A multi-user hologram table | Hacker News Reader

A multi-user hologram table

newatlas.com

32 pointsel_duderino16 comments

I have little faith in this company or approach to holograms. Euclideon has shown off their voxel based renderer for years as something that will revolutionize gaming and real-time production as we know it - but it has done no such thing. It's certainly interesting tech.

The cheesy, clearly photoshopped teaser images here do nothing to inspire confidence - and based on the technology they described - I don't believe it would work the way it looks in the image. They were vague on the details of the implementation - but from what I understand, it works like this. The table has two images projected onto it per person - each in a different wavelength - and the glasses filter out other images. Thus you have stereoscopy per person. I'm not sure if this approach could make color holograms as shown - but I'll give them the benefit of the doubt.

What this won't allow is holograms which protrude from the table. You can make stuff go inside of the table - or seem like it's coming out when you are directly above - but not hover over it as in the teaser images.

They also don't talk about the positional tracking technology other than:

> The computer knows where the glasses are. The glasses have some boxes with little microchips and microcomputers that we build on the side of them.

I'm still wondering where the money is coming from. I would also like to play around for years with this kind of technology.

> Dell hopes that his tables will be used in workplaces for city planning as well as arcades.

As someone with a background in urban planning, these kinds of statements always convince me that nobody knows what urban planning entails.

Funny enough, at community meetings, many planners like to avoid realistic renderings of changes. For example if you want to give people an idea of what sort of size buildings could have under a proposed upzone, you don't want them fixated on the color of the brick facade in the rendering.

I can't really imagine a holographic display helping planning in any real sense beyond getting school kids excited.

It could be used for that exactly, getting the public interested in new development projects. Not sure it would be worth the cost though.

I had worked on earlier model of Microsoft surface table. They had the same pitch that these multi user multi touch devices will be used by education and banking systems etc. they don't work simply because its not solving any problem. Why would people go out of their way to learn new ways of designing, developing and communication just so they can share their work in hologram? What does it solve? Until this is answered AR and VR are not going anywhere.

AR has huge business potential in any job that requires use of both of your hands, especially for tasks where you’re moving around and don’t want to be tied to a monitor.

I agree with you when it comes to jobs that are currently done with monitors on desks.

Is it too late to rename these various stereoscopic systems to something other than 'holograms' ? Because they're not holograms.

Holograms work by interference and have the property that when cut into pieces each piece is a (dimmer and narrower FOV) copy of the whole.

Stereoscopic systems could be called 'stereograms' ?

The new startup marketting language...

Holograms = Augmented Reality

Hover Board = Self balancing wheeled vehicle

AI = algorithms, sometimes big data, rarely machine learning or neural networks

At least in the case of Microsoft Hololens, the name is somewhat justified because the display tech goes beyond stereoscopy and actually projects a light field using waveguides.

Some details on the tech here:

https://forums.hololens.com/discussion/903/how-does-waveguid...

Easy way to tell a faked 3D display: look for imagery lying along a line of sight that doesn't intersect the display medium. For example, the dinosaur's head and the tops of buildings stick out above the display's edges as seen from the viewer's location in the video.

Objects may be in front or behind the display's surface, but the display medium has to lie in front or behind every part of the object.

For all practical purposes, photons travel in straight lines, and require some medium to create or redirect them. It is extremely unlikely that anyone will suddenly discover an easy way to change that fact.

So as I understand correctly this is the same as with 3D movies in cinema, but now the computer knows where each glasses are and is changing the phase of the light for each user.

Interesting tech and although the article is full of shopped images and a video that doesn't tell a thing, this might actually work.

Holograms are just a terrible idea. I can see the use for AR in business, services, and education. But as far as we can throw it, it's pretty useless compared to video.

Total Vaporware. What does "Theoretical Atoms of Light" mean? and what are "Small slip on sunglasses"?

Euclideon has a long history in this space of making relatively scammy claims out of technology that may seem far out to the average person, but is actually fairly bog-standard graphics techniques know by professionals in the field that aren't used because of their crippling drawbacks in the applications that people care about.

For instance, you can read this thread: https://news.ycombinator.com/item?id=2837948 and bear in mind that 2213 days later, the skeptics in that thread have basically been proved out by future results. (Although you can edit out the people who misunderstand octtrees and argue about the storage requirements for them.)

While we can't know what tech they're using for 100% certain, and they may have put their own twists on it, the most likely underlying tech they are referring to is sparse octtrees: https://stackoverflow.com/questions/985893/what-are-sparse-v... They are a real data structure. It is fantastic for static content and has amazing performance characteristics. However, if you are a programmer who has ever done any tree-type stuff, even if you haven't done 3D graphics, imagine trying to "animate" a single point by moving it in that structure. The algorithm to do so is quite simple... it's just impossibly expensive because of the staggering amount of tree traversal you must do. And due to the way the octtree subdivides, you can't even win with a vague handwave at "caching", though motivating that with math would take me longer than I have here. It completely fails to scale to anything like what the world today considers 3D graphics if you want any motion at all.

Further, there were other telltale signs that that was the technique they were using, such as repeated cells in the rendered video that were clearly just repeated cells in the octtree (a trivial elaboration to understand once you understand the data structure). If you looked you could even see them being aligned on powers of two.

Last I knew Euclideon had decided to change their focus to scanning high-resolution 3D images of things like cathedrals and such. I thought that was actually a good idea, it's a perfect fit for what their tech can actually do. Perhaps doing real engineering work and producing a real product of real value wasn't proving as lucrative as fraud. Certainly it's more work. They weren't even trying with that press kit, as others observe the tech described is going to be literally incapable of producing things that look like those concept photos.

You'd think they'd at least change their name, at this point.

I'm also hoping the Euclideon Defense Squad is tired out at this point. (Well, it's not really the Euclideon Defense Squad. It's really the Squad of People Annoyed That Some People Claming Things Are Impossible, What If You're Wrong, They Thought We'd Never Break The Sound Barrier, Didn't They? It just happens to be that the Euclideon was the topic du jour. But that's a bit long. It also tended to miss the fact that even if what they're describing is possible that doesn't mean Euclideon are the people to do it....) They're scammers. It's pretty much proved at this point, I don't even feel like that's a libel risk. They've not produced what they said they will in the past, and if this is their current focus, they will continue to not produce it. Since what they were trying to produce in the past was impossible, their failure to produce it was no great surprise. This table they're describing may be possible, it just won't be as impressive as they are promising. Possibly useful, though. Unless they really were stupid enough to tie it to their voxel technology, in which case awhoop there goes most of the use cases again.

The VO dude sounds like the "Point Cloud Data" dude from Unlimited Details.

Heck the video even shows the same type of scenes...

why all the "designers" who talk about awesome holograms continue to draw their imagined devices like they can display something outside of the boundaries of their screens?

also, interestingengineering.com is a cancer, people who write/edit their articles have understanding of the world on level of three years old children.

Reply on news.ycombinator.com