From: uselton@nas.nasa.gov (Samuel P. Uselton) Subject: Re: Vibrating mirror display (was Re: images out of thin air) Date: Mon, 18 Nov 91 19:26:29 GMT Organization: NAS Program, NASA Ames Research Center, Moffett Field, CA In article <1991Nov17.205109.24670@milton.u.washington.edu> brucec@phoebus. labs.tek.com (Bruce Cohen) writes: >In article yanek@mthvax.cs.miami.edu >(Yanek Martinson) writes: > >> I have seen a fairly simple device of two concave (hyperbolic?) mirrors, >> one with a hole in it, that project an image of a small object inside to >> above the mirrors. The image appears entirely real, it can be viewed >> from any side, and is as clear and three-dimensional that you would not >> believe the object is not there without trying to touch it. >> >> Now questions: >> >> How does it work? It seems incredible to me. There IS nothing at the >> spot in space where the object appears to be. > >If this is the system I am thinking about, one of the mirrors is I *don't* think this is the same system. It sounds to me as if he's asking about something similar to a gizmo I've seen in novelty shops, with a little ball hangin down in front of a cup shaped (parabolic?) mirror. >actually a thin piece of silvered plastic which effectively forms the >cone of a speaker. Nope, the silvered material is stretched over a regular speaker. The vibrations of the speaker (aka sound) vibrates the mirror. Silvered mylar is what I've seen most often, but Larry Sher at BBN had a thicker, tuned plexiglass mirror. >The speaker magnet drives the cone with a waveform >which causes the mirrors focal length to change (linearly for most of >the scan, I believe). >There is a CRT at the prime focus of the mirror ????? >presenting some image, and the CRT scan is synchronized with the mirror. >As the mirror scans, it puts the position of the virtual image of the >CRT phospor at different Z positions in space, so the effect is of a 3 ^^^^^^^^^^^^^^^^^^^^^^^^ Right-O! >dimensional scanning system. > >> What is the name of this system, effect, person who invented it? > >Sorry, I can't remember. > Generically, I've heard it called a vari-focal mirror system. Larry Sher's system (along with some computer wrappings) was marketed by Genisco (remember them?) as SpaceGraph. They sold 6 or 8 immediately after it was put on display at 1981 SIGGRAPH in Dallas. I don't think they ever sold any more. I saw 4 of the installed machines. They were almost never used because the system was to hard to program for what you got. >> Exactly what shape must the mirrors be? > >I believe the moving mirror is approximately spherical, but I could be >wrong on that. It turns out that the mirror is almost flat! (This is one of the reasons I think the original poster is asking about something else.) >I think the shapes are complementary, so that the volume >you see is relatively undistorted by variable magnification effects >off-axis of the moving mirror. > >> It appears to work only for small objects; to achieve the same effect >> for a larger object, would the size of mirrors have to be drastically >> increased? > >Probably. Also, there's a practical limit on the resolution determined The vari-focal mirror systems have the apparent object appearing below the mirror surface, or about on it. So the size is clearly limited by the surface of the mirror (eg 15" woofer). >by how fast you can scan the mirror, and the bandwidth of the CRT, since And, most critical(!), the speed of the phosphor on the CRT! If the phosphor doesn't fade fast enough, you get smeared images in the "depth" direction. Also, phosphors fast enough (to get the *whole* *depth* *range* done in 1/60 th of a second) don't (or at least didn't) come in many colors. All the systems I ever saw were monchrome. >you have to allow time in each frame to scan along 3 axes instead of >just 2. If you want raster images rather than just vectors, a 100 point >resolution in Z multiplies the required CRT scanning rate by 100. There is no "obscuring" by depth since you are really seeing distinct images for each depth superimposed. So raster-style images are hard to understand. >Vectors are cheaper of course, since you don't have to scan, but there's >a limit on how fast you can accellerate the mirror with the coil, so you Typically the mirror is synced at a constant vibration, to match the CRT and its phosphor. Typically number is 60 hz. (= Refrigerator Hum - the first one I saw, at U. Utah, you couldn't talk in the same room with it.) >typically have to sort your vectors for optimum display order, and the ^^^^^^^^^^^^^^^^^ Actually you have to chop the vectors into segments that are "short enough" in depth, and sort the chopped pieces to be displayed when the mirror is in the right place. Also note that the mirror is doing a pendulum style cycle (back and forth) rather than a circular style cycle. These are some of the things that made the system hard to use. >number of vectors may be limited by the mass and size of the mirror and >the power of the coil driver. Not really. The limits are on how much you can draw to the CRT while the mirror is in approximately the right place. > >> It projects the image only a small distance away from the mirror. Is it >> possible to project it to greater distances? > >Same answer as with magnification. > >Incidently, there's an additional problem with the system: the mirror >really is a speaker, and generates a hum at the scan rate. The larger >the mirror, the more power required to drive it, and also the louder the >hum. At less than 20 frames per second, the hum is subsonic, and may be Nobody I know tried scanning that slowly for long. You need to get all the images superimposed on the eye to look like a 3D image and not an animation. Systems I saw were 3 different sources, all used around 60 hz. Besides SpaceGraph and the Univ of Utah models, U. North Carolina at Chapel Hill made one driven from an Ikonas frame buffer. (Anecdote - Henry Fuchs and Fred Brooks came up with this idea at dinner, when Henry was there to interview. They followed up and found out that Sher had patented some part (the tuned plexiglass mirror I think) several years earlier. The Utah system was from about the same time frame - around 1978. Sam Uselton uselton@nas.nasa.gov employed by CSC working for NASA (Ames) speaking for myself