-
An Introduction to Halftone Printing
Understanding Colour
Digital Repro
Mesh and Stencils
Inks and Colour Density
Printing
The Sericol Support Offer
Troubleshooting Guide
AZ of Terminology
Led by trends in design and marketing, advertisers are more
frequently incorporating complexcolour graphics and photography in
their screen printed advertising and point of salematerials. In
printing terms, this has prompted a move away from solid colours
and towardshalftone printing 4-colour halftones, in particular. As
a result, screen printers now findthemselves competing in a
changing marketplace where a greater emphasis is being placed onthe
ability to produce good quality 4-colour halftone prints.
Some screen printers have been producing halftone work for many
years and are perfectlypositioned to capitalise on the growing
demand for this type of work. Indeed, the ability to print 4-colour
halftones has traditionally been held up as a sign of a superior
print operation. Otherprinters, however, are only just coming to
grips with the new and greater demands of thisapplication of the
screen printing process, having recognised the need to move with
the times.
Whether you are an established halftone printer or entering the
market for the first time, The SericolHalftone Printing Manual is
intended to provide an invaluable source of information and
advice.Derived from Sericols long involvement with halftone screen
printing, it places a quarter century ofexperience and expertise at
your fingertips.
For the newcomer, the manual is structured to provide
comprehensive, step-by-step explanations ofall the key areas that
affect the final printed result. It allows you to build up your
knowledge of thedifferent processes at your own pace, in a logical,
easy-to-follow manner. Plain English makes eventhe most complex
concepts easy to understand.
The experienced halftone screen printer will probably prefer to
dip in and out of the text to find theinformation and advice which
is most relevant to their level of expertise. For you, the
manualprovides useful information on how to adapt your current
working practices to improve the qualityand efficiency of your
operation and become even more successful in an increasingly
competitivemarketplace. In particular, the manual spells out the
considerable benefits to be gained fromadopting the latest digital
repro technologies.
The expansion of halftone printing has been accompanied by a
general move towards computer-generated art, coupled with major
developments in digital pre-press technologies. Just as
screenprinters are being asked to produce more and more 4-colour
halftones, to ever higher standards, sothey are being asked to work
in an increasingly electronics-based production environment.
Thesedemands, together with advances in the speed and performance
of hardware and software and thegreater affordability of even high
end digital pre-press systems, are rapidly changing the
workingpractices of the screen printing industry and the graphic
arts industry as a whole. The recent, rapid development of digital
pre-press technology means that there are new skills andnew
opportunities for even the most experienced screen printer to
master. The manual offers anauthoritative introduction to this
increasingly important area of the industry.
The Sericol Halftone Printing Manual is, however, just a part of
Sericols overall customer package.An extensive range of advanced
ink and consumable products is complemented by an experiencedteam
of expert staff, providing a wide range of specialist services
designed specifically to meet thevaried needs of the halftone
screen printer. This combination of products and service makes
Sericolthe obvious partner for any printshop wishing to develop its
own 4-colour halftone printingoperation.
CHAPTER 2
CHAPTER 3
CHAPTER 4
CHAPTER 5
CHAPTER 6
CHAPTER 7
CHAPTER 8
Foreword Index
GLOSSARY
-
S
e
r
i
c
o
l
L
i
m
i
t
e
d
1
9
9
8 CHAPTER 1
-
-
S
e
r
i
c
o
l
L
i
m
i
t
e
d
1
9
9
8
1.1 1.2
Before you consider printing halftone work, it isessential to
have a clear understanding of thefundamental principles underlying
the halftoneprinting process. This opening chapter explainswhat
halftones are; why you would want to usethem; how they are made;
and the variablesyou must control when producing
halftonepositives
Look at a photograph, airbrush illustration, artistspainting or
shaded pencil sketch and you willnotice that it contains smooth
gradations of tone,from the lightest to the darkest areas. This
type ofimage is described as a continuous tone image(often referred
to as contone or CT). Nowcompare a continuous tone image with
aconventional single- or multi-colour screen print:youll notice
that the print contains flat areas ofcolour and no gradations of
tone.
Like most other printing processes, screenprinting cannot be
used to produce continuoustone images, as it is not capable of
laying downcontinuously variable densities of ink. However,through
the use of halftones, it is possible for thescreen printer to
create the illusion of acontinuous tone print.
Fig. 1.1. A halftone is made up of varying size dots tosimulate
continuous tones.
What is a halftone?
A halftone is the result of converting a continuoustone image
into a pattern or grid of regularlyspaced dots (a halftone screen).
The individualdots all have the same density, but vary in size.
On the final print, light reflected by the dotsmerges with light
reflected from the substrate toaccurately recreate the smooth tonal
gradationsfound on the original image.
The tone (lightness or darkness) of a printed areais dictated by
the size of the printed dots. Thedots reflect less light than the
white substrate,which means that larger dots, which cover moreof
the substrate, reduce the overall amount oflight reflected by the
print, resulting in a darkertone. Conversely, small dots cover less
of thesubstrate, so more light is reflected, resulting in alighter
tone.
The relative size of each halftone dot is describedas a dot
percentage. The dot percentage scaleruns from 0% to 100%, where
white (no dot)equals 0% and black (total coverage of thesubstrate)
equals 100%. A midtone area is madeup of 50% halftone dots, which
cover half the areaof the substrate.
Fig 1.2. The amount of the surface of each halftone cellcovered
by a printed dot dictates the tonal effect thelightness or darkness
of that area of the print.
Fig 1.3. Printing different dot percentages together allowsyou
to create the impression of tonal gradation. The lessnoticeable the
individual dots, the smoother the tonalgradation appears.
