Leap Motion: Orion [video] | Hacker News Reader

Leap Motion: Orion [video]

youtube.com

135 pointstiagobraw32 comments

I've just been testing this out.

It works.

Not perfectly, for me, so far - it's a little wobbly, and I heartily recommend clearing your desk before beginning. With a headset-mounted Leap the close distance isn't particularly close, and it DOES work fine at arm's length. Unless what's sitting at arm's length is an expensive condenser mic, in which case you will a) fail to grab the block you're reaching for and b) punch your microphone.

Also, calibration is about as much fun as it is with most optical devices. For anyone who has never had that particular joy - it's not much fun. Tip if you're doing this - rotate the Leap, don't just move it around. Also, I had to iterate through about three allegedly reflective surfaces to get one that worked - ironically enough I ended up using the screen of my Surface Pro.

But it's orders of magnitude better than the last time I tried the Leap, and I suspect with a darker room it'd work even better. No latency I could feel, and I could throw blocks around and build towers in VR fine.

Very impressed indeed.

This evening I'm going to give it another go once the irritatingly persistent IR emitter in the sky goes away, and I'll also be trying their Warlock Battle game, which looks like lots of fun...

The tracking is indeed significantly better post-sundown. I was actually getting some presence in VR. Very impressive.

Is it better than the old LEAP tracking?

I want to like the LEAP, but it's been consistently mildly disappointing.

Yes, significantly.

I was in the same boat. The new tracking's much better.

Doesn't seem to improve all of the old apps, mind. I tried their 3D modelling app with the new Orion software and it wasn't much cop. But for VR it works well.

Is this "just" new software compatible with current devices or new hardware altogether?

I've got an original Leap and it worked fine for me.

It makes reference to needing a mount for the VR attachment, but elsewhere on their site they teach you how to DIY the mount with duct tape :)

> I've got an original Leap and it worked fine for me.

Interesting. What did you use it for? I got one and couldn't find a single useful application.

Sorry, poor contextual phrasing.

I have an original Leap. When I tested it with the new software they've just released, it still worked fine. The new software doesn't require new Leap hardware, and will work with the original hardware.

As for what I used it for when I got it - I bought it for motion capture. It wasn't good enough. The new "Orion" software, on the other hand, might be.

My understanding is both, but the new hardware isn't a consumer device, it's a component meant for integration in a VR headset.

While it's very neat to see these Johnny-Mnemonic style demos, what I'd really like is a realistic replacement for Apple's Magic Trackpad. I'd buy one if it were able to replace my trackpad on an ordinary desktop environment. I don't need to manipulate cubes in 3d space (who does?!) outside of video games. But I'd love to replace my flat trackpad with a skeletal gesture motion-tracker for more mundane tasks.

May I ask what exactly are you looking for in the replacement?

We are actually building a wearable that gives you full hand tracking without the need for line of sight. It's intended for AR/VR applications since it's just as precise as vision based products without all the downsides. But one thing that's personally important to me that we are incorporating into the product is the ability to turn any surface I touch into a trackpad. It basically gives me a portable trackpad anywhere I go and I can sit 10 feet away from my computer and use my leg or arms of the chair to control my computer. I'm curious what kind of usages you need from a product like this.

Well, I don't want to turn arbitrary surfaces into trackpads. I'd just like to be able to perform hand gestures in the air, above my keyboard. I don't need to perform them from across the room or anything like that. So that, to me, is the MVP- simply replacing my trackpad for gestures like pointing, clicking, dragging, zooming in MacOS.

But beyond that...

1. I don't want to wear anything on my hands.

2. I'd like to be able to program new gestures easily. Sort of like American sign language, I want to easily map a gesture to an action.

3. After replacing a trackpad, I'd want to be able to use my hands to manipulate a 3d environment. This requirement is pretty low on my list, because I never do anything in 3D. But it's necessary in the long run, once 3D applications become popular.

Have you tried with BetterTouchTool? Already was cool for gestures, has support for Leap as well:

https://apps.leapmotion.com/apps/bettertouchtool/osx

Exactly. Something replacing the mouse/trackpad I use for hours a day would have real value for me, too.

Have you heard of Ractiv? It's a motion controller like a leap, but allows you to treat your keyboard surface like a track pad. So you can stay on the keyboard all the time. I've been trying to buy one for a year but they never reply so I don't know if they are an active business anymore. Ractiv.com

http://gharbi.me/ractiv/index.php

Have to say this amazed me that they are still selling a four year old device and evolving it by upgrading software.

Glad to see they improved the software. Here's a video someone uploaded https://www.youtube.com/watch?v=nhBaY1UMOJs

I'm interested in ways devices like this could be used for everyday work, automate some IDE task / editor macros. Subtle Gestures could add another dimension to data input without leaving the keyboard.

There was a slightly worrying bit of latency visible here:

https://youtu.be/rnlCGw-0R8g?t=32

Much of that is the projector. The experience within the headset itself is much tighter, less than 35 ms from end to end. (Our tech accounts for 12-13 milliseconds, while a 90 Hz VR display and the GPU add about another 22.)

While the experience of latency varies from person to person, we've found that this level generally doesn't register on human perception. Head tracking needs to be less than 20 ms because you're relying on both your vestibular system AND your eyes to be in sync. If they're not in sync, your brain will interpret this as you being ill, and you will feel sick. With your hands, you only really have your sense of proprioception, so the latency restriction is lighter.

I've had a Leap motion since very early on, and it's been working very well. Especially the software update where they upgraded to a skeletal hand model has made many glitches go away. Both latency and accuracy have been great, so maybe the latency there is because of the animation? I'm not quite sure, but based on how good my experience has been with the first Leap, I doubt this one is worse.

Here's another video: GEOMETRIC - https://www.youtube.com/watch?v=RLEEVGxu_k8

A while back (maybe a year or so) I played with the early alpha/beta support for the whole "stick your Leap on the front of your Rift" thing. It was really cool and seemed like something I'd expect to see in future VR headsets.

I'll have to check out this new software since it seems they've continued to develop it.

The tracking is an order of magnitude better, using the original device. Impressed.

What is happening on the Augmented Reality technology that Leap was originally showing?

I am much more interested in Augmented then Virtual and was hoping to see Leap move more in that area. I haven't seen anything since July 2015.

You're probably thinking of Magic Leap, which is a completely different company from Leap Motion.

Your 100% right I never realized they were two different companies.

Well that seems like a big trademark problem.

Both formed in 2010. Leap Motion is also doing some Augmented Reality. Hard to fight the over $1,300,000,000 in funding that Magic Leap has been able to raise.

Meta2 is launching in 2 weeks and just had its first public demo on stage today at TED 2016. Maybe that has something to do with it. Links:

- https://www.metavision.com - http://www.roadtovr.com/meta-shows-off-their-next-gen-ar-gla...

The demo shows intricate and precise hand and finger joint tracking, seemingly knowing precisely how both hands are positioned in a small volume of space. Maybe we may soon finally be able to push past the current non-vocal computer input speed barrier imposed by physically typing on a keyboard. The spoken word is roughly around 225wpm[1], while few can claim accurate typing speeds of more than 150wpm for extended periods of time.

Example of smaller motor movements that can be performed are finger twitches or trilling piano keys.

[1] https://en.wikipedia.org/wiki/Words_per_minute#Stenotype

Reply on news.ycombinator.com