Print engines

Abstract

A print engine receives image data elements relating to an image to be printed and generates print element control instructions for communication to a plurality of pens (e.g. inkjet pens) the control of the print engine. The print engine includes: a halftoning processor for receiving the image data elements and generating a halftone data set therefrom; and a plurality of print element controllers associated with said halftoning processor, each print element controller being adapted to generate print element control instructions from said halftone data set to drive different printing elements.

Description

TECHNICAL FIELD

The present invention relates to print engines. The invention has particular application in inkjet printers.

BACKGROUND ART

Conventional inkjet printers employ one or more printheads mounted on a carriage which is repeatedly scanned across a scan axis as the print medium is advanced stepwise past the scan axis. The printheads lay down swaths of ink during each scan, between advances of the print medium.

In a conventional high speed printer, each page of a print job is rasterised (converted from a computer output format such as Postscript™ or Portable Document Format™ to a bitmap of pixels) by a dedicated raster image processor (RIP).

The rasterised pages are buffered and then processed by an application specific integrated circuit (ASIC) which converts the bitmap into a halftone image composed of “halftone pixels” (which are not the same as print dots), and in each halftone pixel there is typically specified a number of ink dots of various colours, to give the appearance of a continuous tone image when printed. The most basic colour printer will use combinations of cyan, magenta and yellow (CMY) to make the various colours, or for increased quality true black ink may be also available (CMYK). For high quality images, two additional inks, light cyan and light magenta are also employed to provide increased fidelity particularly with lighter tones such as skin tones (CcMmYK printing).

Since there may be typically six inks available in specifying the halftone image, and each ink is printed from a separate pen on a printhead or from a separate printhead, each printhead has its own controller circuitry which analyses the halftone image and specifies the nozzle firing sequence to cause the printhead to lay down its ink at the correct point on the page, so that the pens in combination reproduce the halftone image.

A continuing goal of inkjet printing technology is to improve printing speeds, which are limited by a number of factors including the time taken to scan across the page, which may be increased in multiples where multiple pass print modes are employed (in which each area of the printed image is covered by a multiple number of swaths to improve print quality).

One way of reducing print times is to employ a page wide array (PWA) of printheads. In PWA printers, an array of printheads extending across the width of the page is maintained in a static position during printing and the medium advances under the PWA, eliminating scanning times. As the number of printheads increases however, the processing requirements of the printer similarly increase.

DISCLOSURE OF THE INVENTION

The invention provides a print engine for receiving image data elements relating to an image to be printed and generating therefrom print element control instructions for communication to a plurality of printing elements under the control of said print engine, said print engine comprising:

• • a halftoning processor for receiving said image data elements and generating a halftone data set therefrom; and
  • a plurality of print element controllers associated with said halftoning processor, each print element controller being adapted to generate print element control instructions from said halftone data set to drive different printing elements.

The term “print element” as used herein refers to an individual print-creating device such as an inkjet printhead or pen. The term also encompasses other types of print-creating components such as dot matrix printheads or indeed any other type of device which can be arranged in arrays whereby each device prints a portion of the image across the page. The printers of the present invention may be adapted to receive such printing components in an array, and so the printing components need not form part of the printer and can be replaceable. The means for receiving the array can be a simple mounting axis on which the components are fixed in position by suitable attachment means.

By associating two or more print element controllers with one halftoning processor, halftoning may be done once for a given portion of the image. Each of the print element controllers may access this halftone data and generate the necessary print element control instructions for the printing elements under its control from the same set of halftone data.

Preferably, said print engine further comprises a plurality of ink printing elements grouped together in a plurality of pen arrays such that all of the nozzles of a given pen array print the same colour of ink.

Further, preferably, said arrays are disposed in use such that at least two arrays are provided to print overlapping swaths of ink of the same colour during a single printing pass, and wherein both of said at least two arrays are under the control of the same print element controller.

This arrangement is particularly advantageous as it allows a single print element controller to carry out the masking for the ink in question and to divide the deposition of droplets for this colour along the swath between the two or more arrays (by generating firing instructions for the two arrays which combine to create the complete mask). By repeating this arrangement it is possible to ensure that all pens of the same colour which print the same portion of an image are controlled by the same print element controller. The controller responsible for each such group of pen arrays can access the halftoning data from a buffer, for example, and generate the necessary masking data and firing instructions once, feeding the relevant control instructions to each pen.

