DK2 Development kit

"jherico" wrote:
My impression is that they tried it and it didn't work out for the reasons I mentioned. Perhaps the issues can be overcome eventually, but I suspect that slide is from before testing or early in the testing. On the other hand it's possible they're just not saying they've solved the problem because of competition.

I'm optimistically voting for the latter in this case- They've really got some super-brains on the team. Currently the demand for AR is pretty hot- and the bar is just set so freaking low.. It's awfully pathetic what companies claim for AR experiences these days. Even if it's initially somewhat gimicky- it is something that would evolve with processor power and software. There's a real need for those stereo cameras even if its just to push a button to see out of the goggles for a sec to type on the keyboard.

"jherico" wrote:

Sure, there's less motion to interacting with a nearby game board than walking around through a convention hall, but that's no guarantee that the problem of mismatch is less. In fact, it's entirely likely to be more. If you're walking down a hallway your brain is probably putting fewer constraints on how it expects the world to move than if you're using fine motor control to make tiny head movements.

Maybe.. I think the real variable is if there's clear tracking points of interest. Remember, it can also be combined with head tracking data and rotation in real-time so there's a variety of inputs to derive a solution..

Come to think of it and this is random speculation, lets say for a static scene at your desk-

once you get an acceptable 'lock'? (place a box perfectly on your desk so it looks right) then you could pretty much rely on JUST head tracking + rotation data which is already close to real-time.

hot-damn mamacita-

"jherico" wrote:

"Tgaud" wrote:
I believed that stereo camera weren't to restitute reality, but to make a positional tracking by analyzing the changes of the images captured

Doing that, particularly in realtime, with arbitrary real world imagery, is an extremely hard problem in computer vision.

Consider: the original Kinect uses distortions in a projected known pattern of light it projects to construct 3D features. The XBox One version uses a time of flight camera to do the same. If these devices could use dual color cameras to reconstruct 3D features they would, because it would be much cheaper in terms of hardware.

its like a laser mouse. It would just detecte movement.
But we would have two camera because we need to take an appréciation of the distance.
so with stereoscopy it becomes easy to triangulate the position of an object to know its distance.
and when you know the distance you can know your movment thanx to a reverse parrallax calculus.

So, The next product will be the consumer Version ?
Oculus, need to out a DK2 with news specs, they can't launch the CK without more test / reviews
The DK 1 look like an Alpha Test
The DK 2 a Beta Test
and the CK the Final version.

I think you can glean a lot about where they may be going from the descriptions in the job openings. Looks to me like they are trying to incorporate a nice low level FPGA/ASIC implementation of computer vision. I would think it highly probable that they are aiming for a camera integrated commercial release but its not entirely clear if it would be for pupil tracking, augmented reality or both. Many job listings are highly slanted towards individuals with strong backgrounds with cameras even for positions where you would not generally expect those skills to be required. Obviously pupil tracking would be handled on device but from the skills they are looking for I think even an augmented reality integration would have most of its higher end stuff handled on board. This is even more the case when you consider that augmented reality is more useful if your mobile, they are developing for android devices and android devices won't have excess power to do the low latency vision integrations that are necessary. The LCD will probably be driven with two sources blended together on the device with something like an alpha channel. Visual information will also be backfeed to the developer-software but it won't be fully integrated in application level rendering. This type of low level onboard augmented reality implementation would allow them to eliminate the 'lag' that is observed when people do things such as attaching a webcam to the rift and walk around drunk.

Look forward to entirely new features in the next version not just stepped up resolution and positional tracking. And I'd bet we are looking at a commercial release later in Q3 or 4 next year.

We may also see the rift skip over 1080 altogether. LG has a 5.5in screen at 2560x1440 that could be a nice match. Pretty sure they have not settled on anything yet. "As a Senior Display Engineer at Oculus, you'll design Oculus's display systems from top to bottom."

The internet has grumblings of sony working on a lower cost head tracking version of the HMZ-T2 http://www.amazon.com/Personal-Viewer-Sony-Hmzt2-Import/dp/B005SCYZGA to go with the PS4 that would be similar to the rift. I for one welcome the competition.

Tbh I'm not much into game development at all, but I'm pretty hyped by the oculus rift even in its early stage. All the demo-videos I've seen these days made up my mind and I wanted to order my rift this weekend.

I'm quite irritated after reading the latest rumours about a dk2. Now I don't know anymore. I don't mind, if the consumer version will be released 2014 or 2015. It's done when it's done.
I'm clearly sure, I don't like to wait till winter 2014/15 and go with a developer kit. But I would be very annoyed, if there's a new upgrade available right after placing my order today... doh... :)
Can nobody clarify this?
Excuse my bad english

They have stated repeatedly that will be no DK2. Despite that old presentation from a while back, in all of their recent interviews, including ones just two weeks back, they (Palmer, Nate) stated the following:

1) there will be know DK2. It is economically infeasible at the present time.
2) They are concentrating on working towards the consumer release which is "multiples of months, not multiples of years away." Which means we won't have to wait 2 years for a consumer rift.

