Hello,
I suppose this post is part-publicity, part-science. My background is in autostereoscopic 3-D, and my employer - a product development firm called Optics for Hire (OFH) - acquired the patents from my former company in the space of 3-D image projection.
OFH has begun making animated movies that explain the underlying mechanisms of some of these patents. Here's one that I think is interesting, and hope you will, too. It describes techniques to create 3-D imagery floating above, say, a conference-room table, with horizontal parallax (or, "theta parallax" might be more appropriate). It is a light field display. Of course, the window-violation constraint must be observed, in that imagery will only be visible where there is a display element along the line of sight from the viewer through each part of the image. Still, when I developed the theory behind this several years ago, imagery 6"-12" tall seemed reasonable.
Anyhow, the YouTube video is here (with an unusually slowed-down recording of my voice), and a relevant patent is 7,364,300.
What do you think? If this is up your alley, other inventions are at the OFH patent page.
Best,
Gregg
A tabletop lightfield display
-
jnhong
A tabletop lightfield display
I think the video is outright misleading. The 3D video image lights up in free space, beyond the boundaries of the rotating diffuser. It does NOT observe the window-violation constraint you just mentioned.
Can you show a video of the real device in real-life action? There have been several different schemes for volumetric displays, many of them can be found in YouTube as well. Rotating LED bars, piston-mounted LED fields, helical screens, string volumes, etc... How does your device compare with these other designs? You might have some compelling advantages with your design if you could actually show it working. Do you have a prototype?
Joe
Can you show a video of the real device in real-life action? There have been several different schemes for volumetric displays, many of them can be found in YouTube as well. Rotating LED bars, piston-mounted LED fields, helical screens, string volumes, etc... How does your device compare with these other designs? You might have some compelling advantages with your design if you could actually show it working. Do you have a prototype?
Joe
-
favalora
A tabletop lightfield display
Joe,
Thank you for your comments.
That was my point. I had meant to say, if this were built, it would observe the window-violation constraint. I am concerned about the scientific integrity of the depiction, which is meant to show what it would look like as a user. (Makers of autostereo 3-D displays have that marketing problem - - how do you show an image that floats in a 2-D medium, when the display is behind it without looking flat?)
No, the technology has not been built.
The company that invented this (and other 3-D displays) did create and sell volumetric displays - Perspecta. I've been in the field since 1988 and appreciate your point. We are not trying to sell this product, we are announcing the existence of this patent. I do think that (if you read the patent) it describes interesting methods.
Gregg
Thank you for your comments.
That was my point. I had meant to say, if this were built, it would observe the window-violation constraint. I am concerned about the scientific integrity of the depiction, which is meant to show what it would look like as a user. (Makers of autostereo 3-D displays have that marketing problem - - how do you show an image that floats in a 2-D medium, when the display is behind it without looking flat?)
No, the technology has not been built.
The company that invented this (and other 3-D displays) did create and sell volumetric displays - Perspecta. I've been in the field since 1988 and appreciate your point. We are not trying to sell this product, we are announcing the existence of this patent. I do think that (if you read the patent) it describes interesting methods.
Gregg
-
jnhong
A tabletop lightfield display
Gregg, I see. I was hoping you did have a prototype to show off. You can see several home-built, school-built, and startup-built devices already out there (on YouTube), so any potential customers will already have a selection to choose from. Your best bet would be to get a DOD grant and then sell this to some aerospace manufacturer. But without a prototype and a basic business plan, it's difficult to visualize success. 
Or you could say, it's just smoked glass and mirrors?
Joe
Or you could say, it's just smoked glass and mirrors?Joe
-
MfA
A tabletop lightfield display
I imagine that 900 RPM rate is mostly to deal with the time sequential nature of the DLP? I'm not sure how you would ever get decent resolution colour images with only a single projector. The only commercial lightfield display I'm aware of uses a ton and a half of projectors (and I imagine is incredibly expensive as a result). AFAICS the only way to make it work with less is to use 1 progressive LCOS (with strobed light) per potential eye and eye tracking.
