GB2025730A - Forming halftone images - Google Patents

Abstract

A system for forming halftone images from an original comprises a read sub-assembly and a record sub- assembly. The read sub-assembly includes c.r.t. scanner 22 and circuit 54, 56 and 58 to generate and store digital tonal density values of, for example, four adjacent points DN, DE, DS, DW in each discrete picture cell in an original 16. Comparison unit 60 generates a 'high detail flag' signal if there is a large difference or ratio between the density values in one cell and if so all the (four) density values are written in memory 68 for transmission to the record sub- assembly. If there is not a large difference or ratio, a single weighted (average) density value is computed and stored. The stored values are transmitted to the record sub-assembly which selects a dot of the required size and shape from a memory in accordance with the single density value or with the average of the (four) values and generates a visible image thereof on a CRT. If the flag signal is present a comparator logic unit generates shift signals in accordance with the difference in tonal density across the cell horizontally or vertically in order to displace the dot accordingly within the cell. The store 68 may be omitted. <IMAGE>

Description

SPECIFICATION Forming halftone images The printing techniques commonly used in the graphic arts industry, i.e., newspapers, books, etc., utilize an dll-or-nothing inking procedure; i.e., ink is either deposited at each point on the printing surface or it is not.

While this technique poses no problems when it is desired to print text, the printing of pictures, such as photographs, introduces the problem of printing of pictures, such as photographs, introduces the problem of printing the continuous tones in the grey scale ranging from black to white. The problem is generally solved by transforming the continuous tones of the original image into halftones. Halftone images are typically produced by a large number of inked dots whose size or spacing relative to one another tricks the eye into perceiving shades of gray rather than the individual dots. This technique works best when the largest dots and the spacing between the dots is small compared with the visual accuity of the eye. Early halftone generating systems used either variable sized dots with uniform spacing or uniform sized dots with variable spacing.

Electronic phototypesetting systems developed over the years have greatly increased the speed of type composition. Many such systems incorporate a halfton e generating capability and in addition to storing typeface characters in a font memory also store halftone dot characters representative of discreete levels of the grey scale. In order to produce halftone image, these systems scan the original image, sample the gray scale value at discrete intervals and convert these values to multibit binary numbers which are then used to access the stored dot character representative of that value from the dot character font memory. U. S. Patent No. 3,806,641 to Crooks discloses a system of this type.

The data processing capabilities of such electronic phototypesetting systems have made possible a drastic reduction in the time necessary to prepare plates for printing material comprised of both text and illustrations. However, the resolution of halftone images produced by such sampled data systems generally falls short of that obtainable when non-digital techniques are employed.

According to the invention, there is provided a photocomposing apparatus for producing a halftone replica of an original image, comprising means for scanning said original image to produce a digital representation of the tonal density of each of a plurality of points of said image, means for comparing the digital representations of adjacent points in an area defining a cell to produce a flag signal when the difference in or ratio of the tonal densities of the point in said cell exceeds a predetermined threshold, arithmetic means for producing a single digital representations when the threshold is not exceeded, dot character storage means for storing representations of a plurality of different halftone dot characters, means responsive to the absence of said flag signal for selecting one of said dot character representations in accordance with said digital weighted value, and means responsive to the presence of said flag signal for selecting one of said dot character representations in accordance with the digital tonal density representations of the points in the cell.

The present invention is directed to a system for extracting digital information from original images storing that information and/or transmitting the information to a remote location, and then utilizing the information to produce a halftone replica of the original image. The invention is particularly useful in a photocomposing system in which digital information representing a large number of pictures is stored in a mass memory such as magnetic tape or disk and used together with other information to compose pages of text and pictures.

A photocomposing system in accordance with the invention is capable of producing halftone replicas of higher apparent resolution for a given mesh value than previously known systems utilizing a pure sampled data approach. Alternatively, a photocomposing system in accordance with the invention can produce halftone replicas of given resolution with lower data storage requirements than previously known systems.

In a preferred embodiment of the present invention, the image to be reproduced is scanned by methods well-known in the art to obtain digital representations of the tonal values of incremental areas thereof.

However, unlike prior systems, the image is scanned at a greater number of horizontal and vertical coordinate points than is desired in the halftone replica. For example, by doubling the number of horizontal and vertical sampling points, four samples rather than one are taken for each discrete area or dot cell interrogated by the scanner. The four samples are then compared with one another to determine the rate of change of image density within the area or dot cell. If a threshold value is not exceeded, the cell is deemed a low detail cell. If, on the other hand, the threshold value is exceeded, the cell is deemed a high detail cell.

