WO1994020301A1

WO1994020301A1 - Light beam image recording and input apparatus and method

Info

Publication number: WO1994020301A1
Application number: PCT/US1994/002890
Authority: WO
Inventors: Richard W. Pease; Redge Johnson
Original assignee: Pease Richard W; Redge Johnson
Priority date: 1993-03-10
Filing date: 1994-03-10
Publication date: 1994-09-15
Also published as: AU6410594A

Abstract

Image input and recording apparatus are disclosed which have a resolution capability of conventional photographic film recorders and which accommodate color image generation by simultaneously or separately scanning beams (102, 104, 202) differing in color and black and white image generation by scanning one beam. A glass plate (38) made of fibers (181) in combination with a lens (34) and a spotsize selector (54) including glass plates (60, 62) set at opposing angles (64) maintains constant beam spot size. Apparatus orthogonality errors uncorrected by optical elements, non-linearities, and beam power drift are corrected using electronics including look-up tables (U4, U6, U9, U10) and a digital frequency synthesizer (U8). Software accepts data in standard formats or formats specifically for the apparatus and preforms image data processing. The input apparatus includes photodetectors (108, 112, 116) opposite the plate (38) from beam sources (100, 200). A method for converting image focus locus shape is also disclosed.

Description

LIGHT BEAM IMAGE RECX-KDING AND INPUT APPARATUS AND METHOD

Technical Field

The method and apparatus of this invention relate generally to the recording and acquisition of images. More specifically, it relates to the inputting or recording of electronic images at extremely high resolution but in an economical time period. The

<l images to be recorded may be on a mass storage device such as a hard or floppy disk, analog storage device, or the like. These images would ordinarily be generated by, or acquired for, a "host" computer located away from and external to the recorder disclosed herein. This image data may have been generated in a standard format such as TARGA or TIFF or may be in a format specifically designed for the present invention. In the latter case, the data may be read directly into the recorder apparatus without the need for reformatting. In the former case, or in the later case if additional processing of the image data is desired, the host generated data may first be read into the recorder's system computer. In this situation, the image data can be first read from the mass storage device into recorder system memory where the required reformatting or image data processing may be accomplished. In either case the data is then sent through the system electronics and optics systems for recording on a recording medium such as film or paper recording medium and the like. The images may be recorded in color using several beams scanned at once (reducing scan time by 2/3) or individually, or the image may be generated in black and white using a single beam. The scanning of the image may be accomplished using laser or other light means.

When used in the image input configuration, the process is essentially reversed with the system scanning an image from an imaging medium and the image data detected by one or more photodetectors. This image data is then input to the system electronics and sent to the system computer where it may be transferred to a mass storage device or directly to a host.

The apparatus and method of the present invention also provide a novel means for converting the complex locus of image points resulting from the image scanning processes, into a format for recording of the image or for projection and enlargement.

In prior art, image input and recording devices are done at resolutions which do not fully exploit the resolution capability of presently available film or paper recording mediums. Consequently, these devices create images which are less suitable for enlargement. Further, those devices which do operate at higher resolutions, are extremely slow. Consequently, there is a great need for the present apparatus which is capable of generating an image at resolutions approximating currently available film or paper recording mediums yet being fast enough that the scanning process may be accomplished in an economical time period. Description of the Prior Art

Prior art devices for scanning an image on film or paper recording medium are well known. However, these devices generally permit the recording of images only at resolutions which are more coarse than the resolving capabilities of the film or paper recording medium on which they are being recorded. Additionally, some of the prior art devices recognize some of the errors which may occur during the scanning process. However, the corrections implemented by these devices are incomplete and insufficient to meet the requirements for generating an image at the resolution limit of the recording film or paper recording medium.

One example of such a prior art device is Arai et al., U.S. Patent N² 4,667,099, which discloses an optical linear encoder for converting scanning light beams incident thereon into a series of light pulses. A conventional linear encoder is formed by a plurality of transparent and nontransparent line shape grids which are aligned parallel to the horizontal deflection direction and equidistantly separated. Arai compensates distortion caused by the optical focusing system by varying the horizontal spacing of the vertical nontransparent strips on the encoder. Consequently, the speed of the pulses from the encoder may be varied by varying the vertical position on the encoder on which the light is horizontally scanned.

-another prior art example is Radl, U.S. Patent 4,284,994, which discloses a laser beam recorder apparatus arranged to minimize line scanning displacement error due to use of a rotating polygonal mirror scanner. The apparatus includes a laser providing a beam modulated in response to electrical input signals representative of the desired image and a rotating polygonal mirror for scanning the image on a film or paper recording medium in one dimension. The scan of the image in the second dimension is accomplished by means of a mechanical system. After reflection by a mirror, the rays return through the scan lenses and are focused to an image point on the film or paper recording medium. The system is limited to a maximum resolution of 2,048 x 2,048 pixels.

Ohnishi, U.S. Patent 4,190,167, discloses a laser computer output microfilmer with a means to avoid data character compression or expansion following a mirror stop/start or speed change operation, or a data line skip operation due to mirror inertia. The video signal generation by the apparatus is delayed for several line scans after such a mirror speed change to allow for the mirror to reach a constant rotational speed.

Minoura et al., U.S. Patent N²4,314,154, discloses a two-dimensional laser scanning apparatus having first and second deflectors. The first deflector for deflecting a collimated beam and the second deflector for deflecting the incident beam from the first deflector in a direction orthogonal to the deflection direction of the first deflector.

Yeadon et al., U.S. Patent N²4,490,608, discloses a position sensor for a flying spot scanner and incorporates an intensity modulated beam of light to expose a photosensitive surface.

Nakauchi, U.S. Patent N²4,870,506, discloses a conversion method for color image copying. The apparatus may be used with a laser beam printer or the like. The signals are photo-electrically read from a colored original by image reading element such as a charge coupled device. The data is then converted to a digital format and temporarily held in frame memory. The digital data is then sequentially read out one frame at a time and processed using a color conversion in a color processing circuit and look-up tables. The data is then reconverted to an analog format and produces light quantity control signals for the three primary colors. The signal is then fed to an exposure head to expose the photo sensitive material.

Brown et al., U.S. Patent N²4,200,830, discloses an apparatus for compensating for the inherent angular errors in a rotative mirror causing variations in line spacing and begin/end points. The system compensates for inherent defects in the angular relationship between facets of a rotating polygonal mirror by providing a mirror element which is pivoted to correct the scanning errors caused by these angular defects. The defects caused by variations in the angles between the facets of the polygonal mirror are also corrected by an electronic circuit which includes a delay device capable of delaying the scanning by a predetermined amount. Hudson et al., U.S. Patent N² 4,180,822, discloses an optical scanner and recorder which includes a means to direct a beam from a laser through an acousto-optic modulator.

Japanese patent ^a 143,013 discloses a method for recording pictures wherein a modulated beam is deflected by rotary polyhedral mirror and condensed into a spot shape on the incident edge face of an optical transmission material such as optical fiber through a troidal lens. Due the curvature center of the troidal lens the focused position is not fluctuated in the incident edge phase with direction even if the direction of deflection beam is fluctuated due to the mirror plane.

With the geometrically increasing computational power of small computers and advances in digital image processing, the capability of enhancing and processing images using computers is enormous. Similarly, the speed and storage ability of such computers is also increasing greatly to the point where images of incredibly fine resolution may be accommodated. Unfortunately, as so often is the case, the transfer of the electronic data in the computer to a "real world" medium, such as film or paper recording medium takes orders of magnitude longer than to actually process the data by the computer. Additionally, the physical requirements of an optical system to faithfully reproduce this image precision are very demanding. Thus, although the computational ability of the computer may afford the opportunity to handle images of great resolution, until now the realization of such images has imposed time and physical considerations making such realizations impractical.

The major advance represented by the present invention provides, for the first time, a practical means for physical realization of high resolution images at the limit of the recording medium's resolution and in a time frame which makes the recording of such images practical.

Consequently, it is a prime objective of the present invention to create an image recording device which is capable of recording an electronic image on a recording medium at a resolution approaching that of the recording medium itself.

It is a further objective of the present invention to allow a high resolution image to be created on the media or in the computer buffer in an economical time frame.

Yet another objective of the present invention is to provide an image input apparatus capable of acquiring an existing image and storing the image. Still another objective of the present invention is to provide an apparatus capable of compensating for the errors inherent in the scanning process.

Yet another objective of the present invention is to provide an apparatus capable of creating or inputting images in color.

Still another objective is to provide an apparatus capable of recording or inputting an image in color wherein all colors may be scanned simultaneously.

Another objective is to provide an apparatus capable of creating or inputting a monochrome image.

Still another objective is to provide a method of converting the locus of focus points from the scanning process into an alternative format field such as flat for recording on film or paper recording medium or other shape, such as spherical for projection by a spherical lens or other shapes as required.

Yet another objective is to provide an apparatus permitting image data generated by an external host computer in a standard format to be reformatted for use by the recording system.

Still another objective is to provide an apparatus to permit image data of any format to be additionally processed such as contrast and color adjustments during both recording and input processes.

A further objective is to provide an apparatus to permit image data of any format to be recorded or inputted based on a plurality of colors, some or all of which are not visible, such infrared or ultraviolet.

Summary of the Invention

A film or paper recording medium recorder apparatus wherein a high quality image may be formed on a recording medium from an electronic image using a light beam of small spot size thereby permitting a high resolution image to be formed thereon, the recorder apparatus comprising; light emitting means for producing a plurality of light beams comprising three colors, such as visible, ultraviolet or infrared, the generation and projection of said light beams forming a plurality of light paths between said light emitting means and said recording medium; focusing means positioned within said light paths for focusing said light beams; modulator means for modulating said light beams in response to said electronic image to be recorded, such that upon modulation, said beams comprise image data to be recorded; beam combining means positioned within said light path for combining said light beams comprising the three primary colors, into a single beam the projection of said single beam forming a combined light path between said mirror means and said recording medium, beam expander and pinhole means positioned within said combined light path; beam moving means for scanning said single combined beam in two dimensions; computational control means for controlling said scanning and said modulation means and the flow of said image data; and an image plane converter means positioned within said light path and adjacent said recording medium such that said combined light beam incident on said converter is transmitted in a substantially parallel fashion through said converter and perpendicular to the recording medium plane such that the image focus locus shape is converted to the shape of the recording medium adjacent to the converter, or to a shape conducive to the formation of an image in a shape to be further relayed to other remote media, producing a beam of nearly uniform spot size at said recording medium surface regardless of the position of said spot on said converter and such that said electronic image is transferred onto said recording medium.

Brief Description of the Drawings

Figure 1 is a pictorial representation showing the major optical components of the invention and the overall operation of the scanning process.

Figure 2 is a pictorial representation of the raster scanning process.

Figure 3 is a pictorial representation showing how the size of the image projection spot may obscure details in the image.

Figure 4 illustrates the prior art film recorders.

Figure 5 is a diagram showing one embodiment of the invention wherein the light from two lasers is used to make three beams, one beam from one laser and two from another, comprising three primary colors and wherein the scanning of the beam is done in both directions using two galvanometer driven mirrors.

Figure 6 illustrates the basic design and operation of a conventional cavity laser.

Figure 7 illustrates the basic principles involved in acousto-optic modulation of the light beam.

Figure 8 illustrates the preferred embodiment method of frame scanning using a galvanometer to drive the frame scanner mirror.

Figure 9 illustrates the preferred method of line scanning using a rotating polygon mirror.

Figure 10 is a top view of the rotating polygon line scanner showing the incident and reflecting beam angles and the effect of the rotating polygon on the focus point to given incident beams.

Figure 11 is also a top view of the rotating polygon line scanner showing the curved focus locus defined by a reflected scanned beam.

Figure 12 is a side view of the rotating polygon scanner showing an active mirror facet and how a beam is formed thereon. The figure also shows that the beam is fully formed on the facet for only a fraction of the time during facet rotation.

Figure 13 is a diagram showing the preferred embodiment of the invention wherein the light from two lasers is used to create three beams comprising three primary colors and wherein the scanning of the beam is done in one dimension using a galvanometer driven mirror and in the other dimension using the rotating polygon mirror and wherein the rotating polygon is between the galvanometer mirror and the recording medium. Figure 14 is a figure showing an alternative embodiment similar to Figure 13 except the frame scanning galvanometer driven mirror is situated between the rotating polygon and the recording medium.

Figure 15 shows another alternative embodiment wherein the frame scanner is a rotating drum located between the rotating polygon and the recording medium.

Figure 16 shows yet another embodiment similar to Figure 5 wherein two galvanometer driven mirror scanners are used to scan in two dimensions but where additionally, a "prescan" lens has been inserted in the beam path between the recording medium and the second scanner to flatten the curved focus field generated by the scanning procedure.

Figure 17 shows an alternative embodiment to compensate for the curved focus field generated by the scanning procedure by placing a movable focusing optic in the beam path to reposition the focus point along the beam path.

Figure 18 illustrates some of the scanning errors which occur during the scanning process including the "bowed" horizontal and vertical scan lines and the uneven distribution of pixels across the scan line.

Figures 19a-d illustrate how use of a lens can cause chromatic aberration and how those errors may be corrected by varying the position of the source of each color.

Figure 20 illustrates acousto-optic modulation with multiple pixels in the sound field.

Figure 21 illustrates how the "scophony" scanning technique is utilized in the preferred embodiment by modulating the pixels in the modulator at the same rate, but opposite direction, at which they are scanned across the line.

Figure 22 shows the spatial filtering technique employed in the preferred embodiment to aid in achievement of small beam size.

Figures 23a-d illustrate the process used in the preferred embodiment to enlarge the focused beam or change the beam shape by a predetermined amount using two rotating plates.

Figure 24 illustrates the principals used in the preferred embodiment to convert the focus field of the beam to the shape required by the recording medium using a fiber¬ optic plate.

Figure 25a-e are electrical schematic drawings showing the preferred implementation of the electronics used in the scanner.

Figure 26 shows how the image input feature of the present invention is used to input a image using three color sensitive photodetectors.

Figure 27 shows an embodiment of the input procedure using a separate lens for each of the three colors.

Figure 28 shows another embodiment of the input device wherein the image is condensed using a drawn fiber optic bundle.

Figure 29 shows an alternative input device wherein the image is concentrated using an internal mirror conical light pipe.

Figure 30 illustrates a top view of the physical layout of the various components of the invention.

Figure 31 shows the side view of the physical layout of the present invention.

Figure 32 shows an alternative physical layout of the apparatus with the lasers positioned at the top of the mounting.

Figure 33 is a block diagram showing the data flow in the system software for handling the image data and wherein the image data is on a hard disk, floppy or other external memory storage device. In this embodiment, this image data would be in a format acceptable to the present recorder apparatus and would be processed directly by the recorder without format conversion by the system computer.

Figure 34 is a block diagram showing the data flow in the system software for handling the image data wherein the image data resides in the system computer memory. In this embodiment, the image data is transferred into system memory and optionally converted into the proper format prior to being sent to the system electronics for recording.

Figure 35 is a block diagram showing the data flow in the system software for conducting an alignment and color balancing procedure for the recorder.

Figure 36 is a block diagram of the electronics used for image input.

Figure 37 is a horizontal view showing the installation and orientation of the modulator and spatial filter for one of the three color beams.

Figure 38 is a pictorial view showing an alternative embodiment of the recorder apparatus of the present invention wherein a single beam is used for scanning.

Figure 39 is another alternative embodiment wherein all three color beams are used sequentially along a single beam path.

Figure 40 is an alternative embodiment for use of the image plane converter wherein the image is converted into another shape, as in a configuration for enlargement through a conventional lens.

Figure 41 is an illustration of the image shape resulting on the flat back of the Image Converter plate, having correct for focus and having straightened horizontal scan lines, but not having corrected the other errors.

Figure 42 shows the orientation of the photodetectors utilized for analyzing test images for color balance, alignment and the like.

Figure 43 shows yet another alternative embodiment of the recorder apparatus of the present invention wherein a single laser, such as Krypton or Helium Selenium, generates the three light frequencies necessary.

Description of the Preferred Embodiment General Background

Photography is the recording of images. Over the last 100 or more years this process has been accomplished by focusing light from real objects onto chemically treated surfaces. Subsequent processing of the surface brings out the image. While photography has become an integral and necessary part of many industries, there are significant problems in the photographic processing of the images, including the composing images from segments of multiple incomplete sources, retouching, layout, and editing. Without the use of computers each of these tasks is a labor intensive process.

There are several techniques currently used to record digital imagery on film or paper recording medium. A common technique is the CRT (Cathode-Ray Tube) recorder. An electronic beam is made to sweep across the face of a CRT phosphor plate as in black-and-white television. The phosphor glows where the beam hits it. Either the strength of the beam or the rate at which the beam sweeps determines the brightness of the pixel. One of three colored filters is placed between the phosphor plate and the film or paper recording medium. The image is recorded in three passes, once each in red, green, and blue. A lens is used to "relay" the CRT face image onto the film or paper recording medium, recording through each filter. The sweeping of electronic beams is a very well developed art, but difficulties include: making of an extremely small electronic spot, moving of the beam quickly, modulating the beam to the 4000 to 1 contrast ratio that the digital image and film or paper recording media are capable of, keeping the beam moving in exactly straight lines, and making a CRT phosphor face plate with no variation in phosphor strength and free from defects or artifacts.

Another technique is to modulate a light beam as to color and strength, and then mechanically sweep the light across the film or paper recording medium. This can be done using two moving mirrors, but that process is slow, especially when sweeping along scan lines. Using a single rotating mirror sweeping the beam onto a curved piece of film or paper recording medium and moving the mirror bed along the film or paper recording medium is used with some success. The rotating mirror sweeps one line per revolution. This process is illustrated in Figure 4.

A third technique is to use three different colored laser beams, modulate each beam separately, and sweep the resulting beam onto the film or paper recording medium using either moving mirrors or a rotating mirror. Figure 5 shows the basic design of such a film or paper recording medium recorder, and further illustrates the use of two independent moving mirrors 67, 68. This figure shows a flat film or paper recording medium target 40, but leaves for later discussion of ways the focus can be made flat. Laser imaging gives a better recording rate and contrast ratio, but lenses between the moving mirrors 67, 68 and the target 40 such as those used to flatten the field, significantly limit the resolution. Such lenses cause difficulties with causing each color of light to arrive at the same place on the film or paper recording medium, due to chromatic aberration in the imaging lens 78 (Figure 16). These lenses 78 are known to the optics industry as "prescan" lenses. This non-intuitive term refers to the placement of the scanning mechanism 67,68 before ("PREscan") the lens 78 in question. The prescan lens configuration is shown in Figure 16.

