Slashdot Mirror
The Plotter Thickens With Volumetric 3-D Display
Gregg Favalora writes: "I wrote back in October indicating that my firm, Actuality Systems, was working on what we considered to be one of the highest-resolution volumetric 3-D displays ever made. What's cool about it is that it sports over 100 million voxels, color, and an embedded graphics processing architecture with 6 gigabits of RAM. And it'll work off SCSI with many existing applications. Anyhow, the news is that it has started working."
"We are still tweaking the optics and finishing the real-time interface, but photos of the display are now at our website. This is taking place in a startup lab environment, so it's not in a pretty package yet. Rather, it's a work in progress, and we hope to be giving public demos in several months." It may still be vapor, but you can almost see Leia appealing to old Ben Kenobi inside that little plastic dome. Howsabout a test sample, Gregg, so we know it's real?
Ford/DERA is working on a 3D monitor that works using MEMS mirrors. Light is projected onto the mirrors, and focused by them onto a point in space so that the image really floats in front of the viewer.. ht m
http://www.dera.gov.uk/html/news/forddera_index
DTI has a 3d monitor, currently available, that uses a lenticular lense placed in front of an LCD monitor. The lens separates the LCD into left and right eye views. The brain puts these together and makes it look like images are floating in front of the monitor.
http://www.dti3d.com
What is this "gigabits of RAM" crap? Do I go to the store to buy a nanohogshead of milk? Perhaps a septapeck or octoliter of beer?
One megabit is 128KB to the rest of us with a clue. One "gigabit" would presumably be 1024 times this figure. 128MB of RAM in a gigabit. 6 gigabits in this display would then be 6*128MB RAM. 768MB of ram, or what typically ships in a low-end server these days.
Next.
http://www.actuality-systems.com/images/
My Freakin Blog
Too bad the fortune tellers can't afford it yet.
--
WolfSkunks for a better Linux Kernel
$Stalag99{"URL"}="http://stalag99.keenspace.com";
--
# Canmephians for a better Linux Kernel
$Stalag99{"URL"}="http://stalag99.net";
The real world is way ahead of you. :) Ceramic scalpels are already in use.
screw that.
Quake.
These are my friends, See how they glisten. See this one shine, how he smiles in the light.
Anyway, VA's stock price is one-squillionth of the ridiculous peaks. Big deal. Do you get some kind of kick out of seeing that? WTF has VA ever done to you? They support k5 - not to mention sourceforge. Do you want *that* to fall over or something?
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
All 2-dimensional and most 3-dimensional display technologies are designed with a single user in mind, viewing the display from one vantage point. While this is fine as long as the user (or small group of users) are performing first-person tasks such as writing a paper, browsing the web, or running around killing people with a rocket launcher.
The moment one tries to bring more than one vantage point into perspective, however, the limitations of a flat display make themselves known. Ever noticed how hard it is to achieve really good results with a 3D modelling package? Ever tried to visualize a complex relationship between dozens or hundreds of objects in 3 dimensional space? It's damned hard using today's display technology.
While demand for these displays will be small at first, it will rapidly grow as they become bigger and cheaper. The first widescale application might well be a third-person an arcade, for example a "model flight sim" where two players sit at a table and dogfight with miniature planes flying in the airspace above the table.
From there, the possibilities are limitless: interactive digital theater in the round; architecture, interior design; and landscaping; there are hundreds of awesome applications for this new toy!
It's only disadvantage is the fact that it is, at heart, a giant moving part. So it will tend to be bulky, power hungry, break down frequently and not like vibrations or drops.
The main drawback of Actuality's approach is the rapidly spinning screen. It occurs to me that this same technology isn't all too far from being implemented with immobile, solid-state electronics.
Picture this: a hemisphere of acrylic, crystal or some other clear material, impregnated with millions tiny triplets of red/green/blue light-emitting polymer. The control circuitry for the LEP "pixels" runs vertically throughout the display and is made of the thinnest wires possible, to avoid obscuring any light from escaping the display. (Perhaps the control circuitry is fiber optic, or perhaps it's made of some sort of electrically conductive crystal.)
The display works on the same principal as an Actuality display--only instead of a rotating screen, we do everything logically, sweeping radially around the display and illuminating all the pixels that lie on a given plane or "slice" at the same time, with the proper colors.
This approach would use far less energy than an Actuality display, would have a beautifully high refresh rate, and would have better brightness and clarity.
It might even be possible to get the light-emitting polymer to emit light of a certain polarity, and coat the surface of the display with a material that is polarized so that at any point on the display's outer surface, the only light allowed to pass directly through that point is light that was emitted in phase with the "slice" which runs approximately parallel to the tangent of that point. Don't despair if this sounds like gibberish. What it comes down to is an ultra-crisp display and ludicrously high refresh rates.