The actual size of the dots will be dictated by thelevel of
definition required and will be influencedby the distance from
which the finished print willbe viewed. Put simply, the less
noticeable theindividual dots in a halftone screen, the
moreoriginal detail can be resolved on the final print
Today, the photographic method has beensuperseded by electronic
scanners andimagesetters, which do away with the need forprocess
cameras, halftone negatives and thescreens themselves.
A scanner is used to capture a digitalrepresentation of the
original image. A pattern ofred, green and blue light is reflected
(ortransmitted, in the case of transparencies) by theoriginal
image. This pattern is captured as a digitalimage and is
subsequently passed on to animagesetter, which can be thought of as
a veryprecise laser printer. An interpreter in theimagesetter
converts the digital image into amatrix of halftone dots (arranged
in regular rowsand columns) and then records each dot on asheet of
light sensitive or thermal imaging film.The matrix of dots (also
known as a raster)represents the halftone screen.
Digital techniques speed up the production ofhalftones
significantly. More importantly, theyallow far more creative and
technical control overthe many variables which affect the final
result.The screen printer has the opportunity to modifythe digital
image extensively before it is output tofilm, using a desktop
computer or workstation tocontrol variables such as colour balance,
imagesharpness, contrast, dot gain and tonalcompression. These
variables will be discussedlater, but for now it is safe to assume
that there isno longer any reason to produce halftones
usingtraditional techniques.
Halftone parameters
When a digital halftone is output to film there areseveral key
parameters which must be specified:the screen ruling; screen angle;
and dot shape.Each can have a marked effect on the appearanceof the
final print.
Screen ruling is a measure of the frequency of thelines of dots
on the halftone screen that is, theactual size of the dots and how
close togetherthey are. The ruling is measured in lines per
inch(lpi) or centimetre (lpcm): thus, you will find 150lines of
dots in each inch on a 150 lpi halftonescreen. In theory, if you
wish to find out thescreen ruling of a single-colour print you can
lay arule and measure the number of dots along aninch or a
centimetre length. However, this is more
An Introduction to Halftone Printing CHAPTER 1
20% dot 80% dot
and the smoother the tonal changes will appear tothe eye.
Ideally, the dots should be too small tobe noticeable, as this will
ensure that there is astepless gradation of tones between light and
darkareas. For example, the dots on a 48-sheet postermay be very
noticeable if you view the posterfrom close range; but when you
view the posterfrom a distance of several metres, the eye is
nolonger able to make out the individual shapes ofthe dots and the
smooth tonal effect is created.
Fig 1.4. Dot size percentage is a relative measurement bothof
these dots cover 30% of their halftone cell.
How are halftones made?
Traditionally, halftones were createdphotographically, by
producing a halftone filmnegative, which was then contact exposed
toanother piece of film to obtain a positive. Thenegative was made
by placing a screen a glassplate carrying a finely ruled grid made
up of lineswhich cross each other at 90 between theoriginal
continuous tone image and a sheet offilm. Light from a process
camera was reflectedfrom the original image, passed through
thesquare windows formed by the crossed lines onthe screen, and on
to the film. This caused theoriginal continuous tone image to be
rendered asa series of dots. The size of the dots depended onthe
brightness of the reflected light, which in turnwas dictated by the
tone of the original: forexample, highlight areas reflected the
most lightand created the biggest dots on the negative and,after
contact exposure, the smallest dots on thepositive film.
To reproduce a full-colour original, multipleexposures were made
with the light being passedthrough red, green and blue filters (on
separateexposures), to produce the film positives (orcolour
separations).
-
STOCHASTIC SCREENS
In recent years, an alternative screening method has emerged,
which offers certain benefitsover conventional halftone screening.
Stochastic screening (also known as FrequencyModulated or FM
screening) takes the opposite approach to halftone screening
instead ofarranging different sized dots in a regular pattern, it
uses dots of uniform size and distributesthem in random patterns.
Variations of tone are achieved by adjusting the concentration
ofdots in a certain area the dots are spaced further apart in
highlight areas and clusteredtightly in shadow areas.
The stochastic approach can help to produce smoother tonal
gradations, as the use of similarsize dots means that prints do not
suffer from tonal jump (see Dot shape page 1.6). Therandom pattern
of dots also eliminates the problem of moir (see sub-section
dealing withmoir, page 1.7).
Another benefit is the use of very small dots (the equivalent of
using very high screen rulings),which can reproduce finer details
and smoother tonal transitions. However, until recently, thisproved
to be a double-edged sword for screen printers, as the largest dot
size available 20microns was too small to be supported by even a
150 threads per centimetre (tpcm) mesh,with its 25 micron opening.
Today, stochastic screens can carry maximum dot sizes up to
100microns, so the problem no longer exists.
Stochastic screening is still relatively new. Some of the
specialist software developed toproduce these types of screens is
still not capable of producing top quality positives (althoughthis
will change) and an extremely accurate (and expensive) output
device is essential for bestresults. If you do want to print with
stochastic screens, therefore, you will need to find a reprohouse
which has high end software and imagesetting equipment and a staff
who are qualifiedin stochastic screening for screen printing
applications.
Fig 1.9
Conventional halftone
screening
Stochastic
screening
1.3 1.4
easily achieved using the screen ruling indicatorwhich is
included at the back of this manual.
The higher the screen ruling, the smaller and lessvisible the
dots. Thus a halftone screen with ascreen ruling of 133 lpi (54
lpcm) is finer than ahalftone screen with a screen ruling of 65 lpi
(25lpcm). Generally speaking, higher screen rulingsgive sharper and
more detailed prints withsmoother tonal changes. The caveat is that
theyare harder to print well, being more susceptible todot gain
(see Dot gain page 1.6) and darker areasmay well fill in
completely, sacrificing shadowdetail. Also the smaller dots are not
alwayssupported by the mesh. Higher screen rulings alsorequire more
digital information to be captured bythe scanning device, which can
slow down theprocessing of the digital image.