With the introduction of computers, it has become practical to convert the images into digital values, manipulate the images in the computer, and reconvert the values into either dynamic displays on a screen or into paper or film or paper recording medium images. Display screens and computer color printer output have much lower resolution than photographic film or paper recording medium. Photographic films and papers currently have enough resolution for high quality photographic publishing.

Those who use high quality imagery are the ones that have the greatest need to edit, layout, and retouch their material. The more useful the computer is to manipulate the images, the more difficult it is to get the source imagery into the computer and the finished images out.

"Film recording" is the process of transferring images from computer to photographic film or paper recording medium. In this process the computer causes imagery to be recorded on photographic film or paper recording medium using monochrome or colored light, exposing the film or paper recording medium according to the requirements of the edited digital image.

There are two distinct types of color film or paper recording medium output from computers, known as "continuous tone color output" and "color separations". In continuous tone color output, color film or paper recording medium is exposed by colored light. With color separations, a separate black-and-white exposure is made for each of three or four colors. These separations are then used externally to create plates for printing, where each plate is for a different colored ink on the same image. The colorization of imagery using color separations is done outside of the film or paper recording medium recording process, so the recording of color separations is a monochrome process. A recorder that can make continuous tone color images can also make color separations. The remainder of this discussion is directed toward the continuous tone color process, but the production of color separation negatives is within the scope of this invention.

Before a computer can be used to manipulate imagery, there must be an image source. Imagery can be originated in the computer by building up image portions from fonts, patterns, figures, created figures, and free-hand drawing. This approach is not practical for many uses where the imagery desired involves people, existing products, scenery, or other complex or individually recognizable items. Such images can be loaded into the computer either by using a video camera at low resolution or by a slower, more meticulous process of image input. Here every part of a photograph or other two (or in emerging technology three) - dimensional object is examined by a photometer-like sensor to see how much light the object reflects or transmits. Most image input is done by sweeping the sensor, either physically or optically, across the input material in a series of parallel lines. In other devices, multiple sensors are used in parallel. In either case many points along the line are sensed for color content and the results are held in the computer. In the image input process the color and light values of the entire object are in this repeating-line pattern, called a "raster." Raster scanning is illustrated in Figure 2. Each of many regularly spaced points along the raster scan line in the computer is called a "pixel" 300, standing for Picture Element. As shown in the figure, an image comprises a series of vertically spaced, horizontal lines 301a, 301b, 301c of pixels 300. An image 302 is formed by "turning on" pixels using a variety of intensities to represent shades of grey.

This process gives the most voluminous representation of the image of the source object possible. No reduction in data is made for patterns or the repetitive value of adjacent pixels. It is sometimes more convenient for the computer to deal with patterns and with data that summarizes objects such as lines, letters, or areas all the same color. Much less data is then required to store the image. The process by which patterns are recognized in raster images and are converted to vector imagery is called "vectorizing". A full science called "pattern recognition" that encompasses vectorizing has developed, but is not sufficiently far developed to find much use in the computer manipulation of images of real objects. Other techniques are also used to reduce the volume of raster pixel data and are called "compression techniques". These techniques do not usually convert the image into typical vector values, such as lines or letters; but rather simply record and summarize repeated pixels and pixel patterns thus reducing the redundancies.

The converse of vectorizing is the conversion of vector imagery, as originated in the computer, into raster images, possibly overlaying imagery of real objects. This process is called "rasterizing", and is very common. The process of recording imagery from the computer is done using a raster scanning technique. In those instances where no outside real imagery is involved and the source material is created in the computer using vector formats, it is possible to record the imagery on film or paper recording medium in vector fashion, where a beam of light draws the image objects. However, due to the dynamics necessary to move the light source from place to place on the film or paper recording medium in random fashion, the process is not significantly faster than raster output, unless the image is very simple. This is because in raster output the path of the light is predictable and well ordered, and only the intensity of the light changes in a random fashion. Therefore, most film or paper recording medium recording and image input operates with data in a raster format.

Different colors are perceived because the eye has elements that are sensitive to different wavelengths of light. The eye discriminates the intensity of light in three different overlapping wavelength bands; greatly overlapped bands in the green and red, and a widely removed band in the blue. Additional information on this topic may be found in Television Handbook. K. Blair Benson, p 2.4, herein incorporated by reference. It therefore takes three distinct stimuli of the eye to detect the range of visible colors. The eye is also sensitive to low light levels on a non-chromatic basis, using a slightly different physiological mechanism. Light may be detected in very dim situations, even single photons in some circumstances, but the light must be at least at the twilight level to distinguish color. Even though color cannot be distinguished in low light levels, the light is nevertheless made up of various wavelengths. This important fact allows the computer to ignore a value of total light level and simply constitute all light values as separate intensities of three colors. Should all three color values be at the same level, the image appears white, grey, or black. While it may seem redundant to keep dim colors separately in the computer, this storage scheme simplifies image manipulation when bright colors are also considered.

There are several other electronic representation schemes of color, especially as used in television, but the colors seen on the screen or on film or paper recording medium or prints typically have three color components. On video monitors, where signals of various colors add to make white, these components are usually a blue having a wavelength from 460 to 475 nanometers, a green with wavelength from 510 to 550 nanometers, and a red at 605 nanometers or longer. For printing, where signals of various colors subtract from one another, these components are usually a cyan of wavelength from 485 to 505 nanometers, a yellow of wavelength from 565 to 580 nanometers, a magenta which is a complex mixture of blue and red, and sometimes a totally light absorbing black.

Color film or paper recording medium is sensitive to these three kinds of light. Chrome film or paper recording media, such as Kodachrome, Agfachrome, Ilfochrome, Fujichrome or Ektachrome, are also known as "color reversal" media. Such film or paper recording media give a positive image, create blue where blue is sensed, green where green is sensed, and red where red is sensed. Color negative film or paper recording media, however, create yellow where blue is sensed, magenta where green is sensed, and cyan were red is sensed. It nevertheless still is the red, green, and blue that are sensed, and these three colors are therefore sufficient to write on common film or paper recording media, slide or print, transparency or paper, color or black-and-white, positive or negative.

Color sensitivity curves for the film or paper recording media that might be used in a film or paper recorder were studied. The results of the study show that a blue at 476 nm or shorter wavelength and a red longer than about 620 nm wavelength will suffice. A green of 514 nm wavelength will work for all instances, but there is a 0.5% propensity for this wavelength to expose the blue sensitive layer of one paper examined. Greens at 528 and 532 nm wavelength are ideally suited. A green at 543 nm will expose the red sensitive layer to a very small degree on some papers. Any of these greens can be used if there is a mechanism by which the exposure is slightly modified to compensate for these overlaps.

Studies have shown that the eye can distinguish about 500,000 colors and hues. The eye is more sensitive to changes in hue in bright colors than in dim. Very good rendition of color can be accomplished with 64 evenly spaced light levels for each of red, green and blue, for a total of 262,144 colors (64*64*64). However, in circumstances where the colors change slowly over a large area, such as in the photograph of the sky near the horizon, using only 64 shades of each color can sometimes cause an effect called banding, where a line can be seen in the image between adjacent values of color. Using 256 light levels for each of red, blue and green, where the values are spaced across the dynamic range of the film or paper recording medium on a logarithmic scale, completely eliminates this problem. With 256 levels there are 16,777,216 (256*256*256) colors possible which has proven to be adequate for the finest of photographic rendition. In the preferred embodiment, 256 light levels are used for each of the three colors. However, a contrast range of 0-256 is inadequate. It is important to note the distinction between a contrast range of 0-256 and the number of steps over which the range may be varied. If the range and the step size were both 256, then the contrast would obviously change 1 level per step size. Alternatively, if the contrast level per step were increased to 10 per step, then a total contrast range of 10*256 or 2560 would be possible. It should also be kept in mind that due to the variation in sensitivity of the eye to different colors, it may be desirable to have a nonlinear contrast gradient between the different steps.

Photographic media are capable of distinguishing light levels in a 4000 to 1 contrast range. A common practice in the computer processing of digital imagery is to assign a light intensity value to each of three colors, red, green, and blue. Using 256 different levels per color, as mentioned, represents more than 16 million colors, more than adequate for any current use. These levels are represented by 8 bits of data for each of the three colors, or 24 bits total. However, if the density is assigned linearly with respect to the 256 intensity levels, then a contrast ratio of only 256 to 1 is effected. This contrast range is inadequate for some photographic purposes. Using 10 bits of data gives a linear contrast ratio of 1024:1, while 12 bits yields 4096:1.

When computers manipulate image data with 8 bits per color, it is not necessarily assumed that the relationship is linear. As the data is displayed on monitors, it does go through a transformation because the monitor does not react linearly to color signals. The same effect is true of every computer input and output device, but the nature of the effect varies from device to device, and separate conversion tables are required for each one.

The preferred approach in the present invention is to assign different levels of intensity to the eight bits of data in an expanded fashion. Within the electronics as discussed below, the eight bits are expanded to 10 bits using a look-up table were a 10 bit number is associated with the 8 bit number. This expansion is used to allow for non¬ linear response to every one of the linear 256 levels found in 8 bits of data and produces a 2^]0 contrast range. Further, the range of the eight original bits and the resulting 10 bits is further expanded in a logarithmic fashion with an analog logarithmic amplifier. The expansion effected by the log amplifier allows 8 bits per color to represent the entire range of both density and color within the film, to 4000:1 or beyond.

Resolution

The average human eye can distinguish detail angularly spaced at approximately 1/2 minute of arc (1/120 degree), as long as the eye can focus on the material and the very center of the eye is used. This value varies widely from color to color, intensity to intensity, and individual to individual. While only the very center of the eye is capable of resolving this detail, the human eye moves around quickly and widely and a person can comprehend far more detail over time than the eye perceives at any one instant.

The rate of information change across space is called the image's "resolution". Unfortunately, because high resolution is difficult to achieve, each of the various image industries (photography, computer graphics, television, motion pictures) uses different techniques to describe the resolution of their products. In television in particular, the values given for resolution are so optimistic that images produced at the stated resolution standards are sufficiently poorly formed that details are impossible to see, even with the best eye.

The most conservative means of measuring resolution is called the "Modulation Transfer Function" (MTF), and is expressed as the distance between alternating pixels. When the MTF is expressed, the level of resulting contrast must also be stated. In television the MTF is expressed as "lines of resolution" when the contrast is only 5%. Even at a contrast ratio of 10 to one, where the detail is ten times lighter or dimmer than the background, the image can look "washy".

In color film or paper recording medium, the film or paper medium will respond to differences in lighting of up to about 4000 to 1. However, if this film or paper recording medium is projected as in a movie, some light gets through the darkest parts of the image with a resulting contrast ratio of about 150 to 1. The contrast ratio range for color prints is higher than 150 to 1. Unless the contrast ratio is expressed, any value for resolution is devoid of adequate meaning. One typical pitfall is where an image forming process, such as large screen video projectors, accepts signals with 1000 lines of detail, but the image spot produced is two or three scan lines wide or is not uniformly illuminated. A diagram of this effect is shown in Figure 3. The resultant image looks very smooth and uniform until fine detail is displayed, and then the contrast drops to the point of unreadability.

In photography, and especially in film or paper recording medium recording, the issue of MTF is obscured by the use of other terms, such as "lines of resolution" or another quantity called "res". Within Eastman Kodak's publications of film or paper recording medium performance, the stated number of lines of resolution is a value approximately three times the MTF at a 100 to 1 contrast ratio. The term "res" in film or paper recording medium recording seems to only reflect the spacing of the pixels written onto the film or paper recording medium, and not to either the size of the pixel or the resolution of the resultant image.

In the context of the present invention, resolution is the spacing of adjacent contrasting pixels in the resulting developed image on the film or paper recording medium, at a 100 to 1 contrast ratio or more. This is the most effective and the most conservative value used in any of the image industries.

Performance

The most basic time requirement of the digital image recording process is the length of time to record a complete image. This time includes the film or paper recording medium advance time, any color adjusting time, and the actual image record time.

The data transfer time is also important. There are several different ways of transferring the digital data to the film or paper recording medium image processor. First, it is possible to design a recorder or image input device in which one image is being transferred while another is being recorded, called the "spooled" approach. In this case the data transfer time is not a factor in the recording speed. Second, it is also possible to record on film or paper recording medium as the image is being transferred, pixel by pixel, called the "real time". Care must be taken to closely match the data transfer rate with the film or paper recording medium recording process rate. Third, when the recorder cannot begin until the entire image is received, then the throughput of the film or paper recording medium recorder is the record time plus the transfer time, called "sequential". Finally, one can import images to the recorder on removable media such as disk or tape. This approach requires time to change the disk, but once changed, image processing runs at the slower of the transfer time or image record rate. This approach is called "pluggable media". Using "pluggable media", the image manipulation process can proceed while previous images are recorded. It is difficult to adequately express the overall performance of a film or paper recording medium recorder or image input device without expressing the data transfer time, which of the four approaches above applies, and the record time.

It is important to keep in mind of how voluminous the data is in an image. Assuming a raster image made up of pixels, the number of pixels in an image is determined by multiplying the number of lines by the number of pixels per line. In the case of television the value is about 400 by 460, or 184,000 pixels. Bear in mind that the image formed and transported in television cannot display this number of pixels at a 100 to 1 contrast, or even at a 5% (or 100 to 95) contrast except in new units.

In photography, the MTF at 100 to 1 contrast ranges from about 50 pixels per millimeter for good color film, to as high as 3000 pixels per millimeter for experimental black-and-white film. However, the human eye viewing a photograph at 18 inches, can distinguish about 15 pixels per millimeter. To allow for enlarging the photograph and for close examination, it is usually important to be able to record film or paper recording medium at or near the film or paper recording medium's MTF. Given a pixel density of 125 pixels per millimeter, the practical maximum that color film or paper recording medium can record, then there are 12,000 x 16,000 pixels in a 4x5" photograph, or 192,000,000 pixels. For pixels of 256 levels per color, there are 16,777,216 (256*256*256) total levels requiring 24 bits (I²⁴ = 16,777,216) or three bytes per pixel, for a total of 576,000,000 bytes per image.

Given this volume of data, one can see that the record and data transfer times can be significant. To record a 4x5" image with a recorder whose pixel rate is 1.0 microseconds per pixel (or one million pixels per second), it would take 10 minutes at 125 pixels per millimeter. The actual time would be somewhat more because of the need to start new lines and make other adjustments. However, this pixel rate, until the current invention, has been very difficult to achieve. Typical data rates for inexpensive film or paper recording medium recorders are on the order of 15 microseconds per pixel on each of three passes, one per color. To record a 4x5" image at film or paper recording medium resolution (125 pixels per millimeter), we would need several hours. Consequently, with existing methods, images are recorded at a much lower resolution when possible.

These factors show that prior art film or paper recording medium recording in high resolution is a slow, expensive process. The present invention provides a means for recording a high resolution image much faster than current recorders.

Most film or paper recording medium recorders utilize the technique of scanning. The term "scanning" actually refers to sweeping a beam of light or image element across a surface or space. In the digital photographic industry, however, this term is also taken to mean the specific process of sensing and digitizing an image from film or paper recording medium, paper or real objects. The genesis of the name is that most, but not all, input devices do scan the film or paper recording medium or paper, but with a charge coupled array, scanning may not actually occur. In the present context, "scanning" is used to refer to the regular sweeping of a beam of light, and the term "image input" to the processes of converting existing imagery into a digital format.

The formation of raster images using lasers has been practiced for 15 years to varying degrees of success. Much of the laser printer industry relies on scanned laser beams. More spectacular, but less successful, applications involve the formation of television or other video images on large screens. Additional details of the basic technology for this process may be found in Linden et al., U.S. Patent 5,136,426 herein incorporated by reference. The present invention addresses the correction of problems in these processes that are inconsequential in television and video quality images but are critical when recording or image inputting at high resolutions.

Lasers - General Characteristics

A "LASER" (acronym for Light Amplification by the Stimulated Emission of Radiation) is a device that creates and amplifies light. In the present context, the meaning of "laser" is somewhat more specific, referring to a device where the light is made to reflect back and forth in a cavity formed by mirrors through some excited material that amplifies the light as it passes through. There are lasers that don't use a cavity, but they are not useful for the present invention. Figure 6 diagrams the basic cavity laser design. The basic laser components are the laser cavity 70, completely silvered mirror 71 and partially silvered mirror 73 which transmits beam 72. Diode lasers are acceptable for the present invention and work basically the same way, but on a very small scale.

Cavity lasers emit light out of the cavity through a partially reflective mirror (most common), around a mirror that is smaller than the beam, through a hole in the mirror (used in welding and materials working), or by a dynamic device that reflects or refracts light out on demand from some electronic signal described in some detail in Pease, U.S. Patent 5,097,480. Light that is amplified in such a cavity has some unique characteristics, including coherence, monochromaticity, and nearly parallel beams.

Laser light is produced from the far ultraviolet through the visible and to the far infrared and microwave, and can be issued in pulses or in continuous waves. For the present invention, the laser light of interest is coherent, continuous wave, monochromatic, preferably visible, high quality light beams. Common gas lasers, such as Helium Neon, Argon, Krypton, and solid state and diode lasers are all candidates. For the present invention, the lasers may include appropriate red, green, and blue wavelengths and must produce a continuous beam, not a pulsed beam. Not excluded from this invention is the use of beams that are not visible, as for use with new papers that develop visible color images based on multiple infrared excitation beams.

Laser Light Modulation

As the beam is scanned across film or paper recording medium to record an image, its intensity must be varied or modulated. There a two broad general methods of modulating a beam. The first is to modulate the source of the beam itself. The second is to modulate the beam after it leaves the source. At the source, the beam may be modulated by varying the power of the laser. To modulate the beam after it leaves the source several methods are available including electro-optical, acousto-optical, mechanical chopping, and moving mirrors. Within the scope of this invention the term "modulation" refers both to the process of varying a laser input power to change the output power, as in a Diode laser, or to modulating the light after it leaves the source.

To modulate the beam's intensity at the rates necessary to match the performance of existing film or paper recording medium recorders, the gas laser cannot simply be turned on and off. Rather, it is necessary to produce a continuous beam and then modulate the resulting light. Diode lasers, however, can be modulated by changing the input power supply.

The practical ways of modulating an existing light beam more than a few times a second are limited to electro-optic and acousto-optic modulation. Electro-optic modulation can be accomplished on thin beams, ribbons of light, or on two dimensional surfaces. An electric charge is used through some material to rotate the plan of polarized light, and the resultant beam passes through a second polarizer, allowing the light through to the extent that it is polarized correctly.

With acousto-optic modulation, acoustic "sound" waves are either propagated through a transparent material, or on the surface of a reflective material. The use Of reflective material is not as efficient, and the effect is currently only used for testing surface wave devices. With the solid transparent material, the continuous sound pattern sets up areas of compression and rarefaction within the material. Light reflects off of these waves as if they were a solid, acoustically-induced diffraction grating. Electro-optic and Acousto-optic scanning and deflection by Gottlieb, Ireland, and Ley is an excellent text that describes these effects and is hereby incorporated by reference. Figure 7 illustrates the acousto-optic modulator 80. It will be noticed in the Figure that the incident beam 81 strikes the crystal 82 at a slight angle 83, known as the "Bragg angle". This angle 83 depends on the frequency of the audio energy (sound), the speed of sound in the crystal and the wavelength of the light. While the modulator 80 will disperse light if the beam enters at other angles, when the beam enters the crystal at the Bragg angle, most of the dispersed light is directed in a single beam 84a emerging at the negative Bragg angle. When there is less sound or no sound, the unmodulated beam 84b simply travels on through the crystal 82. It is the modulated beam that is captured for use as the light source for film or paper recording medium input and recording. In Figure 7 multiple modulated beams are shown. This illustrates that the angle can be changed slightly by varying the sound frequency, but only one beam is produce at one time. The angle-changing effect can be used for moving the output beam and will be discussed later. As discussed in patent 5,097,480 hereby incorporated for reference, surface wave and electro-optic techniques are also useful. As discussed below, the acousto-optic modulation technique is the preferred method.

There is an opportunity to use a diode laser, which has the advantage of being able to produce light in proportion to a rapidly varying input power source. Unfortunately, current diode lasers have several drawbacks in this application. The first is that the beam quality is not as high as the gas lasers described above, and the second is with the available wavelengths. Diode lasers are most efficient in the short infrared. They efficiently produce visible light in the deep red, but as the wavelength shortens into the red, orange and yellow, efficiency and beam quality begins to suffer. In the green and blue, the power, reliability, and beam quality of current diode laser technology falls below limits acceptable for this application. With diode lasers it is possible to modulate the light at high frequency simply by turning the power on and off, but if we were to use diode lasers for red but not the other colors, we would need two different kinds of modulation. The present invention does not preclude the use of diode lasers, either for red or the other colors, including UV and infrared but the preferred embodiment uses gas lasers for all three beams.

Laser Light Optical Characteristics - Chromaticitv

In the preferred embodiment any of the candidate lasers emit light at one or more frequencies, each with a bandwidth of less than one nm. For the present invention, the light bandwidth spectrum only need be narrow enough so that one colored beam does not expose a layer of film or paper recording medium sensitive to another color, achieved as a matter of course within all candidate lasers.

Most lasers produce light on only one spectral line at a time. For helium neon lasers used for red in the preferred embodiment, this wavelength (called a "spectral line") is 632.8 nanometers (nm). The argon ion laser can produce many spectral lines at one time, including several in the ultraviolet between 300 and 400 nm, blue at 454, 458, 465, 472 and 476 nm, blue-green at 488, 496, and 501 nm, green at 514 and 528 nm, an infrared line, but no visible red lines. Some of these spectral lines are more available than others. The most easily produced is the 488 nm line, but it is at a poor wavelength for use in film or paper recording medium recording because it can excite both the green and blue sensitive layers. The best line for green is the weak 528 nm line, but this line is difficult to produce in the same cavity with blue lines. Any of the blue lines will serve, but the preferred embodiment would use the relatively strong 458 or 476 nm lines. A further advantage of the 458 line is that it is approximately as sensitive to shifts in laser input power as is the 514 green line, so that if there is a shift in laser power, both lines vary equally. However, 458 nm is a long way from the green and red lines, making it more difficult to design a lens that focuses all three beams to the same point.

Laser Light Characteristics - Coherence and Speckle

A potentially undesirable characteristic of laser light is its coherence. Coherence means that each wave front of the light leaving the laser is synchronized with its neighbors. All photons travel in lockstep. Should the wavefront be reflected off of some surface, then some reflected coherent wavefronts will interfere with other reflections, causing areas of reinforcement and of interference. Thus reflected coherent light looks beaded or sparkly. This effect is called "laser speckle" and is fully described in Laser Speckle and Applications in Optics by M. Francon, incorporated by reference.

There is an advantage to coherence, and that is that the focusing of coherent light can be as mush as twice as tight, using the same optics, as the focusing of non-coherent light. The present invention makes use of this property since the very small spot size achieved, would be much more difficult with non-coherent light.

Light Characteristics - Diffraction Limits

If we were to just consider the broad application of the bending of light beams by transparent media of different optical density, we would have newtonian optics. This approach is valid for eye-glasses and most other every-day optics, but ignores the particle nature of light. It is impossible to focus a beam of light to an infinitely small spot because of the way the wavefronts of the light bend around the edges of lenses. There is a relationship between the size of a beam, the distance at which it is to be focused, the wavelength of the light, and the minimum spot size at the distance. Any modern text on diffraction optics describes this phenomenon.

The equation for minimum spot size is

1.27 λf Bd =

D

where

Bd = Resulting spot size λ = Wavelength of light f = Distance from final tocusing optic to resultant spot

D = Diameter of beam at final focusing optic 34 in Figure 1 The factor 1.27 is unique for gaussian beams of coherent light, while for non-coherent light the factor is more like 2.44 , thus quantifying the extra focusability of laser coherent light mentioned above.

For a spot of red laser light (633 nm) 250 mm from the film or paper recording medium plane, a distance necessary to allow for beam-moving optics, we would get a 6 micron spot using beam 33 mm in diameter at the last focusing optic.

= 6.09 * 10^"5

33

It is now seen that for small spots at a reasonable distance for beam manipulation, very large beams on the final focusing optic are needed, a factor that drives the design of the present invention.

For a laser cavity, the coherent nature of the light causes it to emit from the laser as if all of the light originated from an almost infinitely tiny source somewhere within the cavity or outside the back end. This virtual source volume is less than one wavelength of light across.

As the beam emerges from the cavity, it diverges slightly, depending upon the diameter of the emerging beam and the distance from laser's virtual point source to the laser's output. Thus a wider beam can have a lower divergence. Long lasers have a lower product of divergence * beam diameter than short ones, but the relationship is not necessarily linear. Diode lasers have a very wide divergence, sometimes tens of degrees. However, after focusing, the beam from a good diode laser can theoretically have nearly the same characteristics as a gas laser of the same wavelength. Such beams are now only achieved in diode lasers in the infrared and long reds.

The present invention utilizes this virtual point light source effect because the scanned spot size is small, approaching the diffraction limit of the final focusing optic, and would be much large were we to use non-coherent light.

Scanning Approaches

To use a raster technique to record or input film or paper recording medium imagery, the image spot must move in both horizontal and vertical directions. One can move either the film or paper recording medium or the light beam, or both. Some implementations use one technique for one dimension and a different one for the other.

With raster scanning, where straight evenly spaced lines are drawn next to another, the speed of the scan in one direction is orders of magnitude faster than the other. The faster scan moves the beam along a line (called "line scanning"), and the slower scan moves the next line down from the first (called "frame scanning"), and so on.

The most obvious, and slowest, scanning approach is to move the film or paper recording medium. In laser copiers, printers, and faxes, this approach is used for the slower of the two scan dimensions, moving the print line down the paper.

If one is to hold the film or paper recording medium still in one or both dimensions, then the light beam must move. There are several effective ways of moving light beams.

Scanning is done with electro-optics, where a beam of light can be bent somewhat as it passes through a charged field in some special liquid or solid. Nitrobenzene is commonly used. A disadvantage is that thousands of volts are need to bend the beam small fractions of degrees.

The acousto-optic modulator can also move beams of light, as shown in Figure 7. When used as a modulator, the frequency of the sound is fixed, and the light reflects off the sound field at a fixed angle. If one were to vary the frequency, the reflection angle changes proportionally. However, the angle and its variation depends on the wavelength of the light, and is not more than a degree or so. A secondary difficulty that affects the present invention is that the efficiency of the modulator's ability to reflect light at constant acoustic power varies non-linearly with the frequency, being most efficient at some tuned center frequency, determined by the dimensions of the modulator crystal and its sound generator.

The most common way of moving light beams is with moving mirrors. There are two common practices for mounting mirrors for beam scanning. These are on-axis and off-axis. On-axis mirrors are mounted on a shaft that is rotated giving one sweep per revolution, or on a shaft were the mirror moves back and forth through a small arc, up to 90 degrees. There are three common approaches to these small sweep mirror motors; the galvanometer, the stepping motor and the resonant scanner. In the galvanometer, an armature reacts to a voltage and a torque is applied against a spring. Figure 8 shows such a mirror configuration. The rotation of the mirror 30 by galvanometer 32 tracks almost linearly with the applied voltage, and can be accurately positioned to any point within a few milliseconds. Accurately sweeping through a continuous track is easily accomplished, and complete accurate sweeps with return can be accomplished up to about 300 sweeps per second using very small mirrors. Use of a galvanometer driven mirror is the preferred method of frame scanning in the present invention.

Stepping motors, with or without gear reducers, can accomplish the same thing and may be one implementation of the present invention. However, these less expensive motors require more electronic controls and the resulting mechanical configuration takes longer to stabilize between lines and is not quite fast enough to meet all objectives of the present invention. However, stepping motors are one method of moving the mirrors and are within the scope of our invention.

With resonant scanners, the mirror sweeps back and forth on a sinusoidal scan at a fixed frequency. The external configuration is indistinguishable from the concept shown in Figure 8. The sinusoidality of the sweep makes it very difficult, but not impossible, to give evenly spaced pixels along a sweep line, but the sweep repeat frequency is not changeable. Also, as the speed varies it is necessary to change the intensity of the beam to cause the same exposure. For sweep rates over 100 per second, the mirrors are also very small.

The process of choice for very high speed scanning is the off-axis mirror, in the form of a regular polygon of mirror 36 material that rotates, as shown in Figure 9. Incident beam 56 is reflected to beams 57a and 57b by a facet 37. As the polygon rotates, the incidence/reflection angle changes from 58b to 58a. This in turn causes the reflected beam to sweep out a line from 57b to 57a. This process is known as "line scanning" and use of the rotating polygon is the preferred method of scanning in the preferred embodiment. One scan is performed for each facet per revolution of the polygon 36. If there are 25 facets, then 25 scan lines can be performed per revolution. This is precisely the case in the embodiment of the device described in Linden, et al., U.S. Patent N²5,136,426. While this approach works for scanning at television resolution, there are a number of problems when used in a film or paper recording medium recorder. First, the facets 37 may not be spaced accurately enough around the polygon 36. Second, the facets 37 may not be precisely parallel to the spinning axis 59. Third, with the axis off the facet center, the virtual reflection plane of the polygon 36 moves in and out as the edge and center of the facet 37 come into use. The leading edge of the facet 37 gives a different optical center than the trailing edge, unless the beam strikes directly at the spinning axis. This last effect greatly complicates the design of a field flattening lens used after the scanner, even for one color. Figure 10 illustrates this problem. Fourth, while speed can be changed, it cannot be changed quickly, and controlling the speed to the tolerances necessary for varying pixel placement on subsequent lines is a significant problem. Finally, very rapidly spinning polygons can act like sirens, and if fast enough, standing waves in the air or gas near the facets 37 cause unwanted diffraction effects on the beam and mechanical erosion of the reflective surface. Polygons operating at these speeds use air bearings as a configuration of preference, making it very difficult to have the polygon spin in a vacuum.

It is one of the objectives of the present invention to use the high speed scanning offered by the polygon, using the optical and electronic configurations discussed below, and overcome these problems, which have plagued the prior art until now. Problems With The Raster Scanning Process

It is decided to build a film or paper recording medium recorder and image input device where the light beam is scanned since it is more reliable and faster to move the beam of light rather than the film or paper recording medium. Further, it is concluded that it would be most desirable to record or input an image with film or paper recording medium in a flat field. A flat field is preferred because film or paper recording medium handling is easier, scratching is reduced, and it is nearly impossible to bend film or paper recording medium in more than one axis (compound curve) without creasing or cracking the emulsion. Bending the film or paper recording medium is done in one axis in some recorders, but this makes the raster scanning of prints or continuous film or paper recording medium quite difficult. The requirement for curved film or paper recording medium plane is illustrated in Figure 11. The figure illustrates the effect when using a polygon 36, but the effect remains no matter how the beam is moved through an angle from a fixed point. As shown in the figure, the locus of the focus points describes curved shape 66.

One design criterion was that the processing time should be under ten minutes for a 4x5" image and comparable times for other image sizes to compete with the current image input and recording devices. Recording should be at the maximum resolution of the film or paper recording medium, which is about 80 to 100 distinct pixels per millimeter. The goal of the present invention was 125 pixels per millimeter, the maximum stated resolution of any candidate color film or paper recording medium, which equates to approximately an eight micron beam spot size at the film or paper recording medium plane.

Performance

Recording 125 pixels per millimeter equates to 3175 pixels per inch, or 16,000 pixels in the 5 inch direction of 4x5" film or paper recording medium.

To manipulate images at high speeds, it is best to use raster imagery. The following are some common terminology definitions of raster scanning used in this description: "Line scanning" is the process of sweeping a beam along one raster line. Devices that do this are called "line scanners". "Frame scanning" is the process of moving from one raster line to the next adjacent raster line. Devices that do this are called "frame scanners". Therefore, scanning may be thought of as a two step process. First, where the beam is scanned horizontally in a line by the line scanner. Second, where the beam is moved one unit vertically to the next line where the line scanning process begins again. The maximum scan rate achievable with the current galvanometric scanner mirrors large enough to generate the eight micron spot size is less than 30 lines per second, or about 600 seconds for a 4x5" film or paper recording medium a the desired resolution. The design of a recorder that would use galvanometric scanners in both directions is presented in Figure 5. One galvanometer 67 would scan in one direction, and this fluctuating beam would then be scanned in the other direction with the other galvanometer 68. The present invention does not preclude using two galvanometers this way, but the preferred embodiment utilizes a polygonal scanner 36 for scanning in one of the dimensions (line scanning). A polygon is not significantly more expensive and is very much faster.

An alternative scanner, the resonant scanner, has generally been rejected because of the problems with speed selectability and the small mirror size required at these scan rates.

Another scanning alternative, acousto-optic scanning, would have the appropriate sweep rate, but two other problems make this alternative undesirable. The first is the extremely narrow and non-linear sweep angle / power characteristics, making it difficult to handle a beam 33 mm (1.25 inch) diameter (the approximate size of the beam at the point where it is incident on the scan mirror). Only using "prescan" final lenses (lenses between the scanner and the film or paper recording medium) would alleviate these problems with acousto-optic scanning, an approach which is undesirable for other reasons. Also this process can only reliably position to 1000 distinct spots, more than an order of magnitude too coarse for the objectives of the present invention.

The method of choice for line scanning in the preferred embodiment is the polygon scanner. Polygons with eight 1.25 x 1.85 inch mirrored facets are commercially available, attached to a motor that can spin well beyond the 1000 facets per second desired for the preferred embodiment. Such commercially available polygons are acceptable for the present invention. Only with face velocities of 500 feet per second would there be problems with excessive noise or standing wave aerodynamic interference. The polygon chosen above for a 1000 lines per second scan rate would turn at 7500 RPM, with a face velocity of about 100 feet per second. While it would seem possible to use more than eight facets, the size of the scanned spot on the scanner and the rate of available data make this unnecessary. Referring again to Figure 10, a narrow beam 56 is shown striking the center of a polygon facet 37. In fact the beam is relatively large at this point, as much as 25 mm. In the preferred embodiment, with a 4x5" image and an eight micron spot 250 mm from the final focussing optic, the beam is approximately 33 mm at the final focusing optic 34 (Figure 1), but becomes smaller as it moves toward the focused spot on the image plane. To be efficient, all of the spot must be on the scanner during the time that pixels are being written. As shown in Figure 12, this is true during only about 40% of the rotation time of each facet (duty cycle), shown in spots 74b and 74c. During the remaining 60% of the time part of the spot falls on the edge between two facets 74a and 74d and is not useful for imaging. The time between active line image is used for switching the optical path mechanics to scan subsequent lines. More facets on the polygon would therefore decrease the available sweep angle.

While the time between facets is significant from a duty cycle standpoint (60%), this time is nevertheless very short. As will be shown later, the use of the polygon gives the opportunity to create a very small spot in the direction of the line scan, but only if the lines are swept quickly. (See section below, Scophony). Should, for some reason, not every facet be used as it is presented to the beam, the mechanism that repositions the line must stop and reset. Moving the film or paper recording medium mechanically is a process with too much hysteresis of motion and inherent inertia to allow for stopping, resetting, and restarting to be practical for equally spaced lines. For these reasons, the process of moving the film or paper recording medium or moving the line scanner along the frame scan axis has been rejected in the present design in favor of a galvanometer frame scanner 32 shown in Figures 1, 8, 13 and 14. Figures 13 and 14 show the two basic arrangements for using a polygon line scanner 36 and a galvanometer frame scanner 32. In Figure 13, the line scanner 36 is shown positioned between frame scanner 32 and film or paper recording medium 40 and in Figure 14, the positions are reversed.

It is also within the scope of this invention to use two galvanometer scanners 67 and 68, as in Figure 5 for both line and frame scanning, or to use two polygon scanners 36 and 76 as both line and frame scanners to move the lines across the image plane, shown in Figure 15. Although this scanning arrangement is within the scope of this invention, it does limit flexibility in the range of scanning and spacing between scan lines. A fabricator could use the approaches to scanning the beam shown in Figures 5, 13, 14 or 15 and still remain within the scope of the present invention. The preferred embodiment comprises a galvanometric frame scanner 32 followed by a polygon line scanner 36, as shown in Figure 13.

This section discusses some of the errors which may affect image quality when high resolution is desired. The solution of these errors, largely ignored or accepted by prior art devices having lower resolution, is indicative of the advancement of the prior art represented by the present invention.

Figures 13 through 16 show devices where two scanners move a single beam across a flat field. To see the errors inherent, the geometry of this process is examined.

First, the focus of the beam is examined. If the beam is large, as for a diffraction limited spot, and this beam is scanned, then the depth of focus of the beam (i.e. the distance over which the beam is in focus) is very limited, a few tens of microns. Since it is farther from the scanner to the corner of a flat piece of film or paper recording medium than it is to the center, relative to a point fixed above the film or paper recording medium center, and since this distance to the corners exceeds focus depth, the corners would be out of focus. In the prior art, there are two approaches in use to correct this problem. The first is the field flattening lens after the scanners, referred to as a "prescan" lens 78, Figure 16. These lenses work well in one or two colors, but have not been balanced yet for three colors well enough to do a 4x5" film or paper recording medium to 8 microns. The other common approach is to place a lens 86 in the light beam before the scanners, as in Figure 17. This lens moves toward and away from the final focusing optic 92 as the beam scans across and down the film or paper recording medium 40, continuously repositioning the focus point along the focus axis. However, moving such a lens 86 accurately at a rate of 500 to 1000 cycles per second to accommodate the invention's scan rate, is beyond the capability of the present art. A practical solution must focus the beam so that the spot size and shape does not vary as the beam moves across the image plane. Details of the solution to this problem by the present invention may be found below in the section "optics".

Next the shape of the scan lines is considered. For initial simplicity, a single mirror that can be moved in two axes, as shown in Figure 18 is first considered. This hypothetical device is restrained such that the mirror can move quickly in one direction, but only more slowly in the other, as in raster fashion. As the beam is swept across the center of the flat film or paper recording medium, the described line is straight 130a. However, at the top or bottom of the image plane the line is not straight. This is because the beam at the top corner is further away than it is at the horizontal center, and at the fixed vertical angle, the height is greater. This gives a dislocation known as "pincushion" 122 as shown in Figure 18. Continuing the hypothetical example, it is further assumed that the more rapidly repeating scan line has the same duty cycle, meaning that the entire line is presented evenly over a fixed scan angle. Under these circumstances we also get the pincushion in the vertical scan plane 126, because at the same angle the distance is greater to the corner than it is to the vertical center.

To make the issue more complex, a realistic scan mirror configuration is presented, as in Figures 5 and 13 through 16. As shown in these Figures, the distance from one mirror to the film or paper recording medium is greater than it is from the other. Consequently, there will still be a pincushion distortion in both directions but the degree of distortion is more pronounced in the line dimension scanned by the nearer scanner and less pronounced in the dimension scanned by the farther scanner. This means that the correction of the problem using some sort of spherical image scan surface is not useful, even further complicating the "prescan" lens design. The locus of points in focus is more complicated than a simple spherical section.

Another set of errors occurs from scanning straight lines from a point source onto flat fields. Because lines at the edges are farther from the scan source, lines separated by the same scan angle are farther apart at the edges of the film or paper recording medium than they are near the center, see 128a and 128b in Figure 18. This effect is also true of the spacing between pixels along a scan line. Pixels being closer at the center of the film or paper recording medium, and more widely spaced at the edges, see 130a, 130b in Figure 18.

To summarize, there are five degrees of orthogonality errors that occur when scanning without field flattening optics onto a flat image plane; Focus, Curved horizontal lines 124, Curved vertical lines 126, Unevenly spaced horizontal pixels 128, and Unevenly spaced pixels 130 along the horizontal scan line.

As a theoretical solution, one could move the pixels in the image buffer to counteract the last four distortions. However, in this embodiment, as discussed above, there can be as many as 750 million pixels per image. With the processing speed of today's small computers, it would be many minutes to correct these errors.

Optical Power Output Requirements

Experimentation with color and black and white film or paper recording medium show that even a 0.5 milliwatt helium neon laser focused to an eight micron spot size is thousands of times too bright to expose film or paper recording medium on a 4000:1 modulated gradient. Consider that an ordinary light bulb produces on the order 0.1 watts of visible light and that this light is more than sufficient to expose film or paper recording medium in a camera through a small lens in far less than a second. An unfocused laser spot from a small helium neon laser can be as bright as the sun, and orders of magnitude brighter when focused. Additional general information on the intensity of focused laser beams may be found in The Laser Guidebook. Jeff Hecht, p 95, herein incorporated by reference. Image Acquisition or Input

In the process of image input, a constant beam of laser light is sent in a tiny spot to the film or paper recording medium, and either the transmitted or reflected light is used to excite photosensors. In this case orders of magnitude more optical power is required because the light is scattered by the film or paper recording medium and must be gathered with a lens or other method, and because the photodetectors are less sensitive to light than film or paper recording medium. Consider that only a few tenths of a watt of light in an eight micron spot may be powerful enough to damage the film or paper recording medium.

Large amounts of optical power are not required for this product, but the other qualities of laser light are essential.

Laser Speckle

Laser speckle is a result of the coherent nature of laser light, and represents excess information in the laser beam. The interference and reinforcement that occurs when viewing laser beams occurs in the eye, and not on the surface viewed. Speckle can also be photographed. Fortunately, the size of the speckle is inversely proportional to the beam diameter at the last focusing optic. Additional information on this property may be found in Laser Speckle and Applications in Optics, M. Francon, herein incorporated by reference. Because of the 33 mm source beam and spatial filter used in the present invention, laser speckle in this application is so small as to not be photographable.

Registration

Assume a design where the three colors of laser light are combined into a single modulated beam preferably using dichroic mirrors, although other approaches are possible as shown in Linden, et al., U.S. Patent N² 5,136,426. It is possible to scan each beam separately in sequence, but this would take at least three times as long. Assuming that there is a focusing element 78 after the scanners as in Figure 16, this element 78 must either move all three beams to the same place on the film or paper recording medium 40 at the same spot size, or the image must be modified in the computer before presentation to the film or paper recording medium to correct for the errors.

If, however, the final focusing optic occurs before the scanners and this optic has the same focal length for all three colors, then it is possible to maintain registration because there are no focusing optics to distort the beam position or focus distance for different colors. Use of a final focusing optic before scanning is the preferred embodiment of this invention, and is shown as 34 in Figure 1.

Using the final focusing optic prior to the scanners gives a unique opportunity to use commercially available standard lenses. Should the final focusing optic not be perfectly color corrected, then some colors will focus slightly farther away than others. With the design as in Figures 13 through 16 the beams are combined after they are modulated and filtered. The total beam path distance from the filter to the final focusing optic can vary somewhat by color, causing all three colors to arrive at the image plane exactly focused. Figure 19a shows how an optic 130 will treat three different colors from the same source. The focal length for each path varies somewhat depending on the wavelength of the light. This process is reciprocal, and by varying the beam path lengths of each of the three color sources, the focus at the film or paper recording medium side is at exactly the same distance. Figure 19b shows how this is accomplished using dichroic mirrors 132, 134 that reflect some colors while transmitting others. Figure 19c illustrates the problem with prescan lenses 136. Even if an expanding optic accurately collimates the beam, the prescan lens will, if the beam is off-center as in scanning the edge of the film or paper recording medium, to some degree separate the collimated colors. There is no practical way of prepositioning the beams to correct for this error because the effect varies with respect to the beam position across the lens. However, in the preferred embodiment, as in Figure 19d, the scanning occurs after the final focusing optic, the colors remain registered with one another because mirrors 138 reflect all light at the same angle, no matter what wavelength. Spot Size

Not all images have the same resolution content. While it is necessary to be able to produce an image at film or paper recording medium resolution at 4x5 inches, say 3000 pixels per inch, many images can be reproduced effectively at a lesser resolution, say 750 pixels per inch. Images at lesser resolutions are smaller in data volume in the computer by the square of the change in pixels-per-inch spacing (i.e. halving the resolution reduces the data volume to one fourth original).

Given a device that writes 3000 pixels per inch, then if only 750 pixels per inch were written without changing the spot size then the resulting photograph would have 90% unexposed film or paper recording medium. A common approach to solving this problem is to replicate pixels so that 4 pixels in each direction are produced based on each pixel in the original. The resulting pixel colors can either be color duplicates or can be averaged between adjacent source pixels. The present invention allows for this approach, but also introduces another approach where the beam is enlarged by a predictable amount. This "enlargement" process is discussed in more detail below in the section on "optics".

Scophony

The Scophony company of England invented in the 1930's a technique for acousto-optic modulation of light using sound waves to create video images. With this process, the relatively slow movement of the sound field through the modulated beam causes part of the beam to be modulated for a particular pixel before the remainder. Figure 20 shows an acousto-optic modulator 140 with a fairly large beam 142 (in actual implementation, on the order of 1 mm diameter) in which the fixed-frequency sound has been turned on and off several times as the sound moves through the beam. Thus, there may be several pixels in the beam at one time. Unless there is some compensation, this detail would be lost. The Scophony technique is, as described in Figure 21, to image the sound field on the image target 40 so that the spot projected 144 has the various pixels moving through it. The beam is focused on the image plane 40 so that the sound image moves through the spot just like it does in the modulator. However, this spot 144 is moved by the scanner 36. Scophony dictates that the sound move through the spot at exactly the same speed as the spot moves, but in the opposite direction. This causes the beam modulation duration to be precisely the time to scan one pixel diameter. Thus, the sound image stands absolutely still, and allows detail to be presented in the scan line at the diffraction limit. However, this approach does dictate that the speed of the line scanner 36 be appropriately fast and stable. This speed is fast enough so that only polygon scanners are currently used, and data presented for recording on a line must be complete before the line begins, because the process must proceed for the entire scan line without interruption or speed variation. These requirements are ideally suited to use of the rotating polygon as the line scanner.

PREFERRED DESIGN Optics

Figure 1 is a block diagram of the present invention showing the major optical and related components of the invention. In the preferred embodiment the 3 colors used by the system are created using 2 lasers. Also, light from other than lasers can be used, such as irredescent, arc, florescent, or light emitting diodes. However, it is within the scope of this disclosure and the appended claims that other configurations of light sources and optical paths may be used, and that the colors of light may be visible, infrared, or ultraviolet. For example, Figure 38 illustrates an additional embodiment wherein a single laser 205 is used to create a single beam path. This embodiment would be most useful when it is desired to generate an image in a single color such as black and white. Alternatively, Figure 39 illustrates an embodiment wherein two lasers 100, 200 are used for three colors, but a single beam path is used to modulate the colors in sequence. In still another embodiment, Figure 43 illustrates how a single laser such as krypton or helium-selenium can be used to produce a beam of all three color wave lengths which may then be split into three separate paths. In the figures, similar numbers are used to refer to similar parts. The discussion below referring to the preferred embodiment of Figure 1 is largely equally applicable to these alternative embodiments. However, in the case of the embodiment shown in Figure 38, if a three color laser such as krypton or helium-selenium is used, it would be necessary to individually scan the three beams sequentially rather than simultaneously as is done in the preferred embodiment of Figure 1.

The advantage of the alternative embodiment shown in Figures 38, 39 and 43 is the cost savings in some of the optical components. For example, when the three colors are combined into a single path as in Figure 39 only one modulator 18 is required whereas in Figure 1, three modulators 18, 20 and 22 are required.

Two colors of laser beams (blue and green) are created from using argon laser 100 and one color (red) with a helium-neon laser 200. Variable attenuators 7 and 8 are used to control the strength of these lasers between image output and image input. Dichroic mirror 2 reflects the blue light beam 102 of laser 100 to mirrors 4 and 6, while the green beam 104 of laser 100 proceeds toward optic 12. A narrow band filter 3 assures that only light of 514 nm wavelength (or optionally 528 nm) green proceeds, removing unwanted blue or blue-green found in the argon laser 100 beam. Likewise, another narrow band filter 5 assures that only 458 nm wavelength (or optionally 476 nm) blue proceeds. Optics 12, 14, and 16 focus beams 104, 102, and 202 to the correct size for Scophony in acousto-optic modulators 18, 20, and 22. The acousto-optic modulators are arranged to take advantage of Scophony as discussed above.

-λfter modulation, each modulated beam passes through a small lens that concentrates the beam through a small perferatiorc ("pinhole") only microns in diameter, in a process called spatial filtering. The spatial filters are represented at 24, 26, and 28 for beams 104, 102, and 202 respectively. Figure 22 shows how this process is employed in one of the three beams. Identical configurations are used on beams 102 and 104. This process removes unwanted artifacts of the modulated beam, filters out reflections off of dust and other optical irregularities, and creates a stable source spot to be focused and scanned onto the film or paper recording medium, unaffected by variations in laser pointing stability. As shown in the figure, beam 202 having a diameter of approximately 1 mm is incident on focusing optics 128 and 129, comprising lenses operative to focus the beam for pinhole 130. The beam then passes through pinhole screen 130 having a pinhole diameter 131 in the order of 8 microns (.008 mm). The hole 131 diameter is sufficiently small that non-gaussian rays are filtered out. Also, spatial filtering simplifies use of non-laser light sources by defining a stable, very small light source point for optic 34 to focus.

The modulated and filtered beams then pass through preferably dichroic mirrors to combine the three colored beams into one beam. Linden et al., U.S. Patent N² 5,136,426 discusses other ways these beams might be combined. Also, since an excess of light is available, simple partially silvered mirrors, non-polarizing beam splitter cubes, or pellicles could also be used. In the preferred embodiment, as shown in Figure 1, dichroic mirror 9 reflects blue but transmits green. Therefore, the blue 102 and green 104 beams are combined by the operation of mirror 9. The combined blue and green beam expands towards dichroic mirror 10 which transmits blue and green but reflects red. Thus, the red 202 beam is combined with the blue and green beams. This 3-color beam expands from its respective pinhole through dichroic mirror 10, spot size selector 54, to lens 34 where it is about 33 mm in diameter. Symmetric achromatic lens 34 is optimized for a 1:1 conjugate ratio, and causes all three colors to be focused to a single point a distance away determined by the focal length of lens 34. In the preferred embodiment, the focal point is approximately 250 mm from the lens 34 and produces a beam spot less than eight microns in diameter.

In Figure 1 the beam next strikes mirror 30 attached to a galvanometer or stepping motor 32 for slow speed frame scanning, and thence onto the facets of a rotating polygon 36. The positions of the two mirrors can be interchanged, but the length of the non-scanned beam path increases somewhat because of mechanical interference with the galvanometer 32 and scanner motor (not shown), meaning that the corrections for the five errors in orthogonality are more severe. As mentioned above, one challenge of using a beam of such small size is the limited focal depth of field which results. This limited focus range would destroy the precision resolution afforded by the present invention if beam focus could not be maintained throughout the beam sweep across the flat film or paper recording medium plane. This presents a major problem since this focus does not lie in a flat plane. Rather, as mentioned above, the focus of the scanned beam describes a complex shape. This shape would be toroidal if two on-axis scanners are used, but with the rotating polygon 36, the shape is slightly flattened on one side from the ideal toroid. Careful examination of this image plane shows that horizontal scan lines are straight and almost exactly evenly spaced. The vertical edges of the scan lines are still curved and the pixels are not evenly spaced across the line. Figure 18 shows the five orthogonality errors of an uncorrected scan. Figure 41 shows the remaining errors if correction was made by the image converter 38 alone.

To couple this image plane onto a flat film or paper recording medium 40, this image plane is machined into a plate of glass 38 made of very small parallel fibers 181 as shown in Figure 24. The image plane converter 38 has a front, source facing surface 180 and other recording medium facing surface 182 shown in Figure 24. In the preferred embodiment, front surface 180, defined as the surface upon which the image light beam is incident, is shaped to conform with the image focus locus points described by the scanning process. Thus, the image beam is in focus at each fiber upon which it is incident. The other surface 182 may be ground to the shape required by the configuration behind the plate. For example, in the preferred embodiment, if a flat recording medium (film) 40 is placed adjacent the other surface 182, surface 182 would be flat. Alternatively as shown in Figure 40, if lens 184 is to be placed behind 182 in order to enlarge the image, other surface 182 would be ground to a shape to accommodate the image plane of the lens 184. The optical fibers 181 of plate 38 carry the focused beam incident on each fiber at its position in substantially parallel fashion to the other surface, through the plate 38 to where the film or paper recording medium (40 in Figure 24) to be recorded or input is found. While in Figures 1 and 40 the other surface of the plate is shown to be adjacent to and essentially parallel to the front of the plate, the other surface could, by extending the fibers, be significantly distant from and oriented at an angle from the front of the plate and still be within the scope of this invention. For example, the fibers could bend as a body 180 degrees so that the other surface is in substantially the same plane as the first surface. This plate 38 then, in effect, corrects for the focus error and one of the remaining errors in orthogonality. Figure 41 illustrate the shape of the focus plane if corrections were implemented by the converter 38 alone (the remaining corrections to the image are made by the electronics discussed below). Converter 38 can be struck by the light at an angle and transmit the light in a parallel fashion normal to the other surface 182. This method converts the curved focus field of the image plane, and carries through the symmetrical small spot size with constant intensity no matter where it strikes the curved front surface of the plate. Thus, the curved focus field is accurately converted into the alternate shape defined by other surface 182.

Alternatively, lens 34 and/or plate 38 could be replaced with holographic elements and still remain within the scope of the present invention.

Plate 38 also guarantees that the spot is always the same size on the film or paper recording medium, whether the beam position is at the film or paper recording medium center or corner, or the beam is bright or dim. Fiber sizes as low as three microns are available. Such plates are commercially available (for other uses) flat on one side and either spherical or cylindrical on the other; but are not suitable for our purposes as is. The curved side must be ground into the correct near-toroidal shape, but since the ground surface does not bend the light rays, the grinding only needs to be accurate to within a few tens of microns, which is the depth of focus of the combined beams. This tolerance is two or more orders of magnitude more coarse than that required for camera, eyeglasses, or other image forming lenses.

The remaining three orthogonality errors are corrected with digital electronics described below, to alter the timing of the pixels and to vary the spacing of the lines. It has been found that these errors may not be corrected using computer programming due to the relatively slow processing speed of such computers.

A photocell 44 in Figure 1 located above the scan plane of polygon 36 detects the presence of a beam generated by lamp or unused laser beam 42. Light source 42 is placed below the scan plane of polygon 36 and has its light directed toward polygon 36 and at such an angle that light from light source 42 is reflected by polygon 36 and onto detector 44 at a point in the polygon's rotation prior to when image data is reflected by the polygon facets onto plate 38. Therefore, the signal from photocell 44 synchronizes the timing of each polygon facet, and allows for dynamic correction of speed and facet construction errors of the polygon 36.

Photodetectors 46, 48, 50 and 52 behind position defining pinholes at 47, 49, 51, and 53 respectively, may be used during the scanning process to detect color balance, overall power, and positional accuracy of the scanned beam. Before every exposure a test image may be sent to these detectors. This image will in turn expose the detectors in red, green and blue. The position of single pixels will be varied in sequential exposures or sets of pixies will be sent and sensed with respect to time. The main computer in the unit will use the sensed return from the photodetectors to compare the response in each color and in each moved pixel. New color tables residing in the electronics may be built by comparing the responses from each color. Changes in the frame scanner position tables in the electronics would be made based on the relative responses of various test image position made on sensors 46 and 48. Changes for the line time delay and clock rate tables would be made in comparison with the responses from sensors 50 and 52.

With the present invention, it is not necessary to blank the unexposed film or paper recording medium during this process because with acousto-optic modulation, no light is directed toward the image plane unless there is acoustic sound present. Notice that sensors 46, 50, and 52 are above the plate out of film or paper recording medium range and sensor 48 is below the plate out of film or paper recording medium range.

While pixel replication or averaging is an option for implementation within the present invention, the preferred embodiment uses another technique implemented by spotsize selector 54 and shown in additional detail in Figures 23a-d. Two flat glass plates 60, 62 are placed in the beam and set at equal but opposing angles 64. As the angle between the plates decreases, the overall beam path increases. This is because as the beam passes through the first plate 60, the distance of the beam normal to its source at 63a in Figure 23c, is shorter than at 63b in Figure 23d. Thus, the overall distance from the source to the film or paper recording medium plane is varied by introducing plate 60 into the beam path and tilting it relative to the axis normal to the path. This effect is shown clearly in Figure 23b-c. The second plate 62 is used to remove the transverse displacement caused by the first plate 60. Correction of this displacement is shown in Figure 23b. Also, the second plate 62 corrects any chromatic aberration introduced by the first plate 60, while changing the overall beam path still more. Wedges displace the converging beam and distort the spot shape in a chromatic fashion, and consequently are not suitable for this purpose.

By rotating plates 60, 62 about their vertical axis, the overall focal length between the final focusing optic and the fiber plate changes. Thus, the image spot may be intentionally expanded (beam focus waist moved to a point having larger diameter) to provide a larger spot at the film or paper recording medium plane. For the smallest spot size, the device is focused with the two plates 60 and 62 in a nearly parallel position. The plates are not used exactly parallel to prevent unwanted reflections from causing ghost images on the image converter plate. When a slightly larger spot size is desired, the angle 64 is widened according to the amount of spot size increase desired. Thus, image formats comprising fewer pixel data (i.e. lower resolution) may be accommodated by slightly expanding the beam to a size comparable to the resolution of the data provided. Additionally, if one wished to vary the shape of the spot on the image plane, this is accomplished by varying the angle of one plate different from the other.

Electronics

The electronic controls must fulfill these requirements:

1) Gather and control the digital data from the System or external computer for image recording and to the System or external computer in image input mode;

2) Correct for non-linearities in the system and for specified color, gamma, and other corrections. Accept correction look-up tables if provided;

3) Accurately position the frame scanner for or paper recording medium scan line increment;

4) Convert the digital data into an analog signal for the modulator drivers;

5) Convert analog sensor data into digital format for image input;

6) Adjust the timing of pixels with respect to line scanner position;

7) Correct for the three orthogonality errors not corrected by the fiber plate, namely the spacing of pixels horizontally and vertically, and correction of the vertical pincushion effect manifested by the varying spacing along horizontal scan lines;

8) Optionally detect and correct variations caused by varying laser power, by drift in frame scanner position, by changes in the relative power of the laser beams affecting color balance, or drift in modulator effectiveness.

The electronics control of the preferred embodiment is shown in detail in Figure 25a-e. As shown in Figure 25a, connection with the computer system bus is shown as 17. In the preferred embodiment, this is a 32 bit interface, however other bus widths are possible and within the scope of this invention. Internally, the data is broken into four, eight-bit sections (bytes). Three of these sections (or bytes), or 24 bits, are used to transmit color (one 8 bit byte each for red, green, and blue) or table data. Thus, 256 grey shades may be provided (2⁸) for each color. The remaining section transmits 8 bits of control information. The content of the control byte determines the routing and usage of the other 24 bit data information. The control bits can be used to reconfigure the use of the system bus to 16 bits of information routed to the appropriate chips by the "Bus Interface Routing Logic with control & Latches" and 16 bits of control and/or data shown as bus 19.

As mentioned, the data is transferred from the System computer to the circuit board through a 32 bit System data Bus 17. The data is transferred in 32 bit words, 24 bits of this are the data for an individual pixel, and 8 bits are control information. Data is transferred into three parallel channels, one for each of the primary colors; Red, Green, Blue.

All three channels work in parallel from the same Pixel Clock, so that the data for all three channels reaches the laser modulators at the same time. Consequently the operation of one channel 150 will be described with the understanding that the description applies to the other two. Data is transferred from the System BUS 17 to the First-In, First-Out (FIFO) buffers Ul such as Cypress CY7C464 where multiple lines of data can be held. This can be a single large device, but in the preferred embodiment is an array of two devices connected to form a buffer of 64K by 8 bits. This FIFO buffer allows the data to be written into the buffer at the speed of the system bus and read out of the FIFO buffer at the rate needed by the modulators to form the image line. Data is accumulated in the FIFO buffer until at least 1 full line of data (Typically, 16,000 pixels) is available at which time data may be accessed from the buffer to the modulators. Data will continue to be accumulated from the system until the FIFO buffer(s) are full (approx 4 full lines of data pixels at 16,000 pixels per line). Data can be read simultaneously with data being written into the buffer(s). This reading is done independently and may be at a different rate than the data being written into the buffer.

Data is read from the 3 channels of FIFO buffer Ul by clocking the READ line with the PIXEL CLK signal. Each PIXEL CLK is delayed by the delay buffer UlOO between the READ line of the FIFO buffer Ul and the Clock line of the RAMDAC U2 (1/3 of Analog Devices ADV7152). In the preferred embodiment, the delay buffer UlOO may be a simple hex invertor such as a 74LS04. The RAMDAC U2 is a complex integrated circuit look-up table which takes each of the three 8 bit data words, and uses each 8 bit data to look up a corrected 10 bit data word from an internal Random Access Memory (RAM) table. The three 10 bit data words (one from each channel) are then converted from digital data to a three voltages by three internal Digital-to-Analog Converters (DAC). The combination of these two functions, RAM memory and DAC converter, is referred to as a RAMDAC U2. The 10 bit data for the lookup function is used to correct the data for such variations as the source of the data, i.e. errors in the source scanner, the film or paper recording medium type being used for output, characteristics of the individual lasers being used, and other such variables. This data is available from manufacturers data, formula, from actual tests, or may be part of the pixel file to be recorded at the image source. This data is read and combined by the internal system computer (not shown) and then down-loaded to this RAMDAC U2 through the system data bus.via BUS INTERFACE ROUTING LOGIC.

The output of the RAMDAC U2 is then characterized and amplified. Film or paper recording medium responds to light stimulation in a non-linear fashion, most closely resembling a log function. In the preferred embodiment for a laser film or paper recording medium recorder, the LOG AMP amplifier U3 is chosen as a logarithmic (LOG) response amp, such as Analog Devices AD640. The Log AMP used in the output path of each color is used to expand the range of the output of the RAMDAC. As such, it functions as a log multiplier of the RAMDAC output. Thus, this allows the expansion of the 10 bit data from the RAMDAC U2 to a 12 or more bit equivalent in order to achieve a contrast ratio of over 4000:1. This could also be accomplished by use of a 12 bit lookup table within the RAMDAC when such a part becomes available. The principle and method of expanding the 8 bit data to a 12 bit (or more) range is an important part of this invention. The color data is provided as 8 bits per color as previously discussed. This allows over 16.7 million possible color combinations ("colors") of the red, green and blue data, but because each of the colors has only 256 combinations, the contrast or range of the individual colors is only 1:256. Typical output materials can record contrasts of 1:4000 or more, so while 16.7 million colors is a sufficient number of colors, the 1:256 contrast range is not sufficient. The previous state of the art was to use more bits/pixel which provided the contrast needed but required significantly more complex and expensive electronics to manage 30, 36, or more total bits (3x (10, 12, 14 bits/pixel)). This method allows the 8 pits/pixel to be expanded to the required 1:4000 or more contrast range using only 8 bit per color processing of the data. The method is implemented by determining the total number of colors required and thus the required bits per color, then the required contrast range is defined. This method then maps the required number of color bits/pixel over the required contrast range via a wide digital lookup table and/or analog expansion (linear or non-linear) of the resultant signal after the digital to analog conversion has been accomplished.

As an alternative, the amplifier U3 can also be a linear amplifier in order to retain a linear response. The amplifier U3 provides additional voltage and/or current drive capacity not available from the RAMDAC U2 alone and protects the RAMDAC U2 from the modulator driver circuitry, from voltage spikes and the like. The output of the amp U3, next proceeds to the modulator driver, such as Isomet 232A-1, which is external to this circuit. This modulator driver amplitude modulates a carrier frequency based on the value of the input voltage from the amplifier U3. The carrier frequency is typically from 40 MHz to 250 MHz. In the preferred embodiment, the frequency is 80 to 125 MHz and the output 23 of the log-amplifier U3 is designed to drive the 50 ohm input impedance of the modulator driver (not shown) from 0-1 volts. The output of the modulator driver is an AM modulated signal used to drive the acousto-optic modulator (see Figure 1) which varies laser light intensity passed therethrough as discussed in detail above. As also mentioned above there would be three input channels 150, one for each of the three colors.

The data path previously described which flows from the System computer Bus 17, to the FIFO Ul, through the RAMDAC U2, through the LOG amp U3, to the modulator driver and to the modulator, is the pixel data (typically 16,000 pixels) which forms a single horizontal scan line. These scan lines are repeated from the top of the image to the bottom. In the preferred embodiment this would be from 4,000 to 20,000 scan lines, depending on the format and aspect ratio. (Thus, the entire picture may be made up of 16,000 * 20,000 or 320,000,000 pixels.)

As discussed above, distortions occur whenever a focused beam is scanned onto a flat field. In a high quality output device such as is the one being described herein, these distortions must be corrected or compensated or the full capability of the apparatus will not be realized. A distortion called "pin-cushion" distortion shown in Figure 18 causes the image to curve in from each side from the top to the center and back out at the bottom. The general shape of the distortion on the left side is ")" and the right side distortion is in the general shape of "(" so the image is most narrow at the center. Therefore the amount of correction required to correct the left side distortion for each scan line from the top to the bottom will vary depending on the vertical position of the scan line. This correction be can be accomplished by using the center line as the horizontal reference points and delaying the start time of the data a specific amount of time for a horizontal line above or below the center (where the image is the most narrow). Matching the delay to the vertical position of the scan line relative to the center line allows the data to be aligned on the left side. Consequently, in order to formulate the correction the vertical position of the scan line must be determined.

The vertical line position is determined by the LINE COUNT counter U5 shown in Figure 25b. This counter U5 is loaded from the System computer via Bus 17 with to a starting value. Various formats will require different start values. In the preferred embodiment the LINE COUNT counter U5 provides a 14 bit output (bus 21) allowing 2¹⁴ (or 16384 values) in order to address the 16,000 possible lines but this can be expanded to any resolution appropriate for the resolution and output media to be used. The value of the line counter U5 then provides the address of the line being scanned.

These 14 bits of LINE COUNT provide an address for a lookup table in which is the amount of delay required for the left side pixel alignment correction for each line. This START DELAY LOOKUP table U6 is implemented using static Random Access Memory (RAM) or one of the varieties of Programmable Read Only Memory (PROM). The number of data bits required for the delay for uniform alignment is determined by the configuration of the specific device. In the preferred embodiment, a 16 bit delay value is obtained for each scan line and may range in value from 0 for the center line to 64K depending on the physical implementation of the recording and optical geometry such as the distance from the final focusing optic to the recording medium and the like. This delay table can also be loaded with different data in order to create masking and other special effects by using the delay to determine where on the current line that data is written.

During line scanning, the 16 bit delay value is loaded into a 16 bit PIXEL START DELAY counter U7 at the beginning of the scan sequence for that line. The counter U7 is decremented (count down) by a specified frequency clock LINE CLK. When the counter reaches zero, the necessary delay has occurred and the pixel data is ready to be clocked out.

Correcting the left side pixel alignment will, however, exaggerate the distortion on the right side. Additionally, another distortion occurs because pixel data can not be clocked at a uniform rate as the scan moves from left to right. The data is spaced further apart on both sides and closer in the center (before correction). Because the left side of the image has now been aligned, the right side distortion is exaggerated and thus the spacing of the various frequencies is not symmetrical on both sides of the image center. It has been determined by calculation and experimentally that 16 changes of frequency (at various time intervals) from left to right are adequate to correct the pixel spacing distortion to the required accuracy for 4x5 inch output and smaller sizes.

Several clock frequencies are required. An engineer skilled in the art of digital electronics could devise an embodiment using fixed clocks. A method unique to this invention, however, is the method of utilizing a Direct Digital Frequency Synthesizer (DSS) U8 shown in Figure 25d. Since this DDS can generate a variable frequency signal, we have implemented this as a variable clock signal generator. This use is unique because DSS chips were designed for generating digital tuning frequencies and communication signals and have not been generally applied as digitally generated clock sources. The use of this method allows for the elimination of the analog components in clock generation which cause drift and require calibration, and the simplification of the circuitry by eliminating the traditional phase-locked loop frequency generators which are not as precise or fast in changing to the required clock rate. This clock is a Direct Digital Synthesizer (DSS) chip (such as Qualcomm Q2220) and these use 20 a more bits as the input. They input directly to the DSS to provide a unique, specific output frequency. The device generates a variable digital clock signal, ranging from a few hertz to 45% of its Vref frequency based on the value of the input. This Vref frequency is shown in Figure 25 as being 50 MHz, but could be implemented at other frequencies. The various clock frequencies for the preferred embodiment are generated by DDS circuit U8. This DDS circuit uses a crystal oscillator to generate the Vref frequency. A 23 bit data word is then applied to the input of the DDS by the output of a lookup table formed by 16 bit address applied to a RAM lookup table U9 and U10 which, using a 50 MHz reference frequency, will generate any frequency from 0 to approx 22 MHz with an accuracy of better than 3 Hz depending on the input data word. The data word representing the desired frequency is obtained as follows:

The 12 Most Significant Bits (MSB) of the 14 bit LINE COUNT data bus 21 is used as the Most Significant Bits of the address for the lookup tables U9 and U10 which provides the data for the determination of the LINE CLK frequency signal J. It has been calculated that each vertical scan line does not require its own unique set of 16 horizontal frequency possibilities. Rather the horizontal lines may be lumped into 4096 groups of 4. Using the 12 MSB of the 14 bits of the LINE COUNT, there are 4096 possible groups of lines. The remainder of the 16 bit address of the 23 bit frequency lookup tables U9 and U10 is determined by the output of a 4 bit counter Ull. This 4 bit counter Ull is set to zero at the beginning of each line. The same 12 MSB of the LINE COUNT form address of a FREQUENCY DELAY LOOKUP TABLE U13 with the balance of the address determined by the output of the 4 bit counter Ull. The output of the FREQUENCY DELAY LOOKUP TABLE U13 is a value which determines the number of cycles for which the clock frequency is to be used. This value from U13 is set by loading the output of the lookup table into a 16 bit FREQ DELAY COUNTER U12. After the counter is loaded, the counter decrements with the pixel clock until the counter reaches zero. The zero output of the counter is used to increment the 4 bit counter Ull which then changes the address of the FREQ DELAY LOOKUP TABLE U13, which then again loads the FREQ DELAY COUNTER, and it again counts down to zero. This process continues until all 16 possible variations caused by the addresses set by the 4 bit counter have been used. The 4 bit counter is reset to zero at the end of each line so the count will always begin at zero.

The changing of the 4 bit counter Ull outputs are also used, as the 4 Least Significant Bits (LSB) along with the 12 MSB of the LINE COUNT'S, of the address for the Frequency Lookup tables U9 and U10. As these 4 LSBs change with the count, the address is changed and therefore causes output data from U9 and U10 to change which in turn changes the output frequency of the DSS U8.

The output of the DIRECT DIGITAL FREQ SYNTHESIZER (DSS) U8 is in the preferred embodiment a 12 bit digital word which changes at the specified rate determined by the address input from U9 and U10. The DSS is designed to have this 12 bit output routed to a Digital-to-Analog Converter (DAC) to produce a sine wave output of the same frequency. However, in this application, only a square wave output is needed. There are two ways to achieve this; the DAC method could be used with the resulting sine wave then converted to a square wave. Alternatively, as is selected for the preferred embodiment, the Most Significant Bit (MSB) of the 12 bit output (pin 18 of Q2220) of the DSS chip U8 which changes from logic "0" to logic "1" at the specified frequency is used as the square wave signal. This method will have a very small skew (approx 12 ns) within each cycle in the frequency but this is not significant in this application. When the output of the DSS U8 is at the desired frequency, then the LINE CLK signal J should be synchronized in order to be useful. In this application, the image is written onto the film or paper recording medium a line at a time. As described above, this line scanning is done by bouncing a modulated laser beam off the rotating multifaceted polygon mirror 36 (see Figure 1). These polygons can be made very accurate in mirror quality and vertical flatness but they can vary in angular position of each facet. The data must be clocked out to the film or paper recording medium only after detecting that the active facet has moved into the correct position 74b and 74c (see Figure 12). As described earlier, this is accomplished by using some of the excess laser light or other light source 42 (see Figure 1) and reflecting it, from below the plane of the polygon 36, off of the active facet 37 to a sensor 44 above the polygon (see Figure 1). The polygon scanner 36 is optionally controlled by a pulsed signal from the control electronics (not shown in Figure 25a-e). In the preferred embodiment, the polygon 36 will rotate one revolution for every 8 control/power pulses. This is convenient, because in the preferred embodiment there are 8 facets, which equates to one per pulse; however, other facet numbers and control frequencies, including a non-controlled free running polygon motor, are implementations that will operate within the scope of this invention.

