How far away are hologram interfaces?
My iMac's LCD display often reminds me of a floating hologram..you know, the kind shown in the movie Minority Report.
Do you think this will ever come to light? Is it something that is more science fiction than an actual user interface?
And do you think Apple would ever head down that path?
Do you think this will ever come to light? Is it something that is more science fiction than an actual user interface?
And do you think Apple would ever head down that path?
Comments
The japs made a hologramme machine a while back, that displayed a pre-recorded image.
As for a computer interface, it will be a long time before anything like that appears. given time it will happen, just probably not in our life time.
There is a project at the Media Lab called HoloTV, only has a few people working on it, but it is designed to do just that: create a holographic television. I used the term "television" loosely. The current setup is huge. An entire laser table is setup, and at the collimating lens there is a device that takes output from a Silicon Graphics Onyx2 which is the source for the holographic image to be projected. When I was there, we were able to get a VW Beetle 3D model projected, in a single color, rotating at about 3fps.
Compared to what you saw in Minority Report, it looked like crap. But you have to start somewhere.
Some of you may recall a video game Sega had in early 90s I believe where it projected so called "holograms" onto a playing surface, and you tried to control them. Keep in mind that those weren't true holograms. It was a trick that has been used for a long time, but it's effective.
Some of my own artwork was holographic. I have a stack of plates in my apartment that contain holograms I made. It's an incredibly limiting and time consuming process. The tools are very sensitive, and we're looking at decades before much happens. Keep in mind that modern holography has changed very little since its inception some decades ago.
linkage here, with references
Holograms record interference patterns of light. The details of such an image are therefore excruciating small, on the same order as the wavelength of visible light (roughly from 400-700 nm).
Let's say that the pixels for a hologram need to be one quarter the wavelength of violet light, 100 nm, to have enough detail to produce proper interference patterns. That's probably not enough detail, but good enough for a low-ball estimate the number of pixels we'd need.
A single inch of display would contain 254000 pixels. A square inch would contain about 65 billion pixels. A wide-screen 19" display would contain about 10 trillion pixels. (Note: American, not British, billions and trillions.)
That's 10 trillion pixels for a monochrome image. Normal holograms don't record colors, but you might be able to apply the old RGB trick for creating visually satisfying color (with no fidelity to the wavelengths in a real scene) by using three layered holographic imagers. So, for RGB color we're up to 30 trillion pixels.
Let's say we'll use 8 bits per pixel. A static image will then requires30 terabytes (that a 10^12 terabyte, not 2^40) of data. A moving image refreshed at 60 frames per second will require a 1.8 petabytes/second, or about 14 petabits per second.
That's 14,000,000,000,000,000 bits per second -- the bandwidth of around ten billion cable modems running full-tilt at the same time.
Holographic data is so enormous because, whether a person looking at a display is using the information or not, the display has to provide a different view of the same scene for every possible angle that the viewer might choose to look at the display at any given time.
Clearly, the interference patterns themselves are not what you'd want to record, store, or transmit -- even if you could apply some data hefty compression. A 3-D mapping of object locations and surface details would be much, much more efficient. You'd still need to translate that data into interference patterns at some point to produce a viewable holographic image, creating some stunning bandwidth and processing issues within the display and display support hardware.
Full-motion true holograms are probably not a viable technology -- not in anything like the foreseeable future. That's not to say that the word "hologram" won't be re-used someday for some essentially very different technology, not based on fine-detail interference patterns.
and from an article in Discover Magazine, Feb 2002. ,
A hologram made by Yves Gentet for the Musée de l'Optique in Biesheim, France, has none of the oil-smear-like inaccuracy of standard holograms. From almost any angle, it captures life perfectly in three dimensions
Quote:
Having invented a camera Gentet found he now needed to invent a holographic film as well.
Gentet's years of teaching himself film chemistry and tinkering alone in his lab have resulted in an emulsion he calls Ultimate. Its basic structure is that of ordinary black-and-white film: an emulsion of silver bromide grains, which are very light sensitive, dispersed in animal gelatin. When light strikes the grains, the silver bromide is converted into metallic silver?the stuff of mirrors. The main difference between Ultimate and other emulsions is that its silver bromide grains are extremely fine?around 10 nanometers across, or 1/10 to 1/100 as big as the grains in ordinary film. The fine grains allow Ultimate to record tremendous detail. They also allow it to record red, green, and blue simultaneously in a single emulsion layer, as interdigitating stacks of tiny mirrors.
