Field-sequential electro-stereoscopic displays require a selection device to alternately occlude and transmit successive fields to each eye. A sequence of images is written using a technique which is similar to that used for planar video or electronic displays. Today such displays are typically produced by DLP projectors or fast LCDs that are part of TV sets now arriving in retail stores. For a flickerless stereoscopic system, the images need to be written at twice the usual planar 60 fields/second refresh rate, because each eye, independently, needs to see 60 fields/second. Therefore, in most stereoscopic video or computer graphics systems the refresh rate is about 120 fields/second.
Improvements in modern field-sequential stereoscopic displays are intimately tied to advances in the art of electro-optical shutters. A perfect shutter for an electronic stereoscopic display would have 100% transmission in its open state, 100% occlusion in its closed state, and the opening and closing of the shutter would take place within the video vertical blanking interval. Interestingly, mechanical shuttering systems may be able to come closest to the performance of the perfect shutter, but they are so inconvenient to use that they could never be used in products on a widespread basis.
There have been five major generations of selection devices for electro-stereoscopic computer graphics and video displays. They are: Devices using mechanical shutters, tethered visors using PLZT shutters, tethered visors using LC shutters, passive polarizing glasses used in conjunction with large LC panels or used in front of projection lenses (like the ZScreen my colleagues and I invented at StereoGraphics a quarter of a century ago), and LC shutters in tetherless glasses.
The first suggestion for a field-sequential stereoscopic system, according to Judge, was by J.C. D’Almeida in 1858, for the projection of stereoscopic slides. Such a system, which is sometimes called the eclipse or the occlusion method, probably consisted of spinning mechanical shutters placed in front of two projectors with similar shutters used in front of the eyes. When the right projector image was transmitted, the left projector image was blocked, and vice versa. The shutters in the selection device used by the viewer must work in synchronization with the projector shutters, so that each eye sees only its appropriate image. If the frequency of blocking and unblocking is high enough, flicker will not be perceived, and a good quality stereoscopic image will result.
There are several references in the patent literature to the concept as applied to movies in the first decade of the century, but it wasn’t until 1922 that the field-sequential technique was employed for motion pictures on a commercial basis. An ingenious design, by Laurence Hammond, called Teleview, was installed, at a cost of $30,000, in the Selwyn theater on Broadway, in Manhattan, for the projection of stereoscopic motion pictures. The projectors for the left and right prints were interlocked with mechanical shutters rotating in front of their lenses. Every audience member looked through a mechanical shuttering selection device. In Hammond’s approach, the shutters were kept in synchronization using alternating current motors. My assumption is that the system was flickerfree.
Motion picture projectors use a shutter, to occlude the projected image while each frame of film is being pulled down. If such a shutter were not used, the audience members would see vertical streaks called “travel ghost.” In addition, for 35mm theatrical projection, whose rate is 24 frames/second, each frame is interrupted once to satisfy the critical fusion frequency constraint. John A. Pross introduced a shutter with this action in 1903, to effectively double the frames perceived to 48 per second. The interlacing technique used for practically all video and some computer graphics displays is the direct descendent of this approach.
Hammond meant the eyes to see images alternately. While the shutter of one eye is blocking the image to that eye, the shutter for the other eye is open. When one projector is pulling down film, or interrupting the projection of a frame, one eye is blocked, at which time the other projector is supplying image to the other eye.
The early history of field-sequential stereoscopic displays is intimately tied to mechanical shutters and their limitations in terms of noise of operation, size, weight and convenience. Their performance can be good, with good speed of transition, superior open state transmission, and dynamic range. Alas, the mechanical shutter can only go so far in terms of its development. The most widely used and interesting mechanical selection device was the one invented by Caldwell and described in a patent issued in 1942.
Caldwell’s device, which was manufactured by Bausch & Lomb, was used on Evans & Sutherland vector graphics computers for molecular modeling applications. It was a spinning cylinder, resembling a beer can, with slits cut in it. The user adjusted a potentiometer knob on the side of the device in an effort to synchronize the shutter with the vector fields as they refreshed, which varied with the complexity of the object displayed. Some people who used the device called it the “eyelash clippers”. The system was made by Terabit for about a decade and it cost $10,000.
The advent of modern electro-optical shutters allowed field-sequential electro-stereoscopy to become a commonly practiced medium. These shutters have no moving parts and are usually constructed of an electro-optical cell sandwiched between two linear polarizers whose axes are crossed, as shown in its generic form. When a voltage is applied to the cell, it becomes isotropic and has optical properties similar to that of a sheet of glass. In this case the crossed polarizers block the passage of light. The shutter is closed. When the voltage is removed or reduced, the axis of incoming polarized light passing through the cell will be rotated through 90°, and the axis will be parallel to the axis of the outgoing polarizer, and light will be transmitted. The shutter is open. In this manner, by controlling voltage to the cell, the shutter may be closed or opened.