Further, preferably, the arrays are disposed in use such that all arrays which print the same ink onto the same portion of image are under the control of the same print element controller.

In a preferred embodiment, the print engine further comprises a memory for storing halftone data produced by said halftoning processor, wherein said memory is accessible by each of said print element controllers to retrieve the halftone data required by each print element controller.

In another preferred embodiment, the engine includes a plurality of memories, each associated with one of said print element controllers for storing halftone data produced by said halftoning processor, wherein each memory is accessible by the associated print element controllers to retrieve the halftone data required by that print element controller.

In this way, the relevant portion of data can be held locally at each print element controller to accelerate the accessing and processing of the halftone data to firing signals. The print element controller preferably accesses this data a number of times as it prepares firing instructions for a number of distinct groupings of printing elements, and therefore the use of local dedicated buffering allows the data processing rate to be increased.

The print engine may also include a communications link to which said halftone data is sent, each of said print element controllers being connected to said link.

Said link is preferably a bus and said print element controllers are preferably addressable via said bus.

Preferably, said halftone processor packetises said halftone data and addresses different packets of halftone data to different print element controller(s).

Preferably, each print element controller is adapted to retrieve halftone data from a store repeatedly and to generate therefrom on repeated retrievals print element control instructions for different sets of printing elements.

The invention also provides a printing system comprising a print engine as defined above, and further comprising a print medium advance mechanism for advancing a print medium past said print engine.

In another aspect the invention provides a wide array printing system comprising a plurality of print engines as defined above, each for printing from a set of printing elements which sets are disposed in an array across a width defining a print area, with each print engine responsible for halftoning and printing a portion of said print area.

Preferably, each of said print engines further comprises a plurality of printing elements arranged in a plurality of pens, whereby said pens combine to extend across the width of the print area.

Further, preferably, said pens include a pen grouping comprising at least two pens arranged to print the same ink at the same point across the width of the print area, said at least two pens being separated from one another in the direction perpendicular to the direction spanning the width, wherein both of said pens receive print element control instructions from the same print element controller.

Most preferably, in this embodiment, said pens include a plurality of said pen groupings, wherein each grouping is under the exclusive control of a single one of said print element controllers.

It should be noted that this arrangement does not exclude the possibility of having two or more groupings under the control of the same print element controller. What is important in relation to this preferred feature of the invention is that the bars of any one grouping are mutually controlled by one controller, so that the masking of a subset of halftone data can be prepared by that controller to fully define the firing instructions for the nozzles of the grouping.

In another aspect the invention provides a scanning printing system comprising a print engine as defined above which includes a plurality of printing element sets, each set being under the control of one of said print element controllers, wherein said sets are disposed on a carriage for scanning across a print medium while printing thereon.

Preferably, each set of printing elements exclusively prints one or more inks, such that the print element control instructions for printing each ink are generated by a single print element controller.

It should be appreciated that each print element controller may generate print masks and firing instructions (e.g. in an inkjet system) for the nozzles of a number of different coloured inks, but that this preferred features specifies that for any given ink (e.g. cyan), the entire print mask is specified within one controller, even if the cyan print mask is shared between two pens, such that both pens will be controlled by the same controller.

In a particularly preferred embodiment of a scanning printer, each print element controller is responsible for controlling a set of nozzles divided into four pens. Most preferably, a first print element controller is responsible for controlling a set of pens printing three primary colour inks and a black ink, and a second print element controller is responsible for controlling a set of pens printing light tones of two of said primary colours, a pigmented black ink and a fixer.

This system is particularly advantageous as it enables a family of modular components to be used. For example, where a basic printer might have a halftoning processor and a set of printhead pens printing cyan, magenta, yellow and black inks, under the control of a print element controller, the present invention can utilise the same components and the associated software in the print element controller and add a second print element controller with a set of printheads printing light cyan, light magenta, pigmented black and a fixer.

