Rochester Institute of Technology RIT Scholar Works
Theses Thesis/Dissertation Collections
5-1-1998 An Analysis of available unsharp masking techniques used with mid-range PMT/drum scanners Eric Neumann
Follow this and additional works at: http://scholarworks.rit.edu/theses
Recommended Citation Neumann, Eric, "An Analysis of available unsharp masking techniques used with mid-range PMT/drum scanners" (1998). Thesis. Rochester Institute of Technology. Accessed from
This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusion in Theses by an authorized administrator of RIT Scholar Works. For more information, please contact [email protected]. An Analysis of Available UnSharp Masking Techniques Used With Mid-Range PMT/Drum Scanners
by
Eric Louis Neumann
A thesis submitted in partial fulfillment of the
requirements for the degree of Master of Science in the School of Printing Management and Sciences in the College of Imaging Arts and Sciences at the Rochester Institute of Technology
May 1998
Thesis Advisor: Professor Joseph L. Noga School of Printing Management and Sciences Rochester Institute of Technology Rochester, New York
Master's Thesis
Certificate of Approval
This is to certify that the Master's Thesis of
Eric Louis Neumann
With a major in Graphic Arts Publishing has been approved by the Thesis Committee as satisfactory for the thesis requirement for the Master of Science degree at the convocation of May 1998
Thesis Committee:
Joseph L. Noga Thesis Advisor
Marie Freckleton Graduate Program Coordinator
C. Harold Goffin Director or Designate An Analysis of Available UnSharp Masking Techniques Used With Mid-Range PMT/Drum Scanners
I, Eric Louis Neumann, prefer to be contacted each time a request for reproduction is made. I can be reached at:
P.O. Box 261 Itasca, IL 60143-0261 [email protected]
Graduate Student
May 1998 Acknowledgements
The completion of this thesis would not have been possible without the assis tance, support, and patience of many companies and individuals that I owe my gratitude to. I must pay special recognition to Professor Joseph L. Noga for his support and mentorship over the years that I spent both as a student and employee of RIT
The School of Printing Management & Sciences, RIT Prof. Marie Freckleton Prof. Len Leger Ms. Grace Gladney And the rest of the SPMS faculty and staff
Screen Mr. David Mitchell
Howtek Mr. Joel Hofmeister
Optronics Mr. Don Rogers
FujiFilm USA Mr. Allen Dunn Ms. Jan Mullen
Family and Friends Mr. Ethan Crist Mr. Carl Ogawa My Mother, Susan Elmore
in Table of Contents
List of Tables v
of List Figures vj
Abstract vjj
Chapter 1
Introduction \
Chapter 2
Theoretical Basis of Study 3 Endnotes for Chapter 2 24
Chapter 3
Review of Literature 25
Endnotes for Chapter 3 27
Chapter 4
Statement of Project Goals 28
Chapter 5 Methodology 30
Chapter 6
Results of Evaluation 33
Endnotes for Chapter 6 58
Chapter 7
Summary and Conclusions 59
Bibliography 66
Appendices 69
IV List Of Tables
Table 6.1 - Controls of USM on scanners evaluated 34
Table 6.2 - Mask colors resulting from color sensitivity filters 55
Table 7.1 - Scanning times 62 List of Figures
Figure 2.1 - O'Brien Effect 9
2.2- Figure Anatomy of a peaking signal 11
Figure 2.3 - Direct screen mask exposure 13
Figure 2.4 - Direct screen separation exposure 13
Figure 2.5 - Photomechanical unsharp masking 15
Figure 2.6 - Optical unsharp masking 18
Figure 2.7 - Digital unsharp masking 20
Figure 2.8 - Hybrid unsharp masking 22
Figure 6.1 - USM controls in Adobe Photoshop 4.0 35
Figure 6.2 - USM controls in DT-S Scan 3.4 36
Figure 6.3 - ResEdit interface to DT-S Scan 3.4 38
Figure 6.4 - Screen USM effects 41
Figure 6.5 - Directional characteristics of DT-S 1030AI 42
Figure 6.6 - USM controls in Trident 2.0 44
Figure 6.7 - Howtek USM effects 49
Figure 6.8 - Directional characteristics of ScanMaster 4500 50
Figure 6.9 - USM controls in ColorRight 5.0 51
Figure 6.10 - Optronics USM effects 56
Figure 6.11 - Directional characteristics of ColorGetter Eagle 57
VI Abstract
Unsharp masking (USM), also known as detail enhancement, is a process of combining an unsharp representation of an original image with the original image to obtain the effect of greater detail. USM can be performed photomechan- ically with additional exposures, electronically with the color scanner, and digi tally with the aid of a post-processing program. Electronic methods of USM per formed during the scanning process offer productivity benefits over both the photomechanical and post-processing methods. Mid-range PMT/drum scanners offer several methods of unsharp masking from which to choose. These meth ods, optical USM, digital USM, and hybrid USM each have advantages and dis advantages which are identified in this study. The study also offers an extensive reference of the available USM techniques for identification by the mid-range scanner operator. Three different midrange scanner/ interface applications are evaluated to identify their unique USM methods and each is evaluated for ease- of-use as well as the effectiveness of it's unsharp masking function. Multiple scans from each scanner/interface combination were completed and analyzed at high magnification. It was expected that more directional limitations would have been evident in the optical method, however it is shown that it's effectiveness does not suffer. Each of the USM techniques used on midrange PMT/Drum scanners has its own merits.
vn Chapter 1
Introduction
Unsharp masking is a necessary function in the process of making color separa tions. It is necessary in order to compensate for the visual loss of detail caused by the printing reproduction process. The concept is not new. It has been part of the color separation process since long before the introduction of the electronic color
scanner. The unsharp masking technique was first introduced with the photome
chanical color separation methods known as indirect and direct screen color sep
"unsharp" aration. An additional exposure created an mask that was combined
with the original in a successive exposure to provide detail enhancement in the
reproduction of an image. Color separation programs used on desktop systems
today (e.g. Adobe Photoshop), now provide a way to simulate the combination
of a digital unsharp image (filter) with the original to provide detail enhance
ment in the reproduction as a function of post-processing. Unsharp masking per
formed during the electronic color scanning process offers increased productivi
ty over photomechanical and post-processing techniques.
Mid-range scanners are the most recent entry into the electronic color scanner
market. These scanners and their accompanying software interfaces have the
benefit of more than twenty-five years of color separation advancements. Among
these advancements are the refinements to the unsharp masking process. There
are now several options available to the manufacturers of mid-range scanners
and in some instances to the users of the mid-range scanners.
1 This study provides both an extensive reference to the unsharp masking tech
niques available on mid-range scanners as well as an analysis of these tech
niques, identifying their advantages and disadvantages. Chapter 2
Theoretical Basis of Study
It is first necessary to provide the definitions for the terms mid-range scanner and unsharp masking. The definitions and explanations of those terms as they were used in this study follow.
Definition of Mid-Range Scanners
Today electronic color scanners have been segmented into three different classifi cations; high-end scanners, desktop scanners, and mid-range scanners.
Unfortunately the boundaries between these classifications is not well defined.
As such, there are essentially two methods to define scanners today: either by their technical definition or by their marketing definition. Technically, the scan
ners' classification may be defined by their components and specific technology.
Marketing classifications may be defined by their cost and target markets. A
"mid-range" scanner classified as by its technical definition could very well be
"high-end" "desktop" classified as either or depending on the market.
The technical definition. In the early 1970's, before any of the confusion, all scan ners were known as electronic color scanners and primarily used photo-multipli er tubes (PMT). With the introduction of charge-coupled device (CCD) scanners in the early 1980's a distinction became necessary. The result was the identifica
"high-end" "desktop" tion of the and scanner classifications. High-end scanners were drum-based, had PMTs and onboard separation computers, and used pro- prietary processing techniques. Desktop scanners were flatbed, had CCDs, required a personal computer and software for separations, and generally used open system (non-proprietary) processing techniques. Each system had its bene fits and limitations. The next logical developmental step was to link the existing high-end scanners to the desktop computers and software, capturing the benefits of both systems and eliminating many of the limitations. The link provided the necessary translation between the high-end and desktop processing techniques.
As this hybrid link between the systems became more in demand, a new classifi cation was realized. In the early 1990's manufacturers began to build desktop compatible scanners with many of the high-end components, creating what is
scanner.1 now referred to as the mid-range
The marketing definitions. The lines drawn between the scanner classifications become increasingly unclear. By some marketing definitions, any scanner that
"desktop" connects to a personal computer can be considered a scanner (e.g.
Crosfield/Fuji Celsis). At the same time, several manufacturers are producing flatbed CCD scanners that exceed the high-end both in quality and price, refer
"high-end" flatbed scanners (e.g. Linotype-Hell Topaz and Scitex ring to them as
"mid-range" EverSmart). Many flatbed, desktop scanners are marketed as scan
"low-end" scanners (e.g. PixelCraft ners to distinguish them from the flatbed
Prolmager 4520RS and Howtek ScanMaster 2500). Mid-range drum scanners
"desktop" link to the often get labeled as drum scanners simply because they
desktop (e.g. Screen DT-S1030AI). For the mid- purposes of this study, the following technical definition of the range scanner will be used: Mid-range scanners are drum-based, use PMTs, require a personal computer and software to perform many of the separation func
and techniques.2 tions, generally utilize open system (non-proprietary) processing
Definition of UnSharp Masking
Unsharp masking (USM), also known as detail enhancement, edge enhancement, detail contrast, and sharpening among other names, can be defined as the emphasis of existing detail in an image. This is not to say that USM adds actual detail to an image, rather it enhances or exaggerates the detail that is present.3
Reasons for UnSharp Masking
There are several factors that can result in loss of visual sharpness of an image from original scene to final printed sheet. While it is impossible to place detail where there is none, or reclaim detail that has been totally lost, in many instances
USM can help to compensate for a considerable amount of this loss.
"Soft" images - Soft images are those in which environmental conditions, such as
Over- fog, mist, or smoke, can effect the visual sharpness. or under-lighting of a particular scene, whether natural or studio, can also effect the sharpness of the image. In some cases, if the original scene captured on film was not sharp, USM can be used to help. To the contrary, in cases where an image is beyond "soft", such as when an image is even slightly "out of focus", USM will not be able to
loss.4 adequately compensate for detail
5 Print contrast - The human eye perceives detail through contrast. This contrast can be reduced by several factors present in the printing process. The most obvi ous of these is tone compression, a necessary step in printing, which reduces the range of densities from highlight to shadow on an original image to a range that is achievable on the printing press. The difference between the lightest and dark
est densities (contrast) on the printed reproduction are far less than those of the
original copy (transparency or reflection print). This, compounded with the
choice of stock brightness and ink absorption qualities, can result in a repro
original.5 duced image with considerably less contrast than the
- or Type of original All originals, whether 35mm slides, color negatives, prints,
original artwork can be classified as one of two types of copy; transmissive copy
(light passes through the copy) or reflective copy (light bounces off the copy).
an effect on image There are several reasons why the type of original may have
detail. The first issue is the generation of the copy. Most often transparencies rep
are made from resent the first generation of the image from the camera. Prints
negative). Similar to the electro such a first generation transparency (or color
generation loses some of photographic process, (photocopying) each successive
the fine detail (e.g., type serifs are the first to be lost when photocopies are made
considered when of other photocopies). Another factor that must be scanning
photographic process of a print, or reflective copy is halation. In the making
phenomenon wherein the light that is duplicating a transparency, halation is a
reflected off of the base material used to expose the photographic emulsion is
it. The visual effect is (paper or clear-base) back through the emulsion, re-exposing
greater amounts of halation than duplicate a less sharp reproduction. Prints exhibit 6 even with the addition transparencies, of an anti-halation coating. It is evident,
then, that transmission copy'1 scanning copy is preferable to scanning reflection
variables - There are Printing many characteristic variables in the printing process
while that, controllable, cannot be eliminated and will cause the loss of visual
sharpness. The most notable of these variables is dot gain. When ink is trans
ferred to paper the integrity of the halftone dot's shape and size are not retained
and dot gain occurs. Hard, sharp dots become soft, irregular dots due to the ink
absorption into the substrate. Another variable that may be encountered is slur.
This can be identified as the elongation (or dragging) of the halftone dot, most
often caused by unmatched surface speeds of the plate and blanket or blanket
and impression cylinders (generally due to improper cylinder packing). Closely
related to, and often confused with slur, doubling is identified by a second print
ing of the halftone image on the sheet. It is most often the result of slippage between the cylinders and becomes more probable as a press ages.
Each of these variables results in a reproduction with decreased sharpness.
However, it is important to note that increasing USM does not compensate for dot gain, doubling, and slur. It only helps to increase perceived detail lost because of these variables. Also, though misregistration also has an effect on the reproduction sharpness, it is not a controllable variable. Increased USM should not be considered an antidote to proper registration, yet, as more quick printers are attempting four-color process on small duplicator presses which are not known for their ability to hold registration, they are discovering that increased
sheets.7 amounts of USM can result in a greater quantity of acceptable 7 - Resolution Resolution in its forms has an effect on many reproduction sharp
ness. A continuous tone original holds much more fine detail than a halftone
reproduction. The primary reason for this involves the resolution of the original
vs. the resolution of the reproduction. The smallest image element responsible
for detail in the original is the photographic grain. This element is miniscule is
comparison to the - smallest element of the reproduction a halftone dot, which is
hundreds to thousands of times larger than the photographic grain. Finer ele
ments are able to sustain greater detail8. In addition, a higher resolution screen
can hold greater detail in ruling reproduction (e.g., commercial printing at
150LPI compared to newspaper printing at 85LPI). The amount of fine detail that
is captured is a result of the scanning resolution in conjunction with the main
aperture of the scanner. The main aperture selection is based on magnification
and screen ruling requirements.
There are other issues that effect sharpness. The reproduction size of an image
(magnification) will have an effect on sharpness. Also, due to the higher quality
optics used in PMT/drum scanners, they effect sharpness with their ability to
9 capture greater fine detail than CCD scanners.
How UnSharp Masking Works
Unsharp masking does not add detail to a reproduced image. It is actually the creation of an optical illusion which increases contrast in distinct areas of the image fooling the human eye into perceiving greater detail. This concept is best explained with the O'Brien Effect, which states that a light/dark border placed
8 between two otherwise similar densities will give the perceived effect of greater
contrast (Figure 2.1). Since the eye perceives contrast as detail, placing exaggerat
ed light/dark at borders the edges of density shifts (supposed detail) the eye is
fooled 10 into perceiving greater detail than actually exists.
B
} similar densities
B
perceived actual densities densities (contrast)
"A" will be perceived as light/dark border (density notably lighter than density "B".)
Figure 2.1 - O'Brien Effect
signals" The result of USM is the addition of "peaking to the separations and
final reproduction. These peaking signals, also referred to as a white line/black
line border, are the visual effect that can be seen under magnified inspection. The
edges of density shifts with USM appear to have a thin white line/black line bor
"white" der. In reality, can be more accurately defined as a density lighter than
"Black" the adjacent light density. would be defined as a density darker than the
adjacent dark density. The reason for this distinction will be described in the next
section. While the black and white borders can be seen under close inspection,
these lines are not distinguishable at a normal viewing distance. Instead, the eye
detail." interprets a noticeable contrast at the edges which is perceived as 9 Controls of UnSharp Masking
The USM signal, or more precisely, the peaking signals, can be manipulated and adjusted in several ways. There are three basic controls of electronic unsharp masking that allow the operator to adjust peak width, peak height, and sensitivi ty of the USM signals (Figure 2.2). Many modern scanners have the ability to make further adjustments to the unsharp masking signal giving even greater control.
Peak width of the signal simply describes the width of the white and black lines
or borders. The wider the line, the more likely that the human eye will perceive a
density change and thus contrast. However, if the line is too wide, the eye will
detect the line as a distraction and not as contrast. Depending on the USM
USM USM method used, the control has different names, including aperture,
12 area, pixel width, pixel radius (or diameter), and kernel size.
white and black lines. In other Peak height of the signal refers to the intensity of
"black" "white" and how (high densi words, how (low density) is the white line
white and black ty) is the black line. The greater the difference between the line,
eye. it is possible to have the greater contrast that will be perceived by the Again,
that the eye perceives the the intensities of the white and black lines too high, so
referred to as lines themselves and not the contrast. The control is intensity,
line control.13 amplitude, amount, and white line/black
USM limits or point, deter Sensitivity, commonly known as threshold, starting
signal. While it is possible to exaggerate mines when to apply the USM nearly
10 every distinct density change, it is not always desirable. USM sensitivity allows the user to distinguish a threshold at which the peaking signals will be applied.
If a density change falls below the threshold it will not receive a peaking signal.
If it reaches or exceeds the threshold level, peaking signals will be applied. It is
most useful in distinguishing image grain (in grainy originals) from actual fine
detail. Some scanners and software offer an additional function to threshold
called grain suppression or smoothing. Those density changes below the thresh
old are not only ignored for peaking, they are actually smoothed out to reduce
artifacts.14 possible grain
(*\ width of white line/ black line border 1 A g) height of white line/ black line border /\ ^^^ fn\ ^^^ ^s) (o\ sensitivity of white * line/black line control -* ?
A
(3)/TJ\ T
Figure 2.2 - Anatomy of a peaking signal
software include the to Additional controls offered by some scanners and ability
tonal areas of the image. For instance, you can limit USM peaking to specific
portion of the image the midtone sharpen only the highlight to midtone leaving
controls allow the user to shadow region unsharpened. Filter or color sensitivity 11 to decide which colors or color pairs will receive unsharp masking or provide
the to exchange data between ability channels (e.g., applying the cyan USM sig
nal to the yellow separation).13
How UnSharp Masking is Accomplished and Controlled
There are three basic methods for performing unsharp masking: photomechani
cal, electronic (during scanning), and digital in post-processing. USM evolved in
the same order. Regardless of the method used, the general idea is the same; cre
ate an unsharp representation of the original image and combine it with the orig
inal image.
Photomechanical. Originally all color separations were made by the photomechani
cal processes known as indirect and direct screen color separation. In both process
es a photographic mask was required for each separation: This mask served sever
al purposes; the mask density provides for the necessary tone compression of the
original densities, the choice of colored filters performs needed color correction,
and the use of a continuous tone emulsion (designed to cause an unsharp expo
sure) combined with a diffusion medium provides for detail enhancement (when
the mask is combined with the separation exposure) (Figure 2.3).
The separation exposure made through the unsharp photographic mask yields a
tone compressed, color corrected, detailed separation with the proper gradation.
"halo" The effect of the combination of a sharp and unsharp image causes a
changes.16 effect at distinct density (Figure 2.4)
12 Light Source
Color Correction Filter
Original Transparency Diffusion Sheet Panchromatic Masking Film
Figure 2.3- Direct screen mask exposure
Light Source
Color Separation Filter
Unsharp Mask Original Transparency Contact Screen Separation Film
Figure 2.4r- Direct screen separation exposure
13 With photographic unsharp masking there is very limited control. There are real
two ways to control ly only the amount of detail enhancement and they are very
"unsharpness" much dependent on one another. The of the mask will have the greatest effect on the width of the white/black borders. The density of the mask
(controlled by exposure) has some effect on the intensity of the white/black bor
"unsharpness" der (Figure 2.5). However, as the mask density increases the of the mask decreases. The controls of photographic unsharp masking are not easi ly adjusted, it would take quite a bit of testing with the photographic emulsions, the developing process, as well as the diffusion characteristics of the diffusion material used in the production of the mask. While not a simple task, USM is controllable in the photographic color separation process (i.e. direct screen color separation). Because manipulation was not easy, most traditional color separa tors would identify the variables in the photographic color separation process
that produced acceptable amounts of detail enhancement and use those settings the majority of the time.
14 Sharp Original
Fine Detail ( )
Mask Exposure Unsharp Mask
Original
J L
Mask (negative)
black border Separation
white border
Figure 2.5 - Photomechanical unsharp masking
15 Electronic. The introduction of electronic color scanners with integrated color computers allowed for the automation of the unsharp masking process. Within the electronic USM category exists a subset of methods that is available on elec tronic color scanners: optical, digital, and hybrid. The scanners themselves have experienced a notable evolution in the more than 25 years since their introduc tion, and as well, the USM methods have evolved. Early electronic color scanners
were solely PMT based, used proprietary processing languages, and had the
scanner (referred to as ability to produce color separation films right on the
direct scanning). In the last decade electronic color separation has been inundat
ed with desktop technologies - first with CCD based desktop scanners and color
high-end com separation software, and then with a hybrid technology utilizing
com ponents such as PMT's coupled with a desktop software interface, most
scanners. CCD scanners do not perform USM monly referred to as mid-range
digital method described electronically, rather they utilize the (post-processing)
below.17 of this The mid-range PMT/drum classification of scanners is the focus
will be presented in depth in the next section. study and it's USM technologies
in a digital form that Digital (post-processing). With the ability to capture images
the color separation process has allows for various types of image manipulation,
an image without USM is taken on an entirely different approach. Capturing any
- to be performed as a func now a viable choice allowing USM post-processing
manipulation programs such as tion in a number of color separation and image
mathematical function Adobe Photoshop. These programs perform USM as a
calculation.18 software locates edges within and /or In some cases the simply
mathematically. In more advanced situations the images and exaggerates them 16 software creates a digital actually unsharp mask (filter) that is then combined
with the normal much - image, as in the photomechanical process the results
being quite similar (in theory and practice).
USM Techniques used with Mid-Range Scanners
There are three primary techniques of electronic unsharp masking that have been
adopted by manufacturers of PMT/drum based mid-range scanners. These tech
niques are most commonly referred to as: optical USM, digital USM, and hybrid
USM (ak.a. digital on-the-fly).
Optical USM. The first electronic color high- scanners (today known as analog
end scanners) utilized special optics within the scanner to accomplish detail
enhancement. In addition to the optics (aperture) that capture the resolution of
the image as a signal, an additional aperture is used to capture a signal for USM.
The apertures, known respectively as the main aperture (or scanning aperture)
and USM aperture are different in size. The USM aperture is often several times
the size of the main aperture. While the main aperture captures the fine detail of
the image, the USM aperture captures an unsharp representation of the image.
As is done in both the photomechanical and digital (post-processing) forms of
separation.19 USM, the two signals are combined to yield a sharpened (Figure 2.6)
The controls of optical unsharp masking are both mechanical and electrical. The second generation of electronic color scanners had some additional hardware including an additional aperture and an additional PMT for unsharp masking.
17 I I I I scan lines U JTTTTl i i i i i i i i i
Main Aperture
i . . i i
USM Aperture
Main Aperture Signal
USM Aperture Signal (inverted positive signal)
black border Separation
white border <
Figure 2.6- Optical unsharp masking
18 The USM aperture primarily controls the width of the white line/black line bor
it der, is the difference between main actually the (scanning) aperture and the
USM aperture that determines the border width. The USM PMT basically cap
tures the lightness and darkness values at the edges of density shifts (as seen
through the USM aperture), the signal from the USM PMT is then amplified
which effects the of the intensity white line/black line border. Processing of the
USM signal is accomplished in a portion of the color computer dedicated to
detail enhancement. Additional control of detail enhancement through the color
computer include such functions as thresholding and smoothing. In most scan
ners the USM PMT can be filtered (red, green, or blue) in order to control the
way the USM effects certain colors.
Digital USM. Some manufacturers of new midrange scanners have opted for a
digital USM method very similar to the digital USM described earlier. For all
intents it is basically the same function. The only difference is that it is accom
hardware.20 plished with the scanners interface software and/or The benefit of performing USM at the scanners workstation versus performing the USM later in other software is primarily for speed. The detail enhanced image is effectively
"sharpened" before the final image file is ever written to the host computers file storage (hard disk or removable media).
Unlike photographic and optical unsharp masking, digital USM is not truly a function of image capture, but rather the product of image processing. While almost every image manipulation program offers some controls for detail enhancement, the most well known of these applications is Adobe Photoshop.
19 pixel matrix
UnsHarp Signal' (achieved with low passfilter) Sharp Signal
Unsharp Signal (inverted positive signal)
black border - Separation
white border <
2.7- Figure Digital unsharp masking
20 Photoshop serves as a very good representation of digital unsharp masking in
color separation. While it is possible desktop to apply digital USM to any digital
image file, it is preferable to perform this form of USM on an image captured
with no USM applied during scanning (i.e. "USM Off"). To apply USM to an
already detail enhanced image would compound the effect and produce an
undesirable image. Like most effects in Photoshop, Unsharp Mask is a filter that
is applied can to the image, it be applied to the entire image or limited using the
masking functions within Photoshop.
low- The digital unsharp masking function is implemented by first performing a
pass filtering operation on the original image. Low-pass filtering smooths an
image by reducing high-spatial-frequency details such as edges, sharp lines, and
points while leaving low-frequency information unaffected. The low-pass image
is then brightness-scaled to a desired level and combined with the original
detail.21 image. The resulting image contains sharpened edge (Figure 2.7)
Hybrid USM. A method that was originally used on digital high-end scanners
which uses a combination of the scanners optics with special mathematical func
mid- tions to perform USM is now one of the more popular methods used in
range scanners. It is sometimes mistakenly referred to as "digital USM", but it is
distinctly different than the Digital USM method(s) described previously. Other
On-The-Fly" accurate terms for this method include: "Digital and "Electronic
USM".21 The term "Hybrid USM, comes from the fact that it combines qualities
from both the Optical USM and Digital USM techniques.
21 scan lines
Unsharp Area Signal (simulated with Main Aperture data) Main Aperture Signal
J L
Unsharp Area Signal (inverted positive signal)
black border Separation
white border
2.8- Figure Hybrid unsharp masking
22 In hybrid USM the scanner does not utilize a USM aperture, rather in fabricates a
USM area from buffered information from previously scanned samples of infor mation obtained from the main aperture. When the scanner has gathered enough information it will create a USM signal by averaging several samples of informa
tion. As is the case in other USM methods, the USM signal is then combined with
the signal from the main aperture to yield the sharpened separation. (Figure 2.8)
23 Endnotes for Chapter 2
1 Joseph. Color Noga, Separation Systems, JPRT-409. School of Printing, RIT 1997
2 Chrusciel, Edward and Rogers, Don. What To Look For In a Color Scanner, Graphic Arts Monthly. October 1993, p.120-121
3 Agfa-Gevaert N.V. An Introduction to Digital Scanning. 1994, p.30
4 Southworth, Miles. Color Separation Techniques. 1989, p.6-20
5 Dainippon Screen Ltd. Scangraph Technical Guide. 1985, p.17
b Noga, Joseph. Color Separation Systems, JPRT-409. School of Printing, RIT 1997
7 Dainippon Screen Ltd. Scangraph Technical Guide. 1985, p.17
8 Dainippon Screen Ltd. Scangraph Technical Guide. 1985, p.17
9 Blatner, David and Roth Steve. Real World Scanning and Halftones. 1993, p. 171
10 Dainippon Screen Ltd. Scangraph Technical Guide. 1985, p.17
11 Agfa-Gevaert N.V. An Introduction to Digital Scanning. 1994, p.30
12 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996, p.4-55
13 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996, p.4-55
14 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996, p.4-55
15 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996, p.4-56
16 Noga, Joseph. Color Separation Systems, JPRT-409. School of Printing, RIT 1997
17 Green, Phil. Understanding Digital Color. 1995, p. 134
18 Agfa-Gevaert N.V. An Introduction to Digital Scantling. 1994, p.36
19 Linotype-Hell AG. NezvColor 3000, User's Guide. 1993, p.4-4
20 Agfa-Gevaert N.V. An Introduction to Digital Scartning. 1994, p.36
21 Baxes, Gregory. Digital Image Processing. 1994, p.343,346
22 Linotype-Hell AG. NezvColor 3000, User's Guide. 1993, p.4-7
24 Chapter 3
Review of the Literature
there is little written on Astonishingly very the topic of unsharp masking in graphic arts related literature. Most of what is written on the subject is either marketing or operational text. There are no known resources that have made an attempt to compare or even define the available USM methods used in electronic scanning. As of the introduction the mid-range scanner is a relatively new event, the combination of USM methods and mid-range scanners in any resources has not been attempted.
The resources that were identified for this study consisted primarily of four dif ferent categories: excerpts from color separation references (text books on the entire process of color separation); excerpts from operational manuals of various color scanners and software; excerpts from color and imaging texts (not neces sarily directly related to color scanners or color separation); and other appropri ate references. In addition, technical staff from each of the three scanners to be used in this study were consulted on various issues.
' Color separation references, primarily R.K. Molla's Electronic Color Separation provided the basic definitions of the concept and mechanics of the unsharp masking procedures. Molla's text is the only reference encountered that offered
subject of masking. While more than a few passing paragraphs to the unsharp this reference identifies that there are options available in unsharp masking, it
25 does not go so far as to identify the distinctions among them. In addition, at the date of publish of this reference mid-range and desktop scanners were not wide ly marketed.
The operational manuals for the scanners to be used in this study provided the
ability to make the necessary distinctions of the unsharp masking techniques
used with the various mid-range scanners. In addition, several manuals for other
high-end, mid-range, and desktop scanners and software were consulted. Some
manuals provided only simple operational instructions while others, such as
2 Linotype-Hell's NezvColor 3000 User's Guide offered a more extensive explana
tion of the concepts and the technology.
3 Color and imaging texts, including John Yule's Principles of Color Reproduction
and further confir offer the technical perspective of unsharp masking theories
mation of the concepts. While well respected references in the color and imaging
industries most of these texts fail to close the gap between theory and practice.
consulted includes articles from Other very helpful source material that was
literature from the mid- graphic arts and imaging periodicals and marketing
of these references were short range scanner manufacturers. Unfortunately most
a better of the claims on fact and long on claims, but did provide understanding
regards to their scanners (USM tech that are made by the manufacturers in
correspondence with the manufac nique). By a similar token, phone and e-mail
more definitive identifica turers provided further confirmation of the claims and
tion of the unsharp masking methods utilized.
26 Endnotes for Chapter 3
1 Molla, Dr. R.K. Electronic Color Separation. 1988
User' 2 Linotype-Hell AG. NezvColor 3000, Guide. 1993
3 Yule, John. Principles of Color Reproduction. 1967
27 Chapter 4
Statement of Project Goals
It is not the intent of this study to determine which of the USM methods used with mid-range scanners is best, rather, the goal is to analyze and compare the methods to identify how each is different. Each method appears to have it's
advantages and "quality" disadvantages; where as the of USM increases, so does
the time and processing memory requirements. Optical USM, while considered
the fastest method, supposedly experiences directional limitations. Digital USM
is not limited however since a directionally, it is function of post processing it is said to be slower and notably requires more memory in the host computer (scan ner workstation). Hybrid USM appears to be the best of both worlds, seemingly not limited optical directionally like USM, and not requiring the same memory requirements of digital USM, however it may not be quite as fast as optical USM.
The primary objectives of this thesis are three-fold: To provide a thorough refer ence of unsharp masking techniques available with the focus on those methods used on mid-range PMT/Drum scanners. To evaluate three distinct interface applications with the focus on the unsharp masking functions/tools. To evaluate the effectiveness of three distinct unsharp masking methods used on mid-range
PMT/Drum scanners; optical USM, digital USM, and hybrid USM.
The first objective will be the result of the theoretical basis of this study. The sec
inter- ond objective will be presented as the results of in depth evaluation of the
28 face applications. The final objective will be the result of extensive testing of the interface applications and scanners; in the process an attempt will be made to prove or disprove the following statements:
1.1) Optical USM is the fastest USM method in use on midrange scanners.
1.2) Optical USM is limited in its effectiveness directionally (lowest quality).
2.1) Digital USM is the slowest USM method in use on midrange scanners.
2.2) Digital USM is not limited in its effectiveness directionally (highest quality).
3.1) Hybrid USM is a fast USM method in use on midrange scanners.
3.2) Hybrid USM is not limited in its effectiveness directionally.
29 Chapter 5
Methodology
Prior to performing experimentation on the mid-range scanners used, research was conducted to provide the most accurate descriptions of the technology eval uated. While a basic of description the three known methods of unsharp mask ing has been presented, all scanners are different; it would not be accurate to say, for all instance, that scanners that accomplish unsharp masking optically yield similar results. For this reason, as many specifications of the scanner and inter face software that can be gathered from the manufacturer's technical engineers are presented.
Once the specific technology had been identified, considerations for scanning were made. The first issue identified was the selection of the originals to be scanned. Two types of images were selected to help illustrate the possible limita tions of the USM methods; a full-color transparency and line-work. The color transparency represents the most common type of original scanned and includes
areas of high and low frequency (fine detail). The line-work will allow easy
analysis of directional limitations.
to the default As every scanner is distinctly different it was necessary identify settings of each scanner in regards to color separation as well as unsharp mask
"normal" correlation to a similar ing. A setting in one program has little setting in another program.
30 Not only are the unsharp masking methods used by each of the scanners differ ent, the controls that they offer to the operator are also quite different. In some cases all the possible settings are accessible through the interface software, in other cases hardware adjustments are required. The degree and number of set tings available are also very different. Therefore, it was necessary to scan a wide range of settings on each of the scanners; from minimum USM settings to maxi mum USM settings to establish their parameters.
Experimentation was performed in the Electronic Color Imaging Laboratory
(ECIL) in the School of Printing Management & Sciences (SPMS) at the Rochester
Institute of Technology (RIT). Equipment within the lab(s) that was utilized included 3 different mid-range PMT/drum scanners. Digital proofing was accomplished at FujiFilm USA - Graphic Systems Division, in Itasca, Illinois
Input (Scanners/Software) Screen DT-S1030AI / DT-S Scan 3.4 Howtek ScanMaster 4500 / Trident 2.0 Optronics ColorGetter Eagle / ColorRight 5.1
Processing (CPU) PowerMacintosh 8600/200 (65MB RAM)
Output (Proofing System) FujiFilm FirstLook Digital Proofer
settings completed digital With several different scans of varying USM being
to the individual controls as proofs of each scan were generated to help identify
scanners. This allowed for a more well as the similar USM levels between the
31 accurate analysis of the unsharp masking methods. Detailed analysis of the unsharp masking settings were accomplished through normal viewing condi tions as well as under high levels of magnification. The observations of the
author and other industry professionals were recorded to provide a summary of
the evaluation results.
32 Chapter 6
Results of Evaluation
USM Controls on midrange scanners evaluated
of Each the three midrange/PMT scanners used in this analysis utilize unique scanner interface software. The controls for gradation/tone, color correction, and unsharp masking are quite different from one application to the next. The unsharp masking controls of each scanner interface software were evaluated for their controls as well as their effectiveness. Each of the three scanner interface software applications evaluated had a set of preset unsharp masking levels in addition to custom settings that could be accessed to further manipulate and customize the effects of unsharp masking.
The USM controls between each of the scanner interface software applications used showed little resemblance to one another at first inspection. However, when each of the individual controls was identified, similarities between the applications became apparent. In Table 6.1 the similar controls of each applica tion can be identified. It is also necessary to note that although there may be like controls among the applications, the manipulation and implementation of each is very different. For instance, a level "10"of any of the controls in one applica
"10" tion is not equal to that of a level in another program.
A linework image was scanned on each of the three scanners for identification of possible directional limitations of the three unsharp masking techniques used. In 33 USM Control Screen Howtek Optronics (a.k.a.) DT-S1030AI ScanMaster ColorGetter (w/ ResEdit) 4500 Eagle
Width Aperture (APT) Radius Diameter (Pixels) (Fringe)
Height Whiteline (HG) Sharpen Sharpen Amount (Intensity) & Blackline (UT)
Limits Grain (GT) Threshold Threshold (Sensitivity)
Smoothing n/a Smooth Smoothing
Color Control Green Filter RGB Control RGB Sensitivity (Cross Coupling)
Tonal Control n/a Region Control Light-Dark (USM start/stop) Cutoff
Signal Contrast n/a Shape n/a
Chromatic n/a Mono/Color n/a Borders Edges
Separate WL/BL Whiteline (HG) Light/Dark n/a control & Blackline (UT) Edges
Table 6.1 - Controls of USM on scanners evaluated order to see the effects of USM, an extended range was forced upon the image
(0.0 highlight and 3.0 shadow). In addition, very high levels of USM were applied so that the effects would be more visible.
34 Photoshop 4.0
Although Photoshop was not evaluated, it is helpful to use this application as a reference because it uses technology and terminology that is familiar to many users of desktop scanners and applications. In addition to the preset levels that
are often used by the novice users, Photoshop offers a fair amount of manipula
tion of unsharp masking with just three basic controls (Figure 6.1). Amount con
trols the intensity of the white and black borders; it has a range of control from
1% to 500%. Radius controls the width of the white and black borders, with a
of range of control from .1 to 250.0 pixels. Threshold controls the sensitivity
unsharp masking; its sensitivity ranges from 0 to 255 levels. Since Photoshop
also allows for area and density masking it is quite easy to apply different
amounts of unsharp masking throughout the image effectively. It is also some
times suggested that different levels of unsharp masking be applied to each of
the individual CMYK channels.1
Unsharp Mask
Last Filter MT Fade... 1>:*F OK
Rrtistic ? ( Cancel J Blur ? Brush Strokes ? 13 Preuiem Distort ? Noise ? El 200*5 H PiHelate ? Render ? Rmount: 50 % Sharpen ? 1 Sharpen Sketch Sharpen Edges ? Radius: 1.0 pixels Stylize Sharpen More ? L3 TeKture ? Unsharp Mask... | Threshold: leuels Uideo ? i- Other ?
Digimarc ?
Figure 6.1 - USM controls in Adobe Photoshop 4.0
35 DT-S Scan 3.4
The DT-S Scan interface is the least complicated of those tested. It also offers the least amount of control of unsharp masking. Since the Screen scanner uses opti cal unsharp masking, the controls are primarily the aperture and the signal strength of the unsharp masking PMT. There are only six levels of USM to choose from (Figure 6.2). The limited control of unsharp masking makes it easier for the less experienced operator to attain acceptable results, however, the more experienced operator may desire more control than is offered. While not publicly
divulged by Screen, their engineers have identified an alternate method for expe
rienced users to further modify USM on the DT-S 1030AI. The preset USM levels
in DT-S Scan 3.4 are: Off, Very Soft, Soft, Normal, Sharp, and Very Sharp.
Window | =U ColorSetup (RGB) _ Color Setup U Trimming Setup ST HD SD o
t Zoom In Prescan 8+ R: 0.00 3.90 |] * o: Zoom Out Prescan 86- ? G: 0.00 |3.90 |] * o: Default Color Setup ? * B: 0.00 3.90 J MI^
Output: [245 ] 0
Tone curue: Standard Color: | Standard |
? OFF rSet Default ] Uery Soft Sc)ft
N Dtmn srlarp u Figure 6.2 - USM controls in DT-S Scan 3.4 36 Until the more recent versions of DT-S Scan, the applications engineers from Screen responsible for performing installations of the DT-S1030AI scanners found it necessary to modify the USM portion of the application using the resource editor ResEdit (by Apple Computer). The reason for the modification was primarily due to the fact that the default settings were targeted to the "over-saturate" "over-sharpen" Japanese market (with a tendency to and the "soft" scans). Even the level was too sharp. With the addition of ResEdit, further control of the USM controls in DT-S Scan can be achieved. There are four controls that effect unsharp masking; aperture (APT), white line (HG), black line (UT), and grain (GT). (Figure 6.3). Off, Normal, and Very Sharp were used for the preliminary scans of this research. A series of 21 scans were made on the Screen DT-S 1030AI using the DT-S Scan software interface and the Res-Edit resource editor for the purposes of (Figure evaluating the individual controls and their effects on unsharp masking scanner/software for 6.4). Finally a linework image was scanned to evaluate the directional limitations. Aperture. The Screen DT-S 1030AI scans with two apertures simultaneously. They pre-determined the are the Main and USM apertures, which are matched pairs, by the personnel at Screen USA manufacturer. Despite many attempts by the author, of their apertures or the ratio of the Main to were unwilling to divulge the size interviewed the author recommended that the USM aperture. Screen engineers by manipulate USM. There are seven resolu aperture pairs not be altered in order to on the DT-S1030AI. tion dependent apertures available for transparency scanning 37 JSR DCS U dctb DITL DLQG OOIO 100 1 oi io ioio if8 OOOI 1114 FREF Goma GT15 HG15 HG30 0101 I 101 DD D D 01 10 IOI0 OOOI I I 10 A 0 1 00 0000 ICON ics4 ics8 MENU MTYP 0010 IOC I OOI 0 1001 0010 100 1 01 10 1010 W % OOOI I I 10 mm h PF.EF 0 100 0000 Nega PAT* OFVs Pit ppat PREF prof DCD 01 10 1010 01000000 01000000 0100 oo< RPM SSAS STR SYRS 2.0bl e.o.s 7.0... TMPL vers a Figure 6.3 - ResEdit interface of DT-S Scan 3.4 For the purposes of evaluation, the default aperture (#2) was used. A larger aper ture (#1) and a smaller aperture (#3) were selected using ResEdit, against the rec ommendations of the Screen engineers. The default aperture for resolutions between 279 and 441 spi is aperture #2 and was automatically selected for the 300 spi resolution used in this study. As expected, the smaller of the three aper tures produced an image with more fine detail and narrower white line/black line borders, however proofs showed the effects of subject moire (interference patterns) in the image. The larger aperture produced an image that was smoother overall, with less fine detail and wider white line/black lines. It was 38 found that altering the aperture pairs had a notable effect on USM and should be considered as a viable option by the operator. White Line/Black Line. The white line (HG) and black line (UT) controls, which are individually accessible with ResEdit, determine the intensity of the white line/black line borders. Separate control of the white line/black line should pro vide a broader level of control for USM. Eleven intensities from 0 to 10 can be selected for both white line (HG) and black line (UT). The default (preset "Normal") value for white line (HG) is 0, while the default value for black line (UT) is 6. With the use of ResEdit it was noted that the there is no variation between any of the preset levels for white line (HG) which the author found quite odd. The set tings for black line (UT), however, do increase incrementally. It was observed upon evaluation of the scans that the higher the HG value, the lesser the white line intensity, while the black line remained unaffected. With the black line (UT) control, the higher the value, the greater the black line intensity, while the white line remained unaffected. It is also interesting to note that the zero (0) value of white line is the greatest intensity level and the maximum (10) value of black line is the greatest intensity level. Grain. The grain (GT) control serves as the threshold setting for DT-S Scan. Grain determines how sensitive the application is to density shifts and whether or not a white line/black line border should be applied. The greater the number selected, the less the image will be sharpened. Eleven intensities 39 from can 0 to 10 be selected for grain (GT). The default (preset "Normal") value for grain (GT) is 0. it was Again, perplexing to observe that there was no variation between any of the preset levels for grain (GT). As anticipated, scans revealed that higher GT values yielded a smoother image, leaving the white line/black line (HG/UT) intensities relatively unaffected in areas with the greatest density shifts. Lower GT values allowed a greater portion of the image to be sharpened. Directional Characteristics. As previously stated, optical USM is applied in the direction of the drum rotation during the scan . The highest levels of USM were evaluated at high magnification on two different images; the color transparency and the linework. Due to this fact, it was predicted that the scanned images would exhibit some noticeable directional anomalies. The color image appears to show some directional limitations, but surprisingly the linework appears to pro duce a consistent white line/black line border in all directions (non-directional effect), (Figure 6.5). 