Oculus enables Asynchronous Timewarp for Rift release!

[Silverforce11]

https://developer.oculus.com/blog/asynchronous-timewarp-on-oculus-rift/

http://techreport.com/news/29914/oculus-sdk-1-3-brings-asynchronous-timewarp-to-windows

tl;dr: The Oculus PC SDK v1.3 implements Asynchronous Timewarp (ATW) on Windows. With the latest drivers and hardware, we reduce judder, deliver consistent low latency, and improve efficiency. All apps benefit from this without having to do anything special.

Asynchronous Timewarp (ATW)
We can deal with varying workloads and background apps by decoupling timewarp from the rendering loop and running it independently and asynchronously. By relying on CPU and GPU preemption, we should be able to pick up the latest available application frame and timewarp it before each display scanout at 90Hz.

Interestingly they use the CPU for what is essentially a GPU compute task.

Windows and several recent GPUs have had basic support for GPU preemption, but when we first tried to use it, it did not work well in practice. Over the last year, we have worked closely with Microsoft, NVIDIA, and AMD to change OS GPU scheduling, GPU command processor microcode, and GPU kernel driver design to enable ATW. Graphics driver VR extensions in the form of AMD’s Liquid VR and NVIDIA’s VRWorks were developed to support this.

I don't know how NV GPUs get parallel compute execution while graphics is rendering, but this is AMD's implementation in their latest drivers for the Oculus Rift.

http://techreport.com/news/29917/radeon-software-crimson-edition-16-3-2-is-rift-ready

https://community.amd.com/community/gaming/blog/2016/03/28/asynchronous-shaders-evolved

Time for some tech sites to go in-depth and measure Motion to Photon latency with the Rift since both LiquidVR & VRWorks are functional!

Ars released their Rift review today:
http://arstechnica.com/gaming/2016/...ift-expands-pc-gaming-past-the-monitors-edge/

Mastering nausea. It wouldn’t matter how comfortable the Rift is on your head if it made your stomach feel uncomfortable.

Oculus has spent years hammering a list of best practices for keeping players’ stomachs happy in virtual reality, and developers have by and large taken them to heart. For some games, this means only moving the player’s virtual perspective slowly or quickly “blinking” from one position to another with a quick fade to black.

AP journalists at CES noted that some Rift apps were more an on-rails experience, to prevent rapid head motion. The above is a another trick, fading in/out scenes when players move their heads fast.

Simple air hockey game Shufflepuck Cantina, for instance, featured a distracting stairstep judder as I moved my head around its alien environs. Running around the world of Albino Lullaby felt a little disorienting as well without any fixed focal point to stay centered on. This effect will probably vary from person to person, too—if you’re the type to get sick watching first-person games on a 2D screen, you might be in trouble.

 

[thesmokingman]

And so it begins...

 

[Mahigan]

AMD VR Master Race.

 

[monstercameron]

AMD VR Master Race.

yet still it's the 970 and 980 that are getting bundled in vr rigs.

 

[cytg111]

Damn .. I am gonna have to pick up the Vive AND the Rift .. and problary ps4.5 too towards the end of the year!

 

[Mahigan]

yet still it's the 970 and 980 that are getting bundled in vr rigs.

http://store.hp.com/webapp/wcs/stores/servlet/ContentView?eSpotName=Phoenix-VR&storeId=10151

http://m.marketwired.com/press-release/amd-takes-83-share-global-vr-system-market1-driving-leading-vr-experiences-consumers-nasdaq-amd-2105536.htm

They're all going to switch once their customers start complaining of Nausea associated with judder on NV equipped VR rigs.

There's still this notion that NVIDIA are superior to AMD. That notion is quickly being erased as developers and Game Engine makers rush behind LiquidVR as opposed to NVs solution.

I think NV have Epic UE4 and Unity. Which isn't bad but AMD have landed Ubisoft, EA/DICE, Crytek and others.

Basically AMD is well ahead of NV in the VR market.