In another aspect the invention provides a method of generating print element control instructions from a set of image data elements representative of an image, said method comprising the steps of:

• • preparing halftone data from said image data elements;
  • making said halftone data available to a plurality of print element controllers;
  • with each of said print element controllers, preparing print element control instructions from said halftone data set using a print mask effective to cause a set of print elements under the control of said print element controller to print said print mask.

The method of the invention preferably also involves the step of printing said mask with said print elements.

Preferably, each different part of said image is set of one or more colour planes of the image in an image swath.

In another aspect the invention provides a computer program comprising instructions in machine readable form which when executed within a printing system including a halftoning processor and a plurality of print element controllers connected to said processor are effective to cause the system to carry out a method as defined above.

The invention further provides a computer program carrier comprising an encoded computer program as defined above.

The carrier for the computer program may, for example, be an optical or magnetic carrier (such as a floppy diskette, a compact disk or a digital versatile disk), an electrical signal, an electromagnetic signal, or an electronic circuit encoding the program logic.

BRIEF DESCRIPTION OF DRAWINGS

The invention will now be illustrated by the following descriptions of embodiments thereof given by way of example only with reference to the accompanying drawings, in which:

FIG. 1 is a block diagram of the architecture of a first printing system according to an embodiment of the invention;

FIG. 2 is a representation of an image file sent to the printer for printing;

FIG. 3 is a representation of a rasterised version of the image of FIG. 2;

FIG. 4 shows the slices of the rasterised image of FIG. 3 as sent for processing to the three individual halftoning processors of FIG. 1;

FIG. 5 is an enlarged view of the slices illustrating the order of processing of the raster bitmap pixels by the three processors using a matrix algorithm;

FIG. 6 is an enlarged view of the slices illustrating the order of processing of the raster bitmap pixels by the three processors using a forward error diffusion algorithm;

FIG. 7 is an enlarged view of the slices illustrating the order of processing of the raster bitmap pixels by the three processors using a first serpentine forward error diffusion algorithm;

FIG. 8 is an enlarged view of the slices illustrating the order of processing of the raster bitmap pixels by the three processors using a second serpentine forward error diffusion algorithm;

FIG. 9 is a diagram showing the relationship between the print element controllers of the system of FIG. 1 and the printbars under their control;

FIG. 10 is a diagram illustrating the scalability of the architecture of FIG. 1; and

FIG. 11 is a block diagram of the architecture of a scanning printing system according to the invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

FIG. 1 shows the main components of a printer according to an embodiment of the invention, which receives print jobs from a computer directly connected to the printer via a parallel port 10. The same printer could be used to print from a local area network connection or from a wide area network connection, e.g. over the Internet.

Print jobs are received by the printer in any one of a number of supported formats, such as in the form of Postscript™ files (although the skilled person will be aware that the nature of file received as a print job can vary according to prevailing standards and the use to which the printer is to be put).

A raster imaging processor (RIP) 12 receives the incoming print jobs and converts them to a rasterised bitmap, in known manner. The output of the RIP in the embodiment being described is a contone image file. The print job is processed on a page-by-page basis with each page being rasterised in turn and forwarded via a bridge 14 to a RAM buffer 16 which stores multiple pages to allow print jobs to be sent quickly to the ASICs for further processing and avoid delays due to real time processing of the images which might otherwise occur.

The bridge provides a link between the RIP or the RAM buffer to a common I/O bus which connects to three parallel halftoning ASICs 18,20,22. Each ASIC is essentially identical and comprises a dedicated halftoning processor which is designed to convert a contone bitmap input to a halftone output specified in the ink colours provided in the printer. In the illustrated embodiment the printer is a CMYK printer and so the halftone image will specify each pixel of the halftone image as a combination of dots of one or more of the four CMYK colours. In addition to the cyan, magenta, yellow and black inks, there are printheads for printing a fixative (F) before the ink is laid down and a post-printing treatment T after the ink is laid down. The fixative and any other pre-printing or post-printing treatment, can of course be substituted with inks of other colours.

Each ASIC 18,20,22 is dedicated to processing a particular parallel slice of the image, where the image is sliced into three equal strips. The strip boundaries run along the direction of the page advance, so that for any particular page printed by the PWA of printheads (described below) each ASIC is responsible for specifying the dots printed in one third of the page.