40 Aperture (APT-#3) White line (HG-9) Black line (UT-9) Grain (GT-9) Figure 6.4 - Screen USM effects 41 '**//shed & Figure 6.5 - Directional characteristics of Screen DT-S 1030AI 42 Trident 2.0 The Trident interface software offers a tremendous amount of control of USM. Upon initial inspection, the USM interface appears to be overwhelmingly com plex. It includes more than ten controls for USM alone (more if every single indi vidual control is counted). Trident offers two levels of control within the applica tion: Production Controls and Expert Controls (Figure 6.6). The Production Controls are the most basic and offer several preset USM levels in addition to four basic controls (Sharpen, Smooth, Shape, and Radius). The Expert Controls allow for greater manipulation (including Sharpen, Threshold, Smooth, Shape, and Radius, Region Control, and RGB Control). There is some repetition of con trols between the two levels; all ranges of control are from 0 to 100% (except for Sharpen and Radius in Production Controls which are 0 to 500%). USM Off, Default Sharp, and Sharpen More 85 were used for the preliminary scans of this research. A series of 21 additional scans were done on the Howtek ScanMaster 4500 using the Trident software interface for the purposes of evaluat ing the individual controls and their effects on unsharp masking (Figure 6.7). Finally a linework image was scanned to evaluate the scanner/software for directional limitations. 43 D== Production Controls n . UnSharp Masking Gradation Tone USM UnSharp Mask : [y] Qn Region Control ; Wk Sot: 1 Trident Defaults Dot Bate S | Trident Defaults Gradation ) Trident Defaults Too*: | Trident i Sharpen : USM: 1 Trident Defaults 35mm 100-250SS j 35mm 1 25 1 -and upS5 Threshold 35mm 251-500SS L 40 35mm 501-750SS 20j Total Ink 35mm 751-1250SS Default Sharp 0 li i ^ Max Black Refl Sharpen 0 | 20| | 40 | 60| | 80| [ 1 00 Refl Sharpen More r RGB Control - Sharpen Less 25 <) Mono Edges O Color Edges Sharpen More 55 R I 100-0[ [ 100.o| Sharpen More 85 F titer I ttr.tr ol USM OFF 0 Light Edges 0 Dark Edges Shape : Smooth : * i ioo| Jk. Shape : 45.000| .A Radius : 60.000 1 Figure 6.6- USM controls in Trident 2.0 Aperture. As the size of the aperture can have some effect on the detail of the image, scans were performed with three different aperture settings. The Howtek ScanMaster 4500 scans with a single aperture. The default aperture at the speci fied resolution is 83p, scans were also done with a 64p and 102p aperture. While it is obvious that there would not be any optical USM characteristics previous experience would lend one visible from altering the aperture, to believe that the smaller aperture would still increase the amount of fine detail captured during the scan. Upon observation of the scans, there was no per ceptible difference in the detail characteristics of the image. Therefore it 44 would seem that changing the aperture on the ScanMaster 4500 does not effect detail enhancement. Sharpen. The Sharpen field governs the amount of sharpening applied to a select ed tonal It is the intensity control of Trident, which determines the white line/black line contrast. This application does not offer separate control of the white line/black line borders, however the user may select to only apply the white or black borders instead of using both (Light Edges/Dark Edges). Scans of increasing Sharpen levels were performed. It was determined through observation that the higher the Sharpen value the more intense the white line/black line. Additional scans were performed to test the effects of the Tight Edges/Dark Edges option. At the level of sharpening used there was no percep tible change between the three scans (white line/black line, white line only, and black line only), it is possible that if higher sharpening levels were applied that the expected effects would be more visible. Threshold. The Threshold number refers to the minimum difference between dark and light areas that must exist before unsharp masking is applied. The higher the Threshold setting the less sensitive the application is to density shifts, requir ing a greater density shift in order to apply any white line/black line effect. Scans revealed that higher Threshold values produced a smoother image, with line/black line borders. Tower only the areas of greater contrast receiving white sharpened with Threshold values permitted a greater portion of the image to be 45 white line/black line borders. Scans with lower Threshold values also appeared to have more grain and noise artifacts. This control performed as expected and described above. Smooth. Trident's unsharp masking function also allows you to smooth images. Increasing the Smooth value helps to diminish the film's graininess. Smooth enables softening of areas rejected for sharpening by the sharpen threshold. Evaluation scans showed that a higher Smooth value resulted in more smooth ing of lower contrast edges. Unexpectedly, it also appears to have increased the overall sharpness of the image. Shape. Shape governs how quickly the colors ramp from gray to white. The tran sition from the white line/black line and an adjacent density at the detail edges is normally harsh and high contrast which causes the eye to perceive the edge as detail. The Shape control allows the operator to control how much contrast is applied at the detail edge. Scans with a higher Shape value appear to have a sharper shift between the areas. also seemed white line/black line edge and the adjacent lower contrast It intensities are effected the from observations that the white line /black line by effect is similar to the manipulation of the Shape control, so the resulting interface application that the author was Sharpen control. Trident was the only preconceived expectations of this exposed to to use this control, there were no control. 46 Radius. The Radius is a measure of how wide the halo effect is at the edges of the tonal It is the measure of the width of the white line/black line border. The wider the border, the greater the likelihood that the human eye will perceive the resulting contrast as detail. The results of the evaluation of the test scans were as expected. The higher radius values exhibited wider white line/black line borders and gave the appearance of greater detail to the image. Tower radius values resulted in nar rower white line/black line borders and less perceived detail. RGB Control. Trident's unsharp masking also allows you to choose mono-colored edges or colored edges between tonal regions. Mono-Edges are made up of a sin gle color, mixed in the ratio you define in the R,G, and B fields. Color Edges are a composite of surrounding colors. Generally, the edge setting is a matter of indi vidual preference. Some images tend to look better with mono-edges. Either Mono-Edges or Color Edges may be selected (only one check box can be active at a time) and work in conjunction RGB Control which applies a weighted intensity to the different (RGB) channels. Mono-Edges is the default selection When Color Edges is selected colors appear to be sharper in the image while neutral areas have less perceptible detail. The weighted intensities of the RGB while how Control were left at their (100%) default values. This control, affecting of specif USM is applied to various hues within the image did not give the level ic control desired. 47 Directional Characteristics. As detail enhancement is a function of post-processing in Digital USM, the addition of white line/black line borders can happen in all directions equally. With this understanding, it was predicted that the scanned images would not exhibit any noticeable directional anomalies. The color image shows no directional limitations, as well, the linework appears to produce a con sistent white line/black line border in all directions (Figure 6.8) . 48 Sharpen (95%) Dark Edges / Tight Edges / Color Edges / Figure 6.7 - Howtek USM effects 49 Qf ^//shed - ScanMaster 4500 Figure 6.8 Directional characteristics ofHowtek 50 ColorRight 5.1 ColorRight's USM controls are not overly complex, but give an adequate amount of control. ColorRight has four preset levels of USM available in addition to the Custom controls (Off, Moderate, Normal, Considerable). There are essentially five controls for USM; pixel diameter, sharpen amount, sharpen threshold, light/dark cutoff, and RGB sensitivity (Figure 6.9). The interface is relatively straight forward, the first three controls being common to most USM interfaces, including Photoshop. Off, Normal, and Considerable were used for the preliminary scans of this research. A series of 18 additional scans were done on the Optronics ColorGetter Eagle using the ColorRight software interface for the purposes of evaluating the individual controls and their effects on unsharp masking (Figure 6.10). Finally a linework image was scanned to evaluate the scanner/software for directional limitations. Ouerall Gradation... JI_ Color Cast Remoual... Diameter (pixels): 5 HE- Scan Geometry... Sharpen Amount: 50 % Inuert Sharpen Threshold: 0 % ? Smoothing H Flip V Flip Light Cutoff: 5 % Dark Cutoff: 95 % Sharpen ? ? Off Red Sensitiuity: 30 % Moderate Green Sensitiuity: 59 % Normal [Reset Considerable Dlue Sensitiuity: 1 1 % IT) (T) Custom... Figure 6.9 - USM controls in ColorRight 5.1 51 Aperture. Typically, the selection of the aperture will have an effect on image detail. The Optronics ColorGetter Eagle uses a single aperture and simulates the USM aperture with buffered information from preceding scan-lines (see Diameter below). The image was scanned with the default lOOp aperture as well as the 50p and 200p apertures. It was observed that the smaller aperture captured more fine detail, to the point that even small text was almost readable, however, the image suffered from some subject moire (interference patterns). When the