Hacker News new | past | comments | ask | show | jobs | submit login
An Iron Man Like 3D Hologram Controlled by Leap Motion and Three.js (badassjs.com)
159 points by acrum on March 7, 2013 | hide | past | favorite | 38 comments



As cool as this is, it's not really a hologram. This is an example of Pepper's Ghost, which is also what's used to project images into your field of vision using Google Glass.


Classic HN, pedantic comments about this not being a hologram.

You know what it is? It's a really cool demo of controlling a perceptually 3D image using a motion controller. Let's have, say, a discussion about the LeapMotion[1] (which is awesome, I have one) or the fact that it's using Javascript (this looks like it is using the packaged WebGL demo[2]).

The LeapMotion is quite nice to use with WebGL / three.js.

[1] https://github.com/leapmotion/leapjs

[2] https://www.leapmotion.com/


It's not pedantic because it has a material effect on what we're lead to believe this is. This is hacker news. A hacker knows (and cares about) the difference between a "3D hologram" and Pepper's Ghost illusion, and that distinction is important in understanding what has been achieved.

This is definitely a cool hack, but it does not involve 3-dimensional projection. This is a 2D image cast on to a 2D projection plane with no stereoscopic effect. That makes the term "3D hologram" a misnomer at best, and a gross misrepresentation at worst.

Because of this mislabeling, we're all sitting around talking about the inaccuracy of the title, rather than the fun part. That's the fault of the person who titled the article, not the person pointing it out.


Too late to edit my top comment, so I shall reply here: I apologize for perpetuating the problem, because now there's several top comments discussing how parent was actually correct, and the real discussion is yet again derailed.


The first thing he said is that it's cool. And he's right, it's not a hologram; some of us care about that, and would really like to see actual holograms. It's not an attempt to take away from the achievement here, rather a desire not to exaggerate it.


Down-voted your comment because you are confusing the truth with being pedantic.

This demo is neat. No question about it. I like when people play with ideas like this.

The demo isn't any more a hologram or 3D than if I took four LCD's, arranged them as the sides of a cube and turned down the lights. Other than not seeing through the LCD's, the images would be the same.

Again, as the person you are accusing of being pedantic said, this is no hologram and it isn't 3D in almost any imaginable way.

Notice how he moves from angle to angle quickly with the camera? That's because there are no volumetrics here. It's just a flat 2D image effectively seen through a 45 degree partial mirror.

It's a 2D image that appears to float in space. And it's cool.

Many, many years ago I thought about the idea of installing a screen spinning or translating (harder mechanically, easier imaging) at high speed in a vacuum chamber with transparent walls and then using lasers to project slices of a true 3D scene onto it. Challenging, but it would be really interesting to see how something like that works out.


It's not even "perceptually 3D" though. It's a 3D rendering projected in 2D.


How did you get a leap motion? What do you think of it?


I'm not ajacksified, but I have a Leap, which I got by joining the developer program. I love it.

I've been writing about it at http://leaphacking.com

Leap has announced that the devices will be available to the public on May 13.

I've been hacking around using JRuby and JavaScript. It's quite nice to work with, and so far my biggest issues are thinking through just how to handle user movement (i.e. what hand/finger combinations and movement should equate to what action).

It's a nice compliment to the XBox Kinect, too.

Gesture recognition has been added so I need to spend a bit of time with that today.


I was lucky enough to give a presentation at an Intel "Perceptual Computing" meetup (their version of the LeapMotion), and I met a dev advocate from LeapMotion.

I think it's very interesting; the API is solid, and it's really easy to use with Javascript because you can interface with it directly through a Websocket server. That, plus three.js, means you can do some interesting things with examining 3D models. It isn't a far stretch to see medical and industrial applications, as well as interesting games (Infinity Blade, laptop edition?)

There are a few companies working on this at once[1]; I think it's definitely something to keep an eye on. It seems like a natural evolution of webcams.

[1]http://news.ycombinator.com/item?id=5339266


Honestly, part of why I come here is for instant debunking of hype. So I appreciated it.


