Support NeoGAF

[-SD-]

Oculus Rift DK2 + mounted Leap Motion controller + code magix by Tomáš Mariančík (SightLine: The Chair developer).

Interacting with virtual reality scene using your fingers with Leap Motion
https://www.youtube.com/watch?v=TRhoQ6o4mI8

No gloves or other wearables needed. I really hope the Oculus Rift consumer version comes with tech like this.

 

[oakenhild]

That looks pretty fantastic.

I wonder what the latency is on the leap motion controller? Any lag from what your hands are doing to what you are seeing is going to be pretty obvious and disorienting.

 

[BeatResetsTheWorld]

The leap needs a far larger sensor range. It's basically a cone from the center of the device that extends up about a foot and half "reliably", but that is pretty cool

 

[Krejlooc]

It doesn't work nearly as well as it looks when you're using it. It's using a 3D Scan representation to model your hands very quickly, but the skeletal tracking is just as slow and imprecise as Kinect. Leap Motion's best use with a rift, at the moment, is as a very fast, black and white pass-through camera.

EDIT: They seem to realize this, however, because the next version is including an offset RGB camera so that the 3D Scan passthrough can be in color.

As passthrough video, Leap Motion is far better than any mounted webcam I've tried. It actually fills the entire FOV, instead of appearing as a tiny floating window, and, due to the nature of it's 3D scan, you get stereoscopy for free (and it's not inherently tied to the physical separation of dual webcams).

 

[Durante]

The leap needs a far larger sensor range. It's basically a cone from the center of the device that extends up about a foot and half "reliably", but that is pretty cool

They are already planning a leap version specifically for HMDs.

I'm also quite positive about the tech, although others aren't (see above). It's just so seamless, especially if integrated in the HMD. And software-induced latency issues can be solved in software.

 

[2+2=5]

Cool but hands should be visible by the sensor if i understand correctly, this limits movements i can do with hands, and i suppose it uses lots of CPU.

 

[kadotsu]

Now we only need good haptic gloves.

 

[greyskull]

We need physical feedback though. Otherwise there's a disconnect between what your brain expects and what it gets. That's the tough part.

 

[Krejlooc]

Cool but hands should be visible by the sensor if i understand correctly, this limits movements i can do with hands, and i suppose it uses lots of CPU.

Skeletal tracking does, but 3D Scan does not. 3D Scan is very low latency and doesn't require a lot of your processor. It's a "dumb" operation in that it's just raw depth information, which isn't tracked in any way.

 

[Krejlooc]

Now we only need good haptic gloves.

We need physical feedback though. Otherwise there's a disconnect between what your brain expects and what it gets. That's the tough part.

Dexmo

aiming for $200 for both hands.

 

[Deadstar]

We need physical feedback though. Otherwise there's a disconnect between what your brain expects and what it gets. That's the tough part.

Yes, combine this with gloves that restrict making you feel like you are holding and feeling something.

 

[Krejlooc]

Yes, combine this with gloves that restrict making you feel like you are holding and feeling something.

See above

Mechanical gloves to track finger orientation is a far better solution, in terms of latency and accuracy, than optical/skeletal tracking. I'm very high on 3D Scan tech - I think it's wonderful tech and has a big place in VR in that it's perfect for capturing an entire scene at once, in 3D, with real depth. But not as an input.

 

[Pie and Beans]

I think whatever ends up as the definitive VR control system has to be a hybrid.

VR controls have to accept that for exploring any sort of environment, they will have to make do with an analogue control stick to control your movement through an area since otherwise VR becomes an adventure in being stuck in a room like a blind person bumping into real life walls or investing far too much money into insane omni-directional treadmill systems.

Perhaps have one hand as the free "tracked to the digit" one, the other holding a more traditional motion controller. The sooner the realities are accepted, the faster we can all get through the frustrating "hurling shit at the wall" wild west period :/

The STEM system remains my favourite because of its analogues. Sony needs to hurry up and get those Move controllers with analogues on out to market alongside Morpheus.

 

[Krejlooc]

otherwise VR becomes an adventure in being stuck in a room like a blind person bumping into real life walls

Look into redirected walking. We're considering using it in Half Life 2 VR for those who don't have ODTs.

 

[EVIL]

Wonder if the dexmo tech could be shrunken down to an actual glove (Even tho I hate the idea of having to put on gloves for VR). Maybe with cloth that stiffens when you run an electrical pulse trough it instead of these visible mechanisms at the back of your hand.

Look into redirected walking. We're considering using it in Half Life 2 VR for those who don't have ODTs.

Don't you need a fairly large room for this? assuming you are talking about manipulating the crude human sensory input, making them think they are walking everywhere but actually make them walk circles or patterns within a fixed space.

 

[Krejlooc]

Wonder if the dexmo tech could be shrunken down to an actual glove (Even tho I hate the idea of having to put on gloves for VR). Maybe with cloth that stiffens when you run an electrical pulse trough it instead of these visible mechanisms at the back of your hand.

For the kickstarted version, they've said it'll likely look very similar to the one they're demoing, only it'll be made of metal instead of printed 3D plastic. Right now, apparently, you can crush the thing in your hand it's so fragile.

 

[Krejlooc]

Don't you need a fairly large room for this? assuming you are talking about manipulating the crude human sensory input, making them think they are walking everywhere but actually make them walk circles or patterns within a fixed space.

It can be done in small rooms, but the effect gets better the larger the radius of the circle they walk. I've tried it in a 10' x 10' living room before.

 

[Pie and Beans]

Look into redirected walking. We're considering using it in Half Life 2 VR for those who don't have ODTs.

Once again, the reality of the situation for any sort of baseline safety is going to be establishing an area/zone to play VR in for a long time where no-one can be harmed by your arm movements, and also sacrifice the tracking of your lower limbs and putting that back onto buttons/sticks/touchpads. Hell, in a strange way, VR should come with a seat-belt system for that purpose, like an entire body version of the Wiimote safety straps haha.

As a bleeding edge "isn't this cool?!" tech demo, the omni-pads and redirected walking are fun what-ifs, but market reality has to become the focus at some point.

 

[SerTapTap]

Looks great in theory, but considering the poor reactions I've heard to the Leap motion in general I really doubt it's viability.

I really don't think stuff like this will ever completely replace controllers either, even in VR. consider a game like Mega Man X. Part of the fun is all the kickass stuff you can do in the game with a button press that really wouldn't work if you had to actually act out what you were doing on screen. Stuff like this is very interesting, but at the same time very limiting if you think about it--suddenly your character's abilities are limited to what a human can do sitting down with their hands in full view of a camera.

 

[Krejlooc]

Once again, the reality of the situation for any sort of baseline safety is going to be establishing an area/zone to play VR in for a long time where no-one can be harmed by your arm movements, and also sacrifice the tracking of your lower limbs and putting that back onto buttons/sticks/touchpads. Hell, in a strange way, VR should come with a seat-belt system for that purpose, like an entire body version of the Wiimote safety straps haha.

As a bleeding edge "isn't this cool?!" tech demo, the omni-pads and redirected walking are fun what-ifs, but market reality has to become the focus at some point.

If we're talking about market reality, I am of the belief that using analog sticks to walk will never be a viable option in VR. It will always make people sick, unless we're willing to go down the path of electrically stimulating your cochlea and that sounds way scarier than tasking someone with walking in a circle in the room. Your body needs to physically move when you expect it to, or else you get sick. That's the bottom line.

 

[Turin Turambar]

It only works if you are looking at your hands, so not very useful.

 

[mcz117chief]

still no tactile feedbeack, atleast when I play a Move game I feel something in my hands. I really don't get the hype for this at all, well... I do kind of do but I really don't share it... maybe it could be used for some control panel game or something but I can't really think of anything more than that, the Kinect and Move do the same thing.

 

[EVIL]

It can be done in small rooms, but the effect gets better the larger the radius of the circle they walk. I've tried it in a 10' x 10' living room before.

oh wow that's pretty reasonable. I always imagined that's how the holodeck worked and if you would view the users without the simulation they would be walking in circles and meeting at the right times.