An output from the control circuitry can, optionally, be used to direct the speed of the polygon. In the preferred embodiment, simplicity suggests that this speed be fixed. However, in alternative embodiments, flexibility can be enhanced by using a control code from the system to set up the polygon scan rate. It is, however, important to remember that the Scophony dynamic pixel focusing process depends upon a predetermined polygon scan rate. If this rate is changed, it could be necessary to re-orient the optical configuration depending on the resolution desired.

The slower of the two scanners, the vertical or frame scanner, is in the preferred embodiment, either a galvanometer or a stepping motor driven mirror. The galvanometer is purchased with control circuitry to position the mirror based on a digital value, in the preferred embodiment this digital value is supplied by LOOK UP TABLE U4 Figure 25b. LOOK UP TABLE U4 outputs a 16 bit binary value based on the line count data available from the line count bus. This 16 bit binary value is utilized by the galvanometer circuitry to position the galvanometer at the correct setting for that line. The positioning requirement is to move the mirror exactly one evenly spaced line width vertically during the time between polygon facets when the pixel line writing is not active. This requires use of correction circuitry. Uniform step increments and/or constant step rates and increments will not accomplish the precise, even spacing required for these reasons:

1) The data, while buffered deep enough for one line, cannot be guaranteed to be available for the entire image, so the vertical frame scanning must be able to be stopped at any line to accommodate the data transfer rates and availability.

2) There must be contingency for not scanning with every succeeding polygon facet.

3) The vertical increments are not exactly evenly spaced, as per the earlier discussion of orthogonality errors, and variation in positioning must be available to correct this error.

4) The control of the frame (vertical) scanner is in the form of a 16 bit value that specifies the required position. The scanner will then respond with a signal F (F_SCANNER_AT_POSITION) which confirms that the mirror is at the required position and is stable. The alternative method is using a stepping motor in which a relative number of steps is supplied by the lookup table U4 rather than the absolute position used with the galvanometer. The stepping motor control (indexer) would then provide the signal F (For F_SCANNER_AT_POSITION.) As the galvanometer is the more easily implemented of the preferred embodiments, this is the approach shown in Figure 25, but the alternative could be fashioned by one skilled in the art of digital control circuitry. The reason that lookup table U4 is necessary is that one of the errors in orthogonality not corrected by the fiber plate 38 is the variation in the spacing between vertical lines given a constant deflection angle. This table U4 corrects for this third error in orthogonality. As previously mentioned, the near-toric shape of the plate 38 corrects for the first two errors, namely the error in focus length and the bowing of lines from the line scanner sweep.

The timing of clocking the data to the polygon line (horizontal) scanner, requires and accurate determination of when each active facet (face) is in position. This Facet Pulse Detection Circuit (U14, PDD1, R4, Cl) shown in Figure 25c is a simple photodiode detector PDD1, such as UDT sensors UDP451, biased to produce a high gain signal when illuminated by light source 42 (Figure 1) and is familiar to those skilled in the art. This circuit produces a pulse when the facet is in the correct position. This pulse activates the SET input of a SET/RESET Latch U135. This output is passed through logic which synchronizes the signal to only allow complete cycles of the LINE CLK signal C to be used to decrement the PIXEL START DELAY counter U7. When the correct amount of delay has expired then the LINE CLK signal J pin 18 U8 is passed through another synchronizer circuit UllOa, e, f and U145 on Figure 25e which then becomes the PIXEL CLK. This signal decrements the counter which represents the current pixel being addressed. The current pixel is specified by the 16 bit PIXEL COUNT counter U15. This can be loaded via system bus 19 with the number of pixels per line. The output of the PIXEL COUNT is distributed as bus 23. The output bus 23 is also used to specify the address of the current pixel when digitizing an image for input to the computer.

If an image is to be output which does not have the same number of pixels as the number of pixels to be output, it is possible to expand (horizontally) the pixels output. This is accomplished using an optional DIVIDE BY N counter U16 Figure 25e as shown in the PIXEL CLK signal path. This counter U16 is loaded with a number, X, which is:

X = number of pixels output per line number of pixels in source image This will then divide the number of PIXEL CLK, signal H, by this number and thus expand the lower resolution data to fill the horizontal output line. The vertical lines should also be repeated or interpolated to maintain symmetry. This may be accomplished by the system computer reloading that line data and rescanning the line thereby duplicating that line. This DIVIDE BY N counter U16 can be removed if the number of source pixels matches the number of output pixels.

This schematic shows the functional elements. This schematic is sufficient to disclose the methods used for the operation and correction phases required for the preferred embodiment. There are items not shown such as; the System bus interface routing logic with control and latches (which depends on the System CPU chosen), the logic needed to load table values into the variance look-up tables, interface and combinatorial logic, or required delays, level shifting, timing, and other incidental logic which would be familiar to, and could be fabricated by, those skilled in the art.

In summary, there are 5 types of orthogonality variations or distortions which affect an image scanned as described in this invention;

1) Focus of a focused beam as it moves from the center to all edges.

2) "Pin-cushion" distortion (bowing of the edge) on the vertical edges of the scans away from the center.

3) "Pin-cushion" distortion, bowing of the horizontal scan lines away from the center of the image.

4) Uneven vertical spacing of horizontal scan lines from either side of the center of the image.

5) Uneven spacing of pixels on horizontal scan lines. These have been corrected by;

The fiber optic plate, ground with a near-toric shape corrects the distortions listed in nos. 1 and 3 above and was discussed in detail earlier.

The variable frequency clock with multiple frequencies per scan line (composed of items identified on the schematic Figure 25a-e, as U5, U8, U9, U10, Ull, U12, U13 correct the spacing problems of the pixels (#5 above) and aligns the trailing edge of the scan (#2 above).

The "bowing" of the starting edge (#2 above) is corrected by a variable start time for each line and corrected by items identified on the schematic Figure 25b, as U6, U7.

The uneven vertical spacing (#4 above) is corrected by use of an accurate frame scanner with digital input and by line look up table U4 and line counter U5 on schematic Figure 25b.

.As has already been mentioned collaterally above, in addition to the optical and electronic configurations, the recorder apparatus of the present invention comprises a system computer. Resident within the system computer is a software package. The purpose of the system computer and associated software is to provide an overall management of the image recording and input processes as will be discussed below. It must be kept in mind that the details of these configurations discussed below pertain to the preferred embodiment. Those skilled in the art of software and computer engineering will recognize that there are many alternative configurations which will achieve the same result.

System Computer

The System Computer controls the initialization, setup, calibration, as well as the beginning and ending of the image process. In the preferred embodiment, it is a computer that is an integral part of the film or paper recording medium recorder and in computing power is comparable to widely available micro-computers. The electronics shown in Figure 25 are found on cards which plugs into the system computer data bus 17. The system's responsibilities are to advance the film or paper recording medium, load and reset tables, direct the scanners and modulators so that sample light sweeps can be made of the sensors 46, 48, 50 and 52 shown in Figure 1 before an exposure to assure calibration. These sensors are used to correct for any long term stability errors in the lasers or in the positioning orthogonality. Colors are corrected by sensing the relative power of each of the three colors. There are several approaches to the implementation of corrections required based on this information. These are: to return the information to the System Computer which will then build new LUTs for download for the next image, to accomplish these correction functions in the computer in the device, or to load correction offset registers for the various values. Such registers are not shown because in the preferred embodiment there will be a complete computer within the image recording and input device that will exercise these between-frame controls and create new corrected tables as necessary.

Note should be taken that while there is a computer within this imaging device (the system computer), it cannot currently serve by itself to do the dynamic timing corrections for pixel and line timings. This is because the pixel rate is much faster than stored- instruction computers are currently capable of performing. The computer is important, however, to control the more complex but less time critical events between frames, such as film or paper recording medium movement, lookup table generation, image format conversion, image acquisition and transmission, diagnostics, and user interface instruction formatting and diagnostics.

The electronics described above resides preferably on a card set that is inserted into a system computer's bus interface. In one preferred embodiment the electronics on the card have a direct connection to the image source, a network or data disk controller, that is also shared with the bus of the system computer. This interface is faster than another embodiment where the data from the image data source goes through the memory of the computer, but is more difficult to implement. As it is simply a cost tradeoff in implementation, the designer does not prefer one embodiment over the other; and therefore the software necessary to drive both electronic implementations is described.

Figures 33-35 illustrate both the control actions necessary and the data flow and storage required to record an image on film or paper recording medium. Heavy lines and shaded blocks refer to data flow, while thin lines and unshaded blocks show control actions and sequences. The level of detail given shows the overall concept of the software at a level to imbue in one skilled in the art the processes necessary to implement a film or paper recording medium recorder according to this invention.

The flow control shown in Figure 33 is that which would be employed when the image data from the host is already in a format compatible with the recorder system apparatus. Alternatively, the flow control shown in Figure 34 will be used when the image data from the host is in a industry standard format such as TARGA or TIFF and must first be converted to a format acceptable to the present recorder apparatus. Figure 35 illustrates the control procedures used in the color alignment process and in conjunction with sensors 46, 50, 52, and 54 in Figure 1.

Software for Pre-formatted Image Data

As mentioned above, Figure 33 describes the software which would be utilized when the image data from the host computer is already in a format compatible with the recorder apparatus.

Beginning in the upper left-hand corner of Figure 33, at the block 430 labeled "START". This function is initiated by the operator at the beginning of the work-day. Actions implied include loading the System program and assuring that all image data files are available.

Next, the system checks that all covers are closed 401. This is done by inquiry to the interface card Figure 25, where a large AND gate (not shown) develops a signal corresponding to all optical or mechanical interlocks being closed. After this, the program directs that the Lasers and line scanner be started 402, and waits for a signal from the interface indicating that these components are up to speed and functioning. Further, the frame scanner 32 Figure 1 is exercised to its maximum positions while the position feed-back is examined to assure that the frame scanner responds to position commands in a reasonable period of time.

The operator next is queried as to the type of film or paper recording medium to be exposed or scanned 403. This indication drives the selection of default timing and color tables 405 to be loaded from an external disk. The appropriate tables are loaded 404 from disk and developed into memory in the system computer. The first of these tables is the color correction table 421. It has 256, 10-16 bit entries for each of three colors. The red, green, and blue 8 bit image data values are used as indexes into the corresponding color table entry RAMDAC U2 Figure 25a. Found at each location is a value that is converted eventually to a voltage in the modulators (18, 20, and 22 in Figure 1) that drive the intensity of that color for any given pixel. This look-up is performed in the data interface card (Figure 25a-e) at high speed.

There are several sources of non-linearities in the process of moving color imagery to the modulators, including: film or paper recording medium color imbalances; film or paper recording medium response non-linearities, usually approximately logarithmic in nature; variations in the color response of different film or paper recording mediums to the laser light used; gamma corrections for either the source image or the film or paper recording medium; custom variations applied to the image by the creator or retoucher; variations from recorder to recorder, or by time within the recorder. All are corrected with the 421 color balance table procedure discussed below.

Given the width of variation, the color correction codes that are looked up in the table may vary considerably. To summarize, there will be no missing codes on table input 421, but there will perforce be missing codes within the looked-up values, since the value range is wider.

Next, the default scan control table block 400 on Figure 33 is loaded 406 from disk into system computer memory. Values that occur for each scan line shown respectively in the table outline in Figure 33, include: 1) The actual position of the frame scanner for the line.

2) A value that indicates the dynamic time delay from facet detect to image line start.

3) A value that indicates the dynamic time for each pixel at the beginning of the scan line.

4) A value that indicates how long (how many pixels) this pixel timing is in effect.

5) A value that indicates the next dynamic time for each pixel after the number of pixels indicated above have been written.

6) A value that indicates how long this next pixel timing is in effect.

7) More similar values, up to 16 spacing and timing variations on each individual scan line.

This default scan control table 400 is created in memory. Both this table and the color table 421 will be written to the image interface electronics 417 (Figure 25) after information about the particular image and optional dynamic corrections and alignments have been received and the tables have been updated to reflect the changes, if any.

Next, the operator/network is queried to indicate the image file to be recorded or input 407. For this electronic interface set, the image file 409 contains general information about the image, such as its identity, film or paper recording medium type and size to be used, composer's color corrections if any, its size in lines and pixels, and possibly pixel aspect ratio. Following the general information, the actual image is organized in 32 bits, 8 bits of red, 8 bits of green, 8 bits of blue, and 8 bits of operation code. The operation code bits are used to indicate end-of-line, end-of-frame, or other actions. These other actions could include operation codes that load other tables to the card or to control scanning. This file format is specific to use with this electronic set. As mentioned above, using this software configuration, the film or paper recording medium recorder does not dynamically change a standard image format into a recorder acceptable format, as the write time of a file in this format is faster than any conversion process could keep up with. Use of the software configuration shown in Figure 33 assumes the image data file created by the host computer was created in a format compatible with the present recorder apparatus electronics. Should the film or paper recording medium recorder be required to operate directly on standard format image files, then the data is handled using the Figure 34 software configuration, with a considerable attendant performance penalty.

Once identified, the header of the file is read 407, and the line width, line count, and pixel aspect ratio is recorded 408.

If the header indicates that there is an included color correction table, it is read in 410, and the default color correction is mathematically merged 411 with it so that the default table 421 now contains entries that reflect both the old and the new set of corrections.

Given that the color correction feature is implemented, (either it is automatic for every image or that the operator or some other factor requests it), the automatic alignment and color balance routine is performed, as described below in the annotations 413 for Procedure "A" Figure 35. This routine does a dynamic determination of the color balance of the lasers in the current state, determines the edge of the image plane vertically with respect to any variations in frame scanner settling position, and recalibrates the horizontal position of the beginning, middle, and end of the scan line to compensate for any instabilities in line scanner speed or facet detector response.

It should be noted here that components have been identified that do not require dynamic correction, but they are perforce more expensive than less well specified components which may well require this dynamic correction routine. Again, either embodiment is acceptable, and it is up to the implementor to decide whether the additional time and complexity of the alignment and color balance capability is more or less expensive than the cost of finer scanners that don't need the correction. Components thus affected are:

1) Helium Neon laser. Inexpensive lasers can vary in power with variations in line current

2) Argon laser. It should be noted that even the most inexpensive Argon lasers typically have dynamic power stability correction capability. To assure that the balance between the selected green and blue is maintained, the detection circuitry should be sensitive to either one or the other color, but not both. This is to assure that a change in color balance between the two lines such that total power does not change is detected and corrected.

3) Galvanometric frame scanners vary in cost based on their long term positioning accuracy and repeatability. Another approach to vertical scanning is the use of stepping motors. Such motor assemblies are accurate enough, but quite slow, subject to hysterisis, particularly since gears are necessary to position to the small step distances. More complex software is required, since one must tell the motor of the number of steps or microsteps required, as opposed to a value representing an absolute position. While stepping motors for frame scanning are not the preferred approach, the assembly might be somewhat less expensive and therefore not precluded in the preferred embodiment.

4) Polygon line scanner speed control circuitry effectiveness varies with expense. In the preferred embodiment, the best polygon control circuitry would be used because significant variations in polygon scanner speed can occur during a single scan without it. Variations in speed do not effect the positioning of the beginning of the scan line under our design where an optical facet detector is used on the active facet, but the width of a line or the spacing of pixels along the line could be affected. The control sequences continues starting at "B" of Figure 33, with the next action of the software downloading the color table 414 and timing table 415 to dynamic look-up memories in the interface electronics 417 (in Figure 25).

At this point the interface electronics (Figure 25) are ready to accept imagery. The first imagery 416 is optional, and it will write the identification of the image on to the extreme edge of the film or paper recording medium. This allows sorting of images after they are developed.

Now the actual image is recorded or input 418. The data path from the image source 409 (usually disk, preferably removable optical disk) travels both through the system computer's data bus 17 (Figure 25) and directly to the card. Controls on the card allow the direction, gating, and timing of the data from the source into the FIFO buffers Ul (Figure 25). These controls are driven by flags in the FIFO that indicate when the FIFO can accept data, when there is enough data in the FIFO to write a line, and when the FIFO is full 419. These flags and controls are absent from the electronics when the implementation shown in Figure 34 is used.

When this split data path is used, the next control block initiates a read from the disk through the interface electronics 418. The data flows directly to the interface. However, any status flags, seek address, data lengths, and error conditions flow to the system computer via its I/O bus.

The next block is the end-of-file detect, determining when the input/record action is complete. When complete, end-of-image actions are performed, such as advancing and/or rewinding film or paper recording medium 420, and control returns via "D" (Figure 33) for the next image.

If not end of file, a status inquiry 419 is made of the interface to determine whether there is room for another line's worth of data in the FIFO. If so, another data transfer is initiated. If not, a wait 426 mechanism is invoked to await a condition 428 when the FIFO can accept another data line. Three different mechanisms can be implemented. Each of the three would be familiar to those skilled in computer interface and software design. Our embodiment allows for any of the three, with no particular preference. They are:

1) Wait and post. The operating system suspends operation in this program, possibly dispatching some lower priority task, awaiting an interrupt. Such an interrupt would come from the interface, and the operating system would post the completion of the data transfer event, and the data transfer task would become eligible for dispatch by the operating system to transfer more data, as shown in the returning arrow to the set-up-path block.

2) Spin. In this implementation, the program makes continuous looping inquiry of the interface card, leaving the loop and returning to set-up-path block as soon as the FIFO Ul (Figure 1) is flagged as available for data. This is the simplest implementation, but one that allows no background tasks.

3) Timed wait. Here either the system waits for a known time that guarantees buffer depletion, or waits for a shorter time and queries the interface electronics again, as in the spin implementation above.

As mentioned above, when the image data file created by the host computer is created in an industry standard format such as TARGA or the like, the image file must first be converted into the proper format compatible for the present recorder apparatus. In this situation, the software shown in Figure 34 would be used for the conversion and recording process.