Electro-optical shutters, such as the Kerr and Pockels Cell, have been known since the 19th Century. The Kerr Cell is described as a transverse device since the optical path is at right angles to the field (or electrodes). The Pockels Cell is a longitudinal device since the ray path is parallel to the field lines and perpendicular to the electrodes. The Pockels Cell resembles the modern LC shutter in this regard. Kerr Cells use a liquid encased between transparent plates and Pockels Cells use a solid crystal. With regard to its use of a liquid rather than a solid crystal, the Kerr Cell resembles a modern LC device.
There are some patents dating from the 1930’s to the 1960’s, for stereoscopic video display systems citing the use of Kerr cells. The use of electro-optical shutters for movies, on the other hand, in this time frame, doesn’t appear in the literature. Interest in field-sequential technology for motion pictures waned after Teleview. One reason for this is that Polaroid and others began to manufacture good quality sheet polarizers. With the introduction of the sheet polarizer, stereoscopic image selection using polarization became convenient and cost effective, and field-sequential technology for motion picture applications was obsolete.
The first mention I can locate for the use of electro-optical shutters in a field-sequential stereoscopic system is in a 1938 patent by Cawley, who recommends the use of a Kerr Cell. (If any reader can find an earlier reference, please let me know.) A 1947 patent by Reynolds, also using Kerr Cells, may be the earliest reference for a field-sequential electronic display in its modern form.
Kerr and Pockels cells require high voltage and would be expensive to use. A significant step was taken by Land and others, at Sandia National Laboratories, with the development of the PLZT shutter (Lead, Lanthanum, Zirconate, and Titanate).
The first manufacturers to use PLZT devices for field-sequential stereoscopic displays were Honeywell, for a video product and Megatek, for a field-sequential computer graphics product. They were offered for sale in the early eighties. These displays flickered because they refreshed at 60 fields/second, with each eye seeing only 30 fields/second, which is quite a bit below the critical fusion frequency. One of the improvements I made at StereoGraphics was a way to run the display at 120 Hz to eliminate flicker. This advance, made in 1981, constitutes one of the major steps in the advancement of modern stereoscopic displays.
The Megatek product was based on work by John Roese of the Naval Ocean Systems Laboratory in San Diego. Both Honeywell and Megatek products used visors connected by a cable to an electronics box used to drive the shutter in synchronization with the video field rate.
The PLZT shutter is made of a solid ceramic wafer into which interdigitated electrodes are embedded. It resembles other electro-optical devices in which the material which produces the electro-optical effect is acted upon by an electric field, and it uses the arrangement of polarizers described above.
Several hundred volts are needed to drive the device, a drawback to many less adventurous users. I’ve measured PLZT devices manufactured by Motorola, and they had a 3% peak transmission in the open state, and a slow transition from closed to open. However, for the transition from open to closed, the device is fast, just like an LC shutter. These results are at odds with the literature which usually gives a 15% transmission and high speed. However, the device does have a good dynamic range, measured at 250:1, not the 2000:1 reported. Whatever its drawbacks, the PLZT made possible the first flickerless field-sequential electro-stereoscopic video display, as noted, and was used in products shipped for the next three years for both video and computer graphics applications.
Roese, who was the first person to suggest the use of the PLZT shutter for stereoscopic applications, may have also been the first person to suggest the use of the LC shutter for this application. The specific device Roese was thinking of using, confirmed from an examination of the patent file, was an early kind of twisted nematic LC cell. It’s slow for a stereoscopic application, taking tens of milliseconds to make the transition from closed to open. The requirement is for a transition on the order of one millisecond, since computer graphics displays typically refresh in less than one millisecond, and video displays using the extant CRT in slightly more than a millisecond.
A breakthrough in the technology of LC devices came with Fergason’s introduction of the fast surface mode shutter. In the surface mode device, the bulk of the molecules in the LC material do not have to move, as is the case for the twisted nematic device. The electro-optical effect occurs entirely at the surface layer where LC material comes in contact with the alignment layer coated to the inside surface of the glass cell containing the LC. Since molecular movement at just the surface is required, not throughout the entire bulk of the LC material, the surface mode device is an order of magnitude faster than the twisted nematic device.
Surface mode shutters, constructed using polarizers as described above, require only a few tens of volts, open in 2 to 3 milliseconds, close in about .2 milliseconds, and can have a transmission of 32%. They may be made on the same assembly line as the ubiquitous twisted nematic part, and they can be relatively inexpensive to manufacture. However, they have a poor dynamic range, of about 20:1. That means that the transmission of the light passing through the shutter in its closed state is only 1/20th that of light passing through the shutter in its open state. Shutters for electro-stereoscopic displays CRT required a dynamic range a few hundred to one to isolate the two channels of picture information from each other.
A headband visor using surface mode devices was introduced in 1984 by StereoGraphics. In order to obtain suitable dynamic range two LC cells in optical series were employed following a suggestion made to me by Donald Glazer, Nobel Laureate and inventor of the bubble chamber. There was some sacrifice in brightness because three polarizers were used, in an arrangement of parts that consisted of polarizer, LC cell, polarizer, LC cell, and polarizer. Compared to a shutter using a single part, such shutters are heavier, bulkier, more expensive to make, and they use more power. Shutters made of LC parts in series would not have been suitable for use in light weight, tetherless eyewear such as CrystalEyes, which will be described below.
Another approach using cells in optical series to producing a high dynamic liquid crystal device was taken by Haven. This shutter is similar to the push-pull device manufactured by StereoGraphics. Haven’s device has been commercially employed in eyewear with active lenses, and StereoGraphics Corporation has used the push-pull device in large liquid crystal panels. In these devices the cells are electrically driven out of phase with each other, to improve performance by compensation of optical artifacts.
Large LC panels of the surface mode type, also known as pi-cells, used as a stereoscopic selection device, were announced in 1983 by Tektronix, which attempted to seed the market with gifts to opinion leader end users, but units weren’t sold for a year or two after that. (A judgment against Tektronix was made in favor of Fergason awarding him ten million dollars for Tektronix’s infringement of his invention of the pi-cell.) StereoGraphics Corporation, which licensed from Fergason, shipped to customers a version of the large modulator, called the ZScreen in 1986. In this approach the modulator covers the display screen and produces circular polarized light which changes its handedness at the monitor’s refresh rate. Viewers look at the display wearing glasses with circular polarizing lenses, which allow for head tipping without producing the ghost image usually associated with linear polarizers. Although users look through shutters made up of large LC panels and polarizing eyewear, which form a shutter, for all the world the technique appears to be identical to that employed in movies, using passive polarizing glasses and sheet polarizers covering the projection lenses. And indeed, this is the basis for the RealD projection system.
HIGH-DYNAMIC-RANGE SURFACE MODE DEVICES
For a period of a couple of years the large LC panel looked like the answer to the quest for a selection device for use on workstations. Most of the users were in the field of molecular modeling. However the large panel is expensive to manufacture and has optical problems not easily solved. A major undesirable attribute of the product is the relatively low dynamic range of about 40:1, which was even lower off axis.
An alternative to the large panel and passive glasses is active eyewear using LC shutters and batteries for power. In 1983 my associates and I prototyped tetherless battery powered eyewear using the PLZT goggles StereoGraphics’ was manufacturing. The goggles were plugged into a box worn on the belt. The box was a battery powered radio receiver which picked up synchronization information from a transmitter, to allow the shutters to be driven in sync with the video field rate. The device worked, but it was cumbersome. Using small LC shutters mounted in cordless eyewear, we thought, would be a better approach. It would eliminate the cost of manufacture associated with the large panels. Toward that end we set out to improve the surface mode device, and invented the high dynamic range LC shutter the achromatic shutter which is made up of a polarizer, a liquid crystal cell, a quarter wave retarder, and a polarizer.
The concept of using wireless active eyewear incorporating shutters synchronized to the display refresh rate by means of a wireless link is not a new one. It was first proposed, as far as I am able to determine, by Hope in a patent issued in 1971. To the best of my knowledge, a product based on this invention never appeared in the marketplace. Hope’s concept used a radio transmitter broadcasting synchronization information to eyewear which used an electromechanical shutter. The shutter is made up of two ruled gratings. When the rulings coincide, the shutter is open, and when one of the gratings moves by the width of the pitch, the shutter is closed. The movable shutter element is actuated by an electric motor which has mounted on it an eccentric cam, as shown in the patent drawing. The motor’s speed is controlled by the transmitted synchronization signal.
I lead the design team at StereoGraphics which created CrystalEyes, active eyewear which uses electro-optical and infrared (IR) rather than electromechanical and radio technology. It was the first product of its kind to reach the market. CrystalEyes eyewear weighed 3 ounces and consisted of a compact package of electronics and electro-optics. An infrared sensor in the eyewear receives synchronization signals from an IR emitter stationed near the display monitor.
CrystalEyes began shipping in July 1989, and fostered widespread acceptance of electro-stereoscopy for computer graphics. Key components, such as monitors, computers and graphics boards capable of operating at 120 fields per second, were available from a number of vendors, as were a large selection of stereo-ready software for many applications.
This product has been responsible for the widespread acceptance of stereoscopic viewing for movies and especially for television. The Xpand product is a direct descendant, as are all other shuttering eyewear, which have become the basis for modern stereoscopic television from every major TV set manufacture. What was once a niche product from a relatively unknown company, StereoGraphics, may become one of them most important electronic product categories in consumer electronics. The new offerings are of identical concept to that which my colleagues and I at StereoGraphics created two decades ago. I estimate we shipped well over 100,000 pairs of CrystalEyes to industrial customers, but now active eyewear products will be shipping in the millions to people all over the world for entrainment applications. It’s hard to put into words what a satisfying experience it has been to have played a major part in this transition to stereoscopic imaging.