mm and the display was still clear. The main difference would be FOV and binocular overlap. But both displays always looked sharp to me.
Funny how different peoples eyes are. My ipd is very wide but almost symmetrical within 0.5mm so auto ipd works for me. I definitely notice when it's off even by a fraction as I need to see the gauges clearly in sims. The previous hmd I had I could not get clear what ever I tried to adjust it to manually. I also use prescription lenses in the hmd. Anyway with a clear view I can use vr for hours when before it was very tiring to even do a short race due to the low ppd.
@@OldNTired68 The differences in people's faces, eyes, and sensitivities can be pretty extreme. I think this is a reason that there is so many highly emotional love/hate comments about almost every headset out there.
Interesting and useful video! Thank you very much!
Another great video Mark. Keep them coming.
Where's your FOV? You can't "Moore's Law" your way out of an optics challenge. The broader consumer market will continually improve display resolution but optics is pretty much a VR only problem.
Clarity > FOV. I like being able to win the eye test game in VRChat.
Anyway, FOV is hard, in order to keep clarity you have to push a LOT of pixels, which has a HUGE VRAM and bus bandwidth overhead, it goes up in exponents.
Width * height * color depth in bits / 8 bytes. (Pimax Dream Air requires 273 MB JUST for the frame buffer, for one frame. Bus has to 'MOVE' all that data in 1/90th a second too. Double/triple that if your buffer is set to 2 or 3 frames. Doesn't include primary monitors which also impact frame buffer size).
To get around that with brute force, you need a good GPU, which reduces the number of potential customers. Offset R&D and manufacturing by raising prices, since you can't scale up due to too few people having the supporting hardware.
(This is one of the reasons GPU upscalers and frame generation are becoming so popular, but those technologies aren't really ready for VR, their flaws very apparent on displays 1 inch from the eye)
Panels get too big and it makes it VERY hard to make an optical stack, & headset gets clunky. And clunky headsets only really work in sim-pits. It gets hard to 'whip' your head around in fast paced shooters. In order to reduce optical distortion panels HAVE to be smaller, which drives up cost a LOT.
Lens tech never really trickled down (makes sence considering cinimatography lens tech never really trickled either)
Patent hoarding is stifling innovation. I'm damn glad we're going to pancake and aspheric stacks. Trying to go back to an Index or any Fresnel lens stack is like wiping Vaseline over my eyes. Just a mess of blur.
(Also, Quest 3 / 3s is LCD with LED backlight, not true MiniLED. Not too far from Quest Pro's displays, just no FALD or Quantum Dots)
DFR must be added to all PCVR applications, and DLSS and FSR must be better implemented so VR can forgo the 6 to 40 grand for a professional GPU. I don't see the brute-force hardware approach to VR problems ever being affordable for consumers.
Many of us would LOVE it if DFR was a priority, but there is no profit incentive for any of the VR development shops to add this. Where it matters most in complex racing and flight sims, the MSFS development staff have shown no interest and the iRacing development staff have shown no interest. If the larger more profitable titles don't add support the smaller shops will never consider it. With no incentive, there is no momentum.
I'm on the fence. The Quest 3 is the first headset I owned that I feel is visually good enough, but I can still see pixels, and of course blacklevels suck. I can definitely understand we want resolution at least to the point of where the Apple Vision Pro is. After that, I think FOV needs to be the main priority.
My Pimax 8KX has the widest FOV but the clarity is lacking.
I just bought the Pimax Crystal Light and the clarity is nice but the FOV is not as wide as the Pimax 8 KX.
Which do you prefer using? I've been on an 8KX for a while because I'm not convinced I would be OK losing so much FOV to go to something else.
@ I prefer the Pimax Crystal Light because of its excellent clarity. What I don’t like about it is the narrow FOV though.
@@LeonardBrownArtbyLeonard Is it like the equivalent of small FOV on the 8KX or even worse than that?
@ I believe it’s equivalent to the small FOV on the 8KX.