The polarized-light technology is probably impractical, but we should have the manufacturing technology for the basic display within 20 years, maybe sooner if this nanotechnology hype ever goes anywhere.
Why use SCSI instead of AGP? Hmm.. Maybe cuz it's easier to get SCSI cables than AGP cables ;-)
--
SCO employee? Check out the bounty
Back when virtual reality was the future (late '80s), TI had a similar display. It was a fast rotating helix made of a transparant material. Due to the rotation every position inside the dome this thing rotated in was "filled" with material only once during every circle. If you fired a laser at that position when it was filled, it would glow in green or red, depending on the laser. And naturally only that position would glow, because the rest of the dome was empty right then.
So by rotating and timing the laser one could display volumetric data. Resolution was very low (a cube that occupied 1/5 of the whole height of the dispay consisted of about five voxels in each direction), but it looked pretty cool anyway.
Price was somewhere between $10K and $50K. TI intended to build a large version for air traffic controllers, so they could walk around a virtual sky in a dome and "see" the planes. Never heard of it again.
The display by Actuality Systems seems to use the same basic principles: rotation and timed illumination. I hope that this time we'll really see these things on the/a market.
And yes, I want one.
memomo: free web based language trainer DE-EN-ES-FR-IT
I think you're talking about Sega's "Time Traveller".
It just used a parabolic mirror to make the image float there.
me
voxel = volume element = one pint in 3D
So thats what they've been serving me at the pub!
Barkeep, A round of voxels!
---
satire, n: 1) witty language used to convey insults or scorn; 2) a form of humor lost on most slashdot moderators.
obl. disc. site is hosed, so I'm speculating:
You'll need a vacuum too. To get 50hz, you'll need a 300rpm rotation. If it's a helix, you'll be stuck with a gigantic propellor, and if it is flat, you have a centripetal pump. So you need vacuum.
(side note: couldn't we just use one of those vapor-trail detectors and two or three low power lasers to acheive 3-d vector display? The idea being that neither beam would be powerful enough on its own to ionize the vapor, but at their intersection, they would combine to have enough power -- or do we run into the limitation that you can't add quanta?)
Also, both display technolgies have the drawback of being non-occluding. You can't display solid objects, because the front face will be transparent, letting the back shine through.
This will of course be fine for air traffic control, and for things like displaying MRI scans doctors are already proficient at reading layered data, but it will be a hassle for many potential uses.
Of course, the ability to select what is displayed will help.
Uh, nope.. a 'voxel' is a point or some higher-dimensional sample of some volume.
Voxels are somtimes thought of as a cubic volume - i.e. the point sample is linearly interpolated in each of the X,Y,Z axes out to some threshold value determined by some means, often the density of that voxel. This is what you are talking about.
However, Voxels are also often represented using a technique called 'spatting', which is indeed, simply drawing a set of (usually semi-transparent) sprites - one per sample, on screen in back-to-front order. This is what the previous poster is talking about.
Voxels can also be represented as isosurfaces based on interpolated density values, vector fields, multiple 2D planes generated by 'slicing' the volume as well as others.
Voxels are certainly not inherently cubic in nature, and stating the only representation of a sampled 3D volume is a set of little cubes is mistaking the definition of a voxel for a rendering method.
I gots ta ding a ding dang my dang a long ling long
3d Technology Laboratories was demoing their 3D Volumetric display at ACM1. Their approach is to shine 2 different wavelength lasers through a glass or plastic cube (dopant revealed only under NDA). Where they cross, voila. A pinpoint of light. The demo display is a 2-inch glass cube. See the website for pictures.
For a really cool demo, get a camcorder on a spinning mount to match your products, then do a time lapse of a plant germinating. This would allow you to do a frame every second or two in high resolution, making the capture process easy, then you can avoid having to do any hidden surface removal for playback. You could also do the math and do all that to compress it for the finished demo.
It would be a VERY cool, high resolution demo that wouldn't be replicable on ANY other type of display out there.
Ok, get a camcoder, a pivot point, some potting soil and seeds... and make me a very cool demo. (I want to see this if you actually do it).
--Mike--
No doubt, what an idiot. Their site has been /. since 11PM last night, and it still is slowed to a crawl.
Someone you trust is one of us.
(Note: I was hoping for a +1 funny with my original comment, but I guess nobody can figure out I was being facetious. Yes, even though they really are just a 3D extension of the "floating clock" concept, these displays are actually a very clever idea. "Makes me yawn" was a joke.)
