Exclusive: Fove's VR HMD At CES 2016, First Look

[kcarbotte]

Tom’s Hardware met up with Fove Inc.’s Chief Technical Officer, Lochliam Wilson, to get a quick hands-on impression of what eye tracking can bring to VR.

Exclusive: Fove's VR HMD At CES 2016, First Look : Read more

 

[sephirotic]

I was disapointed when I got to know that neither Oculus nor Vive were going to use Eye tracking on their first models. I'm much more interested in fove. After the annouced 600USD for Oculus I may be considering waiting more for the launch of Fove.
Not to mention it has a higher resolution.

Eye tracking is an essential part to make the immersion in VR much more efficient, specially the improved resolution only on the area where you are looking, this would allow true 4k, or even bigger resolution displays being viable even with gpu's as weak as a GTX960, although, a high VRAM for 4k may still be needed.

My only concern is how compatible different brands of HMD VRs would work with different games.

 

[Astrophysics]

This is the headset that i will be waiting for.
I have a 3x 25" screen eyefinity setup on a 290x, and it does all i need for the time being.
having played with a a couple of developer HMD's, my inclination is to believe that without eye tracking and foveated rendering, VR is simultaneously missing a feeling of nuance and depth, and currently requires more horse power than is strictly necessary.
The degree to which the area outside the center of your field of veiw can be "fuzzier" than what you are looking at is amazing. You only think the rest is a sharp image, but that is your brain creating a seamless montage, rather than the entire field of view being very sharp.
Eye tracking will simultaneously bring better performance, and better realism. Need Fove.

 

[f-14]

there's a problem with that VR head set it's covered in plastic infront of your eyes, put some glass in there for a window to virtual reality..... some ones going to walk into a door, off a cliff, into a table, infront of a bus on a long green light.
it's not a matter of if, it's a matter of when if it hasn't happened to one of the developers on lunch breaks already.

 

[8R_Scotch]

The biggest obstacle for Fove is software support and compatible games and the like. The tracking technology is compelling, it both adds to immersion and reduces hardware requirements, two big necessities in VR (Palmer said we'll only get lifelike graphics with something like 16k displays the size of a smartphone, that's a tall order, but considerably less so with foveated rendering).

My bet is that Fove won't get the needed media traction, developer support to make this an equal competitor to the likes of the Rift and Vive, but it's technology might be incorporated in future generations, Fove might be bougt or may license the technologies it's developing. Otherwise it'll be limited to more niche users, but likely won't go mainstream.

This tech is exciting, but it's just like hand motion traking, not enough on it's own, but an awesome addition to a more robust VR kit with more consumer visibility and developer interest.

 

[Bloob]

As long as they'll embrace standards, like OpenVR or similar, while simultaneously offering their own API for the eye-tracking -stuff, they'll have a chance.

Eye-tracking really feels like something essential for a great VR experience (so we can hit those high resolutions and FOVs).

 

[clonazepam]

The end-game is probably the digital distribution side, so I expect to see an article about them being acquired this year. Their tech in a Gen 2 or 3 unit from one of the big players.

There's another company doing 5K screens at 60Hz with 210 degree FOV, but I was turned off almost immediately when they stated they had their own engine and planned to be the one-stop shop for all of the software and hardware. They do have the Walking Dead license, so there is that. Actually, that's a bit scary. Was that sort of a pun too? I heard those are a thing these days.

 

[kcarbotte]

The end-game is probably the digital distribution side, so I expect to see an article about them being acquired this year. Their tech in a Gen 2 or 3 unit from one of the big players.

There's another company doing 5K screens at 60Hz with 210 degree FOV, but I was turned off almost immediately when they stated they had their own engine and planned to be the one-stop shop for all of the software and hardware. They do have the Walking Dead license, so there is that. Actually, that's a bit scary. Was that sort of a pun too? I heard those are a thing these days.

The other company you are thinkning of is Starbreeze with the StarVR headset.
I had a chance to try that out in September and talk to the devs.
It's not a 5k screen. It's two 1440p screens which add up to 5k. The company is also sonly using 60Hz panels because 90Hz panels that size didn't exist yet. They have no intention of releaseing until those are avaialable.

 

[clonazepam]

The end-game is probably the digital distribution side, so I expect to see an article about them being acquired this year. Their tech in a Gen 2 or 3 unit from one of the big players.