-
favalora
A tabletop lightfield display
MfA -
The "turntable"'s rotation rate of 900 rpm is related to the visual update rate of the 3-D image, not of the DLP. For example, to create a 3-D image with a visual refresh rate of 60 volumes/sec:
60 (vol/sec) * (60 sec/minute) * 1 (revolution/volume) = 3600 rpm
By this arithmetic, 900 rpm (did I really say 900 rpm in the patent??) would result in a 15 Hz 3-D image.
Assuming it's proper to refer to each image of the sequence as a "view" (and a different word might be better, since in many cases multiple neighboring "views" are visible to a retina if the number of "views" is high) - - then the DLP bandwidth is related to how many views are generated per turntable-rotation. E.g., a DMD might be running at 16 kHz, and you can divvy up those views however you want.
This is different than our Perspecta display, in which the volume is swept each 180-degree turn of a (vertically-oriented) screen. The factor was then ... * 0.5 (revoluion/volume), and since it had an update rate of 30 Hz, the screen spun at 900 rpm. But I digress.
g
The "turntable"'s rotation rate of 900 rpm is related to the visual update rate of the 3-D image, not of the DLP. For example, to create a 3-D image with a visual refresh rate of 60 volumes/sec:
60 (vol/sec) * (60 sec/minute) * 1 (revolution/volume) = 3600 rpm
By this arithmetic, 900 rpm (did I really say 900 rpm in the patent??) would result in a 15 Hz 3-D image.
Assuming it's proper to refer to each image of the sequence as a "view" (and a different word might be better, since in many cases multiple neighboring "views" are visible to a retina if the number of "views" is high) - - then the DLP bandwidth is related to how many views are generated per turntable-rotation. E.g., a DMD might be running at 16 kHz, and you can divvy up those views however you want.
This is different than our Perspecta display, in which the volume is swept each 180-degree turn of a (vertically-oriented) screen. The factor was then ... * 0.5 (revoluion/volume), and since it had an update rate of 30 Hz, the screen spun at 900 rpm. But I digress.
g
-
MfA
A tabletop lightfield display
The video mentioned 900 RPM.
Hmm, it seems I underestimated the rate the most expensive DLPs can get nowadays. TI's 51.2 Gbps DLP with 5 degrees separation could do ~800*800 15 bit colour at 60 Hz, not bad at all ... although obviously you'd have to build your own projector.
With the 16K 1024*768 DLP you could do ~400*400, still acceptable. Now build it
Hmm, it seems I underestimated the rate the most expensive DLPs can get nowadays. TI's 51.2 Gbps DLP with 5 degrees separation could do ~800*800 15 bit colour at 60 Hz, not bad at all ... although obviously you'd have to build your own projector.
With the 16K 1024*768 DLP you could do ~400*400, still acceptable. Now build it
Last edited by MfA on Mon Mar 22, 2010 2:41 pm, edited 1 time in total.
-
favalora
A tabletop lightfield display
Yeah, we had to build our own 3-chip (3-DLP) projector for Perspecta back in 2001. But these days there might be a few companies who OEM high frame rate DLP-based projectors. (Or maybe I'm imagining things.) There are firms, like ViALUX, who make good fast driver boards, at least. But, yes, the optics is usually a separate project.
g
g
-
favalora
A tabletop lightfield display
Hi -
On a related note, we did prototype a different type of 3-D display that uses a vibrating lenticular array. Here is a YouTube video about it (though the imagery here is obviously from an artist). I have photos of the prototype which I might be able to share.
About halfway through we show the optical layout.
Best,
Gregg
On a related note, we did prototype a different type of 3-D display that uses a vibrating lenticular array. Here is a YouTube video about it (though the imagery here is obviously from an artist). I have photos of the prototype which I might be able to share.
About halfway through we show the optical layout.
Best,
Gregg
-
MfA
A tabletop lightfield display
Wouldn't that effectively work as a rumble pack in a mobile device?