Software for Industry Standard Format Image Data

Figure 34 gives the software and data flow for image scanning and recording procedure when the image data was generated in an industry standard format. In this configuration, the image data is read first into system computer memory where it is processed before being sent to the recorder electronics for recording. Since the interface electronics does less, the software and system computer must, perforce, do more. With the additional software, there is an opportunity to accept standard image files, such as TIFF and TARGA, and reformat the data as necessary in the system processor, albeit with a significant performance penalty.

While Figure 34 is more complex, almost all of the differences are additions, not changes. Here we will therefore treat the functions common to both configurations in a cursory manner, concentrating on the additions and the few differences.

As in the complex electronics discussed above, the safety is checked 451, the lasers and scanners are started 452, the film or paper recording medium type is queried 453, and the color and timing default tables are moved 454 into the system computer memory 455, and 471.

The first difference occurs when the image file is read 457, and this difference reflects the ability to handle standard formats (TIFF TARGA and the like), including the decoding of image compression and decompression algorithms. After the line width and count, and pixel aspect ratio are acquired 458, a determination is made as to the image file type 478. If the file contains a color correction table, the table is read in at this time, and the new corrections are blended in with the default corrections 461. If the file is compressed, such as run-length-encode, JPEG, or fractal, then the correct decompression algorithm is selected 462.

Based on the characteristics of the image file and of the decompression algorithm, if any, a large image buffer is established in system computer memory 465. The buffer 465 is contiguous to aid in performance, and is established in blocks equal to one scan line. Each block is four 8-bit bytes per pixel in length (one eight bit byte each for red, green, blue, and a control byte). This data format is the same as that required by the image file in Figure 33 software. The larger the buffer, the higher the performance of the film or paper recording medium record or scan process. When the horizontal scanner presents a facet before there is data available for it (see description of facet detector system 42, 44 Figure 1 above), the momentum of the frame scanner must be stopped and reestablished. This process can take tens of facet times. Thus, the larger the buffer, the less frequently it will be necessary to stop and start the scanning process.

The next blocks, beginning at "B" in Figure 34, are duplicates of those in Figure 33. The next difference occurs in the process of inscribing the film or paper recording medium image ID into the output buffer 470. This process can be used both for image input and recording, since when inputting an image, the ID will be picked up in the electronic copy of the image. For the Figure 33 software configuration, the image was developed directly in the interface electronics, while in the Figure 34 software configuration, the image is developed into the image buffer 465 in the system computer memory.

While the logic of writing the complete image to the interface electronics between the two approaches is the same, the actual control must be amplified considerably in the current case. Therefore, we will consider the remainder of the process different and describe it in more detail.

The image data is read 472 in blocks from the data file 459. The data is held temporarily in its input buffer while end-of-file is sensed. If present, a flag is set for later action after output is complete 465. Next the input buffer is moved to the internal image buffer 470, applying any decompression or pixel re-orientation necessary 474 as dictated by the format of the input file, and as set up previously when the file header was processed.

Next a read-halt condition is checked 475. If not found, another image block is read 472 from the input file. The conditions checked for include:

1) Not enough space in the input buffer for another input block,

2) End-of-file on the last read.

The next actions are described in Figure 34 as occurring sequentially with the buffer-filling read, but could, under some multiprogramming operating system, be issued as a separate process, thus occurring simultaneously or overlapped with the read operation. Either approach could be used in the preferred embodiment with the sequential operation, as shown, being easier to implement, but potentially slower.

The output action is to output 476, a complete line block to the interface electronics 469, which will set a flag on the bus, when the action is complete, allowing another transfer.

If the buffer is empty, as set previously 473 on the input, then control passes via "D" in Figure 34 to process another image. If there is more image, then more data is read.

Alignment And Color Balancing Routine

Figure 35 describes the preferred software for accomplishing a self-alignment and color balance operation. This software is referenced in Figures 33 and 34 blocks 413 and 464 respectively as "perform align and color balance procedure 'A' ".

As discussed above, the routine is optional and requires optional electronic and optical equipment. When more expensive optical components are used, this routine becomes less critical, but with inexpensive components, it may be cost-effective to add the equipment necessary to exercise this routine. Either approach falls within the purview of the preferred embodiment.

The theory of the alignment routine is this:

1) A test pattern is sent to the film or paper recording medium recorder but the pattern covers the optical sensors above and below the fiber plate.

2) The pattern is repeated for each of the three colors

3) The sensors (46-52 in Figures 1 and 42) record in digital fashion a sensed value, recorded into an input buffer once per pixel output time, thus doing a partial scan of the test image.

4) The three resulting input images are examined to determine where the highest level sensed occurs in the image field. The position of the highest level sensed is compared to the ideal position, and the default timing table is adjusted to compensate.

5) The three color levels are compared with each other to see if they occur in the same position. If not, appropriate adjustments are stored for re¬ alignment, or made by dynamic means that make minor readjustments (mechanism not shown) in the positioning of the beam combiner mirror. If they occur at the same positions, their relative intensity is compared and if not optimum, changes are made to the default color table.

The test pattern is used and sent one line at a time to the interface. Line one is set and read 501, 502.

The test pattern 517 consists of four M x N full intensity pixel groups in each image, repeated for red, green and blue. The four groups, shown as 6 x 7 pixels 517 in Figure 35, are positioned at the very top and bottom of the image so as to scan beyond the edge of the fiber plate 38 (Figures 1 and 42). The sets are at the extreme left edge, centered, extreme right edge, and extreme bottom centered.

The active line number is set 504 into the interface electronics 505 and the image line is sent 506. Since the content of each image line is the same, it is not necessary to read each succeeding line from the image file, but this action if performed makes the software somewhat simpler.

The sensor data is transferred 507 from the interface electronics 508. This buffer records one value for each pixel time written to the film or paper recording medium image plane. The data is in a buffer in memory that will contain at once, all lines recorded from each of the three color images 517.

Next, unless the last line of the color has been written 510, the line number is increased 509 and the process is repeated. If last line, and if more colors are to be written, the line number is reset and the process continues. If the writing was complete 511 then control passes via "F" In Figure 35 to the buffer analysis software.

The remainder of the routine does not involve the electronics, but rather analyzes the buffer contents and adjusts the default timing and color tables.

At F on Figure 35, the process of evaluating the returned pattern is accomplished.

First, the returned data is examined 512 to find the brightest pixel for each sensor. This is done by adding the return values for red, green and blue for each position. The position of the brightest pixel for each sensor is compared 513 with the desired location, i.e. the location that would be expected if the system was perfectly aligned.

Also, the relative locations of the most intense pixels in each of the three colors is noted 513 to determine if there is misalignment between the colors. As the correction ability of the spatial filter assemblies 24, 26 and 28 shown in Figure 1, eliminates variations in the pointing ability of the lasers, dynamic registration corrections are not included in the preferred embodiment, but are within the scope of this invention. This correction could be applied to beam combiners 9 and 10 (Figure 1), in the form of motorized gimbals or piezoelectric adjusters, or any other dynamic adjustment easily designed by those skilled in the art of optical engineering.

If there are displacements between the desired and actual position, adjustments are made 514. For discrepancies in the frame direction, that is misalignments for sensors 46 and 48 (Figures 1 and 42), the frame position entries in the timing table 400 or 450 are modified. For discrepancies m the line position as indicated by sensors 50 and 52 (Figures 1 and 42), then changes are made to the offset and pixel spacing values in the timing table 400 or 450.

Next, the relative values of red, green and blue in the brightest pixels are compared 515 to assess color balance. Should there be discrepancies, then color table 421 or 471 is modified 416. This concludes the actions of routine "A" and control returns to the calling software.

Image Input

Image input is accomplished with this device by using multiple photodetectors. The image input process is shown in Figures 26-29. Each is masked with a dichroic filter that allows only one color to pass. The sample is preferably placed on the output side of the fiber plate. A raster scan is made of the plate using the three beams at an intensity set by the modulators. Some variation in intensity may be necessary to correct for defects in the sample film or paper recording medium or other errors. This is accomplished by using a digital image as the sample for detection. Such an image can be created by running an input scan without any sample in place and calculating compensating adjustments for each pixel. Figure 26 shows an arrangement of detectors 108, 110, and 112 where a large lens 106 gathers light dispersed from the raster scanned spot, and the three colors are then separated using dichroic filters 114 and 116. Figure 27 shows another arrangement wherein separate lenses 117, 118, and 119 gather the light for each color. Figure 28 shows another configuration wherein the input light 105 is gathered with a tapered fiber plate 111 and focused onto lens 115. Still further, Figure 29 illustrates where input light 105 is gathered with a light pipe 113 and focused by lens 109 and through dichroic mirrors 114 and 116 onto detectors 108, 110, and 112. These and other implementations would all qualify within the scope of the invention. The output of the photodetectors 108, 110, and 112 is coordinated with the input scan timings and either a new digital image is made in the system computer's memory or the sample correction image is replaced in real time.

Image Input Electronics

Figure 36 gives an overview of the separate electronics that receive and decode the sensors used to acquire digital image data. On the left are shown three sensors 108, 110, and 112 for red, green, and blue respectively. These sensors are equivalent to the ones shown in Figures 26-29. The sensors 108, 110, and 112 are driven by drivers 161, 162 and 163 that give an analog signal to logarithmic amplifiers 164, 165 and 166 that optionally contracts the scale of the input to fit a digital linear image scale.

The scaled analog signal drives three 12 bit A D converters 167, 168 and 169. The extra width of these convertors is necessary so that input non-linearities can be sensed for compression to 8 bits. Three FIFO buffers 170, 171 and 172, one each for red, green, and blue respectively, hold the data. Both the A/D converters 167-169 and the FIFOs 170-172, are clocked by the pixel clock from U115b in Figure 25 to delay buffer 176 in Figure 36. As shown in the figure, the pixel clock signal is used directly to activate A D converters 167-169. However, the pixel clock signal is delayed by buffer 176 prior to latching the data from the A/D converters into the FIFO buffers. The purpose of this is to allow for the time required by the A/D converters to process the pixel data.

The FIFO buffers 170-172 hold at least two lines of data. The pixels are clocked into the system computer's memory, or directly onto the image storage device (ie. floppy or hard disk, etc.), depending upon whether the Figure 33 or 34 image recording software is used. In either case, the data passes through look-up tables 173-175 to correct for anomalies from the source film or paper recording medium type, internal color correction, gamma, or other custom corrections from the user. The pixel data from the input circuit is transmitted to the system computer or other storage device, by bus 17.

The width of the communications bus 17 to the system computer could be less than 32 bits, meaning that all three colors cannot be extracted at one time. In the preferred embodiment, the width could be limited to 8 bits by a user selection, as scaled and corrected by log amps 164-166 and LUTs 173-175, so that data from all three colors could be fetched from the FIFOs 170-172 and LUTs 173-175 in a single 32 bit data transfer. The user could, however, at a performance penalty, opt for the wider sense range.

Not shown are the details of the electronics necessary to load the LUTs 173-175 from the system bus 17 interface, but these electronics are similar to the same processes used in the Figure 25 electronics and are well understood by those skilled in the art.

Image Input Software

This software is a simple extension of the Figure 33 and Figure 34 software. This software will load the LUT, and coordinate the writing of white imagery to the record interface and read the result on the input card.

Physical Layout

Figures 30, 31, 32 and 37 show possible physical layouts for the system computer 350 and optical components of the present invention. There is no preferred physical layout for these components and the layout may be freely altered to meet the requirements of the user. Those configurations shown in Figures 30-32, and 37 are meant merely as representatives of the many possibilities. Additionally, as also mentioned in the text, numerous modifications of software may be used and still accomplishing the recording and inputting process and remaining within the scope of the claims.

It is apparent that numerous other modifications and variations of the present invention are possible in view of the above teachings. For example, there are numerous configurations of input optical schemes which may be used. Additionally, the design of the preferred electronics given here is just one example of how the various timing requirements of the system may be accommodated. Those skilled in the art of electronics design will recognize that numerous changes in the design could be made and still accomplish the goals of the present invention. Still further, and as mentioned in the text, the preferred method is to use a correction plate having a flat other surface to accommodate use of the recorder with film or paper recording medium placed adjacent the plate. However, the shape of this plate other surface may be ground to accommodate other optical configurations. For example, the plate back could be ground to accommodate the image locus of a lens placed beyond the plate second surface to simplify projection and enlargement of the image passing through the plate.

It is to be understood that the above description is in no way intended to limit the scope of protection of the claims and it is representative of only one of several possible embodiments of the present invention.

There has thus been shown and described an invention which accomplishes at least all of the stated objectives.

Claims

We Claim:

1. A film or paper recorder apparatus wherein a high quality image may be

formed on a recording medium from an electronic image using a light beam of small spot

size thereby permitting a high resolution image to be formed thereon, the recorder

apparatus comprising;

light emitting means for producing a plurality of light beams comprising a plurality of colors, the generation and projection of said light beams forming a plurality of light paths between said light emitting means and said recording medium;

modulator means for modulating said light beams in response to said electronic image to be recorded, such that upon modulation, said beams comprise image data to be recorded;

beam combining means positioned within said light path for combining said light beams comprising said colors, into a single beam the projection of said single beam forming a combined light path between said beam combining means and said recording medium;

scanning means positioned within said combined light path for scanning said single combined beam in two dimensions;

computational control means for controlling said scanning and said modulator means and the flow of said image data; final focusing means positioned within said combined light path for focusing said combined light beam; and

an image plane converter means positioned within said light path and having a front, source facing surface and an other recording medium facing surface such that said focused combined light beam incident on said converter source face is transmitted through said converter such that the image focus locus shape is converted to the shape of said other surface, producing a beam of nearly uniform spot size at said

other surface regardless of the position of said spot on said converter source

surface.

2. The film recorder apparatus of claim 1 wherein said scanning means comprises a frame scanning means positioned within said combined light path for scanning said single combined beam in a first dimensional direction and line scanning means positioned within said combined light path for scanning said single beam in a second dimensional direction, generally perpendicular to said first dimensional direction.

3. The film recorder apparatus of claim 1 wherein said image plane converter

means comprises a near-toric surface on the side toward the source and substantially flat on said other surface, and wherein said converter means other surface is placed adjacent said recording medium such that said converter is operative to correct focusing errors resulting from said scanning of said beam and such that said electronic image is transferred to said recording medium.

4. The film recorder apparatus of claim 1 wherein said image plane converter means comprises a near-toric surface on the side toward the source and substantially

spherical on said other recording medium facing surface, and wherein a lens is placed between said other recording medium facing surface and said recording medium such that said converter and said recording medium are located at the foci of said lens and such that said converter is operative to correct focusing errors resulting from said scanning of said beam and wherein said image plane converter means is operative to produce a corrected image for transmission to said lens for relay of said image to said

recording medium.

5. The film recorder apparatus of claim 1 further comprising photodetector means positioned adjacent said image converter means and electrically connected to said computational control means and operative to measure said light beam incident thereon such that the recording parameters may be adjusted sufficiently to allow for registration and color balance and to determine the proper exposure using test beams on said photodetectors.

6. The film recorder apparatus of claim 1 further comprising means for correcting chromatic discrepancies in said final focusing means.

7. The film recorder apparatus of claim 1 further comprising a beam expanding means situated within said combined beam path for expanding said beams focus spot by a predetermined amount and for varying the shape of the spot wherein said beam expanding means comprises a pair of translucent plates within said beam path and through which said beam passes and which may be pivoted relative to each other thereby varying the beam path length between the source and the recording medium and thereby expanding the size of said beam.

8. The film recorder apparatus of claim 2 wherein said line scanning means comprises at least one rotating polygon scanner.

9. The film recorder apparatus of claim 2 further comprising a sensing means for

determining that said rotating polygon line scanner is in position to transmit said image

data to said recording medium and for communicating positioning status to said computational control means.

10. The recorder apparatus of claim 2 wherein said line and frame scanning means are situated between said final focusing means and said recording medium, thus avoiding the introduction of errors of focus and of registration between said plurality of colors when said beam passes through said focusing means.

11. The film recorder apparatus of claim 1 further comprising an attenuator means positioned within said light paths for adjusting the intensity of said light beams.

12. The film recorder apparatus of claim 1 further comprising filtering means positioned within said light paths for more precisely selecting the desired frequency of light for each of said colors.

13. The film recorder apparatus of claim 1 further comprising spatial filtering means positioned within said light paths for filtering said light beams to the desired beam spot size wherein said spatial filtering means comprises at least one lens and an opaque screen having a pinhole therethrough and wherein said pinhole is located at

approximately the focus point of said lens such that said beam is focused to a small point at the location of the pinhole thereby optically redefining the position of the beam source and wherein any modulation sidelobes or beam or external scattering are blocked by the pinhole screen thereby eliminating from said beam any sidelobes caused by said modulation or any scattering and wherein transmission of said beam through said lens causes said beam to expand to the proper size at said final focusing means.

14. The film recorder apparatus of claim 1 wherein said image plane converter means corrects for linearity errors in the scan process and geometry using said image plane converter.

15. The film recorder apparatus of claim 1 wherein said computation control means is operative to correct for spacing and linearity errors in said first dimensional direction.

16. The film recorder apparatus of claim 1 wherein said computation control means is operative to correct for spacing and linearity errors in said second dimensional direction.

17. The film recorder apparatus of claim 1 wherein said final focusing means comprises a lens.

18. The film recorder apparatus of claim 1 wherein said final focusing means comprises a holographic element.

19. The film recorder apparatus of claim 1 wherein said image plane converter means is a holographic element.

20. The film recorder apparatus of claim 1 wherein said light emitting means is a plurality of lasers.

21. The film recorder apparatus of claim 1 wherein said light emitting means is a plurality of light emitting diodes.

22. The film recorder apparatus of claim 1 wherein said light emitting means is an incandescent lamp.

23. The film recorder apparatus of claim 1 wherein said light emitting means is a florescent lamp.

24. The film recorder apparatus of claim 1 wherein said light emitting means is an arc lamp.

25. The film recorder apparatus of claim 1 wherein said light emitting means is a gas lamp.

26. The film recorder apparatus of claim 1 wherein said beam combining means

is at least one dichroic mirror.