I think your predictions about 3D UIs are a little over-reaching. 3D interfaces have been around in research labs for years (using OpenGL images on 2D displays), and the big problem is not displaying the images, it's giving the user an easy way to manipulate the images in 3-space. (Your example of using the scroll wheel to represent depth sounds pretty clumsy, really.) The key to 3D UIs rests not with the display, but with the input device, and that's not an easy problem to solve at all. Just look at all the spiffy new 3D-gaming input devices that constantly fail to catch on. It's hard to build a workable 3D input device, and until someone does, volumetric displays will likely remain rather passive devices instead of interactive ones.
Free Hans!
I'd like to see an actual 3D image with no glass case and no rotating display screen. Now that would be something. This just makes me yawn. A neat toy ... big deal.
Free Hans!
a previous post mention that the screen was spinning at 600rpm. A glass plate spinning at 600rpm would have to be very well balanced. if it was at all off centered then the thing would mostlike fall over and start spinning around until it broke. I imagine this thing is bolted quite well to their desk.
A rabbit in the hand is worth 4 in the cage
read the whole thing!
Please respond quickly or I will take my offer to kuro5hin. Unfortunately, I do not think they will accept. Darn their scruples and lack of commercial filth!
Yours truly,
Vinnie Vendor
Director of Lies^H^H^H^HMarketing
Evil, Inc.
--
--
I like to watch.
Maybe something involving waves in fluid like Scophony would work.
I think (without knowing the toy in question) that yours is just relying on persistence of vision as the ring goes round. You can buy (or build) little clocks consisting of a column of LEDs on the end of a stick (or pendulumn), and as you wave the stick around, the LEDs change their display. Persistence of vision makes you see the brighter LED image for longer, so it looks like the message is written in the air. I think there's a version called the SpaceWriter.
:-)
Point is, with the SpaceWriter system the LED has to physically go through the location where you want the pixel. That's the difference - Actuality's one uses projection to do it, so the LEDs stay fixed in the base of the unit. If you had a pillar of LEDs flying around, (a) it'd be difficult to get it to move fast enough, and (b) it'd get in the way of viewing the image from all sides.
They're definitely missing a trick anyway by not using a version of the Princess Leia film!
Grab.
The heart of Actuality's display technology is a high-speed image projection system which illuminates a swiftly rotating proprietary screen. As the screen sweeps out a cylindrical volume, the projector sends out a sequence of 2-D "image slices." These slices, when computed properly and projected in the correct sequence, serve to create a volume-filling 3-D image. Your persistence of vision does the rest.
Here's how the system works in a deeper level of detail: your application (say, an MCAD system) provides the Actuality display with data via the Actuality API. This geometry information is rasterized and placed into a three-dimensional matrix of memory in the display unit. A high-speed projection system rapidly flips through the 3-D memory in a series of 2-D steps, which we call slices. These slices, when computed properly and selected at the proper times, perceptually combine into a sharp, volume-filling, true 3-D image.
I have no pants and I must scream
hrm from 90 voxels, to 100 million. in about 6 months. can we expect a 10 billion voxel display for chrismas? :P
Rate me on Picture-rate.com
"and dear god does this website suck now." -- CmdrTaco
I don't know what you were trying to say, but reading it made my head hurt :P
Voxels don't have anything more to do with sprites then pixels do. And there is no reason you couldn't make a sprite out of Voxels the same way you could out of pixels, it would just be a 3d one
To clarify:
A pixel is a picture-element, one of the squares of color that make up a digital image. If you have a 640x480 picture, then you have a matrix of 640 pixels by 480 pixels arranged in a grid, each with a specific color value.
A voxel is a volume element, instead of squares of color, you have cubes, and you build your picture the same way you would build something out of legos.
A sprite is an image that moves around the screen programicaly... Like a video game character, in fact, the term 'sprite' is used almost exclusively when talking about video games. A sprite can be a picture, (like a picture of Mario) or a volume (imagine a 3d Mario built out of blocks). It really doesn't matter.
A Voxel is not a set of sprites stuck together, there is no version of the term that means this, and whoever told that to you was totally wrong.
Rate me on Picture-rate.com
"and dear god does this website suck now." -- CmdrTaco
...what I really want is an atomic vector plotter.
I know where I can get one, but I just can't get myself to enjoy Vogon poetry that much, no matter how hard I try...
That was a 2-d image projected off the surface by the mirror mentioned in the last comment. Looks kinda cool, but the image gets a bit distorted and there's no real 3rd dimension to it. IIRC, it was a pretty crappy game too. You can get the same optical effect from that table-top mirage thingy, the one you put some coins or something in the bottom of. The image is projected above the device by a pair of parabolic mirrors. Here's the first link that I got from Google. There's an explanation a bit more than halfway down the page.
Bugrit! Millenium hand and shrimp!
