This invention pertains to the field of inkjet printing systems, and more particularly to a method for reducing banding artifacts associated with bi-directional multi-pass printing on an inkjet printer.
A typical inkjet printer reproduces an image by ejecting small drops of ink from a printhead containing ink nozzles, where the ink drops land on a receiver medium (typically paper) to form ink dots. Inkjet printers typically reproduce color images by using a set of color inks, usually cyan, magenta, yellow, and black. It is well known in the field of inkjet printing that if ink drops placed at neighboring locations on the page are printed at the same time, then the ink drops tend to flow together on the surface of the page before they soak into the page. This can give the reproduced image an undesirable grainy or noisy appearance often referred to as “coalescence”. It is known that the amount of coalescence present in the printed image is related to the amount of time that elapses between printing adjacent dots. As the time delay between printing adjacent dots increases, the amount of coalescence decreases, thereby improving the image quality. There are many techniques present in the prior art that describe methods of increasing the time delay between printing adjacent dots using techniques referred to as “interlacing”, “print masking”, or “multi-pass printing”. These methods often involve advancing the paper by an increment less than the printhead width for each printing pass. As a result, successive passes or “swaths” of the printhead overlap, which has the additional advantage that it can help to reduce one-dimensional periodic artifacts referred to as “bands” or “banding” that can result due to clogged or misdirected ink nozzles. See, for example, U.S. Pat. Nos. 4,967,203 and 5,992,962. The term “print masking” generically means printing subsets of the image pixels in multiple partially overlapping passes of the printhead relative to a receiver medium.
Another attribute of modern inkjet printers is that they typically possess the ability to vary (over some range) the amount of each ink that is deposited at a given location on the page. Inkjet printers with this capability are referred to as “multitone” inkjet printers because they can produce multiple density tones at each location on the page. Some multitone inkjet printers achieve this by varying the volume of the ink drop produced by the nozzle by changing the electrical signals sent to the nozzle or by varying the diameter of the nozzle. See for example U.S. Pat. No. 4,746,935. Other multitone inkjet printers produce a variable number of smaller, fixed size droplets that are ejected by the nozzle, all of which are intended to merge together and land at the same location on the page. See for example U.S. Pat. No. 5,416,612. These techniques permits the printer to vary the size or optical density of a given ink dot, which produces a range of density levels at each location, thereby improving the image quality.
Another common way for a multitone inkjet printer to achieve multiple density levels is to print a small amount of ink at a given location on several different passes of the printhead over that location. This results in the ability to produce a greater number of density levels than the nozzle can fundamentally eject, due to the build up of ink at the given location over several passes. See, for example, U.S. Pat. No. 5,923,349.
In U.S. Pat. No. 5,790,150, Lidke et al. disclose a method where multiple passes are made over the page while fractionally advancing the page. In each pass, the pattern of dots in the data swath is constructed with sufficient spacing between the dots such that the printhead can be scanned across the page at a velocity that is higher than the firing frequency limit of the nozzles.
In U.S. Pat. No. 6,310,640, Askeland discloses a print masking method in which nozzles at the ends of the printhead print with lower duty than nozzles near the center of the printhead, thereby reducing the possibility of banding artifacts occurring at the boundaries between successive printed swaths.
In U.S. Pat. No. 6,206,502 Kato et al. also discloses a method for reducing the duty for nozzles at the ends of the printhead. This method involves using a page advance which is smaller than the number of nozzles in the printhead divided by the number of passes, so that there is a region at the ends of the printhead where the passes overlap for an additional pass. The goal of this is to hide artifacts that can result at the boundaries of the printhead due to page advance errors, etc. Vinals et al disclose a similar method in U.S. Pat. No. 6,375,307 for a single-pass printing configuration. These print masking methods are sometimes referred to as “fractional print masking” in the literature.