There's another company doing 5K screens at 60Hz with 210 degree FOV, but I was turned off almost immediately when they stated they had their own engine and planned to be the one-stop shop for all of the software and hardware. They do have the Walking Dead license, so there is that. Actually, that's a bit scary. Was that sort of a pun too? I heard those are a thing these days.

The other company you are thinkning of is Starbreeze with the StarVR headset.
I had a chance to try that out in September and talk to the devs.
It's not a 5k screen. It's two 1440p screens which add up to 5k. The company is also sonly using 60Hz panels because 90Hz panels that size didn't exist yet. They have no intention of releaseing until those are avaialable.

Nice, looking forward to hearing more about all of the players in the field.

 

[none12345]

Not going to touch these head sets until they meet at least these sets of requirements.

1) Peripheral vision is rendered. I don't want tunnel vision. This is a complete show stopper for me, if only binocular vision is rendered, i will not buy the headset.

2) 8k in smart phone sized screens(i don't want to be looking through a screen door, even 8k isn't enough but it wont be horrible) I may consider a 4k screen, i will NOT consider a 1080p screen. Approx 1000 ppi that close to your eyes is what i would consider minimum.

3) Total latency from draw call to photons leaving the screen is <=11 milliseconds.(90fps). 90fps is not enough if its for instance triple buffered, that would be a 33ms latency, and that is NOT acceptable. This obviously covers more then just the headset, it covers the entire system.

4) 2 screens, 1 per eye are used. OR they can make a single screen that renders 2 subscreens at the same time so you do not have a latency differential between right and left eyes. This is not a show stopper for me, but it is highly desired. Its pretty much required to meet the peripheral vision requirement tho; either that or a very wide curved screen.

5) eye tracking is used. This is not a show stopper for me. However, with the other requirements above, its pretty much mandatory to reduce the processing required to a reasonable level. There is no way you could render the above at full resolution with current gpus, even with quad gpus you couldn't. Nor are you likely to be able to with the next gen cards. However with eye tracking you might be able to do it with current cards, and probably with next gen cards.

 

[kcarbotte]

Not going to touch these head sets until they meet at least these sets of requirements.

1) Peripheral vision is rendered. I don't want tunnel vision. This is a complete show stopper for me, if only binocular vision is rendered, i will not buy the headset.

2) 8k in smart phone sized screens(i don't want to be looking through a screen door, even 8k isn't enough but it wont be horrible) I may consider a 4k screen, i will NOT consider a 1080p screen. Approx 1000 ppi that close to your eyes is what i would consider minimum.

3) Total latency from draw call to photons leaving the screen is <=11 milliseconds.(90fps). 90fps is not enough if its for instance triple buffered, that would be a 33ms latency, and that is NOT acceptable. This obviously covers more then just the headset, it covers the entire system.

4) 2 screens, 1 per eye are used. OR they can make a single screen that renders 2 subscreens at the same time so you do not have a latency differential between right and left eyes. This is not a show stopper for me, but it is highly desired. Its pretty much required to meet the peripheral vision requirement tho; either that or a very wide curved screen.

5) eye tracking is used. This is not a show stopper for me. However, with the other requirements above, its pretty much mandatory to reduce the processing required to a reasonable level. There is no way you could render the above at full resolution with current gpus, even with quad gpus you couldn't. Nor are you likely to be able to with the next gen cards. However with eye tracking you might be able to do it with current cards, and probably with next gen cards.

Most of your complaints have already been addressed. The screen door argument holds no ground.

1) 110-degrees of view is a lot more than you likely think it is. You really don't see the sides of the headsets at all. Tunnel vision is not something that you percieve in the current crop of VR headsets. It's there's but you really don't notice it like you would expect to.

2) there is no descernable screen door effect in even the 1080p display used in the PSVR headset. The higher resolution in both the Rift and Vive is even better in that regard, but its not present in any of them really. It's not at all like a the DK2, for a couple reasons.
First, they are all using low persistance OLED displays that just work better than the display in the DK2. And in the case of the PSVR, the screen has three subpixels (Red, Green, Blue) per pixel which makes a massive difference in clarity.
It has a lot to do with the optics being used as well.

3) Latency for VR is a problem that is being actiively worked on by everyone invloved, from the headset makers, to the game developers, to the teams at Nvidia and AMD making the graphics cards and drivers. There are already steps in motion to reduce latency to the levels you expect. Tripple buffering is being bypassed by both AMD and Nvidia so that the latest tracking data is always delivered to the system without worrying about things like tripple buffereing catching up.

