Sony VR Headset/Project Morpheus/PlayStation VR

[Shifty Geezer]

Games targeting PSVR are targeting a fixed spec and a minimum requirement set by Sony. They won't look as pretty as PC's offerings, but devs can completely nail the framerate and stability. Everything points to PS4 providing a very solid (high framerate, stable, accurate head tracking, reasonable display, etc.) VR experience, not least of which are positive hands-on reports.

Where Oculus on a monster PC is a Dodge Charger SRT Hellcat, PSVR is a Skoda Fabia (What Car Magazine's car of 2015). Which if you're wanting to get from A to B and haven't limitless quantities of money to spend is perfectly fine. What PSVR is not, as some might suggest, is a wonky jallopy limping along on three wheels. Sony have said they don't want to get into VR until they can provide a decent experience, and they've been looking at this since before Palmer Luckey launched his Kickstarter as the recent interview says.

 

[Davros]

So ps4 vr games will run fine they just wont be that good looking ?
There would for example be no chance of something like ryse on ps4 vr ?

 

[Shifty Geezer]

"Good looking" is relative. We've been told generally by VR game developers that their game on PSVR is equivalent to OVR's results. Think of it as the difference between Ultra High and Medium settings. If "Medium Settings" make your skin crawl, then you may consider PSVR games to not be good looking.

Spend 13 minutes watching this. Seems to have some gameplay in between trailers.

 

[Davros]

thanks, they look good
are they genuinely running on a ps4 or a pc with a morpheus attached (i know eve is also on pc)
I'm wondering as quite a few games on ps4 are 30fps how will it manage ? or will 30fps be fine on a vr display

 

[MrFox]

30fps doesn't work for VR.

Devs who make VR games will have to optimize until it's a stable high frame rate. This is one of the reasons the PSVR games we have seen so far don't look as good as Uncharted 4 or Until Dawn, etc.... It absolutely must be at least 60fps and it's then doubled to 120fps to keep the latency as low as possible. Some games will be 90fps or 120fps native, too, with simpler graphics or simpler shaders, etc... Obviously nobody expects a PSVR game to look anywhere near an i7 PC with a 980ti, but it was said to be very close to a 970.

The difference in image quality should be something like the difference between a PS4 and PC. But PSVR has an efficient advantage where it can render at 60 and reproject at 120, which PC headsets cannot do. It's not magical but devs claimed it works for many types of game. It helps reduce the GPU deficiency versus the more powerful PC rigs. The other mitigating factor is the external processor of PSVR which reduced the overhead of the relatively weak PS4 CPU.

 

[Shifty Geezer]

thanks, they look good
are they genuinely running on a ps4 or a pc with a morpheus attached (i know eve is also on pc)
I'm wondering as quite a few games on ps4 are 30fps how will it manage ?

It just does. What you see (gamplay shown) is update 120 Hz in PSVR, 90 Hz minimum. In the game, it renders 60, 90 or 120 Hz, and the headset updates 90/120 Hz with the latest tracking info.

The gameplay shown so far is the gameplay actually running on a PS4. Other content like the Robinson footage may be a little optimistic. But ultimately, games look good, like this gen and not ropey PS3 titles.

 

[orangpelupa]

I'm still concerned with the PS move though... (that was discussed a few pages back, especially about its communication latency)

PS move is really precise and responsive for non vr. But is it good enough for vr? Or sony will switch tk 'overdrive mode' or something when on VR? Then how about the battery life

 

[manux]

I'm still concerned with the PS move though... (that was discussed a few pages back, especially about its communication latency)

PS move is really precise and responsive for non vr. But is it good enough for vr? Or sony will switch tk 'overdrive mode' or something when on VR? Then how about the battery life

Move latency isn't that huge. Something like 17ms or so. That should still give you close to 60Hz if you sync move with display.

It's not like rendering latency approachs 0ms anyway. Even in ideal case you would have 1 frame displayed and 1 being rendered and that's in 120Hz rate... whatever move is lagging you can most likely use same reprojection algorithm sony uses for 60->120Hz conversion. i.e. Correct the movement at the end of rendering pipeline.