In U.S. Pat. No. 6,238,037, Overall et al. disclose a print masking method for a multilevel inkjet printer in which the print mask contains a set of threshold values. A dot will print at a given location on a given pass if the multitone code value for that pixel is greater than the threshold for that pass. This method requires that if a dot gets printed at a given pixel on pass N, then it also must receive dots on passes 0 through N−1.
In U.S. Pat. No. 6,454,389, Couwenhoven et al. disclose a print masking method suitable for multilevel inkjet printers that can produce multiple sized ink drops.
In U.S. Patent Application Publication No. 2007/0201054, which is incorporated herein by reference, Billow et al. disclose a print masking that utilizes a print mask having a plurality of mask planes, each mask plane corresponding to a multitone code value. This approach has the advantage that dot patterns printed in response to different multitone levels can be independent from each other.
The method of Billow et al will now be described in more detail to illustrate print masking. Turning to
A swath data generator 30 then receives the multitoned image signal i and generates a swath data signal s, which controls the volume of ink printed by an inkjet printhead (or printheads) 40. The process of print masking is contained within the swath data generator 30. Prior to multitoning, each pixel contains a numeric code value (typically on the range {0,255}) for each color channel that indicates the amount of the corresponding colorant to be placed at the given pixel's location in the image. After multitoning (at the output of the image preprocessor 20), the image is represented by multitone code values, where the range of pixel code values has been reduced to match the number of density levels that the inkjet printer can produce. For binary inkjet printers, the possible multitone code values will be either 0 or 1, indicating whether to print 0 or 1 drops of ink. Multitone inkjet printers will accept multitone code values on the range {0,N−1}, where N is the number of possible multitone code values, and is normally the number of density levels (or number of drops) that the multitone inkjet printer can produce at a given pixel.
Turning now to
Often, the mask height Mh will be equal to the number of nozzles in the printhead, although this is not a fundamental restriction, and a mask height of lesser or greater value can be used. One of the mask planes is selected for a given pixel according to the multitone code value of the multitoned image signal i, as shown in
xm=x % Mw (1)
ym=y % Mh (2)
where x is the pixel column number and y is the pixel row number of the current pixel being processed, Mw is the mask width, Mh is the mask height, and the “%” symbol indicates the mathematical modulo operator. The value of the swath data signal s is then determined by selecting a mask element 62 from the chosen mask plane according to:
s=MaskPlane(i,xm,ym) (3)
Turning now to
Turning now to
Often inkjet printers are configured to print in a bi-directional print mode, where ink is applied as the printhead moves in both rightward and leftward directions. A common problem with inkjet printers that utilize bi-directional multi-pass printing is that they can be susceptible to banding artifacts caused by differences in the order of ink laydown and the timing between ink laydown on different passes. These differences can cause systematic variations in the produced color due to interactions between the ink and media. For example, consider
It can be seen that there are differences in timing between ink laydown on different print passes, both across the page, as well as down the page. For example, consider a first overlap region 104 where ink is first applied during the rightward first swath 101, and then during the leftward second swath 102. In a left portion of the first overlap region 105 there will be a relatively long time delay between the times that ink is applied on the first swath 101 and the second swath 102. This is due to the fact that the printhead must travel all the way across the page, then turn around and come all the way back across the page. However, in a right portion of the first overlap region 106 there will be a relatively short time delay between the times that ink is applied on the first swath 101 and the second swath 102. This is because the printhead needs to travel a shorter distance before it turns around and comes back. The reverse is found to be true in a second overlap region 107 where ink is first applied during the leftward second swath 102, and then during the rightward third swath 103. In a left portion of the second overlap region 108 there will be a relatively short time delay between the swaths, whereas in a right portion of the second overlap region 109 there will be a relatively long time delay between the passes.