Each ASIC outputs halftone data corresponding to the strip of image which it processes. This halftone data set is converted into pen firing instructions by a corresponding set of thirty-three printhead controllers arranged in three groups 30,32,34. Since the same halftone data for a given strip can be used by all of the eleven printhead controllers in a group, the halftone data for each strip is held in a buffer 36, 38, 40. Each of the printhead controllers can access the halftone data in the buffer and carry out the necessary masking operations to generate firing instructions for the printing pens (indicated generally at 42) under their control. The printhead controllers are ASICs which analyse the halftone data for the slice under their control and generate from the data a set of firing instructions for the pens covering that slice (i.e. converting the halftone data into ink drops to be printed by its associated printheads). In this way, each halftoning ASIC can drive multiple printhead controllers.

Thus, in the illustrated embodiment has three print engines. The first print engine comprises halftoning ASIC 18, buffer 36, printhead controllers 30, and the group of pens 44 indicated in slice 1. The second print engine comprises halftoning ASIC 20, buffer 38, printhead controllers 32, and 26a-h, the group of pens 46 indicated in slice 2, and the third print engine comprises halftoning ASIC 22, buffer 40, printhead controllers 34, and the group of pens 48 indicated in slice 3.

In a variation on this embodiment, the common buffers 36,38,40 are omitted, and the printhead controllers are connected to the halftone processor in daisychain configuration, i.e. with a common bus carrying the halftone output and each printhead controller accessing the bus to grab data required by that printhead. The printhead controllers can be addressable, and the halftone data packetised, so that the halftone data relating to the printheads under the control of any given controller can be addressed to that controller. Each controller has its own dedicated buffer to store the halftone data addressed to it or required by it.

In this variation, the halftone data packets are sent once along the common bus and stored in the buffer(s) of the controller(s) which require access to that data. Then each printhead controller can repeatedly access the data, apply the print masks in turn for each printhead it controls to that data and supply the firing instructions to the respective printheads.

This alternative arrangement can be advantageous from a signal processing and network traffic viewpoint, since the data is held locally in each controller's buffer and is accessed locally by the controller without other network traffic interference. This arrangement also has advantages for scalability as will be discussed in more detail below.

Before describing in detail the configuration of the pens and the interaction between the printhead controllers and the pens, the initial stages of the printing process will be described.

FIG. 2 shows a simple image sent as a print job to the printer. The image 50 is received by the RIP 12 and is rasterised to provide a pixelated image 52 (FIG. 3) of differently coloured pixels 54,56. (The image shown in FIGS. 2 and 3 is black and white for simplicity but it will be appreciated that a coloured image will give rise to a large number of different colours of pixels). The rasterised image is then optionally buffered and fed to the halftoning processors 18,20,22, where it is split into three parallel slices 58,60,62, as shown in FIG. 4.

In practice the splitting occurs as the halftoning ASICs capture the image file from the RAM buffer through the I/O bus. The halftoning processors are configured to pick out the pixels in the particular slices assigned to each halftoning ASIC. This may be achieved simply by reading the appropriate DRAM addresses of the buffer. More preferably, each ASIC will determine that for a page of a given width W (W being the number of pixels), it is responsible for W/3 pixels per line. The first ASIC reads the pixels “numbered” from 1 to W/3 in each line. The second ASIC reads the pixels from (W/3+1) to 2W/3, and the third ASIC reads the pixels from (2W/3+1) to W in each line.

By this means the rasterised contone image is divided in three for further processing, and each halftone ASIC 18,20,22 operates on its own set of pixels to generate the halftone data set for that slice of the page, e.g. by using a conventional halftoning algorithm.

FIG. 5 illustrates a parallel process for processing the contone bitmap. In the example of FIG. 5, the halftoning ASICS operate on the pixels using a matrix algorithm which contains no forward diffusion of error terms. Thus, processor 18 will operate on its contone bitmap in the pixel order 1a,2a,3a,4a, . . . without any reference to the other processors. Similarly processor 20 operates on pixels 1b,2b,3b,4b, . . . and processor 22 on pixels 1c,2c,3c,4c, . . . , etc.