,A photocomposing embodiment of the invention is comprised of an image read subassembly which extracts tonal density data from the original image and an image record subassembly which responds to the density data to produce the halftone replica. In the preferred embodiment, the image read subassembly determines whether a cell is of high or low detail. In the case of a low detail cell, the four individual samples are averaged to develop a single density value for the cell which is then placed in a memory, e.g. magnetic tape. In the case of a high detail cell, all four of the cell samples are recorded on the magnetic tape.

The image record subassembly subsequently responds to the information on the magnetic tape to select a single do character frm a front memory which best represents the original four individual samples. In the case of a low detail cell, the stored single density value is employed to select the dot character of appropriate size from the font memory. In the case of a high detail cell, the four samples are processed to select the dot character of appropriate size and shape from the font memory. In addition, the four samples are used to migrate, i.e., alter the position in the cell at which the dot character is placed.The position can be either the "centerof gravity" of the four, in terms of average density, or some other function relating the desired displacement from cell center (AX, AY) to the magnitude and position of the four samples. Experimentation has shown that good results are obtained by calculating AX and AY as follows: AX = K1L (DE - Dw) AY = K2L (DN - Ds) where L equals the distance from the center of the cell to the sample locations; DN, Dw, DE, Dw equal the image densities of four samples oriented at 90" to one another within the cell (e.g., north, south, east, west) and K1 and K2 are constants.

An embodiment of the invention will now be described by way of example and with reference to the accompanying drawings, in which: Figure 1 is a block diagram of an image read subassembly in accordance with the invention; Figure 2a schematically illustrates the division of an original image into a matrix of dot cells; Figure2b schematically illustrates a single dot cell area showing sampling points therein; Figure 2c-2f schematically illustrate exemplary halftone dot characters, varying in size, shape and position which are formed to represent the indicated dot cells in the original image; and Figure 3 is a block diagram of an image record subassembly in accordance with the invention for recording a halftone replica on film.

Halftone replicas of original images are formed by the juxtaposition of discrete dot characters which may seemingly be of arbitrary shape, such as small circular dots or line segments. Multiple dot characters when juxtaposed together can be configured to give an impression of continuous gray sacle tones comprising a halftone image.

Digital memories have been used to store the shapes of such discrete dot characters. More particularly, a font memory is typically provided in a halftone recording system for storing the shapes of various dot characters generally selected so as to cover the entire gray scale from white to black in selected steps. Such dot characters have been utilized in prior art systems, and examples are disclosed in various United States patents, such as 3,806,641 to Crooks, and 3,922,484 to Keller. The present invention similarly makes use of dot characters of various sizes and shapes which are stored in and retrieved from a dot character font memory.

As will be more fully explained hereinafter, the dot characters are preferably arranged in the font memory in order of increasing or decreasing image density. This arrangement offers the simplest access to the dot characters because the apparatus used to scan the original image converts tonal densitto a multibit binary number can be used as a direct address to gain access to the dot character font memory in order to call out the appropriate dot character.

Any halftone generating system must begin by extracting information from the original image to be reproduced. Similarly to prior art systems, the present apparatus extracts the required tonal density information from the original image by a scanning process. Prior to considering the scanning apparatus depicted in Figure 1, attention is initially directed to Figure 2a which illustrates a pattern comprised of imaginary lines running at 450 to the vertical and intersecting one another to form a matrix of cells 12. The cells 12 may be considered as being arranged in rows and columns such that any particular cell falls within a particular row and column. For example, cell 14 is in column 9 and row 3.Although the cell matrix or mesh depicted in Figure 2a is formed by lines extending at 45 to the vertical, it should be understood that other cell matrices can also be employed, as is readily known in the art. The significance of the cell matrix depicted in Figure 2a is that the original image to be ultimately reproduced by a halftone replica is first scanned to determine its tonal density within each of the discrete cell areas. More particularly, as will be discussed in greater detail hereinafter, the image read apparatus of Figure 1 scans an original image which may be borne on a transparency 16 by successively sampling the cells in accordance with the pattern represented by the matrix depicted in Figure 2a.Subsequently, the image record subassembly depicted in Figure 3 produces a halftone replica of the original image by recording an appropriate dot character in each cell.