The differences in the time delays, both across the page and from swath-to-swath, can result in significant differences in the characteristics of the reproduced image. When there is a longer time delay between passes, the ink applied during the first pass will have a longer time to dry or soak in to the paper. This can result in noticeable differences in the density of the printed region. Additionally, there can also be noticeable differences in the image structure characteristics. For example, coalescence artifacts, as well as surface characteristics such as gloss and haze, are often observed to be a function of the timing between when neighboring ink drops are applied. Typically, the swath-to-swath differences at a given horizontal position in the image are much more objectionable than the variations across the page. This is because the swath-to-swath differences produce a periodic artifact where the image characteristics vary for alternating swaths. These artifacts are sometimes referred to as bi-directional banding artifacts since they are inherently related to bi-directional multi-pass print modes. The magnitude of these artifacts is quite dependent on the characteristics of the particular ink and media used in the inkjet printer, as well as print mode attributes such the amount of ink, the number of passes and the printing speed. The magnitude of the bi-directional banding artifacts can even be affected by the size of the image since this can have an effect on how long it takes the printhead to travel across the page. For many combinations of ink, media and print mode, the bi-directional banding artifacts have been found to be quite objectionable.
Bi-directional banding artifacts can be even more severe for the case of printing color images with multiple color inks. Consider the case of a color inkjet printer using cyan, magenta, yellow and black inks. Typically, each ink will be printed using a different column of nozzles in the printhead. Therefore, as the printhead moves back and forth across the paper, there will also be differences in the order that the different color inks are applied. For example, consider the case where a uniform blue image region is to be printed using equal amounts of cyan and magenta inks. If the cyan nozzles are located to the right of the magenta nozzles in the printhead, they will be applied before the magenta drops on a rightward swath, but after the magenta drops on a leftward pass.
Consider the case where a blue image region is printed using the 2-pass configuration shown in
A variety of methods have been proposed to alleviate the objectionable bidirectional banding artifacts. One solution is to use only uni-directional print modes where ink is only printed when the printhead is moving in one direction (e.g., rightward). However, this solution significantly limits the throughput of the printer since it is necessary to wait for the printhead to return back to the starting position before printing another swath.
Another step that can be taken is to slow down the printing speed to give the ink/media interactions more time to stabilize between passes, but this too will significantly impact the printer throughput. Note that while this approach can help reduce the contributions of the bi-directional banding that are due to differences in the timing between ink laydowns, it will not alleviate the contributions that result from differences in ink ordering.
Another way to reduce the magnitude of the bidirectional banding artifacts is to increase the number of printing passes. This will effectively slow down the rate of ink deposition and dilute the impact of the differences in ink order and timing at any given location. However, this will also have a direct impact on the throughput of the printer, so it is not a desirable solution for applications where print speed is a critical requirement.
Another solution that has been proposed is to modify the amount of ink laydown for the rightward and leftward printing passes. One way to accomplish this is to use different color transforms to process the image for the different printing passes. For example, see U.S. Pat. Nos. 6,354,692 and 7,054,034, and U.S. Patent Application Publication No. 2003/0048327. Alternatively, an ink depletion operation can be used to modify the amount of ink that is printed depending on the print direction. One way that this can be accomplished is to modify the print masks as described in U.S. Pat. No. 6,545,773. These methods can help alleviate the component of the bi-directional banding that results from differences in ink ordering. However, they will be ineffective at compensating for the contributions of the bidirectional banding that result from differences in the timing between ink laydowns since these effects will vary from left to right across the page, and these methods do not provide for changing the ink laydown as a function of the horizontal position.
The present invention represents a method for reducing banding artifacts for bidirectional multi-pass printing on an inkjet printer utilizing a printhead with a plurality of ink nozzles comprising defining different print masks to be used for leftward and rightward printing passes such that both the order of ink laydown and the timing between ink laydown on different passes are each substantially constant for a given horizontal position within the image, independent of the vertical position within the image; and printing an input image on the inkjet printer with the defined print masks using a bi-directional multi-pass print mode.
It is an advantage of the present invention that images with reduced bidirectional banding artifacts can be produced for inkjet printers using bi-directional multi-pass print modes.