The type of substrate you are printing on to willinfluence the
optimum screen ruling for a halftoneprint: higher screen rulings
can be used withsmooth, coated, non-porous substrates.
Higher screen rulings can give sharper, moredetailed prints, but
require more skill to print well.
Fig 1.5. 120 lpi halftone screen
Fig 1.6. 55 lpi halftone screen
Halftone angles (often referred to as screenangles) are not be
confused with screen meshangles. (The latter will hitherto be
referred to asmesh angles to avoid confusion.) They describethe
angle of the halftone screen (the lines of dots)as measured from
the horizontal axis.
Changing the angle of the halftone screen canmake the dots more
or less noticeable to the eye.Since the aim of halftone printing is
to reproducethe smooth tonal transition found on the originalimage,
the less noticeable the dots are the better.
When printing a single colour halftone, thehalftone screen is
usually produced at an angle of45 the angle at which the dots are
leastnoticeable to the eye. The dots are mostnoticeable when the
halftone screen is angled at90.
Fig 1.7. Individual dots are most noticeable when the
halftonescreen is angled at 90. The dots are least obvious when
thehalftone screen is angled at 45.
Halftone angles become especially important onmulti-colour
halftone prints. Theoretically, if eachpositive is given the same
halftone angle and eachdot on each of the positives is in exactly
the sameposition, the printed image would be perfect. Inpractice,
even a minute misregister results indistinct interference patterns,
known as moir.You can see this for yourself by laying twopositives
together and rotating the top one to seethe resulting patterns. You
should find that themoir is least noticeable when the positives
are30 opposed.
Fig 1.8. Theterm moirdescribes theunwantedinterferencepatterns
whichare producedwhen regularpatterns, such ashalftone screensand
screenmesh are placedover oneanother.
-
-
S
e
r
i
c
o
l
L
i
m
i
t
e
d
1
9
9
8
1.5 1.6
CHAPTER 1
background at higher dot percentages. However,when you generate
a halftone screenelectronically, using an imagesetter, you
canchoose from several different dot shapes,including diamonds and
ellipses.
The shape of the dot has little effect on theappearance of the
printed image, especially at finehalftone screen rulings where the
dots will not benoticeable to the eye in any case.
(However,elliptical dots may help to minimise moir.)Where dot shape
does have an effect is on thesmoothness of tonal gradation in the
midtoneareas.
With a conventional halftone screen, as the dotpercentage
increases and nears 50%, the cornersof individual dots just touch
the corners ofadjacent dots, forming a checkerboard pattern.This
gives the impression of an abrupt darkeningof the tone at that
point and is known as a tonaljump. The visual effect is to destroy
the smoothtonal gradation from light to dark. Tonal jump
isparticularly noticeable in fleshtones where distinctsmooth
gradations of tone give way to aposterised effect. The image may
also have anoticeably grainy appearance in areas on eitherside of
the 50% dot.
Diamond- and elliptical-shaped dots are a betteralternative to
conventional round/square shapeddots as they reduce the impression
of tonal jump.As the dot percentage increases, the edges of thedot
join in two separate stages: first, the ends ofthe dots join; as
the dot percentage continues torise, the sides of the dot join. In
this way,diamond- and elliptical-shaped dots give asmoother tonal
transition around the midtonearea.
Fig 1.13. At the 50% dot percentage, each corner of
aconventional halftone dot (top) just comes into contact withits
neighbour. This creates a tonal jump (an abrupt darkeningof the
print) as dot values increase in size from 40% to 50%.
Elliptical dots (bottom) are joined in only one direction, sothe
tonal transition is smoother and there is no noticeabletonal
jump.
While you cant ever eradicate moir, you canminimise it by
keeping a 30 angle between theseparate halftone screens. For a
2-colour halftone(duotone) you could position the dominant colourat
45 and the other colour at 15. When you tryto specify halftone
angles for a 4-colour print,however, you will run into an obvious
problem:because halftone dots are arranged at 90 to eachother on
the halftone screen, maintaining a 30angle between each screen will
mean that the firstand the last screens would be at the same
angle,causing moir.
For this reason, the angle of the fourth screen isalways going
to be a compromise. However,there are accepted sets of angles for
both offsetand flexographic printing which are proven tominimise
moir when printing four colours.Producing the four positives at
these anglescauses the different coloured dots to be printed ina
characteristic rosette pattern, with some of thecolours
overlapping.
Fig 1.10. Halftone screens have 2 angles opposed at 90. If30 was
maintained between each colur, the last angle wouldbe the same as
the first.
As explained previously, the halftone pattern ismost noticeable
when the halftone angle is at 90.For this reason, yellow the
lightest and leastvisible colour is often positioned at 90.Magenta
is regarded as being most visible to the
eye, so this halftone screen is set at 45 (37.5when using the
Flexo standard) to render thehalftone dots less obvious. The cyan
and blackhalftone screens are most likely to cause moir, sothey are
set at 30 to the magenta screen.
Fig 1.11. When 4-colour process inks are printed together,they
form a distinctive rosette pattern in the areas wherethey overlap
one another.
Fig 1.12. Left, standard Offset halftone angles (DIN 16547a)and
standard Flexo halftone angles (DIN 16547b), right.
While the use of correct halftone angles canminimise the risk of
moir, there is another majorfactor to consider the interference
pattern set upbetween the halftone screen on the stencil andthe
threads of the mesh you are using. The closerthe halftone screen
ruling to the mesh count, thegreater the risk of moir occurring.
This problemcan be overcome by carefully controlling thehalftone
angles/mesh count relationship, anglingof the screen mesh,
experimenting with thehalftone angles and controlling the halftone
screenruling/mesh count relationship (see, AvoidingMoir on page
1.7).