Is move latency ideal? no. Is it good enough for many and more use cases? very likely. Would some naive implementation be best possible? Quite unlikely

 

[MrFox]

I remember an interview where they said the sensor fusion and prediction algorithm used on the headset applies to the Move controllers too. In fact it's the Move engineers who did the work. It used to be 22ms on PS3, but depending on the accel/gyro rates, it should be in the single digit on PS4 along with a much better Z resolution because of the 2 camera. That's as close to an ideal latency as you can get. The rendering time is dwarfing the Move latency anyway.

The headset is essentially using the Move tech, not the other way around.

 

[orangpelupa]

I caaaaaant wait for psvr! Hopefully the dirty parallel importers in my country able to get their hands fast (or faster) and sell it at correct price (or even lower).

My country got GBA release faster and cheaper than Japan official release lol (thx to the importers).

 

[Arwin]

Move latency for one controller on PS3 is typically 4ms, which is basically identical to DS4. It goes up when you add more controllers, but at 4 controllers it peaked at 21ms iirc. We don't have any PS4 numbers, but I expect them to be equivalent or better.

 

[cheapchips]

I remember an interview where they said the sensor fusion and prediction algorithm used on the headset applies to the Move controllers too. In fact it's the Move engineers who did the work. It used to be 22ms on PS3, but depending on the accel/gyro rates, it should be in the single digit on PS4 along with a much better Z resolution because of the 2 camera. That's as close to an ideal latency as you can get. The rendering time is dwarfing the Move latency anyway.

The headset is essentially using the Move tech, not the other way around.

Some of the early prototypes were literally a Sony viewer hdm with move controllers strapped to it.

 

[orangpelupa]

Yeah and it looks like voodoo ceremony lol. But instead of candles, it was PS move

 

[Deleted member 11852]

Move latency for one controller on PS3 is typically 4ms, which is basically identical to DS4. It goes up when you add more controllers, but at 4 controllers it peaked at 21ms iirc. We don't have any PS4 numbers, but I expect them to be equivalent or better.

Out of curiosity, which latency is this? The Move controller communicating buttons/sticks/accelerometer changes wirelessly to the console or the camera observing Move spatial movements?

 

[orangpelupa]

I just tried summer lesson and ace combat 7 trailer on YouTube cardboard with the max screen size setting.

O_O it makes me notice stuff that was not noticed on screen. Makes me want to move my head (summer lesson trailer is better for this, because the game camera move with the same direction as I move my head).

Also tried battlefield 3 on rails jet scene 3d video.

All on cardboard with 1080p screen. Aliasing is not a problem at all. Small text also readable.

I think psvr if able to remoteplay ps4 to virtual screen like YouTube app. It will be good enough. Case in point, the small dialogue subtitle on bf3 is still readable.


The videos :

 

[hughJ]

I'm still concerned with the PS move though... (that was discussed a few pages back, especially about its communication latency)

PS move is really precise and responsive for non vr. But is it good enough for vr?

www.youtube.com/watch?v=0J5LaWykiIU
Granted this is using a custom implementation of Move, but it shows the basic hardware can be made pretty robust and precise. In terms of latency, hand tracking isn't nearly as picky as head tracking so you can get away with a frame or two of lag (if need be) without it being very disruptive to the experience. If there's any drawbacks of the overall system it might be from instances of total occlusion if you were to face away from the camera, but I suspect that most VR content we're going to see in the coming years will be delivering Sony/Oculus-oriented motion controlled experiences where you're expected to be facing a single direction most of the time. IMO, I don't think there's any cause for concern over the Move controllers.

 

[Arwin]

Out of curiosity, which latency is this? The Move controller communicating buttons/sticks/accelerometer changes wirelessly to the console or the camera observing Move spatial movements?

Spatial movements ...

 

[MrFox]

IGN gave a quick impression of PSVR versus the recent Vive upgrade. They said the Move is a bit more shaky. Maybe the Move will need to compromise a bit of latency for some smoothing filter. The Vive controllers are rock solid (not surprising considering the tech they used).

http://ca.ign.com/videos/2016/01/10/comparing-valves-vr-demos-vs-playstation-vr-ces-2016

 

[orangpelupa]

No written article? The video can't play here

 

[Deleted member 13524]

No written article? The video can't play here

Nop, just a small ~2min interview.