I was confused for a second because Pepper Potts is the love interest in Iron Man. But you meant https://en.wikipedia.org/wiki/Pepper%27s_ghost (which was also the technique used for Tupac's re-animation at Coachella last year).


I loved how the public went crazy over a technology "invented" in 1873 and how people where like "finally holograms".


Someone could make a better version where the Pepper's Ghost display is an autostereoscopic stereo 3D display that uses a either a parallax barrier or lenticular lenses.

http://en.wikipedia.org/wiki/Parallax_barrier

As long as the user lines up with the center line properly, the image would be both "ghostly" as here and also stereo 3D.


What are the requirements to define something as a hologram?


Holography entails the diffractive reconstruction of the wavefront corresponding to a 3-D scene, and optical holograms are generated using the coherent interference of light scattered from a 3-D scene with that from a second reference beam. Holography, by definition, necessitates scene reconstruction via diffraction. In the case of an electro-holographic display, diffraction patterns corresponding to 3-D scenes are computed and displayed on suitable light modulators to achieve wavefront reconstruction. This technology is still in its infancy, so you probably won't see anything truly holographic in the marketplace for some number of years.

As cool as this demonstration is, it's using a 19th-century technique -- "Pepper's Ghost" -- and cannot correctly be called holographic.

Here's work out of MIT depicting the true state-of-the-art in real-time electroholography: http://web.mit.edu/newsoffice/2011/video-holography-0124.htm....


Today's average 3D displays are "stereoscopic": they have 2 views, for the left and right eye. A real hologram shows the whole object, so multiple people can walk around it and see it from every angle at once.

But this isn't even stereoscopic. This is just a 2D image reflected on a piece of glass.


I think you need elements of light interference and diffraction for something to be a hologram.

It's essentially a way to "record" light the same way you can record sound in grooves or magnetic tape.


This is solely a 2D image reflected in glass. The control via finger is novel, but there is no volumetric or holographic effect whatsoever.


That's not the actual real Earth either. We totally got him!


This is one of the coolest things I have seen. I hope that this is the direction that user interfaces goes in down the road, it is much more exciting than 2D.


... and much less efficient/productive.


Well, like anything different tasks will call for different UI's. After looking at the OPs demo I could immediately see this being used with Radiology Imaging systems.


But it is 2D.


This is pretty awesome. This is the first time I've seen something being projected into a prism to generate this hologram effect. Is this a new trend?


It looks like it's being reflected off an inverted four-sided hollow (perspex?) pyramid, rather than projected into a prism, but still a neat idea; sort of like the inverse of the (formerly Microsoft) Polycom 360 degree conference camera.[1]

I'm curious what it looks like viewed from one of the edges - I can't picture it being seamless?

[1] http://www.polycom.com/products-services/products-for-micros...


He moves the camera to another side across the edge in the video- it's far from seamless, it's exactly as you would expect- two images in perspective.


It is a few years old, and it's not a hologram (not 3d).


While its cool, all this has been done with kinekt.

https://github.com/doug/depthjs


this is amazing. the future of the interface indeed. forget iron man this is more like star wars. but more seriously, i'm very interested to see what comes next with this in terms of interaction. i'm not sure how long lived the LEAP device will be, but this sort of holographic interaction technology certainly has a big future!


The device itself is pretty solid, and there are several other companies working on 3D motion detection (such as Intel[1], who has the lofty aim of replacing single webcams with dual webcams.) I think this definitely has a shot at augmenting computer interaction.

[1] http://software.intel.com/sites/default/files/article/325946...


So seeing this, I'm wondering, could we get a "continuous" 3d projection with a cone-like shape instead of 4 planes?


The following is a fake video of a hologram gaming console. I hope it will inspire someone to make it real: http://www.youtube.com/watch?v=iaaHcH9nQmI


Since we've got the concept, we're 90% of the way there, right? I mean, all that's left is the implementation?


Correct.


Yes, those remaining 10% would be the other 90%.



I actually laughed out loud at how incredibly clever and awesome that is. This is such a great idea and kudos to the person who designed this!




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: