WO1994027410A2

WO1994027410A2 - Method for recording and reproducing moving or fixed three-dimensional images

Info

Publication number: WO1994027410A2
Application number: PCT/EP1994/001603
Authority: WO
Inventors: Erwin Sieber
Original assignee: Erwin Sieber
Priority date: 1993-05-19
Filing date: 1994-05-19
Publication date: 1994-11-24
Also published as: WO1994027410A3

Abstract

For recording three-dimensional images, a two-dimensional, specifically controllable optical filter is placed close to a photosensitive surface. Different points or slits of the optical filter are made transmissive in a set time sequence to give a height and lateral inversion image of the object concerned. For reproduction purposes, a device working on the same principle is used.

Description

Process for recording and reproducing moving and still three-dimensional images

Conventional image display devices such as cinema, television or even a drawing on a sheet of paper are characterized in that light rays are emitted in all directions from every point of their (two-dimensional) display surface. These are then naturally perceived by both viewers in both eyes, which creates a two-dimensional impression.

Three-dimensional vision is made possible in that the two eyes of a viewer perceive two different points from two perspectively different images.

A number of different methods are already known for reproducing three-dimensional images. Most of them are designed so that two perspectively different images are fed to the viewer's two eyes via two separate channels. As is known, red / green glasses or polarizing glasses with horizontal vertical polarization or mirror systems are used for this purpose. However, these procedures are not autostereoscopic, i.e. when the observer's position changes, the perceived perspective does not change automatically.

In addition, several methods for glasses-free, autostereoscopic reproduction of images are already known, but they have different disadvantages:

In the method according to patent application US Pat. No. 5,132,839, a special backlighting apparatus with a lens on a screen arranged in front of it projects parallel projections, which are reproduced in the correct height and in the right direction, in different directions via parallel light beams. However, these parallel projections cannot be obtained using imaging technology, but can only be calculated using a computer system.

1 the. And the human brain perceives parallel projections of larger scenes as unnatural, since it usually only receives central perspectives.

The method according to patent application DE 40 04 739 A1 is only suitable for viewing on a predetermined horizontal viewing plane in which both eyes of the viewer must be located, for example in a vehicle in which the viewer always takes the same seat height.

In its publication in Applied Optics 1990, Vol. 29, No. 29, pp. 4341 - 4342 describes A.R.L. Travis, in addition to his already mentioned patent application US Pat. No. 5,132,839, discloses a method in which a vertical slit diaphragm ensures that a viewer's eyes can only see the views associated with their respective position. This process leads to enormous light losses of the picture tube used, the greater the more different perspectives are to be represented, since the slits have to be correspondingly narrower. The patent application DE 41 23895 A1 also corresponds to this method.

A method according to the publication by H. Bergmann in Bild & Ton 1991, Vol. 44, Number 2, pp. 74-79: Three-dimensional image reproduction (overview of the stand) with a liquid crystal mirror system is very complex and leaves only a few depth levels to (five in the example), the light rays in the example, in addition to the reflections on the way there and back, have to penetrate up to eight times an LCD display switched to "transparent". The corresponding light losses are preprogrammed here too.

In a method according to patent applications US 4,717,949 and EP 0 316 465 A1, two or more views are simultaneously shown on a display, with vertical, permanently glowing wires arranged behind the screen ensuring that the eyes of a viewer are only those can perceive view belonging to their respective position. The position of the light wires is precisely matched to the position of the viewer, so that there is only a narrow viewing area. Moreover This spatial multiplication places high demands on the display in terms of resolution

Because of the disadvantages described, all of these methods still do not permit a satisfactory autostereoscopic display

The aim of this invention is therefore to develop a method with which a satisfactory autostereoscopic recording and reproduction of moving and still pictures is possible.

For true three-dimensional image reproduction on a flat (two-dimensional) screen, in which an object can also be viewed from another perspective, for example from the side, by moving the head (one can therefore look around a corner, so to speak) , it is necessary that the light rays emerging from the screen surface can be controlled in any vectorial manner. From each point of the (flat) screen, different light rays are emitted in different spatial directions in color and brightness. This is only partially used today for mono-chromatic Light achieved through laser holography

This problem can be solved by the novel autostereoscopic method described below for a complete 3-D image recording and reproduction system with the aid of vector images.

Definition of the term vector image (see Fig. 1):

When recording a vector image, light rays are emitted in all directions from the various points A, B, and C of an object to be recorded. Some of these light rays pass through a small hole L in the lens plane g distant and hit it at a spatial distance b film F (or camera display) arranged at the points A ', B' and C in the film plane behind it. The image generated on the film

3

ORIGINAL DOCUMENTS A ', B', C (practically a photo obtained with a classic pinhole camera) now forms a vector image of the object A, B, C with respect to the point L.

Definition: A vector image is a central perspective that is inverted in height and in relation to the recorded object.

Such a vector image can also be computed with a computer instead of obtained in terms of imaging technology, the coordinates of the points A ', B' and C being easily calculated from the coordinates of the points A, B and C, the object distance g and the image width b according to the ray set to let

If the position of the hole L on the lens level is changed over time, different vector images of the object with respect to different points L can be obtained in succession in the film level on an image-wise moving film (or from the camera display)

When a vector image is reproduced (Fig. 2), there is now a point-shaped light source L in the light plane at the corresponding point of the hole in the lens plane, which emits light beams in all directions. Some of these light beams pass through the wrong way up and down through the associated vector image reproduced points A ', B' and C in the film plane (or the display) arranged at a distance b

With this arrangement, the vectorial situation of the light rays emitted by the object A, B C with respect to the one point L at the time of the recording can be completely reproduced. The impression thus arises of a virtual object A, B, C arranged at a distance g

If, in turn, the position of the light source L in the luminous plane is changed over time in synchronism with the recording process, and in the film plane the corresponding to the

4 If the vector images associated with position L are provided, the vectorial situation of the light beams sent by the object A, B, C through the lens plane can be completely reproduced. A prerequisite is a sufficiently large film plane in relation to the objective plane.

Light rays are therefore emitted from the film level into different solid angles with different colors and brightness.

Following this basic method, the following system for recording and reproducing three-dimensional images is now conceivable

In front of a camera display, at a distance b in the lens plane, there is a selectively controllable two-dimensional optical filter (Fig. 3, for example an LCD display, hereinafter always called LCD display). On the LCD display, one point (pixel) is switched to permeable by one control electronics, while the rest are impermeable. This can e.g. line by line from top to bottom and within each line point by point from left to right.

In order to achieve recordings on a different size scale, the camera display and the LCD display can also be sized accordingly, in which case e.g. only every second point on every second line is switched to permeable. According to this principle, an almost infinite change in size of the imaging scale is also conceivable. The selection of the correspondingly required pixels of the camera display and their combination to form a vector image can be made on a correspondingly built graphics card in a computer.

If the last point at the bottom right has been switched to permeable, the cycle starts again at the top left. In this way, a current vector image can be obtained at every point on the LCD display on the camera display. These vector images become a reproduction device via a transmission channel (e.g. also via radio waves)

(Fig. 4) transfer:

5 A two-dimensional, selectively controllable light source (for example a television tube, hereinafter referred to as picture tubes) and a two-dimensional, selectively controllable optical filter (LCD display) are placed one behind the other at a spatial distance b (for example a few centimeters) transmits white light beams of the same intensity into the room in a hemispherical shape synchronously to the recording device described above, eg line by line from top to bottom and within each line point by point from left to right

On a color and grayscale-capable LCD display arranged at a distance b, for example a few centimeters in front of the picture tube, the corresponding current vector image straight from the recording device is shown for each point of the picture tube. All points are shown on the picture tube, for example If the cycle has been run through from the top left to the bottom right, the cycle begins again at the top left. This arrangement enables moving and still three-dimensional images to be reproduced

Likewise, as already explained above, the respective vector images can also be calculated by a computer, for example as part of a simulation program. It is also conceivable that the vector images originate from a suitably built video recorder

This system for recording and reproducing three-dimensional images just described, however, still comes up against the limits of the current state of the art in the case of higher graphic resolutions

In order to obtain a moving picture that is known from cinema or television and is still reasonably flicker-free, the picture tubes should run through about 25 complete cycles every second, e.g. from top left to bottom right (a so-called Interlaced mode, interlaced mode, as it is used today on television) And for each light point of the picture tube, a complete current vector image must be shown on the LCD display If, for example, a resolution of approximately 600 lines of 800 pixels is used for both the picture tube and the LCD display, 480000 LCD display points have to be displayed 25 times per second for 480,000 white picture tube points, which is 5760 Billions of pixels per second A channel for the analog transmission of such an amount of information had to have a bandwidth of several thousand gigahertz. With digital image transmission with image compression, as is already common in the computer field today, this bandwidth could be increased by a factor of 10 reduce to 100, but a bandwidth of approximately 100 gigahertz would still be required. In addition, there are problems with regard to the persistence of the picture tube and the inertia of the LCD display

In order to make this method, which is in principle correct for recording and reproducing three-dimensional moving and still images, even for higher graphic resolutions according to the current state of the art, it can be reduced considerably in some points

a) Reduction to only horizontal three-dimensionality,

b) reduction to a small viewing angle within a few degrees to the left and right of the optical axis,

c) Reduction of the frame rate.

In the variants of the above method described below, for the sake of simplicity, an image resolution of 600 lines of 800 pixels is still expected, which corresponds approximately to the resolution of the familiar television picture. Of course, other picture resolutions and other picture repetition frequencies are also conceivable

Re a) Reduction to only horizontal three-dimensionality: Fig 5 The picture tube is not described horizontally line by line and within a line point by point, but column by column vertically and the respective columns e.g. from left to right, whereby one vertical column is completely illuminated and emits white light. All are on the picture tube Columns have been run through from left to right, so the left column continues with the first column. On the color and grayscale-capable LCD display, which is arranged, for example, a few centimeters in front of the picture tube, the associated current vector image is added to the respective light column of the picture tube reproduced for horizontal three-dimensional display only

These vector images for only a three-dimensional horizontal representation can be obtained in the same way as the vector images used to date (Fig. 1), the position of the hole L only being changed from left to right on a horizontal line, for example. This horizontal line makes sense through the optical axis of the system

With only horizontal three-dimensional recording, it is also possible to improve the light yield during the recording. For this purpose, not only a point, but a vertical gap is switched translucently on the LCD display placed in front of the camera display (Fig. 6) By means of a cylindrical collecting lens or a cylindrical lens system, the light rays emitted by the object, for example from point B, in the direction of the permeable gap are bundled again at point B '

As can be seen from Fig. 5, in the case of only horizontally three-dimensional reproduction, the current vector images belonging to the respective vertical luminous lines of the picture tubes still have to be reversed, but are displayed at the correct height. Furthermore, as can also be seen from Fig. 5, correct reproduction is necessary A slight compression of the vector images in the vertical direction is necessary due to the aspect ratio. The reason for this lies in the different distance between the picture tubes and the virtual image on the one hand and between the front LCD display and the virtual image on the other hand. The distance between the viewer and the LCD display plays a role The factor for the required compression can in turn be easily calculated from the geometry using the radiation sets

Definition: A vector image for a horizontal, three-dimensional representation is a central perspective that is inverted but reversed in height with respect to the recorded object and that is more or less compressed in the vertical direction.

A corresponding compression for a medium observer distance is expediently carried out on the recording side by means of a correspondingly dimensioned cylindrical attachment lens. A slight discrepancy in the reproduction is not perceived. When calculating vector images for only three-dimensional horizontal representation with a computer, this can be done Compression in the program

With this arrangement, moving three-dimensional images can also be recorded and reproduced, which, however, are only two-dimensional in the vertical direction.You can look around a corner of the house, for example, by moving the head horizontally, but not by moving the head vertically, e.g. over a wall If you turn your head until the eyes are no longer next to each other, but one above the other, the image loses its three-dimensional effect with decreasing horizontal eye relief

This system for horizontal 3-D display reduces the bandwidth of the transmission channel required for the transmission by a factor of 600 compared to the complete 3-D system described above for an image with, for example, 600 lines

Re b) Reduction to a small viewing angle within a few degrees to the left and right of the optical axis:

As in a), the picture tube is again described vertically in columns. On the again, for example, a few centimeters in front of the picture tubes, there are color and grayscale capable

9

ORIGINAL SUBLA E As in a), the LCD display shows a corresponding current vector image for each column of the picture tube, but no longer over the entire width of 800 points of the LCD display, but only over an image segment of, for example, 10 picture elements on the left and right of the one that is just illuminated Biotube column (Fig. 7). The other points on the LCD display remain unused and are, for example, opaque. This again considerably reduces the channel bandwidth required for image transmission, in the example with 800 pixels in a complete line, of which only 20 are used by a factor of 40.

Overall, the following reduction can be achieved compared to the bandwidth of several thousand gigahertz mentioned above:

e.g. Factor 10 through digital transmission and image compression, factor 600 through only horizontal three-dimensional representation and factor 40 through reduction to a small viewing angle to the left and right of the optical axis. The required bandwidth is thus reduced by a total of 240,000 and, depending on the fineness of the color representation, is of the order of about 10 to 20 megahertz, which is technically feasible today.

It is now advisable to make full use of the LCD display anyway. Several vertical light columns are shown simultaneously on the picture tube, 40 in the example above. Columns 1, 2, 3,..., 800 are no longer illuminated in sequence on the picture tube, but instead columns 1, 21 are simultaneously illuminated , 41 781. Then the

Columns 2, 22, 42, ... 782 illuminated etc. After the passage 20, 40, 60, ..., 800 the columns 1, 21, 41, ..., 781 are continued (Fig. 8 ). This can e.g. can be achieved by a graphic card in connection with a control program.

At the same time, the vector image sections of the image segments associated with the respective light columns are shown on the LCD display.

10

ORIGINAL DOCUMENTS The consequence of this is that the permissible afterglow duration of the picture tube and the permissible inertia of the LCD display can be, for example, 40 times as large. With an image repetition frequency of 25 Hertz, a maximum of 500 brightness changes per second are achieved by an LCD display Must be able to carry out the element (pixel). This value is achieved, for example, by Ferrokπstall displays with an active matrix today and even surpass picture tubes with a correspondingly short persistence

With this parallel spelling of several light columns, the following (welcome) additional effect occurs (Fig. 9)

If the viewer moves out of the reduced viewing area (a few degrees to the left and right of the optical axis), he may still be able to perceive the image correctly spatially, since neighboring viewing areas result from the fact that display Image segments are not backlit by the light columns actually provided for these segments, but by the neighboring light columns, for example offset by 20 columns. A prerequisite for spatially correct vision is, however, that both eyes of the viewer are in such neighboring viewing areas

Analogously, the above explanations can also be used on the recording side to reduce to a small recording angle within a few degrees to the left and right of the optical axis (Fig. 10)

On the LCD display in front of the camera display, several vertical columns are simultaneously switched to transparent. On the camera display, the vector image sections are generated in the respective display segments

to c) Reduction of the frame rate

1 1

ORIGINAL DOCUMENTS The image repetition frequency can also be reduced downwards. The human eye can no longer resolve films that run at 16 frames per second into individual images. For example, if you approach this lowest limit with this system, or even even slightly lower at 12.5 frames / second, so with this system the two eyes of a viewer (see Fig. 5) eg for point A usually receive their light information at different times T1 and T2, since the whole process ren is based on temporal multiplexing. This results in an effective image repetition rate of around 25 Hertz for both eyes

As a result, the requirements for the bandwidth of the transmission channel, for the afterglow duration of the picture tube and for the speed of the LCD display can be halved again. In order to maintain good flicker-free operation, the field-picture method known from television ( Interlace mode) can also be applied accordingly to the vertical illuminated lines

Other system variants:

Other variants for reducing the original, complete 3D system are also conceivable. For example (Fig. 11), 10 dots on the left, right, above and one below a luminous dot of the picture tubes can form an image segment on the LCD display. The picture tube is then described point by point as follows points 1, 21, 41,, 781, then line 21 points 1, 21, 41,, 781, then line 41, so that there is a luminous grid of points on the associated image segments on the LCD - The associated vector image sections are shown on the display

Then the next point grid is written offset on the picture tube and the associated vector image sections are shown on the LCD display again in the associated image segments until all the luminous dots of all rows and columns have been detected on the picture tube. Then it is started again the system is reduced to a small viewing angle on the left, right, above and below the optical axis, without the vertical 3-D display being lost.

12 ORIGINAL DOCUMENTS ke around and look over a wall. This variant could be used for example in CAD systems (Computer Aided Design).

Fig. 12: A variant analogous to the example above is also conceivable on the recording side.

Areas of application:

If one now considers the technical requirements that this autostereoscopic method for recording and reproducing moving and still three-dimensional images with the aid of vector images, they are, depending on the variant, graphic resolution and image repetition frequency, according to the current state of the art realistic and realizable, since even small personal computers with clock frequencies of 75 megahertz and above work. Corresponding graphics cards (graphics subsystems with their own processor) in conjunction with suitable programs can take over the synchronous, chronologically correct control of the camera display, the upstream display and the recording of the vector images on the recording side.

Corresponding graphics cards can also take over the synchronous, chronologically correct control of the picture tube and the LCD display as well as the supply of the LCD display with the respective vector images. A prerequisite is a picture tube with a surface that is not excessively curved and a linear representation without pillow-shaped and clay-shaped distortions. This can be achieved today with appropriate electronic circuits. Alternatively, instead of the picture tube, e.g. a laser scanner or another flat backlit LCD display are also used.

Thus e.g. three-dimensional representations (CAD drawings, chemical molecules, simulations etc.) with sufficient graphic resolution on computer monitors are conceivable. An application of this 3-D method is also conceivable in the field of television.

1 3

ORIGINAL DOCUMENTS Dimensions of 3-fX monitor systems Prototype Monitor Monitor Monitor TV50 TV50cs TV50CS TV90hd TV90HD Ki

Number of channels: 32 16 32 64 64 64 64 128 256 5

Refresh rate 25 [Hz] 5 30 30 30 25 25 25 25 25

Pixel frequency: [Hz] 160 480 960 1920 1600 1600 1600 3200 6400 128

System parameters:

System quality: 70 70 35 70 70 70 70 70 70 140 2 o Horizontal resolution: 640 [pixels] 640 640 800 1024 800 1024 1024 1024 2048 20 z Vertical resolution: 200 [pixels] 480 480 600 768 600 600 600 1200 1200 1200 12> r Pixel spacing hear .: 0.36 [mm] 0.36 0.36 0.36 0.36 0.80 0.80 1.00 1.00 0.50 4, c Pixel spacing vert: 0.72 [mm] 0.36 0.36 0.36 0.36 0.80 0.80 1.00 0.50 0.50 4, z HW derived values: r screen width: 230.40 [mm] 230 , 40 230.40 288.00 368.64 640.00 819.20 1024.00 1024.00 1024.00 8192, ► o foam height: 144.00 [mm] 172.80 172.80 216.00 276.48 480.00 480.00 600.00 600.00 600.00 4800,

M Z screen diagonal: 271.70 [mm] 288.00 288.00 360.00 460.80 800.00 949.47 1186.83 1186.83 1186.83 9494,

Screen size: 11 [inches] 12 12 15 19 32 38 47 47 47 3

Display distance 25.20 [mm] 25.20 12.60 25.20 25.20 56.00 56.00 70.00 70.00 70.00 1120,

Vector angular width 0.82 [degrees] 0.82 1.64 0.82 0.82 0.82 0.82 0.82 0.82 0.41 0,

Stereo angle width in Im: [mm] 228.57 228.57 228.57 457.14 457.14 457.14 457.14 914.29 914.29 914,

Number of display segments 20 40 25 16 13 16 16 8 8

Column spacing [mm] 11.52 5.76 11.52 23.04 51.20 51.20 64.00 128.00 128.00 2048,

Stereo angle width: [degrees] 25.75 25.75 25.75 49.13 49.13 49.13 49.13 84.87 84.87 84, smallest observer distance: [mm] 635.25 635.25 761, 25 468.83 765.63 961.63 1185.63 592.81 592.81 4512, largest viewing distance: [mm] 4200.00 2100.00 4200.00 4200.00 4200.00 4200.00 4200.00 4200, 00 8400.00 16800,

Claims

Claims:

Claims are made

1 for autostereoscopic methods for the reproduction of three-dimensional images, in which a two-dimensional, selectively controllable optical filter is located at a spatial distance next to and parallel to a two-dimensional, selectively controllable light source, with different points of the light source being illuminated in sequence and simultaneously a vector image (upside-down and reversed central perspective of the object to be displayed) is shown on the optical filter at the respectively illuminated point of the light source

2 for autostereoscopic methods for reproducing horizontal three-dimensional images, in which a two-dimensional, selectively controllable optical filter is located at a spatial distance next to and parallel to a two-dimensional, column-wise controllable light source, various vertical columns of the light source being illuminated in chronological order and at the same time as the respectively illuminated column of the light source on the optical filter, a vector image for only horizontally three-dimensional representation (inverted but high-night central perspective of the object to be displayed, which is more or less compressed in the vertical direction) is displayed

3 for autostereoscopic methods according to claims 1 and 2, in which at the same time or almost simultaneously several points or vertical columns of the light source are illuminated and on the optical filter simultaneously in image segments associated with the respectively illuminated points or vertical columns vector images or sections of vector images being represented

4 for autostereoscopic methods according to claims 1 to 3, wherein the light source is a cathode ray tube

1 5 ORIGINAL DOCUMENTS 5 for autostereoscopic methods according to claims 1 to 3, wherein the light source is a laser projector with a screen

6 for autostereoscopic methods according to claims 1 to 3, wherein the light source is a backlit liquid crystal display

7 for autostereoscopic methods according to claims 1 to 3, wherein the optical filter is a color and / or grayscale liquid crystal display

8 for autostereoscopic methods for recording three-dimensional images, in which there is a two-dimensional optical filter that can be controlled at certain points next to and parallel to a two-dimensional recording display, with different points of the optical filter being switched through in sequence in time , while the others are impermeable, and at the same time a vector image (upside-down and side-reversed central perspective of the object to be recorded) is recorded on the transmittable point of the optical filter on the recording display

9 for autostereoscopic methods for recording horizontal three-dimensional images, in which a two-dimensional, column-wise controllable optical filter is located at a spatial distance next to and parallel to a two-dimensional recording display and in front or behind a cylindrical converging lens or a parallel to the optical filter cylindrical lens system, whereby various columns of the optical filter are switched to be transparent in chronological order, while the others are opaque, and at the same time a horizontally three-dimensional vector image (inverted but high-angle central perspective of the object to be picked up) is taken upside down

10 for autostereoscopic methods according to claims 8 and 9, in which a plurality of points or vertical columns of the optical filter are switched through simultaneously,

16 while the others are impermeable, and at the same time vector images or sections of vector images are recorded on the recording display in image segments assigned to the respective transparent points or columns

11 for autostereoscopic methods according to claims 8 to 10, in which the recording display and the optical filter are dimensioned so generously with regard to the geometric dimensions and the number of pixels that by using only a part of the available pixels, reductions and enlargements of the recorded image section (macro and telephoto) compared to the method according to claims 8 to 10 are possible

12 for autostereoscopic methods according to claims 8 to 11, wherein it is in the recording display is a color and / or grayscale-capable recording tubes

13 for autostereoscopic method according to claims 8 to 11, wherein it is in the recording display is a color and / or grayscale-compatible liquid recording recording display

14 for autostereoscopic methods according to claims 8 to 11, wherein the optical filter is a Flussigkπstall display

1 7

ORIGINAL DOCUMENTS