4) 2 screen per eye is already the norm. Both Oculus Rift and HTC Vive has one screen per eye. As does the StarVR headset I mentioned before. PSVR is limited to on screen because the console can only output to at 1080p. Both GPU companies are working on techniques to dedicate one GPU to each eye as well.

5) Eye tracking will be a big boost in VR performance, but we're not going to see this happen in 2016. The first round of VR hardware will not have eye tracking. Fove seems to be on track to be one of the first to market with this ability.

 

[clonazepam]

I'm just picking up bits of info where I can. Where does Fove sit as far as patents related to eye tracking? I feel like Oculus and HTC didn't just leave it out for no good reason. There's quite a few patents related to eye tracking, not just how to track eye movement, but defining what to actually do with that information. Sony and Google, as well as at least a dozen others have just this one aspect patented left and right.

I saw Apple might be very weak on this front, but might have a bargaining chip or two.

I'd actually love to read a more thorough analysis of the existing patents, how they might relate to one another, and how this aspect may shape the future of VR.

I appreciate your activity in comments, Kevin Carbotte.

 

[Alodar]

Assuming (sort of a large assumption) that the following article is true, it already exists:

http://uploadvr.com/smi-eye-tracking-foveated-rendering-exclusive/

 

[alidan]

I was disapointed when I got to know that neither Oculus nor Vive were going to use Eye tracking on their first models. I'm much more interested in fove. After the annouced 600USD for Oculus I may be considering waiting more for the launch of Fove.
Not to mention it has a higher resolution.

Eye tracking is an essential part to make the immersion in VR much more efficient, specially the improved resolution only on the area where you are looking, this would allow true 4k, or even bigger resolution displays being viable even with gpu's as weak as a GTX960, although, a high VRAM for 4k may still be needed.

My only concern is how compatible different brands of HMD VRs would work with different games.

i would likely turn eye tracking off, much like i do depth of field, the effect is far to slow and gets in the way more often that it adds to visuals, if i can have it only work in cinematics i do.

 

[Tyler_42]

there's a problem with that VR head set it's covered in plastic infront of your eyes, put some glass in there for a window to virtual reality..... some ones going to walk into a door, off a cliff, into a table, infront of a bus on a long green light.
it's not a matter of if, it's a matter of when if it hasn't happened to one of the developers on lunch breaks already.

It's for home use?

Like in a chair. Or a Virtuix Omni

lol

 

[Adilaris]

A lot of interesting comments, while I 'm here thinking that someone should case-mod one into an Alien Face-hugger.
haha.

 

[kcarbotte]

I'm just picking up bits of info where I can. Where does Fove sit as far as patents related to eye tracking? I feel like Oculus and HTC didn't just leave it out for no good reason. There's quite a few patents related to eye tracking, not just how to track eye movement, but defining what to actually do with that information. Sony and Google, as well as at least a dozen others have just this one aspect patented left and right.

I saw Apple might be very weak on this front, but might have a bargaining chip or two.

I'd actually love to read a more thorough analysis of the existing patents, how they might relate to one another, and how this aspect may shape the future of VR.

I appreciate your activity in comments, Kevin Carbotte.

Fove wasn't really willing to talk much about the technology behind its headset just yet.
There should be more info coming from the company in the near future though. I'll do what I can to dig into that as Fove is willing to talk about it.
I know they are building thier own eye tracking tech. I have no idea what patents they have.

 

[kcarbotte]

I was disapointed when I got to know that neither Oculus nor Vive were going to use Eye tracking on their first models. I'm much more interested in fove. After the annouced 600USD for Oculus I may be considering waiting more for the launch of Fove.
Not to mention it has a higher resolution.

Eye tracking is an essential part to make the immersion in VR much more efficient, specially the improved resolution only on the area where you are looking, this would allow true 4k, or even bigger resolution displays being viable even with gpu's as weak as a GTX960, although, a high VRAM for 4k may still be needed.

My only concern is how compatible different brands of HMD VRs would work with different games.

i would likely turn eye tracking off, much like i do depth of field, the effect is far to slow and gets in the way more often that it adds to visuals, if i can have it only work in cinematics i do.

The effect is not too slow. The whole point is that it changes with your gaze, so where you are focused the graphics are always rendered properly, and where you aren't focused its blurry to improve performance. With the eye tracking turned on the feature becomes impreceptible.

You wouldn't be gaining from turning it off, you'd be hurting performance, which is of the utmost importance when dealing with VR content.