Rift, Vive, and Virtual Reality

[Anarchist4000]

Perhaps we should spell things out a little more clearly to make sure we're on the same page. What "async" are you specifically talking about here, and in what capacity are you expecting it to impact either general gaming performance or specifically VR?

ATW is the big one, not necessarily concurrent execution. I'm of the opinion it will render most of the current reprojection techniques irrelevant, boost performance, and provide a smoother experience. Why double frames when you could adjust them to varying degrees to hit a refresh target. I see no reason 60-70fps with more/less of a warp wouldn't be superior. Take a single frame and possibly warp it multiple times to achieve a more responsive environment. That in turn would change the performance targets for developers to achieve a quality experience. Effectively 45fps might be the minimum required framerate for a good experience.

While not async, the method likely allows you to bypass MSAA in favor of SSAA or other methods that boost performance and/or increase quality. It would allow you to boost performance by having less AA. If presenting 90 frames and you could render 100 frames in that time you save 10 frames of AA.

To illustrate:

980 Ti
Average frame rendering time: 8ms, 90 FPS
Timewrap rendering time: 1ms <- higher cost, but never needed.

Fury X
Average frame rendering time: 12ms
Timewrap rendering time: 0ms <- infinitely faster (generously speaking), but needed all the time because GPU/driver can't keep up.

The 980 Ti will give you better IQ and experience, thus the better choice.

980ti
Average latency: 8ms

Fury X
Average latency: 2ms

 

[hughJ]

ATW is the big one, not necessarily concurrent execution. I'm of the opinion it will render most of the current reprojection techniques irrelevant

ATW is already implemented with the Oculus SDK.

980ti
Average latency: 8ms

Fury X
Average latency: 2ms

8ms rendering time + 1ms timewarp = an effective tracking latency decided by the degree of prediction made by the VR SDK.
12ms rendering time + 0ms timewarp = a completely missed frame and a reprojected last frame, resulting in visible animation stutter, but still an effective tracking latency decided by the degree of prediction made by the VR SDK.

 

[Anarchist4000]

0ms+2ms = effective 2ms frame time. Different approach where you're talking the last frame and constantly warping until you get a new one. The actual frames would never be presented, only warped. Higher fps would therefore increase accuracy and there would be less possibility of a missed frame. Should also work with forward and deferred renderers along with a handful of performance optimizing features.

EDIT: Clarifying that dropped frames are unlikely.

 

[manux]

At least oculus uses timewarp *always*. ATW is mandatory not optional for oculus. Head position for frame to be shown is corrected in timewarp. Without atw head position might be quite much of assuming head was moved while rendering the frame.

 

[madyasiwi]

I see no reason 60-70fps with more/less of a warp wouldn't be superior.

Timewarp won't help with dynamic objects, specular, reflection.

Fury X
Average latency: 2ms

No, but judderfest for sure.

 

[Ike Turner]

To clear things up a little bit:
Oculus VR uses always on AsyncTimeWarp: https://developer3.oculus.com/blog/asynchronous-timewarp-on-oculus-rift/ which leverages preemption / Async Compute on Nvidia and AMD GPUs.

While

SteamVR / OpenVR currently only uses Interleaved Reprojection which doesn't run asynchronously. if a frame takes longer than 11.1ms the compositor will drop into half-time mode (run the game at 45Hz instead of 90Hz) and reproject every other frame. << Which is why it's worst than ATW. Why for example the SteamVR version of Elite Dangerous is painfull to play on a Vive compared to the Oculus VR version on the Rift..

 

[Deleted member 13524]

Sales for both headsets have pretty much halted during August:

http://venturebeat.com/2016/09/02/vr-adoption-among-steam-users-has-crashed-to-a-halt/


Maybe the opportunity has already been lost..

 

[London Geezer]

Maybe people prefer to see the sun in August, rather than sweat and get bored in a VR set

 

[Deleted member 13524]

Or you could think the other way, which is August is when many people get forced vacations out of their employers so it should be high-season for those curious about VR, as they finally have the time to play with it.
Regardless, the lack of relevant content, very high price of both headsets and Oculus continually trying to split the tiny industry with DRM may have already killed VR.

Perhaps the PSVR will be the salvation after all.

 

[London Geezer]

It's all in jest of course. But in a way, it would be very sad if humans spent their August playing or even thinking of VR games

 

[Shifty Geezer]

Compromise- watch VR August sunsets.

 

[Ike Turner]

Sales for both headsets have pretty much halted during August:

http://venturebeat.com/2016/09/02/vr-adoption-among-steam-users-has-crashed-to-a-halt/


Maybe the opportunity has already been lost..