larger aperture was select ed the image was smoother overall and held less fine detail. Regardless of the aperture selected, the width of the white line /black line remained consistent (unlike optical USM), but the intensity increased with the smaller aperture, which was unexpected. Diameter. This slide control allows for selecting one of the allowable number of pixels over which the sharpening takes place. The minimum is 3, the maximum is 9, and the default is 5. The control simulates unsharp masking apertures. The width of white/black bor Diameter control is essentially responsible for the the ders, it refers to the area (from 3 to 9 pixels) that will be averaged to create the virtual USM aperture (area). unsharp signal; it can be thought of as a was the Evaluation of the scans showed that as the diameter value increased, appeared sharper. white line/black line border widened and the image the white line/black line bor Conversely, as the diameter value was decreased, sharp. der became more narrow and the image appeared less 52 Sharpen Amount. This slide control allows for setting the amount (intensity) of the darkening and lightening that is applied on each side of the tone Sharpen Amount refers to the intensity of the white/black border, ColorRight does not offer separate control of the white and black borders. The control range is 0 to 100%. It was observed that the higher the Sharpen Amount value, the greater the inten sity of the white line/black line border. As predicted, this gave the perceived effect of greater detail in the image. Reducing the Sharpen Amount lessened the perceived detail by decreasing the intensity of the white line/black line border. Sharpen Threshold. This slide control allows for choosing a tone intensity differ ence between tones beyond which sharpening will take place. This restricts sharpening from taking place between similar tone areas and is good for reduc "Smoothing" ing grain in big enlargements. The check box enables softening of areas rejected for sharpening by the sharpen threshold. The Threshold control determines the necessary density shift in order to apply USM. It was observed that the higher the Threshold value, the less sensitive the appli cation/scanner is to density shifts, requiring a greater shift in order to produce the white/black border effect. With Smoothing selected it was noted that those areas within the threshold limits were smoothed, however, because there is no numerical value associated to Smoothing, it is not evident to what degree that smoothing had taken place. 53 Light/Dark Cutoff. Two additional controls give more customization of USM in ColorRight. These slide controls allow for adjusting the level of lightness in the highlight area under which no sharpening will take place and the level of dark ness in shadow the areas over which no sharpening will take Tight/Dark Cutoff allows the operator to set a start and stop point for unsharp masking. It is possible to limit USM to a specific tonal range (e.g. sharpen only the 20%-60% tones). The Tight Cutoff value cannot exceed the Dark Cutoff value (or vice versa). When cutoff values were set at 5-50%, sharpening was limited to the highlight to midtone regions. In the successive scan the cutoff values were set at 50-95%, which limited sharpening to the midtone to shadows. No sharpening was applied to those areas outside of the cutoffs. Sensitivity. Sharpening occurs when there is an abrupt change in density from one area to another. When the colors are of different densities, the sharpening process is able to distinguish between them. When the colors are nearly the same density, there is no way to distinguish between them unless a filter is used to cre ate the mask data. By using appropriate filters, the color of concern appears at a different density in the mask data. In producing digitized data through a filter, the various colors appear as black or a white in the mask data. "This allows for some control over which hues will receive the greatest amounts of unsharp masking (Table 6.2). The cumulative total of all three channels cannot exceed 100%. 54 Red Filter Green Filter Blue Filter White Mask Black Mask White Mask Black Mask White Mask Black Mask Red Cyan Green Magenta Blue Yellow Magenta Green Cyan Red Cyan Red Yellow Blue Yellow Blue Magenta Green White Black White Black White Black Table 6.1 - Mask colors resultingfrom color sensitivityfilters Some general observations that were identified include: With primary Red Sensitivity, the reds (and yellows to a lesser degree) appeared smooth, while the blues were sharper. When Green Sensitivity was dominant, the greens were smooth, while yellow and reds exhibited greater sharpness. When the Blue are the Sensitivity is greatest, blues are slightly smoother, while yellows and reds most sharp. Directional Characteristics. Since hybrid USM takes place via digital processing expected if there were directional anomalies, on-the-fly it was that any they /black line borders would would be minor, and that the addition of white line The color image showed no directional limita occur effectively in all directions. consistent white line/black line tions. The linework also appeared to produce a border in all directions (Figure 6.11) . 55 I Lw kor" Color RenditionChart Threshold Smoothing / (50%) Pixel Diameter (9) Cutoff (5-50) Red Sensitivity Green Sensitivity Blue Sensitivity Figure 6.10- Optronics USM effects 56 tf '*A//shed * - ColorGetter Eagle Figure 6.22 Directional characteristics of Optronics 57 Endnotes for Chapter 6 1 Marguiles, Dan. Electronic Publishing, February 1998, p46-50 2 Howtek, Inc. Trident, User's Guide. 1996, p6-3 3 Howtek, Inc. Trident, User's Guicde. 1996, p6-3 4 Howtek, Inc. Trident, User's Guide. 1996, p6-3 5 Howtek, Inc. Trident, User's Guide. 1996, p6-4 6 Howtek, Inc. Trident, User's Guicde. 1996, p6-5 7 Howtek, Inc. Trident, User's Guide. 1996, p6-5 8 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996, p4-55 9 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996, p4-55 10 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996, p4-55 11 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996, p4-56 12 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996, p4-55 58 Chapter 7 Summary and Conclusions There were three basic objectives that this thesis intended to accomplish. The first objective was to provide a thorough reference on unsharp masking. The sec ond objective was to evaluate three scanner interface applications with specific focus on the unsharp masking. The final objective was to evaluate the effective ness of three distinct unsharp masking methods used today on mid-range PMT/drum scanners, including optical USM, digital USM, and hybrid USM. USM Reference. Due to the notable lack of quality material published on the topic of unsharp masking, it was the author's desire to provide an extensive reference of available unsharp masking techniques. This has been accomplished in part by the accu mulation and analysis of the limited references available on unsharp masking as well as conducting interviews with a number of graphic arts industry profes sionals and educators. Chapters 2 and 6 of this thesis provide a good reference of unsharp masking information for future researchers. While the primary focus of this thesis was optical digital and electronic unsharp masking techniques, including USM, USM, hybrid USM, definitions and descriptions of alternate unsharp masking tech digital USM have niques, including photomechanical USM and post-processing furnishes a comprehensive list been provided. The bibliography provided herein 59 of resources available at the date of publish of this thesis. These references can be consulted for further detail on the subject. USM Interface The three scanner interface applications that were evaluated each have advan tages and disadvantages. They are all acceptable for the purposes of electronic color scanning and are capable of achieving adequate unsharp masking results. DT-S Scan offers the least complicated interface to unsharp masking of the three applications evaluated. Unfortunately, it also offers the least amount of control. The application offers no real customization of the unsharp masking settings other than the selection of six built in intensities. With the addition of ResEdit, a resource editing application developed by Apple Computer, further control of the unsharp masking settings can be obtained, although this method of manipu lating the DT-S Scan is something that only more experienced operators should consider. The specific effects obtained through the manipulation of these settings are not readily obvious by viewing the controls, therefore, a considerable learn ing curve is required in order for the user to become proficient in this method. Trident offers the greatest amount of control of the three applications evalu ated. There are two levels of control within the software that allow less expe acceptable scans. The Production rienced operators the ability to produce well as the Controls offer several preset levels of unsharp masking as ability masking. More experienced operators to make adjustments to the unsharp 60 have access to Expert Controls that allow greater control of the unsharp masking settings. The ColorRight interface offers a good balance of control while maintaining an to understand interface. The controls not easy are overly complex and the effects of adjusting the settings are logical. Because unsharp masking is applied during high-resolution scanning, all three applications allow for high-resolution viewing. The Screen scanner will scan the entire image. The a Optronics scans user selected area applying unsharp mask ing. The Howtek the application will is only that simulate unsharp masking to the low-resolution file as well as perform a high resolution scan. Of the three scanner interface applications that were evaluated, it was Optronics ColorRight 5.1 that offered the best combination of user friendliness (ease-of-use) and control. The application was neither too complicated nor too restricted in it's controls. The DT-S Scan 3.4 application was the easiest to use, but offered very little control. The ResEdit work-around to DT-S Scan allowed for more control, but any true graphical interface was sacrificed, making manipulation much more difficult to perform accurately. The Trident 2.0 application allowed ample unsharp masking control, however, the complex interface tended to cause more confusion than it was worth. If the operator works with the Production Controls only, there are fewer variables and far less confusion. 