 

[Silverforce11]

What i really want to see is some proper tech journalist testing to see whether Liquid VR or VRWorks lives up to their hype at reducing motion to photon latency from the 35-50 ms that's the standard.

 

[Good_fella]

http://www.tomshardware.co.uk/oculus-rift-virtual-reality-hmd,review-33509-7.html

 

[USER8000]

What i really want to see is some proper tech journalist testing to see whether Liquid VR or VRWorks lives up to their hype at reducing motion to photon latency from the 35-50 ms that's the standard.

I expect there to be little or no latency testing though,just saying! :thumbsup:

 

[Silverforce11]

http://www.tomshardware.co.uk/oculus-rift-virtual-reality-hmd,review-33509-7.html

Seems some games are better optimized than others and are able to hit 90 fps minimum, Eve especially. But this again is just a measure of FPS and not actual nausea inducing motion to photon latency. :/

I was expecting to be able to play Dreadhalls on the Rift without any trouble, but it only took around 10 minutes for me to start feeling off. I wasn’t dizzy and I wasn’t nauseous. The feeling is hard to describe, but I didn’t enjoy it, and I actually had to take a break from the computer after that.

I enjoyed this game for about 10 minutes. I was floating around, weightless, making my way through the space station, when suddenly I was struck with an overwhelming feeling of uneasiness, and my arms were perspiring. It wasn’t quite the same as my experience with Dreadhalls, but it was equally unpleasant, and it took me nearly a half an hour to shake the feeling. Again, your experiences are likely to be different game to game.

Edit: And they aren't using the Rift Ready drivers, for AMD 16.3.1 and NV 364.64

http://www.tomshardware.co.uk/oculus-rift-virtual-reality-hmd,review-33509-6.html

Pointless review.

 

[Silverforce11]

http://www.eurogamer.net/articles/digitalfoundry-2016-oculus-rift-review

This is even worse.

I couldn't hack more than 15 minutes of Ethan Carter or Project Cars - both rated as 'intense' - without taking a break.

And it turns out Async Timewarp is not enabled on any game yet, because it's only enabled on 1.3 SDK, so games have to be recompiled on the latest version, along with the latest drivers from AMD/NV to use this feature.

Why launch in such a state, when reviewers are complaining of poor experience?

 

[cytg111]

"suddenly I was struck with an overwhelming feeling of uneasiness, and my arms were perspiring."

I dont know if that is a bad thing or something amazing.

Bad thing cause, well, noone likes to feel ...bad
Good thing cause, the level of submersion required to set off such a response ...damn

 

[Dribble]

"suddenly I was struck with an overwhelming feeling of uneasiness, and my arms were perspiring."

I dont know if that is a bad thing or something amazing.

Bad thing cause, well, noone likes to feel ...bad
Good thing cause, the level of submersion required to set off such a response ...damn

Of course it's bad, VR can make you feel ill, and it's not going to be fixed by game ready drivers or a slight reduction in judder. Take someone on a roller coaster for more then 10 mins and they'll feel ill too. That immersive experience has down sides for many styles of game. At best you'll feel as ill as you did playing something in 3D (as that is what VR is like with a locked camera), at worst you'll feel a lot more ill (with everything moving around, and your brain unable to understand as things like inner ear and sensation of gravity don't match what the eyes are showing).

 

[airfathaaaaa]

What i really want to see is some proper tech journalist testing to see whether Liquid VR or VRWorks lives up to their hype at reducing motion to photon latency from the 35-50 ms that's the standard.

how can you test something so subjective and meanwhile trying to be objective? i mean its impossible to do such a thing you will need to find people with not exp whats so ever on vr and obviously not being sailors :sneaky:

 

[coercitiv]

how can you test something so subjective and meanwhile trying to be objective?

Latency is not subjective.

 

[airfathaaaaa]

Latency is not subjective.

what you see on numbers and what you ACTUALLY see is two very different things...