@@aprile1710 The Crystal is the equivalent of potato on the 8KX, not small.
The quest 3s recently came out with fresnell lenses
The 3S also has a lower resolution display. The HTC Focus Vision also has Fresnel lenses, but they are just releasing earlier tech. Since Fresnel rely on physical grooves in the lens, it's very hard to make them work well with high ppd. That's why aspheric and pancake lenses are being used.
And the Quest 3s is also 300$… they have to cut corners somehow
I don't need "ultra wide" 180°+ FoV, 160 sounds mind blowing, and I think if any manufacturer can hit 140° it would likely be the point of diminishing returns/ "Good enough"/ accepted by the majority level...
I just want something that improves on the Index without going backwards in other areas, even if just a modest bump in FoV, resolution, lens quality, and a slight decrease in size. Apparently that's too much to ask for from anyone besides PiMax, so I'll just keep waiting. That Sominum VR1 sounds kinda promising.
I agree with you about diminishing returns past 140H. Going from 140-180 will likely increase the size/weight and GPU load by a massive amount.
tbf your human skull doesnt allow for much fov because you can see the ingress of your eyes, the outer bit of your eye socket bone. but the peripheral vision effect is most welcome if they get to it tho
Hypervison vr headset showed it can be done lightly
It will be interesting to see what finally comes to market based all their R&D and prototyping! But it still may be many years before we see these in enthusiast headsets.
Great video, very informative, thanks! The low FOV has killed VR for me. I can't buy the new headsets because they've low FOV. I am very sensitive to small FOV.
For hi resolution there was nice test what Varjo did at first prototypes but was unpractical. "flying mirror" for hires display that tracked your eye movement.
If I remember right next proto's was one fixed hires display that overlapped normal vive resolution display for "not sweetspot" display. I think they achieved human resolution (pixels / deg) or close to it in sweet spot. That also needed 2 optical links.
I think their current "consumer" headset use one display / eye for cheaper optics (No need to combine two displays for one eye) and don't need yearly license.
Maybe when we get displays with pix/inc regions we can lower calculating need from gpu and can give better resolution for optics sweet spot.
If UA-cam recommends 30 Mbps to watch 2160p 60FPS videos, why do we need such high bitrate for VR? I don't notice much difference in my quest 3 between 400 Mbps and 80 Mpbs. Genuinely curious.
UA-cam streams HIGHLY compressed video, which means you only get a full frame periodically and then it is mostly only sending the pixels that change.
An uncompressed 4K/60Hz signal is 18Gbps.
Increase 60H to 90Hz = 27 Gbps
Now double that for 2 eyes = 54 Gbps.
a 4K signal is only 3240x2160 or about 7Mp
However the rendering resolution for most headsets is far above that.
For the VR1 it could be about 4 X more pixels. 5600x5600 = 31.4Mp
31.4/7 = 4.48 x 54 Gbps = 242 Gbps
But most VR is only 8 bit rather than 12 bit color.
so drop that down by 2/3 x 242 Gbps = 161.2 Gbps
Now add DSC ( display stream compression ) 3:1 visually lossless compression
161.2 /3 and you are back down to 54 Gbps at 90fps
Display Port 1.4 has a maximum data rate of 25.92 Gbps.
Display Port 2.0 has a maximum data rate of 77.4 Gbps.
I may have missed something but those should be the main bits.
Issue is the degredation of brightness at higher fov. PIMAX tackled this with their 8kx by using comparatively MASSIVE display panels. If you want a compact package and decent fov im not sure the tech is there yet. If it were we would have it.
Very informative. Thank you!
Thanks Mark👍👍👍
We don't need 200 degree FOV in a headset. I'd take a true 140 degree and be comfortable at calling it a day there as a compromise between res and FOV. PSVR2 has a slightly wider FOV than the Index, and is also higher res. It's really weird how that headset basically gets no credit (its even ommited from this video) or generally gets shit on, when you will not find a better headset for the money. Quest 3 is great and all, but you have the issue of compression of the image with that. If a headset had 3k per eye and the 140 degrees with native display and a small form factor, I think we're in a area where you've got a fair acceptable standard to go forward from there.