A few things:

- VR has to be experienced first hand to sell. Unfortunately 90% of the potential users are forming opinions based on other people's opinions etc..
- The first VR experiences must be flawless or the potential user is going to be turned off. So IPD has to be measured, lenses must be cleaned, headset must be correctly setup etc. Unfortunately most of the time one or all of these are forgotten or missed...
- Room-scale VR is 100X better than seated VR (unless racing game or Space/Flight sim obviously). Unfortunately 2 of the big boys are not going to push in this direction for the foreseeable futur (because their tracking system is not robust enough - Oculus, Because they don't have the hardware- Sony)
- Expectations have to be lowered 10X. Current VR headsets have shitty resolutions and lenses! This is a fact for every single one of them (Vive, Rift & PSVR..not talking about GearVR which is in an other league in terms of craziness). Judging anything based on a video you see on YouTube or where ever else is not a good idea at all. Even the mirrored feed on your PC/TV (for PSVR) will look better than what's displayed in the headset because of the lenses and god-awful PPI.
- The hardware is not powerful enough. Don't believe the marketing hype from AMD, NVIDIA (let's not talk about Sony...) etc. Second gen HMD's with 4K displays will probably be here in 18 Months while even a Titan X (Pascal) can barely run every single VR experience/game at it's highest setting with a perfect 90FPs lock. Devs using UE4 aren't helping either. As a matter of fact I bought Redout (WipeOut style game using UE4) last night on Steam to check it out in VR and for the first time ever had to refund a game I digitally bought. The VR mode is a joke and horrendous. These kind of tacked on VR support is what the VR industry doesn't need at all.
- There's not enough software...sure but the problem is that the software currently out isn't that good either. There are only a handful of experiences or game's that are worth something (Valve Lab is simply the best VR software currently out, Valve Destinations, Budget Cuts is probably the VR "game" but still isn't out, Elite Dangerous VR mode is a mess on Vive but also on Oculus simply because the game was not built for it from the start etc etc..)
- The lack of software is normal given that the platform is brand new...
- Seated VR sucks..yeah I'm repeating myself I know
- People are waiting for the big game to make VR a success...well sorry to disappoint them but the big breakthrough may not come from a game, it may be something else, like a narrated story, virtual tour , etc we don't know. As a matter of fact Destinations (Photogrammetry based experiences) are currently some of the best example of VR done right and the experience, although relatively passive is extremely accessible for the average joe (your mom doesn't have to worry about shooting stuff, running or any other gamy stuff that would hinder her experience..)
- Price..obviously but don't expect much change here in the next 2 years.
- AR + VR is the future
- Cables are not that big of a problem.
- Too much hype will kill any product..so stop believing everything you read and try it for yourself. Especially the Vive.
-etc..

 

[Deleted member 13524]

A few things:

- VR has to be experienced first hand to sell. Unfortunately 90% of the potential users are forming opinions based on other people's opinions etc..
- The first VR experiences must be flawless or the potential user is going to be turned off. So IPD has to be measured, lenses must be cleaned, headset must be correctly setup etc. Unfortunately most of the time one or all of these are forgotten or missed...
(...)
- Expectations have to be lowered 10X. Current VR headsets have shitty resolutions and lenses! This is a fact for every single one of them (Vive, Rift & PSVR..not talking about GearVR which is in an other league in terms of craziness).
(...)
- The hardware is not powerful enough. Don't believe the marketing hype from AMD, NVIDIA (let's not talk about Sony...) etc. Second gen HMD's with 4K displays will probably be here in probably 18 Months while even a Titan X (Pascal) can barely run every single VR experience/game at it's highest setting with a perfect 90FPs lock. Devs using UE4 aren't helping either.
(...)
- Seated VR sucks..yeah I'm repeating myself I know
- People are waiting for the big game to make VR a success...well sorry to disappoint them but the big breakthrough may not come from a game, it may be something else, like a narrated story, virtual tour , etc we don't know.
(...)
- Price..obviously but don't expect much change here in the next 2 years.

If these are clear cut facts about VR, then I don't know why they even tried. Might as well just shove the whole thing into the garbage and try again in 10 years.

 

[Ike Turner]

If these are clear cut facts about VR, then I don't know why they even tried. Might as well just shove the whole thing into the garbage and try again in 10 years.

Well, you HAVE to start some where and right now is the right time. We finally have the tech so that you don't puke all over yourself after 5 minutes or feel like you are having a seizure (great positional tracking and low latency displays). Unfortunately the media, HW manufacturers, fanboys are hyping the living hell out of something that isn't yet for the general public..Kinda similar to Kinect in a sense (enormous hype... and then crapfest)... it was the first iteration of what ultimately became HoloLens. VR right now will pave the way to something even better later.

 

[hughJ]

It's an open question as to if (or when) you get an NES without first having the 2600 and inevitable market crash. Iphone without early cell phones and PDAs, DVD without LaserDisc, etc. Without VR as it is right now the billions that are being pumped into R&D would not be occurring, and especially not from software developers. Personally I can say that my current system would not have been purchased, including my last couple video cards, and the ~$500 worth of content (both VR and regular) I've bought in the last couple years.

 

[Anarchist4000]

It seems pretty safe that at least movies will be coming. I'm sure that is a significant chunk of the market companies are currently chasing.

People are waiting for the big game to make VR a success...

Kerbal VR, because that couldn't possibly end badly!

 

[orangpelupa]

The first VR experiences must be flawless or the potential user is going to be turned off. So IPD has to be measured, lenses must be cleaned, headset must be correctly setup etc. Unfortunately most of the time one or all of these are forgotten or missed...
- Room-scale VR is 100X better than seated VR (unless racing game

And it's also really depends on the game presented in the demo.

