 |
||||
    |
||||
 Flexo RIP Package Platesetter & Imagesetter RIP-Kits DI RIP-Kits Epson & HP RIP-Kits CIP3 & Ink Key Software RIP Options RIP Upgrades TIFF Catchers & Interface Cards CTP Systems Screen Printing Systems Film & Plate Processors Free Demo Version Support Contact RTI |
|
|||||||||||||
|
Digitally Modulated Screening is available as an option for any Harlequin RIP v.6 or later.
The digitally modulated screening option examines each pixel it marks to make sure that it produces the best possible output. It does this with an extensive set of tests. It examines each pixel as well as the pixels around it to make sure that no dot is too small to plate or print, no dot is too large to be visible, no non-dot is too small to fill-in and that no sequence or arrangement of dots is going to cause patterning or dot gain issues. It can afford to do this because of the immense processing power in modern high performance CPUs. As a result, it produces prints that contain a very high level of detail, especially in hilite and shadow areas, smooth vignettes that are exceptionally smooth and flat tints that are ‘flat’. Digitally Modulated (DM) screening is so named because it digitally modulates each and every pixel it produces rather than repeating a fixed pattern of dots (as in AM screening) or randomly marking pixels (as in FM screening). This screening option analyzes each pixel it produces to ensure that no dot is too small to plate or print, no 'non-dot' is too small to fill-in and no dot or 'non-dot' is too large so as to be visible. Dots are created in a carefully controlled manner. It cleverly modulates each pixel based on a deep understanding of laser optics, plate technology, printing press behaviour and ink flow to ensure that dot gain is eliminated, resulting in the complete removal of patterning and graininess. The result is a quality print that was previously unachievable, especially on violet devices.
Graininess (1) The eye trying to average out the areas of white and black to achieve an average gray. This is due to the random placement of dots, as opposed to the regular placement of dots as found in an AM screen. If one considers a traditional AM screen, due to the very similar sized dots consistently repeated across the page, the eye can easily average them out to produce a perception of a certain gray level that appears ‘flat’. With the random placement of dots in an FM screen which produces areas of different sizes, the eye struggles to do so. There are two potential solutions to this issue. Either the random placement must be controlled so that it is less random or the visibility of the randomness must be reduced to be so small that the eye can average it out. Techniques for both of these are known. For example if one considers an EDS FM screen, there are known techniques of using different distribution coefficients for different tone values to produce more regular dot placement. The visibility of the randomness can clearly be reduced by increasing the resolution of the screening, thus making the dots smaller. (2) Different areas of the same color differing in size. This is due to areas of a color being different in size. For example in a number of 1x1 pixel dots, the odd 2x2 pixel dots scattered around would show up as ‘noise’, producing graininess. Patterning (1) Different dot patterns gaining in different ways. This is due to the laser size and the resulting dot gain which is made worse on violet plates. For example compare a 2 by 2 pixel square versus a 4 pixel horizontal line versus a 4 pixel diagonal line. For simplicity, we have reduced the varying exposure levels of a Gaussian curve down to two. The darker color is the primary exposed area. The lighter color is a secondary area, which would be exposed on a violet system and not on a thermal system. It can be clearly seen that the diagonal 4x1 block with a boundary length of 16 is quite a bit heavier than the 2x2. In fact the additional weight of diagonal lines versus horizontal lines is one of the main causes of patterning in certain FM screens.
(2) Small areas of white filling in. This is due to the laser size if trying to print a single white pixel, combined with any ink gain on press, whereby the surrounding black pixels gain so much that the single white middle pixel fills in. In this example we can clearly see that the middle pixel is almost completely filled in. If too many such areas are near each other, the result can be a very dark patch which produces significant patterning.
Square 3x3 block with a middle pixel that fills in (3) Small areas of black disappearing. This is caused by as many as three issues: a) Firstly the small area may not be properly exposed. This depends on the CTP device. b) Secondly the small area may be too small for ink on the press to stick to. c) Thirdly the area may be so small that any plate wear can cause it to be removed. Printability
The above issues produce conflicting requirements. To reduce graininess you need smaller dots, to reduce visibility of the screen in general you need smaller dots, to reduce patterning you need larger dots and to print consistently you need larger dots. Shown below is the same sample printed using digitally modulated screening.
High resolution data that guarantees a minimum dot size. The MxN macro dot is produced by taking the full high resolution pixel data (of depth n, which may be 8 bit, 12 bit, 16 bit or any other n-bit) produced by the Harlequin RIP and sampling it down by a factor of MxN (where at least one of M or N is greater than 1). For CTP devices with equal resolution, M typically equals N and is anything from 2 to 6. The resulting sampled pixel is then screened, with the result, either a white (‘0’ pixel) or black (‘1’ pixel) being scaled back up to the full resolution. This effectively means that the output is screened at a resolution equal to X dpi / M and Y dpi / N. However, special treatment is applied to high resolution pixels that are either 100% black or 0% white. These pixels are removed from the sampled data and added back in to the resulting scaled up screened data. The result is all black and white pixels are preserved throughout the page. What this does is produce an image where all the dots have a guaranteed minimum size (apart from the one exception where a sampled pixel overlaps a solid black/white area) and where all outlines are rendered at the full output device resolution. This produces full high resolution objects such as text and graphics, along with highly consistent dots for press reproduction. Since a minimum dot size is maintained throughout, hilite areas can be printed right down to 0.5% or less. This is unlike previous solutions whereby the threshold arrays that make up the FM screen are replicated in a larger MxN array. It also differs from the CT/LW style of output, in that there the CT was RIPped and screened at a lower resolution and the LW data which could include screened graphics was RIPped and screened at a higher resolution. Issues include joining back adjacent areas which were screened at the different resolutions or overlapping areas of CT and LW data. Eliminates patterns in screening caused by dot gain.