Gentet won't say how he makes the emulsion; none of his techniques are protected by patents. "We've perfected the emulsion that everyone has been looking for for 30 years," he says. "It's savoir faire?know-how." But it's not as if he is asking people to take his assurances on faith. If seeing is believing, then standing in his reception area and looking at his walls, you have to believe him.
the magazine stills look impressive, the holograms must be near magic
Originally posted by shetline
Think about it this way...
Holograms record interference patterns of light. The details of such an image are therefore excruciating small, on the same order as the wavelength of visible light (roughly from 400-700 nm).
Let's say that the pixels for a hologram need to be one quarter the wavelength of violet light, 100 nm, to have enough detail to produce proper interference patterns. That's probably not enough detail, but good enough for a low-ball estimate the number of pixels we'd need.
A single inch of display would contain 254000 pixels. A square inch would contain about 65 billion pixels. A wide-screen 19" display would contain about 10 trillion pixels. (Note: American, not British, billions and trillions.)
That's 10 trillion pixels for a monochrome image. Normal holograms don't record colors, but you might be able to apply the old RGB trick for creating visually satisfying color (with no fidelity to the wavelengths in a real scene) by using three layered holographic imagers. So, for RGB color we're up to 30 trillion pixels.
Let's say we'll use 8 bits per pixel. A static image will then requires30 terabytes (that a 10^12 terabyte, not 2^40) of data. A moving image refreshed at 60 frames per second will require a 1.8 petabytes/second, or about 14 petabits per second.
That's 14,000,000,000,000,000 bits per second -- the bandwidth of around ten billion cable modems running full-tilt at the same time.
Holographic data is so enormous because, whether a person looking at a display is using the information or not, the display has to provide a different view of the same scene for every possible angle that the viewer might choose to look at the display at any given time.
Clearly, the interference patterns themselves are not what you'd want to record, store, or transmit -- even if you could apply some data hefty compression. A 3-D mapping of object locations and surface details would be much, much more efficient. You'd still need to translate that data into interference patterns at some point to produce a viewable holographic image, creating some stunning bandwidth and processing issues within the display and display support hardware.
Full-motion true holograms are probably not a viable technology -- not in anything like the foreseeable future. That's not to say that the word "hologram" won't be re-used someday for some essentially very different technology, not based on fine-detail interference patterns.
That's the problem with the whole idea...it's 'pixels squared'. Why not a lcd cube that has cubed pixels instead of square ones?
Here, again, we run into the pixel problem. But my answer is: If pixels are a problem, get rid of them. As it is, we are close to a resolution-independent interface - which means, an interface not described in pixels. Filters and effects don't have to be applied on a per-pixel basis, that's just been the obvious (and, for performance reasons, necessary) operation - but the next version of Photoshop reportedly virtualizes the pixel by allowing the user to specify different shapes of pixel, so some independence is being gained there as well. And wasn't there an early competitor to Photoshop that used resolution-independent filtering?
Now, you have to be able to manipulate those mirrors somehow, because they are, finally, like screen pixels. But the daguerrotype article mentioned that magnetism is used to store amplitude. Ah, yes. Magnetism. A magnetic field can precisely describe an effect over an area simultaneously, and silver (the component used in the hologram's mirrors) is certainly capable of responding to magnetism. What would be required is a means to map logical shapes and colors to perturbations in a magnetic field. Now, this would take some serious doing, if it's even possible - I'm thinking more in terms of basic exploratory research than the next model iMac. But I like the idea.
And, of course, there's a very good chance that I'm sailing with four sheets to the wind. But at least it's a fun ride.
Originally posted by Placebo
That's the problem with the whole idea...it's 'pixels squared'. Why not a lcd cube that has cubed pixels instead of square ones?
First of all, you'd have to overcome problems of both transparency and opacity. "Voxels" (volumetric pixels) representing empty space between objects would need to be nearly perfectly clear, so close to perfect that even looking through thousands of layers of such voxels, other voxels behind them would not be substantially dimmed, faded, or blurred. These same voxels would need to be able to become nearly perfectly opaque to be able to represent real objects that are opaque enough to not let ghostly images of what's behind them shine through.
We're a long, long way from producing a voxel that can change from 99.99% transparent to 99.99% opaque in fractions of a second, that can change color, and a long, long way from being able to wire up such voxels with 99.99% transparent conductors.
Achieve all of this somehow, someway... and then what you'd have is something like a 3-D fish tank. Nothing could appear deeper or further away than the physical depth of the volumetric display. Nothing could appear to project outside of the display. This would be a far cry from what holography is capable of.
http://www.cs.tut.fi/~ira/wave.html
This one more, big problems to solve before seeing an consumer holographic display.
Originally posted by JLL
Another scifi (or sci fact) monitor tech:
http://www.cs.tut.fi/~ira/wave.html
That's pretty cool!