As the halftone images are generated they are sent to the printhead controllers 30,32,34. The printhead controllers then generate firing instructions.

In FIG. 6 a more sophisticated halftoning algorithm is shown which relies on the forward diffusion of error terms in the contone bitmap. The algorithm used is a linear algorithm, and operates across the three slices in turn. Thus, after the first of the processors 18 has processed pixels 1a,2a, . . . ,7a,8a, it delivers the error term to processor 20, at this point processor 18 can start processing the next line 9a,10a, . . . since there are no further dependencies. Processor 20 carries forward the error term from pixel 8a to process pixels 1b,2b, . . . ,7b,8b before handing over in turn to processor 22. Thus processor 18 is one line ahead of processor 20 which in turn is one line ahead of processor 22 and so the 3 processors work in pipeline fashion. Once the pipeline has been filled up the three processors work in parallel. The mechanism for handling over the error term can be achieved either through the same IO bus or a dedicate link between processors.

FIG. 7 shows a variation on this algorithm where instead of a linear left-to-right processing order, the direction of processing is reversed for each row in a serpentine fashion. Thus, while the first row is processed in the order described above for FIG. 6, processor 22 begins processing the second row with the pixel shown as 25 in FIG. 7. This serpentine error diffusion algorithm can achieve higher quality output but it creates dependencies in both directions. Thus processor 18 after it has processed pixel 8 is stalled until it receives the error term for pixel 41 from processor 20. In a similar way processors 20 and 22 are also stalled. Thus this serpentine error diffusion algorithm offers no parallelism and at any point in time only one processor is working while the others are stalled.

FIG. 8 shows a particularly preferred embodiment which overcomes this difficulty by halftoning the individual slices of the page longitudinally as opposed to transversely. Thus, processor 18 performs a serpentine algorithm down and up the columns as opposed to across the rows of pixels, in the order 1a,2a,3a, . . . etc. Once it has finished with the slice, processor 18 delivers a full row of error terms to processor 20. At this point processor 18 can start processing the left-hand slice of the next page of the print job and processor 20 is then free to process its slice in the order 1b,2b,3b,4b, . . . etc. When processor 20 finishes processing the centre slice of the first page it delivers the row of error terms to processor 22.

At this point processor 18 will have finished processing the left-hand slice of page 2 and given the error terms for page 2 to processor 20, which can then start processing the centre slice of page 2 (allowing processor 18 to start processing page 3). Processor 20 similarly hands its row of error terms from page 1 to processor 22, allowing processor 22 to begin processing the right-hand slice of the first page. Thus processor 18 is one page ahead of processor 20 which in turn is one page ahead of processor 22, and the 3 processors work in pipeline fashion. Once the pipeline has been filled up the three processors work in parallel.

The following table illustrates the order of processing of the various slices on successive pages of an n-page print job. In this table, a timeslot is the length of time required for a processor to receive the error terms associated with the pixels of a slice, then to process the slice, and finally to hand over the error terms from that slice to the next processor (where applicable). Thus, from timeslot 3 onwards, each of the processors processes a slice in parallel with the other processors but with processor 18 leading processors 20 and 22 by 1 and 2 pages respectively.

	Processor 18	Processor 20	Processor 22
Timeslot #	(Left Hand Slice)	(Centre Slice)	(Right Hand Slice)
1	Page 1	Idle	Idle
2	Page 2	Page 1	Idle
3	Page 3	Page 2	Page 1
4	Page 4	Page 3	Page 2
5	Page 5	Page 4	Page 3
. . .	. . .	. . .	. . .
. . .	. . .	. . .	. . .
n − 1	Page n − 1	Page n − 2	Page n − 3
n	Page n	Page n − 1	Page n − 2
n + 1	Idle	Page n	Page n − 1
n + 2	Idle	Idle	Page n

In this way the sophistication of a forward error diffusion algorithm is utilised whilst reducing the extent to which each processor is dependent on the results of the others.

