Facsimile and static presentation processing – Static presentation processing – Data corruption – power interruption – or print prevention
Reexamination Certificate
2001-08-23
2004-08-03
Grant, II, Jerome (Department: 2626)
Facsimile and static presentation processing
Static presentation processing
Data corruption, power interruption, or print prevention
C358S001400, C358S001300, C358S001800
Reexamination Certificate
active
06771379
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to procedures and machines for incremental printing of text, graphics or photograph-like images on printing media such as paper, transparency stock, or other glossy media; and more particularly to collection and use of information about the condition of printing elements, in preparation of printmasks for multitask printers.
BACKGROUND OF THE INVENTION
Incremental printers form images through an elegant overlay of micromechanical, electronic, chemical, liquidic and capillary phenomena. It is accordingly in the nature of such systems to be vulnerable to infinitesimally fine perturbations from manufacturing perfection, and so they are.
In fact, it is desirable that incremental printing systems be able to operate satisfactorily with the poorest possible degree of manufacturing perfection, since such apparatus is maximally economical. From its beginnings, the incremental-printing art has undergone continuous struggles with several successive stages of image-quality difficulties, all arising from sensitivity of this basic methodology to fine irregularities in the apparatus.
Meanwhile, in a quest for greater image throughput, printing-element arrays have been made both progressively larger in overall length and progressively finer in resolution. This evolution has continued to aggravate the image-quality problems that arise from tiny imperfections.
(1) Primitive Banding
Earliest apparatus in this field operated on a single-printing-pass basis—that is, one pass of the printhead (printing-element array) over each part of the image—and suffered from image imperfections, sometimes called “banding”, along lines where successive swaths of marks failed to abut perfectly. Early rounds of refinement in the field therefore focused upon techniques for minimizing the conspicuousness of such interswath seams.
That earliest type of banding presented itself especially in pictorial images—i.e. images made up of line drawings, common graphics, or continuous pictures. (Images of lettering, i.e. text, were relatively immune since many choices of text size and spacing could be arranged to print within the height of individual swaths, thus avoiding printing anything across the seams between swaths.)
This banding was mitigated by hiding the edge of each swath within the height of one or more other swaths—or, more specifically, by the introduction of multipass printing: each part of the image receives marks from plural, usually multiple passes of the printhead. Such printing requires allocation of the marks as among the several passes, and this is the function of so-called “printmasks”.
While printmasks have been developed using quite a number of different notations, the underlying idea is quite basic—to define which dots print in which passes. Initially, all the different printmask notations in common consisted of simple, relatively small patterns that were conceptually superimposed over the pixel data of an image.
These masking patterns were stepped or tiled across and down the image to serve in allocating marks for the entire image. Some pioneering patents in the name of Mark Hickman and in the name of Lance Cleveland, among others, canvass the several successively more sophisticated methodologies that developed in this era, in the late 1980s and early 1990s.
(2) Progressively Subtle Forms of Banding
Unfortunately, while multipass printing was reasonably effective in concealing the interswath-seam type of banding it also introduced new and progressively more resilient forms of banding that continued to challenge designers in this field for several years.
Because the earliest masks were regular, they tended to visually beat against the pixel-structure periodicity of images, or against features of the images themselves—producing repetitive moiré-like patterns that were often very conspicuous. In response, pseudorandomness was introduced into a later generation of printmasks.
Surprisingly, these patterns too—being small and repetitive—rather than eliminating patterning altogether, instead yielded bizarre shapes variously described as “worms” or “organ” shapes that seemed to crawl repetitively across the images, particularly in midtones. These stubborn artifacts, a new and different kind of banding, in general represented shapes entirely unrelated to image features.
It was thought that actual randomness would dispel these obtrusive forms, but they later remained even when true randomness was introduced into printmasks. The problem was that masks were still small—and tiled across and down—so that creeping kidneys, and so forth, continued to be generated repetitively and with fixed periodicity, and thus to be conspicuous.
These types of banding effects also persisted when masks were made larger, up to 2 cm (about ¾ inch) wide and tall. In addition, with truly random masking there seemed to appear a new, different kind of image defect: randomness-generated granularity.
Some workers noted that banding was generated basically because individual printing elements were mispointed, or firing more heavily or lightly than nominal, or simply had failed. This observation led artisans to look for relatively simple solutions in the nature of identifying misbehaving elements and straightforwardly reassigning the printing tasks of each degraded element to some other element.
This simple-reassignment type of system turned out to be capable of improving image quality very significantly, but only for a short time. When all the tasks of a failed element were transferred to some nominally healthy surrogate, that healthy element was thereby called upon to do double duty.
In many or most cases this two-times normal loading on the healthy (but, after all, not really new) element accelerated its aging and deterioration. Reassignment of its double-overload tasks to yet another element was soon required, leading to triple overload of that backup unit. Skilled workers in this field could then see that simple reassignment would provide no longterm solution, within the life of a printer.
(3) “Shakes”
The next major developments in masking for incremental printing were introduced by Joan-Manel Garcia, whose patent documents have been mentioned earlier. He devised a powerful conceptual construct within which to generate randomized masks for very large photograph-like images automatically.
Garcia's mask-generating procedures, which he named “Shakes”, can operate in the field—and on the fly if desired—obviating the need for dedicating large amounts of nonvolatile data storage to hold factory-computed large masks. This development accordingly in principle removed or greatly elevated an upper size limit for random masks.
He also noticed, however, that masks significantly wider than 3 or 4 cm (roughly 1 to 1½ inch)—even when repetitively tiled—no longer led to noticeable banding. Garcia explained this observation in terms of spatial frequencies to which the human eye is sensitive.
Hence Garcia's random masks needed only be made about 3 or 4 cm wide (and tall) to avoid perceptible repetition, and these sizes were very readily within the ability of his algorithms to produce. Even if present, the so-called “worms” are much less conspicuous when the stepped masks are of these larger sizes.
Further, into this methodology Garcia integrated important observations about perceptible granularity. More specifically, he recognized that graininess in images can be generated as a form of signal noise in the masking process, particularly as a result of randomness.
Hence Garcia was able to include in his formulations a balancing of desirable randomness that minimizes banding vs. undesirable randomness that raises perceptible grain. Garcia and his colleagues also introduced ways—including one formulation which they called “popup” masks—to mitigate the relatively large amounts of computation entailed in Garcia's classical Shakes procedures.
Another particularly beneficial property of the Shakes techniques is that they attack not only the first-generation problem of failed prin
Boleda Miquel
De la Lama Jesús
Grandmougin Lionelle
Sanchez Salvador
Serra Marc
Grant II Jerome
Hewlett--Packard Development Company, L.P.
Lippman Peter I.
LandOfFree
Streamlined real-time printmask revision, and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Streamlined real-time printmask revision, and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Streamlined real-time printmask revision, and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3316540