It should be apparent that as the number of dot cells (Figure 2a) into which the original image is divided is increased (increasing the mesh value as it is called in graphic arts), the resolution of the halftone replica to be produced will also be increased. In typical prior art systems, in order to increase the resolution of the halftone replica, it is necessary to commensurately increase the amount of data which must be stored and processed. That is, in order to double the resolution of the halftone replica in accordance with prior art systems, the amount of data which must be stored and processed is quadrupled since the number of points scanned in the original image must be doubled in both the horizontal and vertical dimensions. Moreover, to do so may result in a halftone replica which is not readily printable.

In accordance with a significant feature of the present apparatus, each dot cell 12 of the original image is sampled at multiple points and the resulting multiple samples are processed to yield a single dot character which more accurately represents the density of the original image within that cell area than prior art systems.

More particularly, attention is directed to Figure 2b which represents a single cell 12 which is sampled ate multiple points. In accordance with the exemplary embodiment disclosed herein, it will be assumed that each dot cell 12 is sampled at four points displaced from one another by substantially 90 . The four points will be respectively referred to herein as north, east, south, and west. The terms DN, DE, DS, and Dw will be used hereinafterto respectively represent the tonal densities of the north, east, south and west points of a dot cell area and additionally these terms may sometimes be used to designate the particular multibit binary number representative of that tonal density.Whereas prior art systems typically sample each dot cell at a single point, the present system samples each dot cell at multiple points, e.g. the four points illustrated in Figure 2B. As will be seen hereafter in discussing the apparatus of Figures 1 and 3, the data representing the tonal densities of those four points of the original image in each cell is examined to determine whether the dot cell comprises a high or low detail portion of the image. The determination of whether a cell comprises a high or low detail portion of the image is based upon comparing the four density values to determine the rate of change of image density within the cell. This determination can be made by, for example, calculating the ratios or differences between the density values.If the rate of change exceeds a certain threshold, then the date corresponding to that cell is tagged with a high detail "flag" nd the four density values are recorded and preserved for use in developing the halftone replica. On the other hand, is a cell is determined to be a low detail cell, then the four density values are processed to derive a single value which, in the simplest embodiment, comprises the arithmetic average of the four density values. The data derived during the image read process is in most applications of the invention recorded on a memory such as a magnetic tape which is later used by the image record subassembly of Figure 3 to produce the halftone replica.As will be seen hereinafter, in producing the halftone replica, the single density value associated with a high detail cell are processed in the image record subassembly to derive an address to access a dot character of appropriate size and shape from the dot character memory. Additionally, the four density values associated with a high detail cell are used to determine whether the dot character to be displayed in a cell in the halftone replica should be migrated within the cell. That is, depending upon the relative magnitudes of the four density values within a cell, the dot character displayed within that cell can be displaced horizontally (AX) and vertically (AY) from the center of the cell.

The halftone image generating system essentially operates in two modes; namely (1) an imput mode during which the image read apparatus of Figure 1 scans an original image and extracts density data therefrom and (2) an output mode during which the image record apparatus of Figure 3 responds to the density data to produce a halftone replica. The input mode will be considered first and reference will initially be made to Figure 1 which illustrates a photographic transparency 16, bearing the original image to be reproduced. The transparency 16 is placed in front of a scanner 22 which preferably comprises a cathode ray tube 24 capable of producing a spot of light 26 for scanning the transparency 16. The scanning spot 26 is deflected in a manner readily understood in the art under the control of horizontal and vertical deflection coils 28 and 30.The coils 28 and 30 are respectively driven by drive circuits 32 and 34, each of which includes a digital to analog converter and amplifier circuits. The scanning spot 26 is focused by a lends 36 onto the transparency 16. The light penetrating through the transparency 16 is focused by lens 38 onto a photodetector 40 such as a photomultiplier tube which develops an analog signal related to the magnitude of the light passing through the transparency 16 which in turn is determined by the density of the image carried by the transparency. It should be recognized that although it has been assumed the original image is carried on a transparency 16, other types of scanners can be used for scanning opaque film of the like in which the reflected light from the original image causes a photodetector to develop the image signal.Prior to discussing the processing of the image signal produced by photodetector 40, the manner in which the scanning spot 26 is controlled to derive the image signal will be explained.

It will be recalled from Figure 2a that the original image is to be scanned by successively examining the density of the image in each of a multiplicity of discrete cells 12. In order to cause the spot 26 to move from cell to cell, appropriate deflection signals must be applied to coils 28 and 30. Moreover, within each cell 12, the spot 26 must be successively deflected to the north, east, sout and west points depicted in Figure 2b. on manner of sequentially generating the appropriate deflection signals for application to the coils 28 and 30 is to utilize a readonly memory 44 which, in sequential locations, stores digital information representing th deflection signals required to move the spot 26 successively to each cell and within each cell to each point.