Well, it is NOT free standing out in the open air. and it is not animated.
It is contained inside a glass sphere. Maybe the size of a basketball, or smaller. The images are of a small section of a DNA strand, roughly one full twist. The quality of the image is similar to a nicely shadowed but obviously computer rendered diagram in 3D (well duh!) It is definitely not photograde, although that by itself should not be a problem.
The image is shown glowing, but it is in a darkened space, so probably it will not be ready for daylight presentations for a while.
I am amazed that it is done at all, although it will be a while before it progresses beyond the novelty stage.
"It is a greater offense to steal men's labor, than their clothes"
shouldn't it be: "the thick plottens" ?
I saw another version of a 3D display in an arcade years ago, attached to a machine that cost something like $5 to play and seemed a pretty simple game. And this was back in the days when all games cost 25c. Next time I went back to that arcade the machine was gone. Basically, (especially in light of the recent advances in flat-screen technology) will I be able to afford one of these before my kids are my age?
Kurdt
Kurdt
I'm not anti-social. Just pro-technology.
These displays are still not the "ultimate" in 3D because they aren't true 3D displays. Yes, they give an excellent sense of depth, but they cannot emulate real world images of most objects. The technical reason is that the light generated or reflected by an object can only be properly represented by 4 dimensions of data (not counting wavelength, time and external light sources), but these displays can only really produce 3. To put it in laymans terms, these displays can only represent additively transparent (glowing) objects. You cannot visualize opaque objects properly. This is still useful for scientific visualization, medical imaging, etc... but not all that useful for common real-world applications.
Check out the Harvard Medical School Surgical Planning Lab for instance. They're working on (among other things) a system that allows doctors to perform surgery while the patient is inside an MRI machine, so that the surgeons can literally see what's under their knife before they make the next incision. (Right now it's done on a CRT, eventually they want some sort of HUD overlay.) Very, very cool stuff. If unenclosed holographic projection ever happens, they'll be first in line to use it.
News for Nerds. Stuff that Matters? Like hell.
I'm curious, 'cuz I have a "Saturn-5" electronic toy sitting behind me which has a motorized plastic ring with 10 red LED's around it, which flicker as the ring spins. By playing with the timing parameters (2 knobs) I can make various spherical patterns appear. Is this in violation?
What if I replaced the ring with a circuit board with ~200 LEDs on it (I think Radio Shack has all three colors now)? Although communicating with 200 spinning LEDs would be a challenge in itself .. maybe a spinning mirror would be better .. How much of this is already patented?
Oh, well, back to work..
-B
Not that this wasn't entirely predictable.
It's interesting that they made such a big announcement that their product actually works.. ;)
It's even funnier to see what their 2D Test Pattern is. ;)
That will be nice when they actually get it out and there are a few advances in memory. Unfortunately, it will probably not have the cool sort of holographic effects that they had in Star Wars and Star Trek.
But I bet that the folks who do military-grade radars will love it. Imagine being able to view the exact 3D position of an aircraft instead of just looking at the overhead view.
It's just odd to see that they are using SCSI to do the interfacing. SCSI's a lot slower than the AGP port, and you are transfering several hundered times the data.. ;)
Gentoo Sucks
Holography also have other uses; they enable radiologists to interact with the data that have been collected by scanners and they may facilitate the production of "what if' images which some surgeons have found useful in surgical planning. Programs have concentrated on the parts of the body and the kinds of conditions (i.e., tumors, trauma, and vascular abnormalities) that are commonly examined with CT and MR scanners.
Studies were designed to determine if the digital holography systems would allow diagnosis of conditions that are extremely difficult or impossible to detect with existing technology; provide for more accurate and comprehensive diagnosis and understanding of conditions that are difficult characterize fully with existing technology; increase the radiologist's confidence in the diagnosis made; reduce the time required to arrive at a diagnosis; facilitate communication of relevant information; improve surgical planning; and allow for more fully informed patient consent to treatment.
Sure its a cheesy website but it has some pretty useful information on the subject.
privacy 101
360 degrees of Karma
-Vess
For those of you wondering just how the heck this thing works - it uses a (really fast) conventional 2-D projector and a very complex array of lenses and mirrors to project a constantly changing image onto a 2-dimensional translucent screen that rotates at 600 rpm. By changing the image as the screen rotates, the illusion of a 3-D object is created.
More technical info (with pictures) can be found here and a shot of the screen while it's not moving can be seen here.
main(c,r){for(r=32;r;) printf(++c>31?c=!r--,"\n":c<r?" ":~c&r?" `":" #");}
--
"May the forces of evil become confused on the way to your house"
--
"May the forces of evil become confused on the way to your house"
-George Carlin