Yes I don't get the focus on crazy wide FOV right now when we struggle so hard with tradeoffs at only 110-120, all I need is 140-150 with additional pixels to fill that increase, and eye tracking / foviated rendering, with maybe 20% increase in pixel density, especially in the first 90 deg - if that is even possible
Great video! I may send some of my customers here ;-)
I'm just wondering why VR videos, even through a PC, are so blurry ? Everything is a bloody mess.
I'm guessing because I haven't made any of those "yet", but typically a VR video stream would bounce around with the players head and I think many people use some kind of image stabilization which might blur the image. I know some people have made these types of videos sharp, so there is a way to do it properly. Another reason might be that the video stream being shared my be rendered at a lower priority than what the player is seeing. Once again only guesses.
Now is when someone who knows the answer can reply with a correction to my erroneous information :)
I have had the star vr one,pimax 8kx,varjo xr3 and many more... i realy hope there comes a good micro oled headset with a great binoculair overlap and great brightness with a decent FOV
The MeganeX Superlight 8K is close to that, but a little over priced.
Honestly, I think it’s just an expensive aspect to develop for such a small audience that actually cares about it. As an enthusiast, I appreciate a higher FOV, but I would assume most people are just fine with the FOV the Quest 3 offers
finally, a good VR info video
Has anyone played around with simulating a larger fov? Something like simulating the eyes of a cow.
Do u mean a more fisheye look so cramming in say 140 deg view onto a 110 deg display for example?
Cows have wide pupils because they are prey animals and need to see predators along a narrow stretch on the horizon in focus. Predators have vertical pupils so they can see their prey directly ahead of them clearly at different distances in focus. Humans have round pupils that are balanced in terms of what is in focus in both directions.
The key is that this isn't field of view related. This is focal plane related.
A good analogy to this would be a Tilt and Shift camera lens that allows you to get an in focus view along a single axis. The lens still sees the exact same area, but things are in focus along a specific section of it.
Perhaps looking at pixels per degree of each headset would reveal some insights. 😉
more FOV please! I don't mind rift's resolution.
Then find a used Pimax 5K Plus.
super interesting insights!
FOV is difficult to get right optically. Pimax shows you how much of an issue it is. This is because we need to use curved screens on this scale. In addition to this, other market contenders that don't push FOV on their products but instead of absolute res, pushes other companies looking to innovate in FOV to drop that endeavor and compete with the race to the highest res instead of working on FOV development.
much better would to just add microLED’s just outside the peripheral vision that match the color of the edge of the screen.. simply because our FOV isn’t that wide to begin with (would say 120° H and 188°V.. so why have screens that reach 200° + FoV when the display moves with your head anyways.)
Would curved displays inside the headset help with fov in any way?
I read about some company looking into that. I can't remember if it was Samsung or not.
Are pancake lenses superior to aspheric ones refarding required render resolution. Do aspheric lenses require much bigger resolutions (for the same panel resolution) due to distorsion correction?
@@davidzx692 I don't have that answer. This is something that Meta and Hypervision should know, but we haven't had any wider FOV pancake lenses in the wild to look at yet.
The advantage of pancake lenses is that they use reflection rather than refraction as their means of bending light. This allows for more aggressive angles and generally higher lens performance. And it produces much less chromatic aberration.
This is the reason I keep my pimax 8xk, nobody else even to this day offers large FOV which besides wireless is probably the most important thing to feeling detached from reality. I can easily get higher pixel density, the market has no more modern solution for FOV to this day.
120hz,4k,oled and high fov with simple plug and play for pcvr is all we want why cant companies do that at $1,000 bucks
As far as FOV goes, it’s always puzzled me why no one is looking at wrap around curved panels. Well I say no one, I mean public ally. There’s prob some really interesting prototypes in labs around the world.