Thus, the locations of the read only memory 44 can be sequentially accessed in response to an address counter 46. The address counter 46 is enabled by application of a start signal 48 and is then controlled by a timing and control circuit 50. That is, the timing and control circuit 50 will periodically supply a clock pulse to increment the address counter 46. A different location of the read only memory 44 will be accessed to supply digital deflection signals to the drive circuits 32 and 34 to cause the spot 26 to sequentially scan the cells depicted in Figure 2a and the points depicted in Figure 2b.

The output of the photodetector 40 is applied to a sample and hold circuit 54 which is also under the control of the timing and control circuit 50. Thus, each time the spot 26 is positioned at a new point, the sample and hold circuit 54 samples the output of the photodetector 40 to store an analog signal which is then applied to an analog to digital converter 56. The output of the analog to digital converter 56 is in turn coupled to storage registers 58 capable of storing at least four multi-bit numbers, DN, D, DS, DW, respectively representative of the tonal density of the north, east, south and west points of a cell 12.The converter 56 and registers 58 are similarly controlled by the timing and control circuit 50 which enables the registers 58 to accumulate the digital density values of the point of a single cell.

An arithmetic/comparison logic unit 60 is provided to determine, based upon the values DN, DE, DS, and Dw for a cell, whether the portion of the original image within that cell is comprised of high detail or low detail.

This determination is made by comparing the magnitudes of the values DN, DE, D5, and Dw to determine the rate of change of density within the cell area. This determination of rate of change of density within a cell area can be performed in a variety of manners such as determining the ratio of differences between points within a cell. In accordance with one preferred technique, the digital density samples DN, DE, DS, DW are summed to develop a signal Z. The cell is deemed a high detail cell if (DN - ) or (DE - Z) or (D5 - ) or (Dw - ) is > Twhere z equals K (DN + DE + DS + DW) and T defines a threshold constant.Other criteria can be used to determine whether a cell comprises a high detail cell. For example only, the difference between the highest and lowest valued samples can be compared with some threashold; i.e. D maximum - D minimum > T.

From the foregoing, it should be apparent that a cell will be deemed a high detail cell if the density value of any of the four points within the cell is greater than the average of the four points by some threshold determined by the value T. It will of course be readily recognized that the arithmetic capability to make the determination in accordance with the foregoing equations is rather simple and that accordingly the arithmetic/comparison logic unit 60 can be readily implemented either by special purpose logic circuitry or by a properly programmed computer.In any event, if the logic unit 60 determines that the cell is a high detail cell, it generates a high detail "flag" signal on output line 62 followed by the digital density values of the cell DN, DE, Ds, Dw. On the other hand, in the event the logic unit 60 determines that the cell is a low-detail cell, then unit 60 merely outputs a signal digital value L: which, as has been pointed out, is related to the sum of the density values DN, DE, DS, DW for that cell.

The output of the logic unit 60 is applied to a memory write circuit 66 which then records the data in memory 68 which preferably comprises a magnetic tape or magnetic disk drive.

In a typical embodiment of the invention, the single digital value T representing the density of a cell is represented by eight bits, meaning that 256 density or gray scale levels for that cell can be represented.

Inasmuch as in accordance with the preferred embodiment, three of those eight-bit combinations are used as special flags, the eight bits representing Z define 253 different density or gray scale levels. The three reserved eight-bit combinations can be used as follows: code 0 indicates no dot in this cell, code 255 indicates the end of the current row or column of cells, and code 254 indicates a high detail dot cell has been encountered. Thus, the code 254 comprises a high detail "flag" which indicates that the data following represents the digital density values of the four sample points within a cell. In a preferred embodiment, each of the digital density values DN, DE, DS, DW iS represents by six bits. Thus, a high detail cell is represented by only eight bits.However, since high detail cells will occur very infrequently, i.e., only at a sharp edge or transition in the original image, the amount of additional data required to be stored in memory 68 to represent the original image is not much greater than in prior art systems in which only a single point was sampled in each cell. For example, if ten percent of the dot cells in a given image are "high detail", then the total data is increased by only thirty percent over the same image data without the image enhancement yielded by the present invention.