61 The digital USM method and interface is the most efficient for the scanner manu facturers to implement. The digital method is only software, it requires no addi tional hardware on the scanner. It can easily be enhanced with future software upgrades. Many manufacturers who previously used optical and hybrid USM methods have since replaced them with digital USM methods. The Screen DT-S 1040AI, which replaces the DT-S 1030AI uses digital USM, in addition, all of Screen's high-end digital scanners now also use Digital USM. Optronics International, who have recently restructured, eliminated their software (ColorRight) development department in lieu of using a third-party developed interface. The third-party interface happens to be Trident (re-named ColorRight Plus by Optronics), so now Optronics USM method is digital as well. USM Effectiveness were made in the Based on the initial research on USM, several statements statements revolve hypothesis that were to be proven or disproven. These - of the USM data and around two basic concerns speed of scanning/processing of the USM identification of directional limitations in the scanning/processing and conclusions made: data. These statements are presented again here, on midrange scanners. 1.1) Optical USM is the fastest USM method in use scanners. method in use on midrange 2.1) Digital USM is the slowest USM scanners. method in use on midrange 3.1) Hybrid USM is a fast USM possible nor disproven. It was not to These statements could neither be proven distinct dif- has on scan time due to determine the effect that unsharp masking 62 ferences in the technology of the three different scanners. Between the drum rotation speed and traverse (stepping) mechanism it is impossible to truly deduce the effect of USM processing on the total scan time. Screen DT-S 1030AI 0:57 min @ 900rpm (separate calibration) Howtek ScanMaster 1:25 min @ 600rpm (2:07 w/ calibration) Optronics ColorGetter 1:55 min @ 1200rpm (2:00 w/ calibration) Table 7.1 - Scanning times 1.2) Optical USM is limited in its effectiveness directionally (lowest quality). 2.2) Digital USM is not limited in its effectiveness directionally (highest quality). 3.2) Hybrid USM is not limited in its effectiveness directionally. Statement 1.2 was disproven, while 2.2 and 3.2 were proven. It was determined directional limita that none of the three USM methods evaluated showed any evidence of direc tions that could be classified as quality issues. The only any on the Screen tional limitation occurred when scanning the color transparency minimal. DT-S 1030AI, which utilizes optical USM, however the effect was Recommendations for future research to determine how As it was not possible in the confines of this study unsharp it would be of interest to perform masking affects the scanning/processing time, which the specific time requirements a study under controlled conditions during time the USM could be identified. The to of USM scanning/processing ability 63 functions is made difficult because of the differences in scanner technology, how ever this could study now take place due to some changes in Optronics choice of technology. Since now Optronics essentially uses the Trident software this would allow a side-by-side comparison of the hybrid and digital USM techniques on the same scanner. It was originally suspected that directional limitations would have had a more prominent effect on scans from the Screen DT-S 1030AI scanner utilizing optical USM than was actually realized. The prediction of directional limitations was based on observances previously noted on older analog high-end scanners. While some directional effect can be identified, it was expected that this would occur primarily in the scan/spin direction of the image. Further study could be conducted to determine what exactly causes the directional anomalies and how they are best controlled. It might be suggested that the Screen DT-S 1030AI be compared to an analog high-end scanner such as the Screen SG-608. As digital USM on the scanner has now become the method of choice by scanner as with the manufacturers, a comparison between on-scanner methods (such Howtek ScanMaster/Trident combination) and post-processing methods (such manipulation would be as PhotoShop, Tive Picture, and other image programs) methods applied to both PMT and very appropriate. A comparison of these same CCD scanners would also yield some useful information. 64 Conclusion process. Unsharp Masking is a vital part of the electronic color separation management sys Despite improved and simplified scanner interfaces and color in fact it tems the necessity of unsharp masking has and will not disappear, may separation. become more important to understand its role in quality color basic con Enhancements in technology continue to improve the results, but the cept remains the same. 65 Bibliography 66 Bibliography Agfa-Gevaert N.V. An Introduction to Digital Scanning. 1994 BarneyScan Corporation. CIS-ColorAccess, User Notes. 1992 Baxes, Gregory. Digital Image Processing. 1994 Blatner, David and Roth Steve. Real World Scanning and Halftones. 1993 Chrusciel, Edward and Rogers, Don. What To Look Tor In a Color Scanner, Graphic Arts Monthly. October 1993 Crosfield Electronic Timited. MagnaScan Plus, Operators Manual. 1992 Dainippon Screen Ttd. DTS Scan, Reference Guide (ver. 3.3). 1993 Dainippon Screen Ttd. Scangraph Technical Guide. 1985 Field, Gary G.Tone and Color Correction. 1991 Field, Gary G. Color and Its Reproduction. 1988 Green, Phil. Understanding Digital Color. 1995 Howtek, Inc. ScanMaster 2500 Plug-In Module, User's Guide (ver 1.1). 1995 Howtek, Inc. Trident, User's Guide. 1996 Howtek, Inc. TWAIN Source Seamier Module, User's Guide. 1996 Jackson, Tonnie T. Unsharp Masking - Photographic/Electronic, GATF, SecondSight. 1989 Tinotype-Hell AG. NezvColor 3000, User's Guide. 1993 Marguiles, Dan. Electronic Publishing, February 1998, Molla, Dr. R.K. Electronic Color Separation. 1988 Noga, Joseph. Color Separation Systems, JPRT-409. School of Printing, RIT 1997 Optronics, Inc. ColorRight 5, User's Guide (ver .1). 1996 PixelCraft, Inc. ColorAccess, User's Guide. 1993 67 Southworth, Miles. Color Separation Techniques. 1989 Rich, Jim and Bozek, Sandy. Photoshop in Back and White. 1995 Yule, John. Principles of Color Reproduction. 1967 68 Appendices 69 Appendix A - Photoshop 4.0 Screen Captures 70 Licensed to: P.IT SPMS PSW400P.7 1 0053 1 -999-329 Adobefhotoshop* PowerPC7 4.0v4 0.1 Thomas Knoll, Mark Hamburg, Seetha Narayanan, Sean Parent, Greg Cilley, Laura Hoffman, Jason Bartell, Scott Byer, Allen Chan, Jeff Chien, Tom Costa, David DiCiacomo, Andrei Herasimchuk, Charles McBrian, Marc Pawliger, Anapathur Ramesh, Akiko Sonoda, Robert Swirsky-Warner, Doug Ahmann, Doug Olson, Paul Holland, Andrew Coven, John Leddy, Kevin Connor, Russell Brown Protected by U.S. Patents 5,1 46,346, 5,546,528, and 4,837,613. Patents pending. 1989-1997 Adobe Systems, Incorporated. All rights reserved. Adobe, the Adobe logo and Photoshop are trademarks of Adobe Systems, Inc. CALIBRATED PANTONE and PANTONE are trademarks of Pantone, Inc. MacApp 1985-1993 Apple Computer, Inc. - - Mask Last Filter Unsharp facie... 0K )| Rrtistic l Blur [ Cancel j Brush Strokes Distort Preuieu) Noise PiHelate ED 20oxEi Render Sharpen Rmount: \so % ' Sharpen Edges A Sharpen More Radius: |..o piHels mrna Threshold: I" leuels A 71 Appendix B - DT-S Scan 3.4 Screen Captures 72 Standalone driuer software for macintosh DTS Scan Version 3.4 ColorSetup (RGB) I HD Window | R: ||0.00 Color Setup mu |ffl OS Trimming Setup &T G: llo.OO Id 3.90 OS Zoom In Prescan - b: Zoom Out Pi est an 3t: \\oW]ft gjgjg OS Default Color Setup ? Output: KJJUffl \^ZJ& Tonecurue:| Standard | Color: | Standard | D. Set Default Uery Soft Soft Normal Sharp Uery Sharp 73 Appendix C - ResEdit Screen Captures 74 ResEdit ::::::""": 2.1.3 m !! .=:; : a This version \>y: 1 * SumitBaruio s & SamirariBassir: " 5 5 Copyright 193^-199-^ APPfc Corner, ]fc ""KB All rights reserved 1D1 DTS Scan 3.40 siisiiri %? Hi::::; liliiir! ::!:::;: 0*S ALRT APT BNDL CCor cfrg CGLt cicn CNTL 'we o i i niliiii CODE CURS DATA DCSU dctb DITL DLOO DTSS 3D FREF Goma GT1 5 GT30 HG1 5 HG50 10IS D <^... A/z\ :r::::;: ;::;;;;: Hh JJ '<* ''-'.: OFVs PARM PAT* PICT ppat PREF STR" SYRS RPM SEN SEE SNAM SSAS 5TR 20bl 605 7.0 .. VDEF WIND TMPL UT15 UT30 VARN vers wctb 75 iiH HG30s from OTS Scan "DT-SI030" 3.40 ^HN ID HPT ID - 16001 from DTS Scan 3.40 It Size Name Soft" 16000 126 "Very O Number of 7 "Soft" 16001 126 HP List "Normal" 16002 126 11 ***** "Sharp" 16003 126 Sharp" NO. Il 16004 1 26 "Very Max 1 279 Resolution!! ransparentl MaH 1 308 9 Resolutions III ills Horn on Scan 5.10 eflectionl E. Size Name Soft" 0 16000 126 "Very dummy "Soft" 16001 126 2) ***** "Normal" 16002 126 "Sharp" NO. 16003 126 |2 Sharp" 16004 126 "Very Maw |441 Resolution!! ransparentl MaK [523 1 Resolutions eflection) GTJOs from DTS Scon 3.40 ID Size Name dummy 0 Soft" 16000 126 "Very ^j ***** "Soft" 16001 126 "Normal" NO. |3 16002 126 "Sharp" 16003 126 MaH 699 Sharp" 1 16004 126 "Very Resolution!! ransparentl % MaH 91 5 1 a Soft" Dg HG30 "Uery 10 = 16000 from DTS Scan 3.40 Posi CMYKIHPI) CMVK(HP2) CMVK(RP3) CMVK(RP4I CMVKIRP5I CMVKIRP6I CMVKIRP7) 'Soft" IPI1H HG30 ID = 16001 from DTS Scan 3.40 W^m Posi 1 CMVKIRPI) CMVKIRP2) 1 _J CMVK(RP3) 1 1 CMYK(RP4) |o 1 CMVKIRP51 1 1 CMVKIRP61 I4 <> CMVKIRP7I I6 ' Normal" IlJ^ HG30 ID = 16002 from DTS Scan 3.40 es <> Posi 1 CMVKlflPl) CMVKIRP2I 1 CMVMRP3) 1 CMVKIRP4) 1 1 CMVKIHP5I jo j CMVK1RP6) I4 ! CMVK(HP7) I6 1 a 'Sharp" IlJ^S HG30 ID = 16003 from DTS Scan 3.40 =s Posi 0 CMVKlflPl) CMVKIHP2) 1 | CMVKIRP3) 1 J CMVKIRP4I 1 ZJ CMVKIHP5I jo J CMVK1RP6I I4 j 6 CMVKIRP7I a Sharp" IlJ= HG30 "Uery ID = 16004 from DTS Scan 3.40 ss Posi |o l CMVKIRPI) r CMVKIHP2) |0 CMVK1HP3) |0 CMVK(HP4) |0 CMVKIRP5) |0 CMVK1HP6) |4 CMVKIHP7) J6 a 77 Soft" IDs UT30 "Uery ID = 16000 from DTS Scan 3.40 1=1 Posi |2 J CMVKIRPI) CMVK(RP2) |2 CMVKIHP3) [2 | CMVK(HP4I |2 CMVKIRP5) |2 CMVKIRP6) |l CMVK1HP7I |l | B Soft" IPa^SI UT30 IU = 16001 from DTS Scan 3.40 ^i^i^ Posi I4 CMVKIRPI) CMVK(RP2I !4 CMVKIHP3) J4 CMVKIRP4) I4 CMVKIRP5) I4 CMVKIHP6) [3 CMVKIRP7) J3 B "Normal" lO^S UT30 ID = 16002 from DTS Scan 3.40 SSI Posi 6 ZJ CMVKIRPI) CMYKIAP2I 6 ~J CMVKIRP3I 6 1 CMVKIHP4) 6 ~l CMVKIRP5) 6 CMVKIHP6) 5 J CMVKIRP7) 5 ~] B 'Sharp" lp= UT30 ID - 16003 from DTS Scan 3.40 === <> Posi I8 CMVKIRPI) CMVMHP2) Is 1 CMVK(flP3) I8 1 CMVKIHP4) I8 CMVKIAP5) Is 1 CMVK(flP6l |7 1 <> CMVKIHP71 17 1 a Sharp' 3.40 IDS UT30 "Uery ID 16004 from DTS Scan Posi |l0 ^ CMVK(HP1I CMVK(RP2) 1 1 0 CMVK(flP3) | 1 0 CMVK(flP4) |l0 CMVK(RP5) |l0 CMVK(RP6) |9 CMVKCHP7) |9 a 78 Soft" |D GT30 "Uery ID = 16000 from OTS Scan 3.40 ^^m Posi jo | CMVKlflPl) CMVKIAP2) |0 | CMVKIAP3I |0 j CMVKIRP4) JO | CMVKIRP5) |2 | CMVKIHP6I |4 CMVKIHP7I J8 j a "Soft" sn=l r,T3n in = lfiOOl from OTS Scan 3.40 w=^^ <> Posi 0 CMYK(flPI) CMVKIAP2) 0 CMYKIHP3) 0 CMVKIHP4) 0 CMVKIHP5) 2 CMYKIHP6) 4 CMVKIRP7) 8 a Normal" Il_l=l* GI30 " ID = 16002 from OTS Scan 3.40 ^= Posi 1 1 CMVKlflPl) CMVK(flP2) 1 CMYKIHP3) [o CMVKIHP4I |o CMVKIHP5) I2 CMVKIAP6) U <> CMVKIAP7I J8 B "Sharp" Scan 3.40 GT30 ID ; 16003 from DTS Posi 8 CMVKIAP1) CMVKIRP2) CMVKIRP3) CMVK1RP4) CMVKIHP5) CMVKIRP6) CMVKIHP7) Appendix D - Trident 2.0 Screen Captures 80 H i&B Right* : -..:.: sWt - UnSharp Masking = r . UnSharp Mask : \7\ fin HI Reqion Control : Sharpen : Region... *R .... j 35.0| Queue Manager... 