When the halftone data has been generated by the respective halftoning processors, it is retrieved from the buffers by the various printhead controllers responsible for generating printing firing instructions for the pens in that slice (or more particularly, for the nozzles of the pens in that slice). It can be seen from FIG. 1 that 44 separate pens are provided to print each slice of a printed page. In this embodiment, each printhead controller is capable of generating nozzle firing instructions for four separate pens. (In this context, a pen is an array of nozzles provided on a die). There is no reason why a printhead controller could not be responsible for generating the firing instructions for a smaller or greater number of pens. This arrangement is scaleable by replicating the printhead controller/printhead modules. The scalability can be employed to increase speed, image quality, printhead redundancy, etc. The ratio of printheads per controller can be varied, or the width of slice for which a controller is responsible can be varied.

FIG. 9 shows the group of pens responsible for printing a particular slice, and the relationship between these pens and the printhead controllers. While the physical relationship between the pens is reasonably accurate, in that the width of the strip being printed is fully covered by a printbar comprising four pens, and in that 11 different printbars are provided parallel to one another (such that any given point on the print medium passes in turn under each printbar), the disposition of the printhead controllers relative to the printbars is immaterial, and only the connections between the controllers and the individual pens are of relevance to the discussion which follows.

Each pen is identified by a letter and a numeral. The letter indicates the colour ink being printed according to the following code: T=post-printing treatment (this can be of any suitable type); Y=yellow; M=magenta; C=cyan; K=Black; F=fixer (note that the fixer is laid down before the inks, and the last printbar to print on the page is the post-printing treatment T. The numeral indicates the printhead controller which is responsible for generating the firing instructions for the nozzles of the pen in question.

For illustrative purposes, the connections between printhead controller nos. 1, 2, 9 and 11, and the pens under the control of these printhead controllers are shown. The other connections have been omitted for clarity.

The top printbar 70 (as seen in FIG. 9) is the only printbar containing pens printing the post-printing treatment (labelled “T”). All 4 of these pens are commonly controlled by printhead controller 1. However, each of the other printbars 72, 74, 76, 78, 80, 82, 84, 86, 88, 90 includes pens controlled by at least two separate printhead controllers. The reason for this is that each of the other printbars (with the exception of printbar 90) is composed of pens which are mutually redundant with the pens in another printbar. For example, the left-most pen of printbar 72 which is denoted as “Y 2” and the left-most pen of printbar 80 (also “Y 2”) are mutually redundant in that they each print the same ink onto the same spot on a print medium passing below the array of printbars. Such mutually redundant pens are all controlled by the same printhead controller. It's can therefore be seen that printhead controller 2 controls the two yellow pens Y 2 in each of the printbars 72 and 80. The yellow pens Y 3 which complete these two printbars are controlled by a separate printhead controller, namely printhead controller 3. Printhead controller 2 can therefore perform the masking and the print element control instructions generation for the yellow ink in the left half of the strip and printhead controller 3 performs the same task for the right hand of the strip. Each printhead controller thus operates on only the halftone data for the portion of strip for which it is responsible.

If each printhead controller stores the halftone data in a buffer of its own it can be repeatedly accessed to drive different printheads as different print masks are applied to the data. There is no need to re-ask a common buffer for halftone data, which improves the bandwidth of the printing pipeline. In this regard, it is particularly useful to store in a single buffer the halftone data commonly required by a number of different printing elements printing the same ink on the same point on the page (as the page moves past the respective printing elements. In practice, this is advantageously achieved by using a common printhead controller for a number of same colour, same position printheads, and storing the halftone data for these printheads is a single buffer exclusively accessed by that controller.

In the case of the black printbars 78, 82, 88 four separate printhead controllers (numbers 8, 9, 10 and 11) are required, since each set of three mutually redundant pens (for example, the pens labelled “K 9”) should preferably be under the control of a single printhead controller to obtain maximum benefit from the invention. Since each printhead controller can control four pens, each of these printhead controllers also controls a fixative pen (F 8, F 9, F 10, F 11).

It will be recalled that only a single set of halftone data is required for all of the printhead controllers shown in FIG. 9. All of the printhead controllers access the same data in a buffer and perform their respective tasks on the portion of the data corresponding to the pens under their control.

FIG. 10 illustrates how this concept is scalable. When a page is divided into N slices (N greater than or equal to 1), the contone data arriving over a PCI bus 91 is divided between a number of halftone processors 92, 94, 96 (in practice, each halftone processor will select the data relevant to its slice, or the data for each slice will be addressed to the relevant halftone processor).