The DK2 nonsense won't die no matter how many times they are quoted. So I am going to make some speculation of my own.

They have basically solved the latency problem, from graphical and six degree of rotation tracking, anyway.

1) they can solve positional head tracking in any of a number of ways. But they have to keep the costs low, latency of this tracking low and create a solution that won't suddenly break when everyone uses their Hydra, Sexense STEM, or General Haptics controllers (with STEM.) So I bet there is some work to do there. Mainly it is yet another sensor and from what I have seen with Hydra's even one of these will now require some form of base unit (maybe added to the box where you plug plug in HDMI and USB connections to and from.) Balancing costs and reliability is important here because if you get cost wrong the unit suddenly costs 400$ not 300$ and if you get positional tacking wrong (drift/latency, etc) you get more simulation sickness and who wants that for a consumer version.

2) They have to pick just the right final screen. There are a lot of new screens coming out. 1200 resolution in the form factor they use (or close to it) is out. Who is to say that 1400+ or 1600+ won't be? The differences on these levels is hugely important. Whatever they first launch with is what the world is going to see and think VR is good or not with. You, I and the other enthusiasts would probably be okay with 1080p at this point. Notice the speculation, mostly wishing, that there would be a 1080p DK2. Better to wait just a bit longer while they tackle the other issues cleanly and then select the best screen at the last possible moment. If I were them, I'd be testing many kinds of suitable screens and make a decision on none of them until the other parts were solved. Every quarter brings something new in the screen department. Pull the trigger when everything else is ready and you have to. Don't lock yourself into a lower resolution just yet.

3) Drivers, API, etc. There are still some areas of this to nail down. Whatever they do for positional will have to be added to the API. The final resolution rendering will have to be optimized (to avoid scaling, tearing, jutter, etc.). They need to streamline the APIs in other words. Thankfully I think with Carmack leading the way they will get this done. It isn't an overstatement to say in this area, the man is a genius.

4) Software. At launch they will want dozens of playable games. Sure we have demos, and refits and things we can hack together that will keep enthusiasts going. What the paying public wants are things they load up, strap on their shiny new rifts and play, see and experience. They don't want to have to tinker with FOV settings, drivers, etc. The games are coming. VR JAM and share are helping. Everyone's experiments are helping. Some of those experiments are going to be full on games.

5) X factors, product pipelines (distributors, point of sales decisions, etc.) They have to decide how to approach a launch. Get the right endorsements, ads, branding, etc. Sure their word of mouth is freaking awesome at this point but a good marketing campaign can make the difference between a slow to grow sleeper hit, or an out of the park grand slam hit. They have to tool up for the final form. You see how hard it is to create and ship 25 to 30K DK1 units (approx. so far.) Now multiply that by 10 for even a modest launch or a 100 for a huge launch. These things take time.

TL;DR: These things take time. All the problems are solved, or solvable in the near term. Give 'em time. The sweet spot is coming. And it will be sweet, I am a believer and I still have to wait for Christmas for my DK1 unit (don't ask.)

Best guess: (and just a guess):
This time next year I'd bet we are either just getting our consumer rifts, hearing about the upcoming launch or at most a few months away from it.

- Cal

Tada: [http://www.engadget.com/2013/10/18/oculus-rift-john-carmack-interview/]

Oculus Rift's new Chief Technology Officer John Carmack says a new Oculus Rift dev kit is in the works and will arrive in developers' hands before the retail unit ships some time next year.

Thanks Mr. Carmack (and thanks again for all the nice abuse from some strangely unrelaxed people here).

spyro

"spyro" wrote:
Tada: [http://www.engadget.com/2013/10/18/oculus-rift-john-carmack-interview/]

Oculus Rift's new Chief Technology Officer John Carmack says a new Oculus Rift dev kit is in the works and will arrive in developers' hands before the retail unit ships some time next year.

Thanks Mr. Carmack (and thanks again for all the nice abuse from some strangely unrelaxed people here).

spyro

It is possible Carmack may be referring to a dev kit made for the android version of the oculus rift, which we are lead to believe will be different (in terms of compatibility) than the PC version. This would make sense considering the project he is working on in is with the mobile / phone integration side of the oculus.

Disclaimer: At this point all the information surrounding a 2nd dev kit is incomplete, Palmer has yet to make any official announcements, hoping for the best though!

Palmer did say this much though:

We're working on a lot of new tech for the consumer product, we want to ship a new development kit before the consumer version launches. That way, developers can build and test their games on the nearly final hardware that users will be playing at launch. The timing of that dev kit is tightly tied to our progress on these new features.

We are not going to formally announce a new development kit or the consumer version anytime this year. Also, we're working to ensure that content built using the current Rift development kit is compatible with new Oculus hardware, though there will be a bit of integration required to take advantage of the new features, especially for the best experience.

http://www.reddit.com/r/oculus/comments/1oq43j/john_carmack_confirms_new_dev_kit_incoming/

Well in a two month span they went from saying that there would be no DK2 because it would be economically unfeasible to we will put out a DK2 sometime near the release of the CV1. We can only go by what they tell us Spyro. Fair enough?

- Michael T.