Prior to proceeding to an explanation of the image record subassembly of Figure 3 for executing the output mode, attention is directed to Figures 2c-2f which illustrate typical halftone dot characters to be printed on the replica in response to particular original image cells. Consider initially the original image high detail cell of Figure 2c in which the north and south points are considerably more dense than the east and west points.

In order to best represent this original image cell by a signal halftone dot character, a dot character having a shape elongated in the vertical direction as shown is used. Similarly, in Figure 2d a high detail original image cell is represented in which the east and west points have a density considerably greater than the north and south points. The halftone dot character utilized to best represent this original image cell comprises a dot elongated in the horizontal direction. Figure 2e illustrates an original image cell in which the south, west and north points are of considerably greater density than the east point. The halftone dot character used to represent this original cell configuration comprises a large dot displaced along the horizontal axis of the replica cell.Figure 2f illustrates an original image dot cell in which the west and north points are of considerably greater density than the east and south points. A halftone dot character, smaller in size than the dot character of Figure 2e, is shown displaced both vertically and horizontally from the center of the replica cell.

Thus, as should now be apparent from Figures 2c-2f, the size, shape, and position of the halftone dot character within a cell of the replica can be selected to best represent a cell of the original image. As will be seen hereinafter, a character font memory is preferably used to store the size and shape of each dot character. This character font memory is similar to that disclosed in the aforecited Crooks patent 3,806,641.

Although the character font memory can be as large as desired, thus able to store any desired number of different dot character shapes, it will be assumed herein that the character font memory to be discussed in connection with Figure 3 stores characters of three different shapes, as represented in Figures 2c, 2d, and 2e or 2f, with each shape being stored in 253 different sizes. As will be seen hereinafter, the particular dot character to be accessed from the font memory to represent a cell in the halftone replica is determined in part by the previously-mentioned value Z associated with the cell.That is, the size of the dot character used in the replica to represent a cell is determined by the density of the corresponding cell in the original image which of course is dependent upon the sum of the density values DN, DE, DS, DW for that cell. As will be seen hereinafter, the shape of the dot character to be accessed from the font memory to represent a cell in the replica is determined based upon a comparison of the density values representative of the four points within a cell. Similarly, the displacement from the center of the replica cell is determined based upon a comparison of the density values of the four points within a cell.

Attention is now directed to the image record subassembly of Figure 3 which executes the previously-mentioned output mode to create a halftone replica which is recorded on film or some other photosensitive medium 100. The image record subassembly develops the halftone replica based upon the data read from memory 68 which, as previously noted, may comprise a magnetic tape or disk. The data stored in memory 68 is read by memory read circuit 102. In the case of a low detail cell, the memory read circuit 102 will read the value E: can be used as an address to access an address from primary memory 108 which stores the starting address of a related dot character stored in font memory 110. Font memory 110 stores the information required to form and display each different dot character.It will be recalled that it has been assumed that the preferred embodiment is capable of storing dot characters of three different shapes, each shape being available in 253 different sizes. The information necessary to form each of these dot characters (3 x 253) is contained in the font memory 110. The information necessary to form a dot character may be contained within a block of addresses within the font memory 110 and the primary address memory 108 contains the starting address for each block of addresses within memory 110. Thus, in the case of a low detail cell in which the value Z is directly applied to the addressing circuit 106, a starting address will be read out of memory 108 into memory control circuit 112.This starting address then is utilized to read the information from the character font memory 110 required to form that character in a display device such as a cathode ray tube 114. In a preferred embodiment of the invention, the character will be formed on the face of the cathode ray tube 114 by scanning the spot 116 in a regular pattern and by blanking and unblanking the spot based upon the information read from the character font memory. The output of the character font memory 110 is applied to blanking circuit 118 which modulates the intensity of the spot 116. The spot 116 is caused to scan within a particular cell area by signals provided by a cell scanning circuit 120 which supplies analog signals to horizontal and vertical mixer circuits 122 and 124.The outputs of the mixer circuits 122 and 124 are applied to horizontal and vertical deflection amplifiers 126 and 128 whose outputs in turn are connected to deflection coils 130 and 132. A lens 134 focuses the spot 116 on the film 110 which is driven past the face of the cathode ray tube by a motor 140.