I tried Vive with the lab. The demo guys refuse to let me try other games. Refuse to increase the laser height or adjust the angle (I keep losing tracking if I'm standing upright, need to crouch a bit).

Did not get the presence one bit.

Right beside that there's an oculus dk2 being demoed. Sure it have shitty screen but the guy manning it is very friendly and allows various vr games.

Despite of its shitty screen, despite my arms are not tracked, I felt presence!

I instinctively, uselessly fails my arm. I instinctively want to "evade" the virtual objects.

VR is too reliant with many factors to convey the awesomeness.

The demo games need to be specially selected. The demo person need to properly serve the demo goers. The demo itself also need to run flawlessly.

 

[Silent_Buddha]

Sales for both headsets have pretty much halted during August:

http://venturebeat.com/2016/09/02/vr-adoption-among-steam-users-has-crashed-to-a-halt/


Maybe the opportunity has already been lost..

Not surprising. Currently actually impressive games are few and far between.

There's a lot, I mean a LOT of VR shovelware appearing in the Steam store. That is leading to a lot of fatigue with the format among current VR owners that I know and that stream. Most of them still hold out hope, but not many of them are recommending VR to friends who haven't already gotten it due to this.

Despite the potential of VR, it's still just a gimmick currently. Gimmick's don't really work on PC users but does significantly better on consoles (Wii and Kinect, for example). I expect PSVR to remain relevant for longer, but in 2-3 years I expect even on consoles interest will start to rapidly dwindle.

A certain portion of the population won't ever be able to use VR due to nausea issues. Games attempting to mitigate nausea restrict the ability to produce compelling games. Roomscale can only help so much. Shooters that don't induce nausea, for example, are basically on rails shooters (like Time Crisis or House of the Dead)) - only without the rails or movement. And that's a genre that only ever has short term draw potential. Movement beyond roomscale is required if the VR is to become as compelling as traditional digital gaming. Yet movement poses nausea issues. Teleportion movement is a compromise but it also breaks the immersion that VR relies on.

There are game types that can work extremely well. But almost all of them are very niche gaming genres. Adventure games, for example. Stationary on-rails shooters (many of the more impressive VR games/demos are of this sort). Games that put you in a single room where you can do things (like a cooking game, board game, etc.). Plenty of others. Even space simulations like Eve Valkyrie are a niche market, thought a fairly sizeable one. However, Eve Valkyrie can cause nausea in some people.

The technology is impressive, if lacking. But I still fail to see how it will expand greatly as a non-niche gaming platform.

And all of that is compounded by the various hardware providers being unwilling to co-operate with each other. Nvidia and AMD. Oculus and HTC. The less co-operation and sharing between entities such as those results in lower possibilities for something to eventually get the general public interested in VR as more than a short term fad.

Who knows, maybe after this generation fails the 3rd attempt at mass VR adoption in 10 or so years will be able to succeed.

Oh, something that might help with both immersion and nausea would be mechanical systems that can mimic the forces that the body feels. https://en.wikipedia.org/wiki/Motion_simulator However, those systems are inherently expensive and won't help with mass adoption. Without something like that, VR games are going to be extremely limited by trying to avoid anything that might induce nausea in too much of the potential market. Those unidirectional treadmills that some companies are experimenting with won't be enough.

Regards,
SB

 

[Silent_Buddha]

It seems pretty safe that at least movies will be coming. I'm sure that is a significant chunk of the market companies are currently chasing.

Having tried this a bit. I don't see VR movies taking off at all. The potential there for nausea is very high, unless you want to only ever use a stationary camera. The most comfortable VR movies and shorts to watch are those where the camera is stationary. Once any type of camera motion is introduced, nausea rapidly starts to build.

3D stereoscopic movies work pretty well, however, as long as there is some kind of anchored visuals (TV in a room, sitting in a virtual cinema or drive-in, etc.) to frame the movie. However, even then, people that are sensitive to nausea in VR (like me) will start to feel uncomfortable and slightly nauseous after 2-3 hours of continuous viewing. Movies with a lot of pans, quick pans, etc. will induces greater level of discomfort. And the problem here is that the discomfort persists for hours after you stop using the VR headset.

Regards,
SB

 

[hughJ]

If people are getting nausea with traditional movies in VR then they would also be susceptible to nausea with that media on large screens or theaters, so I'm not seeing that as an argument against the uptake of VR as a media viewer. Once the panel resolution is sufficient then VR HMDs will be able to function as a virtual monitor and cinema setup, albeit at a relatively low fixed cost. You're literally the only person I've heard of on the internet from the last 4 years that's complained of nausea from virtual cinema (never mind the issue with sliding windows from the Oculus Home/video UI), so I would be hesitant to paint that as a legitimate concern for the industry's future.

The generalized locomotion issue is a significant problem, but it's an issue that's been taking baby steps over the last couple years and insight into ways of mitigating or working around it are becoming a lot more nuanced than the sort of absolutist no-way, no-how, not-ever. Beyond that we still have GVS which may or may not have some place in the medium to distant future to directly combat the vestibular mismatch.