 

[sankt-Antonio]

Now that the topic came up, I think i have spoiled myself in regards to VR. Last week i was at the VR center at the Automotiv company i work for, they have a VR setup that uses a couple of beamers, wich project the VR on the walls around you. I got regular sized glasses, like the ones you get with a 3d tv, just with some added white dots for head tracking. No cables, nearly no weight.

BAM! VR!

I sat in a real frontseat, with the car around me and the best part was, i could still see my own hands/body. That fucked with my brain, because there is no low resolution screen shutting you off from your own body. I actually "griped" the wheel in front of me, and seeing my hands made that just perfect. That was a holodeck kind of feeling. Damn.

 

[Krejlooc]

oh wow that's pretty reasonable. I always imagined that's how the holodeck worked and if you would view the users without the simulation they would be walking in circles and meeting at the right times.

The main problem is you have to trust people to not walk into walls. The entire thing works on the premise that people will intentionally avoid running into scenery, and the illusion is impossible to achieve in open fields where people can run for long distances in any direction. It really only works well in hall ways.

It also requires an external outside-in tracking system to watch where you are in the room, which is the opposite direction of where oculus wants to go.

The problem of locomotion is not even close to being solved, but I'm having a blast exploring it at the moment.

 

[Pie and Beans]

If we're talking about market reality, I am of the belief that using analog sticks to walk will never be a viable option in VR. It will always make people sick, unless we're willing to go down the path of electrically stimulating your cochlea and that sounds way scarier than tasking someone with walking in a circle in the room. Your body needs to physically move when you expect it to, or else you get sick. That's the bottom line.

I think the strafing sense of movement is fine if someone adapts to it over time, but for me personally I feel 'turning' should be controlled like a dial with a circular touch pad and gestures on the "other controller" so to speak.

Basically its thinking about how best to pilot the "robot" you're in the cockpit of rather than completely trying to transfer consciousness although I get the appeal for limited situation stuff.

 

[grandwizard]

I can't believe after people got fooled by Kinect's marketing TWICE we still have people thinking we are ready for the 'you are the controller' motion tracker age. The technology is not there yet. You will end up with poor/lost tracking, bad feedback and condition requirements that are too tight to be working correctly all the time.

What we are ready for is actual working motion tracking with high accuracy that can track hand position, with some buttons for crucial functions and a stick for player movement. The Razer Hydra and PSMove were the best before, the near future will be Sixense's STEM.

 

[Krejlooc]

I think the strafing sense of movement is fine if someone adapts to it over time, but for me personally I feel 'turning' should be controlled like a dial with a circular touch pad and gestures on the "other controller" so to speak.

Basically its thinking about how best to pilot the "robot" you're in the cockpit of rather than completely trying to transfer consciousness although I get the appeal for limited situation stuff.

That's generally the idea everyone is approaching VR with right now - cockpit simulations are the best. I'm personally not satisfied by that, I want a full VR experience, logistics be damned. Which is sort of why we're exploring every single option at the moment, because it's easy to sit down and rattle off a list of drawbacks and positives for each and every one.

 

[wmlk]

The Leap Motion is really fantastic. I've seen some people do some awesome things with it in a short time at Hack the North.

 

[OnionPowder]

These advancements in technology are so fucking amazing. In 10 years it will be crazy.

 

[Teremap]

Dexmo

aiming for $200 for both hands.

Between this and the OP tech there's just sooo many reasons to be excited for VR right now.

The future is gonna be wild!

 

[-SD-]

It only works if you are looking at your hands, so not very useful.

Isn't that just a question of adding more sensors and broadening the scanning area?

 

[Krejlooc]

Isn't that just a question of adding more sensors and broadening the scanning area?

It will always require a line of sight, and the sort of positional tracking being suggested would require a ring of these around your head, 360 degrees.

 

[Canis lupus]

And get tired in 15 minutes

 

[Krejlooc]

And get tired in 15 minutes

That's a concern of UX/UI design, not the technology for positional tracking. It's not hard to conceive UX/UI designs that focus on tracking our hands and fingers that could be far more convenient than physically craning your hands around to manipulate objects.

Think broader. Imagine, essentially, telekinesis.