You are right. The panels wouldn't be a big problem. That's a good idea.
However, no company in the world has the tooling to build the curved lenses these curved panels would have to be paired with. That is a huge problem.
Somebody must invest in the high-tech machines to create such lenses first. I'm not sure the consumer VR headset market can absorb the cost for that.
why cant they just have an extra low-resolution panel do the extra fov, like glasses, you have your high detail in the middle and you can see around the edge but it is blurry/low rez
HyperVision has a two display per eye solution with lower resolution panels on the outside.
I Want Wide FOV Too! The Pimax 8KX, 5K+, and XTAL 5K and 8K units all gave me eye strain. So Im on the hunt. Im gonna take a chance on the VR1. No way Im going back to Pimax.
It was hit or miss on the Pimax wide FOV headsets. Some had eye strain and headaches, and I know some people who were OK with the optics and loved those headsets.
From what I hear the VR1 is better optically and has better binocular overlap than the Pimax wide FOV lenses do. If you can afford it and can pair it with a 5090, you should be very happy.
@@markkovalcson7243 Yes! it was hit or miss for sure, some had eye strain and some didnt. I loved the wide FOV with the Pimax and when it worked, the immersion was fantastic.
The 8KX also had an issue where if you had low IPD, the image was blurry. Pimax wasnt addressing the issue and touted it was use error. It was not. I had to heavily modify the 8KX to get rid of the blurriness. When I say modify, I had to disassemble the headset and carve out plastic to allow the lenses to come closer together, not because my IPD is very low, but to allow the lenses to line up properly with my eyes, long story on that but it worked!! . I had a huge write up on the mod at the Pimax forum back in the day and provided pictures and a small tutorial. It was a daunting mod and wouldnt recommend trying it unless your a seasoned electronics or small device repair tech.
My mod fixed the slight blurriness experienced with lower IPD users, however the eye strain continued to be and issue, some believe its the canted lens design that XTAL and Pimax implement. Im really hoping that Somnium didnt take that design from the XTAL and Pimax playbook. Its hard to tell from any of the VR1 pics if they have canted displays/lenses. Somnium has a 14 day return window so if it has canted lenses at least I can return it.
Foveated rendering would certainly reduce the GPU requirements for wide FOV displays.
Psvr2 still uses fresnel, you did not mention. what " the oled uses?
Correct. The PS VR2 is a headset adapted for PC use because they were not selling well. Also the 2000x2040 resolution is less than the Reverb G2 that used Fresnel lenses. So as a trajectory, it is near the end of the line.
What I was saying was that to handle new higher resolution displays Fresnel doesn't hold up.
@@markkovalcson7243 This is innacurate. Sony always had designs of making the PSVR2 for the PC, and the smoking gun on that is the USBC Displayport. When they were developing the hardware, USBC Virtual Link was thought to be the industry standard going forward, so Sony left it in, and built the port into the PS5. As we all know, Nvidia had other ideas and stopped using it after the 20 series cards. It has nothing to do with the headset "not selling well" as we don't know any information on that because Sony has been a bit tight lipped about everything recently. We know one thing, though: the headset flys off the shelves every time they have a sale on it.
Also, fresnel has no bearing on if a display holds up well or not. The optics are all about how light passes through and reaches your eye, and if you have a HDR display, you need to allow more light to pass through and you can't get better than fresnels at doing this, unless you produce a really bright OLED to mitigate the offset of light if you want to use HDR with something like pancake lenses. That would drive the cost up to enterprise levels like the Xtal.
What I don't understand is why headset manufacturers don't at least put some leds around the lenses (in the facial interface) that match the colour of whatever is at that part of the screen, so at least it's not completely black outside the lenses. I think that would already improve the perceived FOV significantly!
Yours is the first of 2 comments with this exact suggestion. FWIW, I have seen a DIY attempt at this. I don't remember if it was successful.