Dot shape When halftone positives are producedusing the
traditional photographic process, dotsappear as black circles on a
white background atlower dot percentages, squares at around the50%
dot, and open white circles on a black
135
45
15
45
75105
Y 90K 75
M 45
C 15
C 67.5Y 82.5
M 37.5
K 7.5
90
Dot gain
An important variable that must be controlled onany halftone
print is dot gain. This describesincreases in the size of printed
dots whencompared with the dots specified by the softwarewhich
created the halftone screen. As its namesuggests, dot gain causes
dots to grow larger, sothat they cover more of the substrate and
reducethe intensity of the reflected light. The effect ofthis is to
make the halftone appear darker,especially in midtone areas. Dot
gain also lowerscontrast by darkening highlight tones, and it
caneliminate shadow detail by causing shadow tonesto fill in.
There are two types of dot gain: optical andphysical.
Physical dot gain is the result of changes in dotsize caused by
the pre-press and productionprocesses. When a halftone screen is
output tofilm by an imagesetter, the individual dots mayincrease
slightly in size. Similarly, transferring thehalftone screen to the
stencil may also increasethe size of the dots. However, the
biggestincrease in dot size is caused by the printingprocess
itself. Ink viscosity, press parameters such as screen tension and
squeegee pressure and the absorbency of the substrate will all
affectthe size of the printed dot.
Optical dot gain describes the apparent increasein dot size
caused by the scattering of light in thesubstrate. This causes the
dots to cast tinyshadows, which have a darkening effect.
Total dot gain is the sum of both physical andoptical dot gain
and is measured using adensitometer (see Chapter 3).
Dot gain occurs around the circumference of thedots, so the
larger the dot, the greater the gainwill be. This is why dot gain
is most noticeable inmidtone areas where the dots have the
highestcircumference to surface area ratio. (Figuresquoted for
total dot gain always refer to the 50%dot, precisely because dot
gain is greatest at thisdot percentage).
As previously mentioned, dot gain also variesbetween different
ink systems it can be as highas 20% with some conventional UV inks,
and aslow as 3% with some water-based inks. This,together with the
wide variety of substrates used,
-
makes it impossible to quote a standard dot gainfigure for the
screen printing process. Thus, youhave to determine the dot gain
for your own setup and compensate for it by reducing dot sizes
onyour positives accordingly. (How to control dotgain is covered in
Chapter 3.)
Fig 1.14. This image shows the effects of dot gain filled
inshadows, dull highlights and an overall lack of contrast.
Fig 1.15. Compensating for dot gain brightens the
highlights,reveals shadow details and boosts overall contrast of
theimage.
1.7 1.8
Alternative AnglesAs explained previously, there is no reasonwhy
you shouldnt experiment with halftoneangles to come up with the
optimum result foryour own work. The halftone angles shown onthe
right provide a starting point for yourexperiments. Experimentation
may be necessary to counterspecific production parameters. For
instance, ifyou have used gray component replacementduring the
processing of your digital file (see,Chapter 3) the black screen
will be muchheavier than usual, which may make the dotsmore
obvious. In this case, you may find youachieve a better result if
you switch the
How to select mesh angles
Included in this manual are two sets of positives for each of
the following halftone screen rulings:
55 lpi/21.6 lpcm 75 lpi/29.5 lpcm 100 lpi/39.4 lpcm 133 lpi/52.4
lpcm60 lpi/23.6 lpcm 85 lpi/33.5 lpcm 120 lpi/47.3 lpcm
One set of positives was produced using standard Offset halftone
angles; the other using standard Flexohalftone screen angles. Both
sets include panels which simulate different mesh angles when
printedthrough mesh fixed at 90 to the frame.Assume that you wish
to establish a standard for printing 85 lpi halftones for a
particular ink system. Thefirst step is to determine which set of
standard halftone angles produces the least moir. Prepare ascreen,
stretched with straight mesh of the recommended count for the ink
system you wish to use;place the two 85 lpi positives adjacent to
each other and square to the edge of the frame; expose thescreen
and develop under your normal production conditions.Next, print
through the stencil using your chosen ink at the normal thinning
rate. Study the print to seewhich of the Offset or Flexo angles
gives the best result that is, no moir pattern on the print. Youmay
find, for example, that the Flexo angles produce the best result,
giving a moir-free print for threeof the colours, with moir only
present on the yellow screen. To eliminate this moir pattern, look
at theother Flexo panels marked 5, 15 and 30. If the 15 panel
reduces the appearance of the moir, thenhave the screen for the
yellow printer stretched at 15.
Because the maximum width of screen fabric is 2.2 metres, angled
mesh is not an option for printersusing very large format screens.
One solution to this problem is to print one colour at a
differenthalftone ruling to the others, this will eliminate the
clash between that colour and the mesh. Thedifferent screen ruling
is not usually detectable in the final print. For this reason, if
you have room onyour test screen it is a good idea to include the
screen ruling positives that are either side of the rulingyou
intend to use. Alternatively you could experiment with different
halftone angles (see boxAlternative Angles)
The standard Litho and Offset halftone anglesminimise moir
occurring between filmpositives. However, the screen printer also
hasto consider the possible interference effectcreated between the
halftone screen and themesh.
To reduce the chances of moir occurring, youmust ensure that the
mesh count you use isappropriate for the halftone screen ruling
thatis, the mesh count which is least likely to
causeinterference.
You can generally calculate which mesh countsare appropriate and
which are not bydetermining the ratio between the mesh countand
halftone screen ruling you are using. Tocalculate the ratio, use
the following formula:
Ratio =
(Make sure that both figures refer to the sameunit of
measurement either inches orcentimetres.)
