bobk@mntgfx.mentor.com (Bob Kelley) (12/16/87)
How can I use ray tracing techniques to directly generate
binary holograms which reconstruct into the usual kinds of
ray traced images? I think it would be amazing indeed to
view a reconstruction of the usual kinds of mirrored and
refractive spheres that behaved as if it were really there!
I've been reading a couple of articles:
_Binary Fraunhofer Holograms, Generated by Computer_
A. W. Lohmann and D. P. Paris
Applied Optics Vol 6, No 10 p. 1739
and:
_Minimization of Reconstruction Errors with
Computer Generated Binary Holograms_
R. A. Gabel and B. Liu
Applied Optics Vol 9, No 5 p. 1180
I'm inspired by the cover of Stewart Brand's new book, _The Media Lab_
Any information which might help me understand this technology is welcome.
Bob Kelley
612 SE 19th Avenue
Portland, OR 97214
--
Robert J. Kelley
Mentor Graphics Corporation
...!tektronix!sequent!mntgfx!bobk or bobk@mntgfx.MENTOR.COMwanger@hpfcdq.HP.COM (Leonard Wanger) (12/18/87)
Last year I visited the Media Lab at MIT. The way they were creating the computer generated holograms was to expose a thin vertical strip of the film to an image. Then a new image (rotated by a small amount) and the vertical strip next to the last one was exposed. Repeating this process for the whole piece of film allowed a cylindrical hologram to be produced. This method is a little expensive for ray-tracing considering the number of images to be produced (over a thousand per hologram if I remember!). Also, although the result was impressive, you only can see the three dimensionality in the horizontal direction. Normally a hologram allows you to look around objects (an example is one I made with a magnifying glass in it that allows you to look at the object through the glass or under normal magnification around the side) but since the holograms are created in vertical strips, you can not look over and under objects (normally every point in a hologram has the complete image, in this case all the points in the vertical strips contain the image). Has anyone heard about a computer generated hologram created in squares (opposed to strips)? This would square the number of images needed, but would give the true 3-dimensional effect (in the vertical and horizontal). Len Wanger -- Hewlett Packard Graphics Technology Division
hansen@mips.UUCP (Craig Hansen) (01/06/88)
In article <390006@hpfcdq.HP.COM>, wanger@hpfcdq.HP.COM (Leonard Wanger) writes: > Has anyone heard about a computer generated hologram created > in squares (opposed to strips)? This would square the number of > images needed, but would give the true 3-dimensional effect (in > the vertical and horizontal). There's no particular obstacles, except that 1) it takes a lot longer to make 10000 images than 100 images (100 times longer). 2) binocular vision and head motion parallax is usually left-to-right, so if you're going to go in only one direction, vertical strips are more important than horizontal strips. -- Craig Hansen Manager, Architecture Development MIPS Computer Systems, Inc. ...{ames,decwrl,prls}!mips!hansen or hansen@mips.com
dlleigh@mit-amt.MEDIA.MIT.EDU (Darren L. Leigh) (01/22/88)
In article <1225@mips.UUCP> hansen@mips.UUCP (Craig Hansen) writes: >In article <390006@hpfcdq.HP.COM>, wanger@hpfcdq.HP.COM (Leonard Wanger) writes: >> Has anyone heard about a computer generated hologram created >> in squares (opposed to strips)? This would square the number of >> images needed, but would give the true 3-dimensional effect (in >> the vertical and horizontal). > >There's no particular obstacles, except that > >1) it takes a lot longer to make 10000 images than 100 images > (100 times longer). > >2) binocular vision and head motion parallax is usually left-to-right, > so if you're going to go in only one direction, vertical strips > are more important than horizontal strips. I tuned into this discussion a little late so forgive me if this has already been brought up. Computer generated holograms created in strips (stereograms actually) do not have vertical parallax because they are usually meant to be viewed in white light. Because holograms work by the principle of diffraction, the different wavelengths in the illumination beam are refracted at different angles and can cause an unacceptable amount of color blur. The color blur is caused by blue light from a higher vertical view and red light from a lower vertical view mixing with light of in-between wavelengths from the central view because of their differing diffraction angles. We can get around this problem by using a rainbow hologram, which is made by throwing away all vertical parallax and using only one vertical view. When the hologram is viewed in white light, moving your head up and down will change the color of the image through the spectrum instead of changing the vertical parallax. The computer generated stereograms are made basically the same way which is why you can get such nice color by making multiple exposures at different angles. A computer generated hologram with both vertical and horizontal parallex wouldn't look very nice unless you viewed it under monochromatic light.