It has the additional advantage that it will reduce bi-directional banding artifacts that result from both differences in the order of ink laydown, as well as the timing between ink laydown on different passes.
It is yet another advantage of the present invention that the reduction in bidirectional banding artifacts can be achieved without a significant impact on the printing speed. In some embodiments of the present invention, the impact on the printing speed will be negligible.
The present invention represents a method to reduce bi-directional banding artifacts typically associated with bi-directional multi-pass printing on an inkjet printer by way of a novel method for defining different print masks for leftward and rightward printing passes. The leftward and rightward print masks are defined such that the order of ink laydown and the timing between ink laydown on different passes are substantially constant for a given horizontal position within the image, independent of the vertical position within the image.
Turning now to
The movement of the printhead across the page is typically referred to as a printing pass, or sometimes as a print swath.
An important feature of the present invention is that different print masks are used for the rightward and leftward printing passes. A rightward print mask 124 is used for the rightward printing passes (i.e., the first printing pass 120 and the third printing pass 122). A leftward print mask 125 is used for the leftward printing passes (i.e., the second printing pass 121 and the fourth printing pass 123). The rightward print mask 124 and the leftward print mask 125 are designed in a manner such that the order of ink laydown and the timing between ink laydown on different passes are substantially constant for a given horizontal position within the image independent of the vertical position within the image.
In one preferred embodiment of the present invention, the rightward print mask 124 and the leftward print mask 125 are defined by starting from a conventional multi-pass print mask and rearranging segments of the print mask to form two new print masks. The example shown in
To understand how the method of the present invention satisfies the required condition that the order of ink laydown and the timing between ink laydown on different passes are substantially constant for a given horizontal position within the image, independent of the vertical position within the image, consider a first overlap region 126 and a second overlap region 127. The first overlap region 126 corresponds to the region of the page where the first printing pass 120, the second printing pass 121 and the third printing pass 122 all overlap. The second overlap region 127 corresponds to the region of the page where the second printing pass 121, the third printing pass 122 and the fourth printing pass 123 all overlap.
Consider the scenario discussed above relative to
It can be seen that for the left portion of the first overlap region 128 and the left portion of the second overlap region 129, ink is applied such that the order of ink laydown and the timing between ink laydown on different passes is the same (cyan/magenta/long delay/magenta/cyan) for a horizontal position on the left side of the page, independent of the vertical position within the image. Likewise, it can also be seen that for the right portion of the first overlap region 130 and the right portion of the second overlap region 131, ink is applied such that the order of ink laydown and the timing between ink laydown on different passes is also the same (cyan/magenta/short delay/magenta/cyan) for a horizontal position on the right side of the page, independent of the vertical position within the image. Furthermore, it can be seen that the same would be true for all other horizontal positions across the image as well. Therefore, the abrupt changes in color and/or image structure than can occur for a conventional multi-pass print masking configuration due to the pass-to-pass differences in the order of ink laydown and the timing between ink laydown on different passes are eliminated. This drastically reduces the bi-directional banding artifacts that are often observed for bi-directional multi-pass printing.
It should be noted that the timing between ink laydown on different passes is different for the right side and the left side of the page, and therefore the color and image structure characteristics can be somewhat different from side-to-side across the image. However, any changes in the image characteristics will occur slowly across the width of the image, with no abrupt transitions. As a result, these differences will generally be much less objectionable than the bi-directional banding artifacts associated with conventional multi-pass print masking configurations, such as that illustrated in
A more detailed illustration of how a rightward print mask 124 and a leftward print mask 125 can be defined starting from a conventional multi-pass print mask is illustrated in
Similarly, a leftward mask plane 133 includes 3 segments corresponding to different subsets of the printing nozzles: two identical subsets of active nozzles 136 at the top and middle of the leftward mask plane 133, and a subset of unused nozzles 137 at the bottom of the leftward mask plane 133. The subsets of active nozzles 136 correspond to the segments of the leftward print mask 125 labeled as “2” and the subset of unused nozzles 137 corresponds to the segment of the leftward print mask 125 labeled as “X” in
While the process of defining the rightward print mask 124 and the leftward print mask 125 based on a conventional multi-pass print mask represents a simple method for implementing the present invention, it will be obvious to one skilled in the art that there are many different methods that could be used to form the rightward print mask 124 and the leftward print mask 125 to achieve the desired result that the order of ink laydown and the timing between ink laydown on different passes are substantially constant for a given horizontal position.