Each halftone processor produces a halftone data set and addresses the packets of halftone data to the printhead controllers responsible for printing a particular colour on a particular section of the page wide arrays. The addressed packets travel from each halftoner 92,94,96 along an associated bus or communications link 93,95,97 from each of which a number of printhead controllers 98 are connected.

Each printhead controller 98 has an internal buffer (not shown) into which the halftone data in the packets are loaded for processing by a masking processor (not shown). Thus, the page image is first split into slices and each slice is processed in parallel and sent out along a bus once. Each printhead controller 98 stores the data which it requires for the pens under its control. In the illustrated embodiment, each controller 98 has four pens 100 under its control. Connected between any printhead controller and each of its pens is a pen power supply unit 102 (optional) which in addition to supplying power also performs ink short protection and provides an interface with the pen circuitry.

To generate firing signals for each of the four pens, a printhead controller processor accesses the locally buffered halftone data and applies a first print mask to generate firing signals for a first of the pens 100. Next, the data is again retrieved from the buffer and a second print mask applied for the second pen. The data is retrieved and processed again and again until the requisite firing signals have been sent to all of the pens under the control of the printhead controller, following which the buffer can be cleared. In order that all of the pens fire their ink droplets in synchronicity, each pen may have a small buffer for the firing signals or the printhead controller may have an output buffer storing the firing signals until these are sent to the pens to fire the nozzles.

It can be seen from FIG. 10 that the number of pens used to print a slice can be scaled up (assuming a limited number of pens per printhead controller) by dedicating additional printhead controllers to each slice. Dividing the page into more slices can increase printing speed.

If a printhead controller can process a certain amount of data this can control a wide band of printing area without any redundancy or the same amount of data can be used to control printheads covering a narrower band with greater redundancy.

Many of the concepts described above can also be applied to a scanning printer. In particular, the set of pens and printhead controllers shown in FIG. 9 could, in a perpendicular orientation, be employed in a scanning printer. By mounting the pens on a scanning carriage, a four pen swath can be printed with the multiple redundant pens shown as the array of printbars is scanned back and forth across a print medium (with the medium being advanced between traverses of the carriage).

Even without multiple redundant pens, the concepts of the present invention can be employed in a scanning printer. FIG. 11 shows a schematic architecture for such a printer. A halftoning processor 110 provides a halftone data set to a bus 111 accessed by a pair of printhead controllers 112, 114. Bach printhead controller has four pens under its control as previously described. Shown in this embodiment is a separate dedicated buffer 113,115 for each printhead controller.

Printhead controller 112 controls a first black pen 116, a cyan pen 118, a magenta pen 120, and a yellow pen 122. This printhead controller and associated set of pens can be taken from a simpler four-pen printer. Printhead controller 114 generates nozzle-firing instructions for a second black pen (which may be of a different type of ink e.g. pigmented black if the first ink is dye-based black ink) 124, a light cyan pen 126, a light magenta pen 128, and a fixer pen 130.

To fire the pens, each controller 112,114 buffers the halftone data for its colours as previously discussed and then accesses this data four times in succession applying print masks for the four colours to output four sets of print element control instructions (i.e. nozzle firing instructions), one for each printhead.

As a variation on this ink choice, one could double up a normal four colour printer having a tri-colour ink pen and a black pen to provide two tri-colour pens and two black pens, so that whether printing in colour mode or monochrome, the printing speed is doubled.

As in the page-wide system previously described, the scanning print engine of FIG. 11 takes advantage of re-using a single set of halftone data elements in two different print element controllers.

The invention is not limited to the embodiments described herein which may be varied without departing from the spirit of the invention.

Claims

1. A print engine for receiving image data elements relating to an image to be printed and generating therefrom print element control instructions for communication to a plurality of printing elements under the control of said print engine, said print engine comprising:

a halftoning processor for receiving said image data elements and generating a halftone data set therefrom; and

a plurality of print element controllers associated with said halftoning processor, each print element controller being adapted to generate print element control instructions from said halftone data set to drive different printing elements.

