Electronic Holography: "Holovideo"

MIT Media Lab Spatial Imaging Group

. Overview of our research in holovideo
. Holovideo papers
Links to holovideo-related work
TV of Tomorrow Research - where Cheops began.
Our TVoT page from 1995
Spatial Imaging Group
Holovideo on the cover of Frames, 1993 Sept
. Our sponsors

Overview of Our Research in Holovideo

Five years ago, members of the MIT Spatial Imaging Group created a new visual medium -- electro-holography -- capable of producing realistic 3-D holographic images in real time. Subsequent research led to full-color images, computation at interactive rates, and most recently a scale-up to a 36-MB display system capable of producing images as large as about 100 mm in width, height, and depth. Computing involves tranforming a 3-D numerical description of the object scene into a holographic fringe pattern.

The next photo shows the real-time holovideo display system as it appeared in 1991.
ooops Photo of Interactive Holovideo Display System

The holographic fringe pattern use to create the interactive image in the above photo contained 2 MB of information, and was computed on a massively parallel supercomputer (a Connection Machine Model 2 with 16 Kprocessors). A time-multiplexed scanned acousto-optic modulator transfers the fringe pattern onto a beam of light at a rate of over 100 MB/s.

In our most recent work on real-time holographic display systems, a 36-MB system produces an image that is approximately 140 mm wide, 80 mm tall, and 150 mm deep. The following picture shows schematically the layout of the current 36-MB holovideo display system.

This is what the display setup looks like:

A new approach to fringe computation - called Diffraction-Specific fringe computation - has yielded two types of holographic bandwidth compression. Diffraction-Specific fringe computation is based on the spatial and spectral discretization of the hologram. The hologram is treated as a regular array of holographic elements called "hogels." Each hogel is a small piece of the hologram and possesses a homogeneous spectrum (distribution of spatial frequencies). [This research was performed by Mark Lucente, former member of the Spatial Imaging Group.] The first method of holographic bandwidth compression, called Hogel-Vector Encoding, is a two-step process. First, an array of hogel vectors (one representing the discretized spectrum of one hogel) is computed from the 3-D object scene description. In the second step, each hogel vector is decoded into a hogel (the usable fringe) through the linear superposition of a set of precomputed basis fringes. Hogel-Vector decoding has been implemented using the Cheops Image Processing system and the Splotch Engine - a superposition stream-processing daughter card. The second method of holographic bandwidth compression is called Fringelet Encoding. A fringelet is computed for each hogel, and each fringelet is rapidly decoded to produce fringes. This method is designed to reduce the total number of calculations required per fringe sample, increasing computation speeds by over 100 times. Fringelet Encoding promises to greatly simplify the design and contruction of holovideo displays.

Current Holovideo Researchers:

Our Sponsors

We are grateful to our numerous sponsors, past and present:

© 1995 - MIT Media Lab - Spatial Imaging Group