In a preferred embodiment of the present invention, rightward print mask 124 and the leftward print mask 125 are defined by specifying a plurality of mask planes, each associated with a different printing level, as was taught in U.S. Patent Application Publication No. 2007/0201054. However, the print masks can also be defined using any of a number of other techniques for specifying where drops should be printed on each printing pass, such as by specifying threshold matrices as was taught in U.S. Pat. No. 6,238,037.
The example described with reference to
While the configurations described above for implementing the present invention accomplish the desired result of substantially reducing objectionable bi-directional banding artifacts, they do have an impact on the throughput of the inkjet printer. This is largely due to the fact that there are a significant number of unused ink nozzles in the leftward and rightward print masks. For the case of an even number of printing passes, it can be seen that the number of unused nozzles is substantially equal to number of active nozzles divided by the number of passes. For example, from
The result of having these subsets of unused ink nozzles is that the page advance distance is correspondingly smaller for a given printhead length than it would be using conventional multi-pass print masks with print modes having the same number of printing passes. Essentially, the configurations that have been described are equivalent to adding an extra printing pass. Therefore, the throughput for a 2-pass print mode of the embodiment described above would be approximately equivalent to the throughput for a conventional 3-pass print mode. This would represent a decrease in the throughput to ⅔ (67%) of what it would be for the conventional 2-pass print mode. As the number of passes increases, the throughput impact is less significant. It can be seen that the throughput reduction factor will be approximately equal to N/(N+1), where N is the number of printing passes. For example, for a 6-pass print mode using the embodiment described above, the throughput would only be 6/7 (86%) of a conventional 6-pass print mode. In many cases, the image quality advantages of the present invention will be well worth the reduction in throughput.
For print modes with an even number of printing passes, the throughput impact can be largely mitigated by using an alternate embodiment of the present invention that uses different page advance distances before the leftward and rightward printing passes. This is illustrated in
In the limit, the number of unused nozzles can be shrunk to zero. In this case, the first page advance distance 166 would be zero, and the second page advance distance 167 would be the full width of the printhead. This would produce a throughput equivalent to the conventional 2-pass print masking configuration. However, it would have the disadvantage that it eliminates any redundancy in the printing nozzles since the same nozzle would pass over a given position on the paper in both printing directions. As a result, there would be no way to compensate for artifacts associated with missing or misdirected nozzles, which would eliminate one of the advantages of multi-pass printing. Therefore, it will generally be desirable to use a first page advance distance 166 of at least a few pixels. If the number of unused nozzles is maintained at a small value, there will be a negligible impact on throughput relative to the conventional multi-pass print masking configuration, while maintaining all of the advantages of the present invention with respect to reducing the visibility of bi-directional banding artifacts. It should be noted that this approach for improving the throughput will only work for the case of an even number of printing passes, so there will be an advantage to using print modes having an even number of printing passes when implementing the present invention.