 

[Silverforce11]

how can you test something so subjective and meanwhile trying to be objective? i mean its impossible to do such a thing you will need to find people with not exp whats so ever on vr and obviously not being sailors :sneaky:

I just want to see if either GPU vendors have cracked the sub 20ms head motion to photon latency that is the holy grail of VR.

Last time NV said their best attempt is halving the 57ms pipeline. Still above the target for comfort.

AMD have said they can get to 10ms, but I find that hard to believe.

 

[Flapdrol1337]

http://store.hp.com/webapp/wcs/stores/servlet/ContentView?eSpotName=Phoenix-VR&storeId=10151

http://m.marketwired.com/press-rele...-experiences-consumers-nasdaq-amd-2105536.htm

They're all going to switch once their customers start complaining of Nausea associated with judder on NV equipped VR rigs.

There's still this notion that NVIDIA are superior to AMD. That notion is quickly being erased as developers and Game Engine makers rush behind LiquidVR as opposed to NVs solution.

I think NV have Epic UE4 and Unity. Which isn't bad but AMD have landed Ubisoft, EA/DICE, Crytek and others.

Basically AMD is well ahead of NV in the VR market.

Amd's 290 and newer is better suited to this async timewarp thingy, but most vr systems I've seen had high end nvidia cards.

The 83% is complete nonsense since they count the consoles as vr systems.

 

[digitaldurandal]

what you see on numbers and what you ACTUALLY see is two very different things...

The numbers are actually quantifying what is displayed.. hopefully what is displayed is also what you see.

 

[Ferzerp]

Of course it's bad, VR can make you feel ill, and it's not going to be fixed by game ready drivers or a slight reduction in judder. Take someone on a roller coaster for more then 10 mins and they'll feel ill too. That immersive experience has down sides for many styles of game. At best you'll feel as ill as you did playing something in 3D (as that is what VR is like with a locked camera), at worst you'll feel a lot more ill (with everything moving around, and your brain unable to understand as things like inner ear and sensation of gravity don't match what the eyes are showing).

I think one of the theories is that motion sickness provokes the nausea response because unable to reconcile the conflicting inputs, the brain's interpretation is that you must be poisoned and hallucinating and so the best thing to do is to purge your stomach to get rid of any poison you can.

 

[thesmokingman]

what you see on numbers and what you ACTUALLY see is two very different things...

:\


One would think that the two have a direct bearing on one another.



As mentioned we really need to see this with the new sdk. This will be really bad playing while drunk man. Keep a bucket near you at all times!

 

[railven]

Woof, this is not first impression you want for VR.

Any idea when the recompile games/files can get to reviewers? If I've learned anything recently of reviewers, they don't like to revisit something they don't like.

 

[pj-]

I don't think recompiling with the latest SDK is going to fix anything. The "intense" games are rated that way for a reason.

Unnatural movement (e.g. your virtual "body" is moving when your actual body isn't) is going to cause discomfort for some people no matter how good the headset is. There are ways to mitigate it but it by definition causes a mismatch between your senses, which can lead to nausea. It is a matter of game design, not failings in the hardware or software.

I don't think there are many (any?) reports of people getting motion sickness in "comfortable" rated games.

 

[Azix]

Amd's 290 and newer is better suited to this async timewarp thingy, but most vr systems I've seen had high end nvidia cards.

The 83% is complete nonsense since they count the consoles as vr systems.

its not non-sense. The consoles are going to have a huge impact of VR space and the rumours of a new ps4 with more power (likely using AMD graphics) makes that number more significant.

On the PC side, systems for sale is not market share.

 

[PhonakV30]

Amd's 290 and newer is better suited to this async timewarp thingy, but most vr systems I've seen had high end nvidia cards.

The 83% is complete nonsense since they count the consoles as vr systems.

It's not! Read this nice post from AtenRa , You will see why.Also Here From wccftech :

Console Gamers are More Excited About Virtual Reality Than PC Gamers – 72% PS4/Xbox One Owners Show Interest