If the ratio is 5.0 or above, you can be reasonablysure of
producing moir-free prints, all otherfactors being equal. If the
ratio is below 3.5, youshould choose a different mesh count as moir
isa strong possibility.
It has also been shown that moir is more likely ifthe first
decimal place in the ratio is an evennumber for example 3.4.
If you have to use such a combination, you caneither experiment
with the halftone angles on thepositives, or change the angle of
the meshitself
AVOIDING MOIR
halftone angles for the black and magentascreens. However, if
you are trying toreproduce fleshtones, always try to keepyellow and
magenta 45 apart. If green is thedominant colour in the image for
example ina landscape photograph you may discoverthat it helps to
have yellow and cyan 45apart.Although 45 has been shown to produce
theleast noticeable halftone dots, it can sufferfrom interference
with the mesh. In whichcase, use the Flexo angles
describedpreviously or experiment with your own set ofangles.
Possible Alternative Halftone Angles
C M Y K
172.5 52.5 7.5 112.5
82.5 112.5 7.5 52.5
22.5 52.5 7.5 82.5
20 50 5 80
82 22 97 52
mesh counthalftone screen ruling
-
CHAPTER 1
GEOMETRICSCREENINGA screening method, which is gaining
inpopularity with screen printers, is Geometric.The positives in
this case comprise parallellines, instead of dots, which vary in
thickness toeffect tone changes (see fig. 1.16).
As with halftone dots, the lines must run atdifferent angles for
each of the four printingcolours to avoid primary moir (see page
1.7).Because each geometric positive has only oneangle of direction
(conventional halftones havetwo angles, see Fig 1.10. page 1.5), it
is easier toproduce a set of angles, that avoid primary moir,and
can be used on straight mesh with minimalsecondary moir being
created.
Other advantages include smoother tonalgraduations and reduced
tonal jump. This isachieved because there is no break point as
withconventional dots. i.e. where a dot begins toseparate from
adjacent dots below 50% tonevalues. Commonly used rulings are 55 -
75 l.p.i.,although some printers regularly use 85 l.p.i.
Mostobservers comment that finished prints appear tobe a finer
screen ruling than is actually used but,adversely, others comment
on the grainyappearance of some images.
Can geometric totally eliminate moir?In the lightest areas,
where the lines becomeextremely narrow, they can reach a point
wheresome will print and others may dry in, appearingas breaks in
the lines, possibly causing visibleinterference patterns on high
key images (wherethe majority of colours are less than 50%
tonevalue).
Some repro houses have worked very closely withtheir clients to
achieve the optimum angles for thework they produce. However, most
printers wouldagree that total elimination of moir is unlikelywith
any halftone pattern but the tendency isgreatly reduced with
geometric. Ultimately, toensure consistent and best quality
halftone prints aprinter must fingerprint his press, type of
workand production methods to establish optimumangles and other
parameters such as print orderand dot gain values.
Dot Gain?
Although the halftone pattern is lines, the termdot gain still
applies. Target areas can still bemeasured with a densitometer to
record anygains, or losses, in tone value. As withconventional
halftones, test prints are necessarywith any unfamiliar ink/screen
ruling/substratecombination.
In the interest of good communication betweenprinter and repro
house, we recommend having ageometric positive with agreed step
valuesproduced by the repro house. The resulting valueson the print
can be recorded and used to set up atone curve. This can be applied
wheneverpositives are produced for the same screen ruling,ink,
substrate and production equipment. As aguide a useful set of
percentage values is 5, 10,20, 30, 40, 50, 60, 70, 80, 90 and
95.
Fig. 1.16.
Enlargement ofmagenta positive.
1.9 1.10
-
S
e
r
i
c
o
l
L
i
m
i
t
e
d
2
0
0
0
-
Any screen printer who is contemplating theuse of digital repro
techniques for 4-colourprocess work, must have a clear
understandingof colour: its properties; how the eye perceivesit;
how it is represented on digital displays andthe printed page; and
the ways in which it canbe managed. This information is essential
if youare to make the most of the powerful colourcontrols which
digital pre-press systems offer.This chapter provides a guide to
thefundamentals of colour theory; the RGB andCMYK colour models;
and the relationshipbetween colours on the desktop and final
print.
First things first: colour does not exist it ismerely an
illusion created by a combination oflight, the perceptual
mechanisms of the humaneye and brain, and the ways in which
differentsurfaces affect the light that they reflect ortransmit. To
understand basic colour theory, youmust appreciate how these
factors combine tocreate the colours that your brain perceives.
Additive colour So-called white light is made upfrom various
wavelengths of radiation emitted bya light source, such as the sun.
These wavelengthsare what the human eye perceives as colours andare
referred to collectively as the visual spectrum.(There are many
other wavelengths that falloutside the visible spectrum and cannot
beperceived by the human eye. Examples include:infra-red radiation;
ultra violet radiation; and X-rays.)
Using a prism, you can separate white light intoits constituent
colours. You will notice that themain bands of colour are red,
green and blue.These are said to be the primary colours of
light.The entire range, or gamut, of colour that thehuman eye can
see can be expressed by addingthese three colours of light together
in differentproportions and intensities. For example, addingall
three colours together at full intensity produceswhite. Adding
equal proportions but lowerintensities of all three colours
produces shades ofgray. Adding any two of the primary
colourstogether produces a secondary colour: green andred produce
yellow; green and blue producecyan; and red and blue produce
magenta.(Whenever you add two primaries together thesecondary
colour they produce is brighter hencethe term, additive.) The
absence of any coloursproduces black.
Fig 2.1. Additive colour notice how cyan, magenta andyellow (the
subtractive primaries) are formed in areas wheretwo of the primary
colours overlap.