The spot 116 not only scanned within each cell area in response to signals generated by cell scanning circuit 120, but is also of course moved from cell to cell corresponding to the matrix depicted in Figure 2a. In order to do this, a scan format read only memory 144, similar to the scan format memory 44 discussed in connection with Figure 1 is provided. The scan format read only memory 144 is driven by an address counter 146 which is enabled by a start signal 148. The address counter is under the control of a timing and control circuit 150. The timing and control circuit 150 similarly controls the other elements of the subassembly of Figure 3 in a manner well-known in the art to of course synchronize the positioning of the spot 116 within a particular cell with the character information being read from the memory 110 to control the intensity of the spot within the cell.More particularly, in response to a start signal 148, the address counter 146 is enabled and is incremented to cause the read-only memory 144 to generate X and Y positioning signals which are applied to horizontal and vertical digital to analog converters 152 and 154. The outputs of the converters 152 and 154 are applied to mixer circuits 122 and 124 and therethrough to deflection amplifiers 126 and 128.

Thus, the X and Y deflection signals supplied by the read-only memory 144 determine the cell position of the spot 116. The scan signals provided by the circuit 120 cause the spot 116 to scan within the cell. As previously pointed out, in the case of a low detail cell read from the memory 68, the value Z can be directly used as an address by addressing circuit 106 to access an address from the primary address memory 108 to thus cause the font memory to supply the appropriate information to the blanking circuit 118 to draw dot characters of various size on the face of the cathode ray tube 114.

In the case of a high detail dot cell read from memory 68, a flag detect circuit 160 will respond to the high detail flag and disable gate 104 via inverter 162. Further, flag detect circuit 160 will enable registers 164 causing the digital density values DN, DE, DS, Dwfollowing the high detail flag to be read into registers 164.

Further, the flag detect circuit 160 will enable arithmetic comparator logic unit 166 which will then operate upon the stored density values DN, DE, DS, DW to develop size, shape, and displacement information. More particularly, logic unit 166 will develop dot character size information by calculating the value S in accordance with Z = K (DN + DE + DS + DW) The value E developed by the logic unit 166 is applied via output line 170 to the input of gate 172 which is enabled by the output of th flag detect circuit 160. The output of gate 172 is applied to the input of addressing circuit 106.

In addition to developing the size information Z, the logic unit 166 examines the density values DN, DE, DS, Dw to determine whether a shape different from a circle should be used. That is, in accordance with a preferred embodiment, the logic unit 166 will examine the individual density values to determine whether the dot character should have a shape as represented in Figure 2c or 2d, for example. In an embodiment of the invention, this determination is made by examining the following relationships: DN + Ds > TorD+ 2 rDE+DW > T 2 The first of the two cases represented by the foregoing relationships is represented in Figure 2c.If the relationship is satisfied, that is, if DN approximately equals Ds and if DN + DS iS substantially greater than DE + Dw, then it is desired to select the vertically oriented halftone dot characters shown in Figure 2c. If, on the other hand, the second relationship is satisfied, that is, if DE approximately equals Dw and DE + DW iS much greater than DN + DS, it is desired that a halftone dot character as represented in Figure 2d be created.

Since it has been assumed that the system is capable of producing three different character shapes, it is sufficient for the shape to be represented by two bits. After the logic unit 166 determines based upon the examination of the values DN, DE, DS, DW as to which dot character shape should be presented, it supplies those two bits on output line 174 to the input of gate 176. Gate 176 is enabled by the flag detect circuit 160.

The output of gate 176 is applied to addressing circuit 106. Thus, it should now be apparent that when a high detail cell is detected by flag detect circuit 160, the logic unit 166 operates upon the succeeding density values DN, DE, DS, Dw to determine the shape and size of the dot character to be drawn by the cathode ray tube 114. This determination results in the generation of bits on output lines 170 and 174 which are then used as an address by addressing circuit 106 to access the information from font memory 110 to draw the dot character of selected size and shape on the cathode ray tube.

In addition to determining the size and shape of the dot character to be presented by the cathode ray tube 114, the logic unit 166 determines whether the dot character should be displaced within the cell. This situation is represented in Figures 2e and 2f wherein the dot characters are shown as being displaced horizontally (AX) and vertically (AY) within the cell in the replica. In accordance with a preferred embodiment, the dot character is migrated or displaced within a cell, in accordance with the following formulations: AX=(DE- Dw)andAY=k4(D-Ds) where L represents the distance from the center of a cell to the center of one of said points and K3, K4 represent a constant.

The logic unit 166 provides digital representations of AX and AY on output lines 180 and 182 respectively.