It will be obvious to one skilled in the art that the present invention can be combined with many different print masking methods. For example, in U.S. Patent Application Publication No. 2007/0201054, Billow et al. describe using print masks having a non-uniform duty cycle, meaning that not all of the nozzles in the printhead will print with the same duty (see FIG. 7 and FIG. 8 of U.S. Patent Application Publication No. 2007/0201054). Generally, the duty of the nozzles near the ends of the printhead will be lower than the duty in the center of the printhead, and the masks will be defined in such a way that the sum of the duties for the print masks over all passes will be a constant, independent of the row of the image. This approach can be advantageous for hiding banding artifacts that commonly occur near the swath boundaries. Such masks can easily be rearranged and used to form leftward and rightward print masks according to the methods described earlier. However, this can cause abrupt changes in the duty cycle across the height of the print mask, which cannot always be desirable. Alternately, it is possible to generate leftward and rightward print masks having a non-uniform duty cycle that satisfy the requirement of the present of invention that the order of ink laydown and the timing between ink laydown on different passes are substantially constant for a given horizontal position within the image. However, this requires that the leftward and rightward print masks be defined directly rather than simply rearranging conventional multi-pass print masks.
The method of the present invention can also be combined with fractional print masking techniques, such as that disclosed by U.S. Pat. No. 6,206,502. Consider the arrangement shown in
The rightward printing mask 174 is labeled with several different segments. A segment labeled as “X” corresponds to a set of unused ink nozzles on the top end of the printhead. A segment labeled “1” corresponds to the main printing region of the print mask which prints ink the first time the printhead passes over a given region of the paper. Fractional mask segments labeled “1F” and “3F” correspond to the overlap printing region 179 where the image is printed with more than 2 printing passes. The fractional mask segment labeled “1F” will print using a reduced duty the first time the printhead passes over a given region of the paper, and the fractional mask segment labeled “3F” will print using a reduced duty the third time the printhead passes over a given region of the paper. Likewise, the leftward printing mask 175 is also labeled with several different segments. A segment labeled as “X” corresponds to a set of unused ink nozzles on the bottom end of the printhead. A segment labeled “2” corresponds to the main printing region of the print mask which prints ink the second time the printhead passes over a given region of the paper. Fractional mask segments labeled “2F” and “4F” correspond to the overlap printing region 179 where the image is printed with more than 2 printing passes. The fractional mask segment labeled “2F” will print using a reduced duty the second time the printhead passes over a given region of the paper, and the fractional mask segment labeled “3F” will print using a reduced duty the fourth time the printhead passes over a given region of the paper. Generally, the duty of the fractional mask segments will be defined in such a way that the sum of the duties over all the printing passes will be a constant, independent of the row of the image.
The advantage that the fractional print mask configuration shown in
The invention has been described in detail with particular reference to certain preferred embodiments thereof, but it will be understood that variations and modifications can be effected within the spirit and scope of the invention.
Number | Name | Date | Kind |
---|---|---|---|
4746935 | Allen | May 1988 | A |
4967203 | Doan et al. | Oct 1990 | A |
5416612 | Ingraham et al. | May 1995 | A |
5790150 | Lidke et al. | Aug 1998 | A |
5923349 | Meyer | Jul 1999 | A |
5992962 | Yen et al. | Nov 1999 | A |
6086181 | Majette et al. | Jul 2000 | A |
6206502 | Kato et al. | Mar 2001 | B1 |
6238037 | Overall et al. | May 2001 | B1 |
6254217 | Askeland et al. | Jul 2001 | B1 |
6310640 | Askeland | Oct 2001 | B1 |
6354692 | Ross | Mar 2002 | B1 |
6375307 | Vinals et al. | Apr 2002 | B1 |
6454389 | Couwenhoven et al. | Sep 2002 | B1 |
6545773 | Hudson | Apr 2003 | B1 |
7054034 | Underwood et al. | May 2006 | B2 |
7301668 | Kawanabe et al. | Nov 2007 | B2 |
7715043 | Billow et al. | May 2010 | B2 |
8157343 | Marumoto | Apr 2012 | B2 |
20030048327 | Serra et al. | Mar 2003 | A1 |
20060087527 | de Pena et al. | Apr 2006 | A1 |
20070201054 | Billow et al. | Aug 2007 | A1 |
20080204494 | Jahana et al. | Aug 2008 | A1 |
Number | Date | Country | |
---|---|---|---|
20100013878 A1 | Jan 2010 | US |