Your computers monitor uses an additive coloursystem, known as
the RGB colour model, todisplay colours. The screen is covered with
tinyred, green and blue phosphors which glow (emitlight) when
activated. So, the monitor reproducesthe yellow of a lemon by
activating a combinationof red and green phosphors. The mixture of
redand green light that they emit is perceived asyellow by the
eye.
The brightness of the colours on screen isdetermined by the
intensity of the light emittedby the phosphors and this is dictated
by theintensity of the electrons which are fired at thescreen to
make the phosphor coatings glow.
Subtractive colour A printed image does not emitlight in the
same way as the phosphor on yourcomputer monitor. Instead, the
light that reachesyour eye has been reflected from the surface
ofthe printed material. Crucially, all surfaces act as afilter
reflecting or transmitting certainwavelengths of light and
absorbing (subtracting)others. For example, when light strikes a
tomato,the pigment in the surface of the tomato filters outall the
non-red light, so what you see is a redcolour. Printing inks act in
the same way,absorbing some wavelengths and reflectingothers.
This explains why the primary colours of light donot produce an
additive effect when printed
together. Put simply, each primary colour filtersout the other
primaries, making it impossible tocreate secondary colours. To
illustrate: supposeyou print an area with blue and red ink; the
blueabsorbs all the non-blue wavelengths (red andgreen) and the red
absorbs all the non-redwavelengths (blue and green). Hence, all
theprimary colours would be removed, resulting inblack.
To reproduce colours on printed materials, youmust control the
amount of red, green and bluelight which is reflected by the print.
This can beachieved by printing combinations of thesecondary
colours of light cyan (C), magenta (M)and yellow (Y). These are
generally referred to asthe subtractive primaries, as each colour
absorbsone of the primary colours of light: cyan absorbsred;
magenta absorbs green; and yellow absorbsblue. For instance, if you
want to reproduce a redcolour, you would print with a combination
ofmagenta and yellow inks to filter out both thegreen and blue
light.
Fig 2.2. Subtractive colour - Notice how the primary coloursare
formed where two of the subtractive primaries overlap.
In theory, printing cyan, magenta and yellow inkstogether should
produce black. However, this isnot the case the result is a dark
brown colour.Adding black ink to the subtractive primariesincreases
the possible density range that can beprinted, ensuring clean
blacks, deep shadows andneutral grays. Black is known as the key
colourand is denoted by the letter K - hence the CMYK
colour system and 4-colour process printing.
Colour percentages
By using different proportions of CMYK inks youcan effectively
control the proportions of red,green and blue light reflected from
the print andthe colour of light perceived by the eye. With4-colour
process printing, the four inks are printedas separate, overlapping
halftone screens. In thisway, the proportion of each subtractive
primary isdictated by the size of the halftone dots for thatscreen.
(Process colours are expressed in terms ofthe dot percentage for
each CMYK ink.) Dot size,then, dictates colour as well as tone.
It follows that dot gain can have a significanteffect on colour
also. If the size of a yellow dotincreases by, for instance, 5%
when it is printed,then it will absorb 5% more of the blue light
thatfalls on it. This will result in a warming of theimage, as
there will be less cold blue lightreaching the eye.
The problem is exacerbated by the fact that dotgain values vary
from colour to colour, dependenton the print order. This can have a
marked effecton the colour balance of 4-colour process prints.For
example a colour printed onto virgin paperwill have a slightly
different dot gain value thanwhen overprinting because of factors
such asstock absorpency, ink build etc. Added to this isthe fact
that dot gain is not uniform, so midtoneareas are more likely to
suffer a greater colourchange than highlights and shadows. This
explainswhy the most common example of colour changedue to dot gain
occurs in the magenta content offleshtones.
Colour fidelity is another reason why dot gainmust be measured
and controlled.
2.1 2.2
-
S
e
r
i
c
o
l
L
i
m
i
t
e
d
1
9
9
8
Understanding Colour CHAPTER 2
BLUE
MAGENTA
CYANGREEN
YELLOW
RED
WHITE
YELLOW
REDGREEN
CYAN BLUE MAGENTA
-
C0LOUR GAMUTS
Each colour system is said to have a colour gamut. This
describes the range of colours thesystem can reproduce. As
explained previously, white light contains all the colours that
thehuman eye can perceive.
The RGB colour system, as used in computer monitors, is capable
of reproducing millions ofdifferent colours, but it cannot
reproduce the entire gamut of colours that the eye can see.This is
due to the limitations of the phosphor coatings which emit the
light.
The CMYK colour model has an even smaller colour gamut, due to
inherent impurities in thepigments used to manufacture printing
inks; the substrates they are printed on to; and thevery fact that
it is a subtractive system the inks absorb light so light intensity
(and colourbrightness) is reduced.
In order to print an image that was captured as a colour digital
file, it must be separated converted from the additive RGB colour
model, as used by scanners and monitors to thesubtractive CMYK
colour model, as used in printing inks. This conversion process can
giverise to problems in terms of colour fidelity. (Colour
separation is covered in greater detail inChapter 3.)
The difference in colour gamut also makes it very difficult to
predict how colours on amonitor will appear in print. Even after an
image has been separated, it may still appear verydifferently on
screen to how it will appear in print. (You can prove this by
holding a printedcolour swatch next to the monitor and bringing up
the same colours on the screen. Thechances are they will appear
very differently on screen.)
The common solution is to never trust thecolours you see on the
monitor. Instead, usethe softwares colour information tool(referred
to as the Info palette in AdobePhotoshop) to measure colour rather
thanjudging it by eye. You should also makesure that you use a
reliable proofing system,which provides an accurate
representationof what the printed result will look like.(Both
topics are dealt with in later chapters.)