These lines are coupled to the less significant bit input stages of the digital to analog converters 152 and 154 to thus effectively reposition the center of that particular cell to cause any character to be displayed to be offset or migrated from the predetermined cell position.

In accordance with one further aspect of the invention, an operator controlled translation table lookup means 190 may be optionally provided. The table lookup means 190, when provided, is coupled to the memory control circuit 112 for modifying, under operator control, the address provided by the primary address memory 108 to access the font memory 110. The purpose of the translation table lookup means 190 is to enable an operator to enhance shadow or highlight effects in the halftone replica produced on the film 100 and to essentially linearize the system from original image to the apparent effect of the produced halftone replica.More particularly, it will be recalled that the dot characters are stored in font memory (or at least the starting addresses stored in memory 108) in sequence of increasing or decreasing density so that the value of E can be directly used as an address by the addressing circuit 106. This means that ideally, each value of 2: will be quantized into one of 253 levels, in accordance with a preferred embodiment, and that one of 253 levels will be used to access a corresponding one of 253 differently-sized dot characters. In order to highlight certain effects or enhance certain types of image subject matter, it is not always desired that the relationship between the E: value and the dot character size be the same.That is, under some circumstances, it is desirable to effectively superimpose a desired non-linear equalization curve upon the values of Z. For example, it may be desired that a s value of 28 indeed access a dot character of size 28 but that a value of 32 access some dot character of size different than size 32. The translation table lookup means 190 enables an operator to effectively introduce such an equalization curve. The translation table lookup means operates by responding to a primary address read out of memory 108 to provide a different address for accessing the font memory 110.From the foregoing, it should now be apparent that a photocomposing system has been disclosed herein having a halftone generating capability in which considerably enhanced halftone replicas can be developed requiring only a relatively small increase in data storage. The objectives of the invention are achieved by sampling the original image at a resolution greater than that desired forthe halftone replica.

Thus, in accordance with the preferred embodiment, each dot cell is sampled at four points rathr than one point. Based upon these four samples, the image read subassembly determines whether the cell is a high or low detail cell. If the cell is a low detail cell, then the four samples are combined and the cell data is processed essentially as in prior art systems. If on the otherhand the cell is a high detail cell, then the four individual points are recorded in memory and the image record subassembly utilizes those four sample values to select a particular size and shape of halftone character to be displayed, as well as determining the amount of migration or displacement of the halftone character within the cell. It should be understood that although a specific embodiment of the invention has been disclosed herein, variations within the scope of the invention will readily occur to those skilled in the art. For example only, although it has been assumed that each cell is sampled at four individual points, the invention is not restricted to any particular number of points and indeed a greater number of sample points within a cell can be utilized. Further, although exemplary criteria for determining whether a cell is a high or low detail cell have been set forth, different criteria can be utilized.

Thus, for example, ratios of density values within a cell, rather than differences, can be utilized to determine whether a cell is high detail. Similarly, different but related criteria can be utilized to determine the shape of the halftone character to be selected by the logic unit 166. Further, although a special-purpose hardwired embodiment of the invention has been disclosed herein, it should also be recognized that an equivalent embodiment of the invention can be readily implemented by a programmed general purpose computer.

Although particular embodiments of the invention have been described and illustrated herein, it is recognizedthatmodifications and variations may readily occur to those skilled in the art, and consequently, it is intended that the claims be interpreted to cover such modifications and equivalents. The invention extends to the image read subassembly and image record sub-assembly separately and each including a memory 68 and ta the two sub-assemblies coupled together without the memory 68 whereby a halftone replica of an original image may be produced within storing a digital representation of the image.

Claims (14)

1. Photocomposing apparatus for producing a halftone replica of an original image, comprising means for scanning said original image to produce a digital representation of the tonal density of each of a plurality of points of said image, means for comparing the digital representations of adjacent points in an area defining a cell to produce a flag signal when the difference in or ratio of the tonal densities of the points in said cell exceeds a predetermined threshold, arithmetic means for producing a signal digital weighted tonal density value for said cell from the digital representations when the threshold is not exceeded, dot character storage means for storing representations of a plurality of different halftone dot characters, means responsive to the absence of said flag signal for selecting one of said dot character representations in accordance with said digital weighted value, and means responsive to the presence of said flag signal for selecting one of said dot character representations in accordance with the digital tonal density representations of the points in the cell.

2. Apparatus as claimed in claim 1 further comprising output means for producing a visible representation of a selected dot character at the centre of a cell area when said flag signal is absent or at a position displaced from said centre by an amount dependent on the digital tonal density representations when the flag signal is present.

