Leap Motion sets a course for VR | Hacker News Reader

Leap Motion sets a course for VR

community.leapmotion.com

56 pointsStavrosK33 comments

I've used the leap motion, and so far the accuracy is nowhere near what they show in the video. simply forming the index finger and thumb into an O-shape was enough for them to disappear from existence.

If this is an active demo of their next generation software, I'm very impressed with the improvements, but I suspect its a mock-up.

Nope, it's not a mockup, that's what the new SDK looks like. They really improved it.

Source: I have a Leap.

Yeah, this was probably the early version of leap. The new SDK came out this May and it's actually improved quite a bit since then.

For a hacknight back in April, I tried mapping my arm into a 3D pointing device that recognized a recoil wrist gesture as firing a gun. Sound effects like "pew pew pew" were optional but heavily encouraged.

I initially tried the Leap Motion mounted on the Oculus Rift. As this blog notes, the top-down recognition was iffy at best but things really went to shit the moment the sensor was moved. I was unable to get even gun orientation working in the 6 hours or so that I tried, so maybe the new SDK + sensors can fix that.

I did end up building my demo, but with a alpha Thalmic Myo instead of the Leap. Video: https://www.youtube.com/watch?v=lihaqf3COw8

Shark Punch is an example compelling game design using this type of tech. Hahaha. http://www.rockpapershotgun.com/2014/03/10/shark-punch-oculu...

They had a hacky demo of this at the Leap booth at Unite in Seattle last week. The Leap Motion was taped to the front of the Oculus, and you could (sort of) interact with a floating sphere. It's a neat idea, but the tech has a long way to go if that demo was anything to go by.

I am excited for the advancement of input relating to VR though.

Given that the dk2 uses IR LEDs as markers to augment it's gyro/accelerometer for position tracking (by imaging them with a webcam pointed at the user). The leap dumping IR out of the front of the visor is going to be problematic, a combination of washing out the DK2 tracking camera and messing with the tracking algorithm by adding extra points.

Though maybe the frequency of the IR LEDs is different enough...

For IR LEDs you have not only the wavelength of light, you also have a repitition rate. Typically you modulate the IR LED with a square-wave in the 10s of kHz; this lets you have a much higher instantaneous power with a much lower average power. If done right, it also means that 2 IRs with a sufficiently different clock-rate will interfere with each other less.

Based on our experience so far with the latest Oculus SDK, mounting a Leap Motion Controller onto a DK2 does not effect positional tracking, since the DK2 was designed to have more LEDs than needed for input experiments such as this.

I've been a developing with leap for a while now and have been waiting for something like this to come out for me to justify buying an oculus.

Anyone know if Oculus or Leap developers are making any decent profit these days?

Oculus has sold like 100K headsets at most. They raised 75M in funding then got bought for $2B. Definitely not profitable if you consider money invested to profit made.

Another metric might be burn rate to income. 100000 units x $350 = $35M, so maybe to this point they haven't spend that $35M. Something feels off about this. Oh right, cost of devices. So it's burn rate to profit. 100000 units x $50 = $5M. They may have spent less than $5M so far, but I'd be surprised.

I think "Oculus or Leap developers" was referring to developers making software that uses Oculus or Leap, not the companies producing the hardware. I also assume that the answer must be no (for now) since the install base is still quite small.

I don't think I've ever heard of a /hardware/ company pivoting like this. Pretty cool.

It wouldn't call it a pivot. Same hardware but with an extra attachment. And a software adjustment to handle the new positioning.

It's not really a pivot: it's still tracking hand gestures and it's even validating one of the ways the device was used by customers, which is great because natural interactions are one of the big upcoming battlefields of VR.

This is a killer peripheral for Oculus.

Sadly no one's figured out how to adapt the DK1 mount to fit neatly on a DK2. I have a very nice holder printed off of Thingiverse, but it's a little tricky figuring out how exactly to get it onto the Oculus.

I really hope Oculus actually add some mounting threads or something to the front of the visor for this type of thing in the CV1.

Because of the IR marker LEDs I'd be hesitant to strap a leap on there considering it's dumping IR light out of it to illuminate your hands.

The sync cable should mitigate that. I haven't heard of people having problems when they try it.

Our VR Developer Mount is compatible with both DK1 and DK2. Additionally, we've made the files available on Thingiverse if you want to DIY our version: http://www.thingiverse.com/thing:445866

The field of view makes me nervous. But wow, I hope they pull this off.

Yeah, my first thought was "Does it work if you're not looking at your hands?" Not a deal breaker though. I could see this combined with some hand mounted sensors to track out-of-view gross movement and just augment with the fine movement when in view.

EDIT: found this in the article: "It possesses greater-than-HD image resolution, color and infrared imagery, and a significantly larger field of view."

I imagine that without haptic feedback, so far it makes more sense to be looking at your hands while you're using them in VR. But I had the same thoughts -- would it be better to have a setup where the Leap Motion is on a pendant or something so it's mounted on your chest?

Even given that is has a wider than visual FOV I think you're onto something. The Oculus DK2 already has an IR camera for tracking the rift in three space. Why wouldn't they partner to extend the Leap functionality using the existing camera?

Edit: Spelling

I like their approach for simple apps where you are just working with your hands in front of you. But definitely optical tracking tech can't compete with exoskeleton like gloves and other sensor approaches. Curious if people think otherwise, optical tracking without many markers and many trackers just hasn't been done well in any product that I've come across, save motion capture studios.

I've never seen a truly useful application of hand gestures to control a PC. Great for short demos, though.

Agreed, the only thing that's been done sort of well is full body controls in MS Kinect. But it's only really the dancing games that I've found compelling. Most gamers would rather just play with a controller for most game genres they already enjoy.

From one side yes, it makes me nervous. It doesn't represent the "ideal" VR case where your hands can be detected no matter where they are.

That being said, for what they are aiming to achieve, having the FOV be just larger than that of the Oculus is a huge win. The LEAP takes a few frames to detect a hand when it comes into range, and another few to get the skeletal tracking right. By having a slightly larger FOV than the Oculus, that detection step can happen outside of the user's FOV.

The article says that they are working on new hardware that will have a much larger fov and higher image quality. This current tech demo is using the existing hardware.

They will be bought by FB in less than a year..

But for how much ?

I will bet you $100 to the charity of the winners selection, that they won't.

Cool, but still pretty useless. I'd like to see leap tackle a useful situation, perhaps commercial. I can think of dozens off the top of my head and none involve a consumer controlling a PC interface for which we already have multiple superior options.

Reply on news.ycombinator.com