2. A print engine as claimed in claim 1, further comprising a plurality of ink printing elements grouped together in a plurality of pen arrays such that all of the nozzles of a given pen array print the same colour of ink.

3. A print engine as claimed in claim 2, wherein said arrays are disposed in use such that at least two arrays are provided to print overlapping swaths of ink of the same colour during a single printing pass, and wherein both of said at least two arrays are under the control of the same print element controller.

4. A print engine as claimed in claim 3, wherein the arrays are disposed in use such that all arrays which print the same ink onto the same portion of image are under the control of the same print element controller.

5. A print engine as claimed in claim 1, further comprising a memory for storing halftone data produced by said halftoning processor, wherein said memory is accessible by each of said print element controllers to retrieve the halftone data required by each print element controller.

6. A print engine as claimed in claim 5, wherein said memory is a buffer for temporarily storing halftone data produced by the halftoning processor.

7. A print engine as claimed in claim 1, further comprising a plurality of memories, each associated with one of said print element controllers for storing halftone data produced by said halftoning processor, wherein each memory is accessible by the associated print element controllers to retrieve the halftone data required by that print element controller.

8. A print engine as claimed in claim 1, further comprising a communications link to which said halftone data is sent, each of said print element controllers being connected to said link.

9. A print engine as claimed in claim 8, wherein said link is a bus and said print element controllers are addressable via said bus.

10. A print engine as claimed in claim 9, wherein said halftone processor packetises said halftone data and addresses different packets of halftone data to different print element controller(s).

11. A print engine as claimed in claim 1, wherein each print element controller is adapted to retrieve halftone data from a store repeatedly and to generate therefrom on repeated retrievals print element control instructions for different sets of printing elements.

12. A printing system comprising a print engine as claimed in claim 1, further comprising a print medium advance mechanism for advancing a print medium past said print engine.

13. A wide array printing system comprising a plurality of print engines as claimed in claim 1, said print engines being disposed in an array across a width defining a print area, with each print engine responsible for halftoning and printing a portion of said print area.

14. A wide array printing system as claimed in claim 13, wherein each of said print engines further comprises a plurality of printing elements arranged in a plurality of pens, whereby said pens combine to extend across the width of the print area.

15. A wide array printing system as claimed in claim 14, wherein said pens include a pen grouping comprising at least two pens arranged to print the same ink at the same point across the width of the print area, said at least two pens being separated from one another in the direction perpendicular to the direction spanning the width, wherein both of said pens receive print element control instructions from the same print element controller.

16. A wide array printing system as claimed in claim 15, wherein said pens include a plurality of said pen groupings, wherein each grouping is under the exclusive control of a single one of said print element controllers.

17. A scanning printing system comprising a print engine as claimed in claim 1 and further comprising a plurality of printing element sets, each set being under the control of one of said print element controllers, wherein said sets are disposed on a carriage for scanning across a print medium while printing thereon.

18. A scanning printing system as claimed in claim 17, wherein each set of printing elements exclusively prints one or more inks, such that the print element control instructions for printing each ink are generated by a single print element controller.

19. A scanning printing system as claimed in claim 17, wherein each print element controller is responsible for controlling a set of nozzles divided into four pens.

20. A scanning printing system as claimed in claim 17, wherein a first print element controller is responsible for controlling a set of pens printing three primary colour inks and a black ink, and a second print element controller is responsible for controlling a set of pens printing light tones of two of said primary colours, a pigmented black ink and a fixer.

21. A method of generating print element control instructions from a set of image data elements representative of an image, said method comprising:

preparing halftone data from said image data elements;

making said halftone data available to a plurality of print element controllers;

with each of said print element controllers, preparing print element control instructions from said halftone data set using a print mask effective to cause a set of print elements under the control of said print element controller to print said print mask.

22. A method as claimed in claim 21, further comprising printing with said print elements.

23. A method as claimed in claim 22, wherein each different part of said image is set of one or more colour planes of the image in an image swath.

24. A computer program comprising instructions in machine readable form which when executed within a printing system including a halftoning processor and a plurality of print element controllers connected to said processor are effective to cause the system to carry out the method of claim 21.

25. A computer program carrier comprising an encoded computer program as claimed in claim 24.