3. Photocomposing apparatus for producing a halftone replica of an original image comprising in combination: means for scanning said original image to produce a digital representation of the tonal density of each of a plurality of points of said image; comparison means for comparing digital representations of adjacent points within an area forming a cell for generating a flag signal when the differences in tonal density between said points exceed a certain threshold; arithmetic means responsive to the digital representations of points within a cell for producing a signal digital weighted value for that cell; dot character storage means for storing representations of a plurality of different halftone dot characters; means responsive to each digital weighted value for selecting one of said halftone dot character representations from said dot character storage means; output means responsive to each halftone dot character representation selected for producing a visible represention of that dot character position at substantially the center of a cell area; and means responsive to the generation of a flag signal associated with a cell for altering the horizontal and vertical positions of the visible representation produced in the area of that cell.

4. Photocomposing apparatus as claimed in claim 2 or 3 wherein each of said cells contains four points successively displaced by 90" from one another and wherein the tonal densities of said four points are respectively represented by DN, DE, DS, DW and wherein said means for altering the position of the visible representation includes means for altering the horizontal position by AX proportional to L (DE - Dw) and the vertical position by AY proportional to L (DN - DS) where L represents the distance from the center of a cell to the ceter of one of said points.

5. Photocomposing apparatus as claimed in claim 1,2,3 or 4 wherein said arithmetic means includes means for arithmetically averaging the digital representations of points within a cell to produce said signal digital weighted value.

6. Photocomposing apparatus as claimed in any preceding claim wherein said dot character storage means stores representations of dot characters of different size and shape.

7. Photocomposing apparatus as claimed in any preceding claim wherein each of said cells contains four points successively displaced by 90" from one another and wherein the tonal densities of said four points are respectively represented by DN, DE, DS, DW and wherein said comparison means generates said flag signal if the tonal density of any one of said four points differs by more than a threshold value from the tonal density of any other one of said points.

8. Apparatus as claimed in any of claims 1 to 6 wherein each of said cells contains four points and wherein said comparision means generates said flag signal if the tonal density of any one of said four points differs by more than a threshold value from the average tonal density of the four points.

9. Photocomposing apparatus as claimed in any preceding claim wherein each of said cells contains four points and wherein the tonal densities of said four points are respectively represented by DN, DE, DS, DW and wherein said arithmetic means for producing said weighted value includes means responsive to DN + Ds orDc + Dw where = K(DN + DE+ D5 + Dw).

10. Photocomposing apparatus as claimed in any of claims 1 to 8 wherein each of said cells contains four points and wherein the tonal densities of said four points are respectively represented by DN, Dc, Ds, Dw and wherein the means for selecting the dot character representation in the presence of said flag signal includes means responsive to DN + Ds > > z or DE + DW where z where z = K (DN F DE + Ds + DW)

11.A method for producing a halftone replica of an original image comprising the steps of: scanning discrete areas of said original image to derive a digital representation of the tonal value density of each of said areas; comparing said digital representations of tonal value density of adjacent ones of said areas to ascertain the relationship between tonal value densities thereof; selecting from a plurality of dot characters particular ones of said plurality of dot characters in accordance with the values of said digital representations; displaying each of said selected dot characters on an output medium at a position dependent upon said relationships.

12. Apparatus for producing a digital representation of an original image for use in a method as claimed in claim 11, comprising means for scanning said original image to produce a digital representation of the tonal density of each of a plurality of points of said image, means for comparing the digital representations of adjacent points in an area defining a cell to produce a flag signal when the difference in or ratio of the tonal densities of the points in said cell exceeds a predetermined threshold, arithmetic means for producing a signal digital weighted tonal density value for said cell from the digital representations when the threshold is not exceeded, and means for storing for each cell the flag signal and digital tonal density representations or the digital weighted tonal density value.

13. Apparatus for producing a halftone image for use in a method as claimed in claim 11, comprising means for providing for each discrete cell of the image a flag signal and digital tonal density representations of points in the cell or a digital tonal density value for the cell, dot character storage means for storing representations of a plurality of different halftone dot characters, means responsive to the absence of said flag signal for selecting one of said dot character representations in accordance with said digital weighted value, and means responsive to the presence of said flag signal for selecting one of said dot character representations in accordance with the digital tonal density representations of the points in the cell.

14. Apparatus substantially as herein before described with reference to Figure 1 and/or Figure 3 ofthe accompanying drawings.