3M Threshold: Probe... 36P 60 40 ' _ wm '_ 1 3.0I Descreening... sen 20; Smooth : Scan Info... XI 0 1 ool Focus f> Aperture... 0 | 20|| 40 1 | 60|| 80|; 100 Hide Toolbar *T () Mono Edges O Color Edges Unanchor Toolbar :#:N R | 100.o| G | 100.0| B | 100.0| [x] Light Edges 0 Dark Edges Shape: i \ | 45.0| * Radius : \ \ | 40_0| * j 81 Production Controls =0=^= Production Controls ^^^P Gradation Tone USM m Gradation Tone USM jmj Trident Defaults (**; -i J** $#*U j Trident Defaults -I &t fetid: P Trident Defaults -i ** *: | Trident Defaults " Ora*Htm f Trident Defaults -i GTMlatiMI j Trident Defaults - ramif Trident Defaults Twmtz Trident Defaults _-i USM: Trident Defaults -i tHJM: Trident Defaults 35mrn 100-2509? Update Tables [ 35mm 1251 -and up% 1 35mm O0CB *UCR 251-5009? 35mm 501-75095 35mm ?* mt^ 52o| start Bctf^ii m Tout b*: 751-125098 Default Sharp H* B1oltf^cT[ Star* BIfcf~Sl Max BI**k: Refl Sharpen Refl Sharpen More Sharpen Less 25 Sharpen More 55 Sharpen: | iqq] Sharpen More 35 USM OFF Smooth : 1 )oo| * Smooth: fToo] * .*. Shape: | 45 000 Shape: | 45_000| A. Jk. Radius: | 40.000 Radius: | 60.000 , ? Production Controls tasking Gradation Tone USM m UnSharp Mask: \~~] On S feSet;| Trident Defaults "* \ Region Control: Trident Defaults &tt^k)Xkm:\ Trident Defaults TWM&jj Trident Defaults WHJ USM OFF Update Tables [ Sharpen: o< < I 25.0| Ttal letf^o] 64[ Startup^ Threshold : - Htm 8aefcf~9o"| Start Bt<*{^26] * CM Smooth: Sharpen | too| * UJM 0 | 20| | 40| | 60 1 1 80| | 1 00] RGB Control Smooth | ioo| * () Mono Edges O Color Edges ? R | 100.0| G | IQQ p| Shape | 42.000 A. - Radius | 42.000| 82 Production Controls UnShorfj Mos k i n q Tone USM UnSharp Mask : Trident Defaults 13 Region Control : Trident Defaults Qrwfarfjfft j Trident Defaults " *J 1 Trident Defaults OSM: | 35mm \ Update Tables QOCK #UCR I 72-l H* Slate* f"^|m t*rt Blank W fH] zo\ Sharpen: | 1Qp| LULL 0 j 20| f 40| | 60]j 80| j 100; RGB Control Smooth : | tool * () Mono Edges O Color Edges R | 100 p| G | 100 o] B | 1QQ.p| Shape: | 40.000| Filter Control M Light Edges H *rk Edges Shape : R*dius: | 45 000| | 40.0| Radius : A 1 45.0| Production Controls UnSnarp Maskinn UnSharp Mask: [<] On Ml S*t:| Trident Defaults 3 Region Control : D< B*4b; | Trident Defaults 3 Gravities I Trident Defaults Trident Defaults 35mm 25l-500 I Update Tables Sharpen : QOCtt Threshold 40; 1 3.0| Smooth: 20: % ! 0; 1 -o| Sharpen : | |po| 6|| 20|l 40|| 60|| 80| 100 RGB Control Color Edges h: I I00| () Mono Edges O B R | 100.0] G | 100.o| | 100.0| Shape: 40 00o| | ? Light Edges x] Dark Edges Shape: Radius: 45.000 | [ | 40.0J* Radius : | 45.0| * / 83 Production Controls m^^zr. UnSharp Masking UnSharp Mask: ^ 0n Mfc Set! j Trident Defaults Region Control : Trident Defaults Trident Defaults Trident Defaults *"' 1 35mm 501 -750 | Update Tables Gee* %tm no ' MxBUfcf^o]y mm\mtalk^~~2^[% 40 5.0| 20 L1. _ HIKmm ^mmm 2.0| Sharpen: I iQpj W 0| 20|| 40l| 60|[ 80|; 100 - RGB Control - Smooth | ioo| * (.) Mono Edges '.._) Color Edges ? G [ 100 p| B | 1QQ.p| Shape Filter Control - | 40 .0001 Light Edges Dark Edges M. ^ [x] Shape : Radius | 55.0001 | 40.0| A Radius: / \ | 55.0| y =Q== Production Controls h.. UnSharp Marking Gradation Tone USM H^fj UnSharp Mask: KlOo GB fftfc S*t: | Trident Defaults "* j, Region Control: Dot 6aw: | Trident Defaults ? | 6r*4#tf** j TrTdent Defaults "* j Tow*: 1 Trident Defaults ^ | WtMt j 35mm 751-1250% ^\ J Update Tables } Sharpen: 100 QCC* #WC* 80; 1 ,8-l * Start * Threshold : TH1 tok^320 UCR^ 64] 60 Blac* %. Start Black 26 |* Max ^ 90 ] 40 1 6.0] : 20- Smooth 0 1 2-| Sharpen: | iqq| o'l 20|r~40|| 60!| 80| 100 Color Edges Smooth: 9t (a) Mono Edges O | iqq| 0 B ? R | ioo.o| | 1QO.0| | 100.o| - "ilter Control Shape: | 45.000| Light Edges Dark Edges 1 g] 3 Shape: Radius: 60.000 | 45.0| * L. | ^ Radius: | 60.01 * 1 /A 84 -[J AA Production Controls UnSharn Masking UnSharp Mask: 0 0n We S*t; | Trident Defaults Region Control : D**G*: Trident Defaults 3 Grad*tion | Trident Defaults Trident Defaults USM: j 35mm 1251 -and upSK ? I [ Update Tables Sharpen : Qeot $ucr 32 O] 8.0| ^^ *^^kmm^Bzm*^^m]^^mm Sharpen. | ioo| * CI3 0 | 20|| 40]| 60|| eo] i too ?GB Control Smooth * | 100 (i) Mono Edges O Color Edges R | 100.0 G | 100.0| B | 100.0| Shape | 45.000| Light Edges ^Dark Edges Jk. E3 Shape: Radius | 60 000 | 45 o|* A Radius : \ | 60.0|* / \ 1D1 Production Controls UnSharp Masking UnRharp Mask : Hon tS9i; Trident Defaults Keqion Control : ]' " Dot Gain: Trident Defaults Sr*i**i* | Trident Defaults r j Trident Defaults Default Sharp rabies [Update Sharpen: Qqcr $ucr ' T*t*1 ^f320]* ^^^^CS* Threshold Mm mmk$~so\* U Stortgtmf~26t I "I : 1 20 Ul Smooth : 0 III Sharpen: * I l | ioo| | 20| | 40| | 60| | 80] 1 00 - RGB Control Color Edges Smooth: | ioo| * () Mono Edges O B Jk R | ioo.o| G | 100.o| | 100.0| Shape: | 45.000 [ [x] Light Edges [x] Dark Edges A Shape: Radius: | 40.000 1 | 45.0| * Radius : A | 40.0|* 85 1l.j: Production Controls UnShtiro Masking USM UnSharp Mask : [x] 0n ** *? ** 8**** I Trident Defaults 8ra*lttwt | Trident Defaults TatMrl j Trident Defaults U9W: [""Refl Sharpen More *r | Update Tables ] Qecu #}UCR I 45-0| T>tal Wtfi^| ta*t~iflS* Threshold : Max Black ^ 90[- start * glarttf"^] r^roi Sharpen * LIU j ioo| r~Toi o | io|| 20] | 30|| 4oll 100 RGB Control Smooth * | lOOj () Mono Edges O Color Edges ? R | loo.oj G | ioo.o| ioo.o| Shape | 40.000 Filter Control Light A E Edges 0Dar k Edges Shape: Radius | 60 000 / \ | 40 0 / \ Radius: | 60.0| gHH Production Controls UnSharp Masking a UnSharp Mask : Trident Defaults Region Control : Dot Gala: | Trident Defaults Gradataaa j Trident Defaults Tea*: Trident Defaults f~ U8W: Refl Sharpen j Update Tables 0am #<** fc|J20ljS: Start UCRd 6i Max Black:) 9o| Start Bladtlj 2fe| ii:|:j::::r::rAA ILLL Sharpen [ 100 * 0 | I0|| 20|| 30|| 40|. 100 RGB Control - Smooth [Too] * () Mono Edges O Color Edges B ak | iou.o| G | 100.o| | 1QQ.o| Filter Control - Shape | 50.00o| A Radius | 50.000 50 0 * 86 Production Controls UnSharp Masking Mask- IS UnSharp ^A Qn H Trident Defaults Region Control : **** tt1; | Trident Defaults " Grtvdati|et j Trident Defaults Tw*j j Trident Defaults Sharpen Less 25 3 1 Updal i 25-i T*ai McJ 32o|* Start Utatr'ISl* Threshold: Mtat lt Smooth: Sharpen. | IQOJ * LZM o | 2o|| 4t)1 1 feoll 81i ,0 RGB Control Smooth | iool * () Mono Edges Q Color Edges ? R | 100 o| G | 1QQ p| B | 1QQ p| - Shape | 42.000| r Filter Control Light Edges Dark Edges ? (3 ? Shape : Radius | 42 000 1 -l A Radius : / \ 1 .0| Production Controls I UnSharp Masking UnSharp Mask : ^ On Ink Set! j Trident Defaults -1 Region Control: Oetatii;[^ Trident Defaults -1 ora***tew Trident Defaults '1 Hw;{^ Trident Defaults -1 &H5- Sharpen More 55 -1 Update Tables [ Moo Sharpen Qocr >c* : 80; | 55_0| 60 Thresho hUx Black -j~^p| Start % BU Smooth : 20l 0! Sharpen * II III =J| 0.0| [ ioo| 0 | 20|j 40|| 60|| 8o|j 10 Ji RGB Control - Smooth [ ioo| * (i) Mono Edges O Color Edges A | 100 o| G | IQO.o] B | 10Q.Q | - -Filter Control Shape | 40.000| ? Radius | 45.000| 87 m^m^ M^=g Production Controls UnSharp Masking Gradation Tone UnSharp Mask : \/] Qn " ' Tr'dgn Region Control: in i M.W.W.V.' I ii,TTTTTTTTTTr,-,-,-,-,-,-,-,-.-.-.-n.i1 1 1 1 1 UlWfcmrfa G. a t*tt; j Trident Defaults ?][ tlfiH: ) Sharpen More 85 ? | | Update Tables Q0C8 rf'UCR f]^*' M*x Black ^oJ1 fittari Black Sharpen: | iqq| 5B CM IUJIO | 20| | 40| | 60 | 80 [ 1 00 : | ioo| '!.) Mono Edges O Color Edges R | 100 o| G | ioo o| B | 100 o| Shape: | 35 000 1 [jj Light Edges BDar k Edges Shape : Radius: | 55_000| | 35.o|* Radius : A | 55.o|* AA Production Controls j UnSharp Masking Gradation Tone UnSharp Mask: J On Trident Defaults Region Control: Trident Defaults flr4tt>ttj Trident f"~ Tttft3 Trident Defaults A USH:1 Trident Defaults r| A( Update Tables \ QecR g)UCR :, | 55.0| T*mm{li2o\l* Start UCfifT^w Hax goj*^rTSi^gm BUek:] CM Smooth: oo| Sharpen | ioo| * 0; | 20| f 40|| 60l| 8o| ; 100; RGB Control Smooth | ioo| * () Mono Edges O Color Edges A B R | 100_0| G | 100 0 1 | 1QQ p| Shape Filter Control | 45.000 1 ^ Light Edges [X] Dark Edges A Shape : Radius | 40.000| 88 Appendix E - ColorRight 5.1 Screen Captures 89 ColorGetter* Scanning Software VERSION5.1 Copyright* 1991-1996 Optronics, an Intergraph Division ^K 7 Stuart Road, Chelmsford, Massachusetts 01824 (508) 256-4511 Ouerall Gradation... Color Cast Remoual.. Scan Geometry... Inuert H Flip VFlip Moderate Normal Considerable Custom... 90 Diameter IpiHels): 5 Diameter (pixels!: 5 Sharpen Amount: 50 % Sharpen Rmount: 50 % Sharpen Threshold: 0 %. Sharpen Threshold: 0 % ? Smoothing ? Smoothing Light Cutoff: 50 % Light Cutoff: 5 % Dark Cutoff: 95 7, Dark Cutoff: 50 % Red Sensitiuity: 30 % Red Sensitiuitg: 30 % Green Sensitiuity: 59 7> Green Sensitiuity: 59 % Reset Blue Sensitiuity: 11 % (T) (7) ( Beset Blue Sensitiuity: 11 % (T) (T) [ ] Diameter (pixels): 3 Diameter (piKels): 5 Sharpen Rmount: 33 % Sharpen Rmount: 50 Sharpen Threshold: 0 % Sharpen Threshold: 0 ? Smoothing ? Smoothing Light Cutoff: 5 % Light Cutoff: 5 Dark Cutoff: 95 % Dark Cutoff: 95 Red Sensitiuity: 30 % Red Sensitiuitg: 1 I Green Sensitiuity: 59 % Green Sensitiuity: 1 Blue Sensitiuity: 1 1 7. 001 Blue Sensitiuitg: 98 _ Diameter (piKels): 5 Diameter IpiHels): 5 Sharpen Rmount: 50 % Sharpen Rmount: 50 % Sharpen Threshold: 0 % Sharpen Threshold: 0 % ? n Smoothing Smoothing Light Cutoff: 5 % Light Cutoff: 5 % Dark Cutoff: 95 % Dark Cutoff: 95 % Red Sensitiuity: 1 % Red Sensitiuity: 30 X 3 Green Sensitiuity: 98 % Green Sensitiuity: 59 % Beset Blue Sensitiuity: 1 % HI [T) [ ; Blue Sensitiuity: 1 1 % rn [^ii ^set] Diameter (pixels): 7 Diameter (piKels): 5 Sharpen Rmount: 66 % Sharpen Amount: 50 % Sharpen Threshold: 0 % Sharpen Threshold: 0 % ? Smoothing ? Smoothing Light Cutoff: 5 % Light Cutoff: 5 % Dark Cutoff: 95 % Dark Cutoff: 95 7, Red Sensitiuity: 30 7o Red Sensitiuity: 98 % ^ Green Sensitiuity: 59 7o Green Sensitiuity: 1 % Beset Reset Blue Sensitiuity: 1 1 7o HI [T| [ ] Blue Sensitiuity: 1 % (T) (T) ( 91