Another option is to use a monitor which iscalibrated to show
colours as they willappear on the printed page. These arerelatively
expensive and must be usedunder carefully controlled
viewingconditions changing light levels, flare onthe screen and
even the reflection of theoperators clothing can affect the
on-screencolours. Similarly, different softwareapplications may
represent coloursdifferently. For these reasons, so-calledsoft
proofing (judging colours from amonitor by eye) cannot be relied on
totallywhen colour fidelity is critical.
SUBSTRATES AND COLOUR
In 4-colour process printing, the CMYKinks are printed as
overlapping halftonescreens. This requires that the inks
aretransparent, so that the light can passthrough them, reflect off
the substrateunderneath and bounce back towardsthe eye. In areas
where the coloursoverprint, the colour of the reflected lightis
determined by the combined filteringeffect of all the inks.
The fact that the light is reflected fromthe substrate, however,
introducesanother colour variable namely, thecolour of the
substrate. If the substrateis a colour other than white
(whichreflects all colours equally) it will haveits own filtering
effect on the light. Thismeans that generally acceptedcombinations
of ink colours couldproduce markedly different colours tothose
specified.
2.3 2.4
Fig 2.3. The CIE (Commission Internationale delEclairage) Yxy
colour model indicates the colour gamut ofdifferent colour models.
As you can see, CMYK inks canproduce only a small percentage of the
total colour gamutperceived by the eye.
CMYKGamut
RGBMonitorGamut
-
-
S
e
r
i
c
o
l
L
i
m
i
t
e
d
1
9
9
8
There are many screen printing parameters toconsider when
producing 4-colour processwork, but thanks to recent advances in
thedigital repro process, it is no longer necessaryto tackle all
the key variables on press. A fullunderstanding of colour
manipulation at therepro stage has enabled printers whospecialise
in halftone work to move theemphasis away from the manipulation of
inkformulations at the time of printing andtowards tighter control
over the production ofhalftone film positives during pre-press.
By investing in the latest technology andadopting up-to-date
working practices, screenprinters now have the necessary means
tostandardise as far as possible the 4-colourhalftone screen
printing process. The potentialbenefits, in terms of speed,
quality, efficiency,cost and repeatability, are substantial.
This chapter opens with a tour of the digitalrepro process,
providing an introduction to thevarious technologies and how they
operate. Buthow something works, and how it is of
practical use, are two very different questions.The following
section addresses the practicalapplication of digital systems to
control keyhalftone printing variables, such as colourbalance, dot
gain and tonal range.
The other question to consider is whether ornot to bring the
digital repro process in-house.The final section provides valuable
pointers onwhich parts (if any) of the repro process screenprinters
should be handling themselves, andwhich should be left to
specialist repro houses.
The digital repro process can be split into threestages: input;
processing; and output. Thesestages are shown in figure 3.1.
Whether you planto handle the repro process yourself, or use
aspecialist repro house, a clear understanding ofeach stage is
vital if you are to achieve the highstandard of film positives
needed to producequality halftone prints.
ORIGINAL IMAGES
The first step in the digital repro process is toconvert
original images, such as photographs andartists illustrations, into
digital files. But beforeyou begin, you need to check whether the
imagesare suitable for print reproduction.
Often, you will have little say over which imagesare to be used
the client or their designer willhave made the choice before you
becomeinvolved. However, by understanding thelimitations imposed by
less than ideal images, youare better equipped to deal with the
problemsthat are likely to arise.
If possible, ask for photographic images to besupplied as
transparencies rather than prints.Transparencies possess a wider
possible tonalrange and more saturated colours.
The rich colours used in some paintings areimpossible to
reproduce using printing inks. Acommon solution is to photograph
the work ontransparency film, along with a colour chart. Youcan
then compare the colours of the chart in thescanned image with the
colours of the originalchart and match the colours as closely as
possibleusing image editing software.
Artwork which has been generated in computerimage editing and
illustration programs can betransferred directly to the processing
stage.
3.1
Digital Repro CHAPTER 3
THE INPUT STAGEOnce the appropriate images have beenevaluated,
the next step is to convert them todigital image files, using a
scanner.
During the scanning process, red, green and bluefiltered light
is passed through or reflected fromthe original image and is
sampled by an array oflight sensitive measuring devices either
photomultiplier tubes (PMTs) in a drum scanner orcharged coupled
devices (CCDs) in a flatbedscanner. The brightness and the colour
of thetransmitted or reflected light is used to mapcolours and
tones from the original image to agrid of small squares. These
squares are known aspicture elements (or pixels) and the grid is
knownas a bitmap (hence, bitmapped images). Pixels canappear as
white; black; a shade of gray; or acolour.
Four closely related parameters are specified atthe time of
scanning:
scan resolution
scan size
bit depth
colour model
Scan resolution
Scan resolution is a measurement of the numberof pixels in a
given distance. It is most commonlyquoted as the number of pixels
per inch (ppi) forinstance, 133 ppi. Resolution can also be
quotedin terms of a Res number, which specifies thenumber of pixels
per millimetre. A resolution offive pixels per millimetre, for
example, would bequoted as Res 5.
The actual size of each pixel is determined by theresolution of
the image, in much the same way asthe size of a halftone dot is
determined by thehalftone screen ruling. (Incidentally, there is
animportant relationship between scan resolutionand halftone screen
ruling see, Scan resolutionand sizing). At high scan resolutions
there aremany pixels per inch, so individual pixels aresmall:
details are rendered sharply and tonaltransitions appear smooth. At
low scan resolutionsthere are fewer pixels per inch, so
individualpixels are larger. At very low resolutions, the
3.2
ORIGINATIONFlat copy artwork,
photographic prints& transparencies
DIGITAL ARTWORKCreated on PC or AppleMac
and supplied on disk(Jaz, CD etc.)