27. The film recorder apparatus of claim 1 wherein said recording medium is film.

28. The film recorder apparatus of claim 1 wherein said recording medium is paper.

29. A film or paper recorder apparatus wherein a high quality image may be formed on a recording medium from an electronic image using a light beam of small spot size by sequentially scanning each of a plurality of color beams across the recording medium thereby permitting a high resolution image to be formed thereon, the recorder

apparatus comprising;

light emitting means for producing a light beam the generation and projection of said light beam forming a light path between said light emitting means and said recording medium said light means producing a beam which comprises a plurality of colors;

filtering means positioned within said light path for individually and sequentially selecting each of said colors;

focusing means positioned within said light path for focusing said light beam;

modulator means for modulating said light beam in response to said electronic image to be recorded, such that upon modulation, said beam comprise image data to be recorded;

scanning means positioned within said light path for scanning said single beam in two dimensions; computational control means for controlling said scanning and said modulation means and the flow of said image data; and

an image plane converter means positioned within said light path and having a front, source facing surface and an other recording medium facing surface such that said focused combined light beam incident on said converter source face is transmitted in a through said converter such that the image focus locus shape is converted to the shape of said other surface, producing a beam of nearly uniform spot size at said other surface regardless of the position of said spot on said converter source surface.

30. The film recorder apparatus of claim 29 wherein said scanning means comprises a frame scanning means positioned within said combined light path for scanning said single combined beam in a first dimensional direction and line scanning means positioned within said combined light path for scanning said single beam in a second dimensional direction, generally perpendicular to said first dimensional direction.

31. The film recorder apparatus of claim 29 further comprising an attenuator means positioned within said light path for adjusting the intensity of said light beams, filtering means positioned within said light path for more precisely selecting the desired frequency of light for each of said colors, and spatial filtering means positioned within said light path for filtering said light beams to the desired beam spot size.

32. The film recorder apparatus of claim 31 wherein said image plane converter means comprises a near-toric surface on the side toward the source and flat on the other side and is operative to correct focusing errors resulting from said scanning of said beam, and wherein said image plane converter means corrects for linearity errors in the scan process and/or geometry using said image plane converter.

33. The film recorder apparatus of claim 29 further comprising photodetector means positioned adjacent said recording plane and and electrically connected to said computational control means and operative to measure said light beams incident thereon

such that exposure parameters may be adjusted sufficiently to allow for registration and color balance and determine using the exposure beam on photodetectors.

34. The film recorder apparatus of claim 29 further comprising a beam expanding means situated within said combined beam path for expanding said beams focus spot by a predetermined amount and for varying the shape of the spot wherein said beam expanding means comprises a pair of translucent plates within said beam path and through which said beam passes and which may be pivoted relative to each other thereby varying the beam path length between the source and the recording medium and thereby expanding the size of said beam.

35. The film recorder apparatus of claim 30 wherein said line scanning means is a rotating polygon and further comprising a sensing means for determining that said rotating polygon line scanner is in position to transmit said image data to said recording medium and for communicating positioning status to said computational control means.

36. The film recorder apparatus of claim 29 further comprising spatial filtering means positioned within said light paths for filtering said light beams to the desired beam spot size wherein said spatial filtering means comprises at least one lens and an opaque

screen having a pinhole therethrough and wherein said pinhole is located at

approximately the focus point of said lens such that said beam is focused to a small point at the location of the pinhole thereby optically redefining the position of the beam source and wherein any modulation sidelobes or beam or external scattering is blocked by the pinhole screen thereby eliminating from said beam any sidelobes caused by said modulation or any scattering from and wherein transmission of said beam through said lens causes said beam to expand to the proper size at said final focusing means.

37. The film recorder apparatus of claim 29 wherein said light emitting means is a laser.

38. A film or paper recorder apparatus wherein a high quality image may be formed on a recording medium from an electronic image using a light beam of small spot size thereby permitting a high resolution image to be formed thereon, the recorder apparatus comprising;

light emitting means for producing a light beam, the generation and projection of said light beam forming a light path;

first mirror means situated within said path for separating said light beam into a plurality of light beams corresponding to each of a plurality of colors and forming three light

paths between said light emitting means and said recording medium;

focusing means positioned within said light paths for focusing said light beams;

beam combiner means positioned within said light path for combining said light beams comprising said plurality of colors, into a single beam the projection of said single beam forming a combined light path between said beam combiner means and said recording medium; scanning means positioned within said combined light path for scanning said single combined beam in two dimensions;

computational control means for controlling said scanning and said modulation means and the flow of said image data; and

an image plane converter means positioned within said light path and having a front, source facing surface and an other recording medium facing surface such that said focused combined light beam incident on said converter source face is transmitted in a substantially parallel fashion through said converter such that the image focus locus shape is converted to the shape of said other surface, producing a beam of nearly uniform spot size at said other surface regardless of the position of said spot on said converter source surface.

39. The film recorder apparatus of claim 38 wherein said scanning means comprises a frame scanning means positioned within said combined light path for scanning said single combined beam in a first dimensional direction and line scanning means positioned within said combined light path for scanning said single beam in a second dimensional direction, generally perpendicular to said first dimensional direction.

40. The film recorder apparatus of claim 39 further comprising an attenuator means positioned within said light paths for adjusting the intensity of said light beams, filtering means positioned within said light paths for more precisely selecting the desired frequency of light for each of said plurality of colors, and spatial filtering means positioned within said light paths for filtering said light beams and focusing said beams to the desired beam spot size.

41. The film recorder apparatus of claim 39 wherein said beam combiner means

is at least one dichroic mirror.

42. The film recorder apparatus of claim 40 wherein said image plane converter means comprises a near-toric surface on the side toward the source and flat on the other

side and is operative to correct focusing errors resulting from said scanning of said beam, and wherein said image plane converter means corrects for linearity errors in the scan process and/or geometry using said image plane converter.

43. The film recorder apparatus of claim 41 further comprising photodetector means positioned adjacent said recording plane and said line and frame scanning means and electrically connected to said computational means and operative to measure said light beams incident thereon such that exposure parameters may be adjusted sufficiently to allow for registration and color balance and determine using the exposure beams on photodetectors.

44. The film recorder apparatus of claim 39 further comprising a beam expanding means situated within said combined beam path for expanding said beams focus spot by a predetermined amount and for varying the shape of the spot wherein said beam expanding means comprises a pair of translucent plates within said beam path and through which said beam passes and which may be pivoted relative to each other thereby varying the path length between the source and the recording medium and thereby modifying expanding the size and symetry of said beam.

45. The film recorder apparatus of claim 44 wherein said line scanner is a rotating polygon and further comprising a sensing means for determining that said rotating polygon line scanner is in position to transmit said image data to said recording medium and for communicating positioning status to said computational control means.

46. The film recorder apparatus of claim 39 further comprising spatial filtering means positioned within said light paths for filtering said light beams to the desired beam spot size wherein said spatial filtering means comprises at least one lens and an opaque screen having a pinhole therethrough and wherein said pinhole is located at approximately the focus point of said lens such that said beam is focused to a small point at the location of the pinhole thereby optically redefining the position of the beam source and wherein any modulation sidelobes or scattering is blocked by the pinhole screen thereby eliminating from said beam any sidelobes caused by said modulation or any scattering and wherein transmission of said beam through said lens causes said beam to expand to the proper size at said final focusing means.

47. The film recorder apparatus of claim 39 wherein said light emitting means is a laser.

48. A film recorder apparatus wherein a high quality image may be formed on a recording medium from an electronic image using a light beam of small spot size thereby permitting a high resolution image to be formed thereon, the recorder apparatus comprising;

light emitting means for producing a light beam, the generation and projection of said light beam forming a light path between said light emitting means and said recording medium;

focusing means positioned within said light path for focusing said light beam;

scanning means positioned within said light path for scanning said beam in two dimensional directions;

modulator means for modulating said light beam in response to said electronic image to be recorded such that upon modulation, said beam comprises image data to be recorded;

computational control means for controlling said scanning and said modulating means and the flow of said image data; and

an image plane converter means positioned within said light path and having a front, source facing surface and an other recording medium facing surface such that said focused combined light beam incident on said converter source face is transmitted through said converter such that the image focus locus shape is converted to the shape of said other surface, producing a beam of nearly uniform spot size at said other surface regardless of the position of said spot on said converter source

surface.

49. The recording apparatus of claim 48 further comprising an attenuator means positioned within said light path for adjusting the intensity of said light beam, and spatial filtering means positioned within said light path for filtering and focusing said light beam to the desired beam spot size.

50. The recording apparatus of claim 49 further comprising frame scanning means positioned within said light path for scanning said beam in a first dimensional direction, and line scanning means positioned within said light path for scanning said single beam in a second dimensional direction, generally perpendicular to said first dimensional direction.

51. The film recorder apparatus of claim 49 further comprising spatial filtering means positioned within said light path for filtering said light beam to the desired beam spot size wherein said spatial filtering means comprises at least one lens and an opaque screen having a pinhole therethrough and wherein said pinhole is located at approximately the focus point of said lens such that said beam is focused to a small point at the location of the pinhole thereby optically redefining the position of the beam source and wherein any modulation sidelobes or scattering is blocked by the pinhole screen thereby eliminating from said beam any sidelobes caused by said modulation or any beam or external scattering and wherein transmission of said beam through said lens causes said beam to expand to the proper size at said final focusing means.

52. A film image input apparatus wherein a high quality image may be formed from an image on an existing medium to an electronic image using a light beam of small spot size thereby permitting a high resolution image to be sensed the image input

apparatus comprising;

light emitting means for producing a plurality of light beams comprising a plurality of colors, the generation and projection of said light beams forming a plurality of light paths between said light emitting means and said existing image medium;

beam combiner means positioned with said light path for combining said light beams comprising said plurality of colors, into a single beam;

scanning means positioned within said combined light path for scanning said single combined beam in two dimensional directions;

computational control means for controlling said scanning and acquisition of said image data; and

an image plane converter means positioned within said light path and having a front, source facing surface and an other back, existing image medium facing surface such that said focused combined light beam incident on said converter source face is transmitted through said converter such that the image focus locus shape is converted to the shape of said other surface, producing a beam of nearly uniform spot size at said other surface regardless of the position of said spot on said converter source surface.

53. The input apparatus of claim 52 further comprising a modulator means for modulating said light beams with said image to be sensed such that upon modulation, said beams comprise image correction or enhancement data as the light source for input scan detection.

54. The input apparatus of claim 52 further comprising an attenuator means positioned within said light paths for adjusting the intensity of said light beam, and spatial filtering means positioned within said light paths for filtering said light beams to the desired beam quality spot size.

55. The input apparatus of claim 54 further comprising frame scanning means positioned within said light path for scanning said beam in a first dimensional direction; and line scanning means positioned within said combined light path for scanning said single beam in a second dimensional direction, generally perpendicular to said first dimensional direction.

56. The input apparatus of claim 52 wherein said apparatus is operative to general a flying spot scanner for acquisition of existing color imagery.

57. The input apparatus of claim 52 further comprising a photodetector means located opposite of said existing image medium from said light emitting means and operative to sense the transmitted light intensity through said existing image medium and

transmit said intensity information to said computational control means.

58. The input apparatus of claim 52 further comprising a plurality of photodetectors and wherein each of said detectors is operative to respond separately to one of said plurality of colors.

59. A method of converting an image plane comprising:

providing a film or paper recording medium recorder apparatus wherein a high quality image may be formed on a recording medium from an electronic image using a light beam of small spot size thereby permitting a high resolution image to be formed thereon, the recorder apparatus having;

light emitting means for producing a plurality of light beams comprising a plurality of colors, the generation and projection of said light beams forming a plurality of light paths between said light emitting means and said

recording medium;

beam combining means positioned within said light path for combining said light beams comprising said plurality of colors, into a single beam the projection of said single beam forming a combined light path between said beam combining means and said recording medium,

scanning means positioned within said combined light path for scanning said single combined beam in two dimensions; computational control means for controlling said scanning and said modulation means and the flow of said image data; and

final focusing means positioned within said combined light path for focusing said

light beams;

providing an image plane converter means positioned within said light path and having a front, source facing surface and an other, recording medium facing surface such that said light incident on said converter source face is transmitted through said converter such that the image focus locus shape is converted to the shape of said

other surface, producing a locus shape of nearly uniform spot size at said other

surface regardless of the position of said locus on said converter;

outputting said image data from said control means;

modulating said light beams with said image data;

focusing said modulated beam to the desired spot size;

scanning said beam across said image converter source surface; and converting said image by transmitting said image through said converter such that the beam remains in focus throughout said converter and such that said beam remains in focus when it emerges from said other, recording medium facing surface thereby converting said scanned image into the format determined by said other, recording

medium facing surface.

60. The method of claim 59 wherein said plate is used to convert the image plane of a scanned image

61. The method of claim 59 wherein said plate is used to convert the image of a

two dimensional scanned image.

62. The method of claim 59 wherein said plate is used to convert the toric image plane of a scanned image where the radii of curvature of the two dimensions are different.

63. The method of claim 59 wherein said image is generated using laser light.

64. The method of claim 59 wherein said laser light image comprises a plurality of colors and is scanned simultaneously.

65. The method of claim 59 wherein the step of providing an image converter comprises the step of shaping the output face of said converter for compatibility with a lens for projection of said scanned image through said lens.

66. The method of claim 59 wherein the step of providing an image converter further comprises the step of shaping the output face of said converter to conform to a surface to accommodate recording an image.

67. An image recording electronic control apparatus for controlling the operation and data flow in a laser recorder apparatus wherein electronic image data is transferred from a computer system to a recording medium, wherein a high quality image may be formed on a recording medium to an electronic image using a light beam of small spot size thereby permitting a high resolution image to be formed thereon, the recorder apparatus having; light emitting means for producing a plurality of light beams comprising a plurality of colors, the generation and projection of said light beams forming a plurality of light paths between said light emitting means and said recording medium;

modulator means for modulating said light beams in response to said electronic image to be recorded, such that upon modulation, said beams comprise image data to be recorded; beam combining means positioned within said light path for combining said light beams comprising a plurality of colors, into a single beam, the projection of said single beam forming a combined light path between said beam combining means and said recording medium; scanning means positioned within said combined light path for scanning said single combined beam in two dimensions; and final focusing means positioned within said combined light path for focusing said light beams; an image plane converter means positioned within said light path and having a front, source facing surface and an other, recording medium facing surface such that said combined light beam incident on said converter source face is transmitted through said converter such that the image focus locus shape is converted to the shape of said recording surface, producing a beam of nearly uniform spot size at said recording surface regardless of the position of said spot on said converter; and electronic control means for controlling said scanning and said modulation means and the flow of said image data, the electronic control apparatus comprising: a communications interface electrically connected to said controller for transfer of said image data into said controller in an electronic format;

an input buffer for receiving and storing said image data for processing by said electronics controller;

a first look-up table for receiving said electronic input image data and looking up a corresponding contrast value;

a second look-up table for determining the line positing data and determining a corresponding line start delay time such that image data transmission is altered to account for errors in said scanning process;

a third look-up table for determining pixel clock timing; and

a digital frequency synthesizer for generating a variable frequency output as required by said pixel clock timing and electrically connected to said image data buffer for controlling the transmission of said pixel data to said modulator means such that said pixel image data transmission rate is adjusted according to the image data location within said image such that the image data is clocked at a rate which compensates for variations in pixel positions.

68. The electronics control apparatus of claim 67 further comprising a logarithmic amplifier means for converting and expanding said contrast data into a logarithmic scale.

69. The electronics control apparatus of claim 68 further comprising a sensor interface means for detecting when said scanner means is in position to receive image data electrically connected to said image buffer for controlling the output of image data.

70. An image input electronics control apparatus for controlling the input of image data from an existing image medium to an electronic image using an image input apparatus wherein a high quality image may be formed from an image on an existing

medium to an electronic image using a light beam of small spot size thereby permitting

a high resolution image to be formed, the image input apparatus having; light emitting means for producing a plurality of light beams comprising a plurality of colors, the generation and projection of said light beams forming a plurality of light paths between said light emitting means and said existing image; focusing means positioned within said light paths for focusing said light beams; beam combiner means positioned with said light path for combining said light beams comprising said plurality of colors, into a single beam; scanning means positioned within said combined light path for scanning said single combined beam in two dimensional directions; computational control means for controlling said scanning and the acquisition of said image data; a frame scanning means positioned within said light path for scanning said beam in a first dimensional direction; and line scanning means positioned within said combined light path for scanning said single beam in a second dimensional direction, generally perpendicular to said first dimensional direction; a plurality of photodetectors and wherein each of said detectors is operative to respond separately to one of said plurality of colors; and an image plane converter means positioned within said light path and having a front, source facing surface and an other, existing image medium facing surface such that said focused combined light beam incident on said converter source face is transmitted through said converter such that the image focus locus shape is converted to the shape of said other surface, producing a beam of nearly uniform spot size at said other surface regardless of the position of said spot on said converter source surface, input electronics control comprising:

a plurality of sensor drivers means electronically connected to said sensors for converting a sensed image density from said existing image medium into an analog voltage signal;

a plurality of logarithmic amplifier electronically connected to said driver means for

logarithmically amplifying said analog signal;

analog to digital converter means for converting said logarithmically amplified signal into digital data;

data buffers for storage of said converted image data for processing by said controller;

a look-up table for receiving said digital input image data and looking up a corresponding contrast value; and

a communications interface electrically connected to said controller for transfer of said image data into said controller in a digital input format.

71. A film or paper recorder apparatus wherein a high quality image may be formed on a recording medium from an electronic image using electromagnetic radiation

beam of small spot size thereby permitting a high resolution image to be formed thereon,

the recorder apparatus comprising;

radiation emitting means for producing a plurality of electromagnetic radiation beams comprising a plurality of frequencies the generation and projection of said beams forming a plurality of paths between said radiation emitting means and said recording medium;

modulator means for modulating said radiation beams in response to said electronic image to be recorded, such that upon modulation, said beams comprise image data to be recorded;

beam combining means positioned within said radiation path for combining said beams comprising said plurality of frequencies into a single beam the projection of said single beam forming a combined radiation path between said beam combining means and said recording medium;

computational control means for controlling said scanning and said modulator means and the flow of said image data; final focusing means positioned within said combined radiation light path for focusing said combined beam; and

an image plane converter means positioned within said light path and having a front, source facing surface and an other recording medium facing surface such that said focused combined radiation beam incident on said converter source face is transmitted through said converter such that the image focus locus shape is converted to the shape of said other surface, producing a beam of nearly uniform

spot size at said other surface regardless of the position of said spot on said

converter source surface.