The result of reducing the typical MxN dot is that there is no interference or very little interference between adjacent pixels, especially in the diagonal direction. This is different from other screening methods where there is a substantial amount of interference, especially in the diagonal direction, something which is made much worse with a CTP device with a large laser size and/or when using violet CTP devices (where partial exposure occurs).
No matter what dot pattern is produced by the screening algorithm, the overall dot gain of each dot is similar. Because each dot gain is similar we get a very consistent ‘dot’ reproduced throughout the entire image. This means that no (or very little) patterning artifacts arise and a small MxN macro cell can be used. In fact we typically use a 3x3 or 4x4 macro cell at 2540 dpi which produces extremely high quality results for the former and very high quality results for the latter. A further significant benefit of the reduced MxN macro dot is that shadow areas are kept much more open. This prevents filling in of high density areas and gives more depth to these areas.
One of the additional benefits of this reduction is that plates produced by most equipment is very close to linear, requiring very little calibration. This also greatly reduces banding in vignettes due to stepping in gray levels caused by dot gain. This reduction of the MxN macro dot can be applied either over the complete tonal range (from 0% < tone value < 100% - don’t forget we always treat black and white areas specially), or only in mid-tone and shadow areas (that is not in hilite areas). Not reducing the standard MxN macro dot equats to a printable dot on press. In order not to produce gaps between adjacent dots for systems with small laser sizes, the reduced dots can also be joined up. There are several ways to do this, each further limiting the amount of dot gain interference.
The joining of dots can include making a square of 2x2 MxN macro dots completely solid so that as we go up the tone range we correctly progress to a 100% solid tint. The end result of this significant breakthrough is plate quality that is unbelievable. In fact in initial testing, the quality of plates produced on a violet ECRM Mako device with this screening technology is very similar to that produced on much higher quality thermal plates. This innovative screening produces can produce high quality results on a device that costs significantly less.
Shown above are a blow ups of violet CTP plates. You can clearly see there is dot gain on the plate when using a standard MxN macro dot. This is made worse when put on press with ink gain, producing very strong diagonals (and filling in any small areas of white surrounded by black). In comparison, the reduced MxN macro dot when put on press with ink gain produces an almost perfect dot. The smaller the M and N get in the standard macro dot, the worse this problem gets. However, this is not the case for the reduced macro dot. Eliminates graininess using a 2x1 / 1x2 mega dot. Digital Modulated Screening uses a mega dot solution where each mega dot is composed of 2x1 macro dots, which are in turn made of MxN micro dots. Each macro dot is typically an MxN macro dot which may be reduced in size where the 2x1 mega dot is either in the X or Y direction. A random mix of 2x1 and 1x2 mega dots are used.
The random mix of 2x1 and 1x2 mega dots is completely controlled such that when additional dots are added adjacent to existing dots they are overlapped so as to produce ‘dot chains’ with no clustering at all. That is all ‘dot chains’ produced by this screening option are exactly the same width - M pixels wide or N pixels wide and no blobs (areas wider than this) are allowed to form. Additional mega dots are also completely controlled so as to produce mega dots of similar sizes. If two overlapping mega dots are being used for a certain tone value, slightly longer chains of three overlapping mega dots are occasionally allowed (more so as the tone value increases), but longer chains of four overlapping mega dots are not allowed.
The dot chains are constrained for certain tone values so as to produce a minimum length dot chain and chains with similar lengths that therefore gain (or wear) by similar amounts. The algorithm by definition is symmetrical, so black dot chains are controlled up to 50% and white dot chains are controlled up to 50%. Above 50% (for black dot chains) and below 50% (for white dot chains) where dot chains cannot produce the required tone values, mega dots must group into larger blobs. In this case they are allowed to do so, but the size of the these blobs are tightly controlled so that they grow in size uniformly as tone values increase.
The expectation with this screening is that any device (violet or thermal) that can currently achieve a 175 lpi screen will be capable of producing at least a 250 lpi (or even 350 lpi) equivalent modulated screen and any device that can currently achieve a 200 lpi screen will be capable of producing a 350 lpi equivalent modulated screen. If you would like to test the screening with your device please call or email and we will setup a free demo of the software. Currently the software supports the following machines: On ECRM MAKO devices (including 2, 4, and 8-up systems, as well as the newer MAKO 800) using both Fuji NV and NV2 metal violet plates, this screening technology has consistently produced outstanding prints equivalent to 350 lpi, whilst using a 420 lpi equivalent dot. On ECRM DPX devices using lower quality polyester plates, the screening has been able to produce high quality prints equivalent to 250 lpi, whilst using a 310 lpi equivalent dot. This marks a substantial improvement to the currently recommended maximum 175 lpi screen. In all cases, highlights down to 0.5% and shadows up to 99.5% were obtained, with smooth vignettes, flat tints and exceptional image detail.
If you are using a Harlequin RIP or running the Harlequin RIP demo, you can unlock this screening option free on a 30 day trial. |