SCANNERMaps original and
converts image intodigital information
COMPUTERImages adjusted using
photograph manipulation softwareand combined with other
artwork
in page layout programme
RIP(Raster Image Processor) creates halftone screens
and angles
IMAGESETTERProcesses informationand outputs separated
film
POSITIVESRight reading
emulsion side up(for screen printing)
DIGITAL PRINTERProofing results used toapply final
adjustmentsbefore sending to RIP
CROMALIN PROOFProduced fromfilm positives
Fig 3.1. The digital repro flow path
STENCIL MAKING&
PRINTING
-
individual pixels are so large as to be visible tothe eye,
causing images to look jaggy. This effectis most noticeable in
areas of the imagecontaining straight, diagonal lines, which
appearas sets of distinct steps hence the term,staircasing.
Fig 3.2. At high scan resolutions (left), fine details can
beresolved and tonal gradations are smooth.
At very low scan resolutions (right), individual pixels may
bevisible to the eye and staircasing is evident on
straight,diagonal lines.
It would seem logical, therefore, to scan originalimages at the
highest resolution possible.However, this is not necessarily the
best solution,as the higher the scan resolution, the larger thefile
size. Large files take longer to edit, slow downmonitor redraw and
increase the time taken bythe imagesetter to output the final
films.
Scan size
The dimensions of the scanned image aredetermined by the number
of pixels it containsand the scan resolution. This can be
illustratedusing the following example:
Suppose a scanned image measures 1,000 pixelsby 1,000 pixels and
has a resolution of 200 pixelsper inch (200 ppi). The scanned image
wouldmeasure 5 inches by 5 inches:
1,000200
Now suppose you were to scan the image at 500ppi. The pixels
would be smaller and arrangedcloser together, so the image would be
reduced insize: it would now measure 2 inches by 2 inches:
1000500
In practice, correct scan size and scan resolutionshould be
determined before the scan is made, bymatching them to the intended
size and halftonescreen ruling of the printed image. Altering
theseparameters after the scan has been made can havea dramatic
effect on the appearance of the image.For example, if you were to
simply increase theoverall size of the image, each pixel would
beenlarged until the bitmap fitted the newdimensions. In effect,
you would be lowering thescan resolution the pixels would be
bigger, sothere would be fewer pixels per inch. This mightspoil the
smooth transitions of tone or even createnoticeable jaggies or
staircasing, as describedpreviously.
Reducing the size of a digital image has theopposite effect:
pixels become smaller, so thereare more of them per inch. This
effectivelyincreases the resolution of the image and givessmoother
tonal transitions. However, the file sizemay now be unnecessarily
high for the dimensionsof the scan, as explained previously.
Resampling
Ideally, if you need to increase the size of ascanned image for
any reason, you should rescanthe original to achieve the correct
dimensions. Ifthis is not possible for one reason or another,
youcan use computer software to resample theimage. This involves
adding extra pixels to theimage, rather than changing the size of
theexisting pixels. The most accurate method isknown as bicubic
interpolation, whereby theprogram assigns the colour of a new pixel
byaveraging the colours of the surrounding pixels.
You may also need to use interpolation if yourscanner offers a
low maximum optical resolution(scanner specifications usually state
the unitsmaximum optical and interpolated resolutions).You are most
likely to need to use it whenscanning a small original such as a
35mmtransparency for output at a large size.
Resampling should be treated as a last resort as itdegrades the
quality of the original scan andcauses a marked softening of the
image. This iswhy it is worth investing in a scanner with a
highmaximum optical resolution, so that you canensure the correct
scan resolution without havingto resort to using interpolation
programs.
3.3 3.4
SNAP DECISIONSWhenever you are asked to incorporate scanned
photographs in a print job, check thefollowing:
Does the image contain a wide range of tones?
Some scanners compress the tonal range,heightening contrast. An
original image with awide range of tones helps to minimise
thisproblem.
Is a limited range of tones intentional?
For instance, the client may have chosen to usehigh key or low
key images to accentuate mood.Automatic scanning software may
attempt tocompensate by redistributing the tones. Manualadjustment
using scanner or image editingsoftware may be necessary.
Has the original been screened before?
If the answer is yes, the original must bedescreened at the
scanning stage, or duringprocessing using image editing
software.Descreening will reduce the risk of moir causedby
interference between the halftone screen onthe original image and
the halftone screenapplied to the scan.
Is the image sharply focused?
Scanning or post scanning software can help tocompensate for a
small amount of blur, but outof focus images will not reproduce
well. Use apowerful eyeglass to check for sharply focuseddetails.
(Even the sharpest images may benefitfrom a small amount of
software sharpening.)
Were individual photographs taken on thesame film stock?
Different films have different colour balanceswhich can lead to
some images appearingwarmer or cooler than others. This
isespecially noticeable when photographs ondifferent film stocks
are used next to each other.If mixed film stocks are batch scanned
(severaldifferent images scanned at once) the images willgenerally
require some editing at the processingstage.
Are you scanning from a duplicatetransparency?
Duplicating an image tends to boost contrast(sometimes
unacceptably), alter colours andreduce overall sharpness. Avoid
using dupes ifpossible.
How big is the original image?
The larger the original, the larger it can beprinted without
sacrificing sharpness or smoothtonal transitions. Enlarging a
photograph tends tolower its colour contrast, so ensure that
picturesfor large scale output have rich, saturatedcolours to begin
with.
Are the originals damaged?
Fingerprint marks can usually be removed bycareful cleaning
prior to scanning and minorscratches can be removed in an image
editingprogram.
Size = = 5
Size = = 2
-
-
S
e
r
i
c
o
l
L
i
m
i
t
e
d
1
9
9
8
Bit depth
Each pixel in a digital image contains colour andtonal
information. However, due to the inherentlimitations of the RGB
colour model, the pixelshave a smaller colour gamut (a smaller
range ofreproducible colours) than continuous tone colouroriginals,
such as colour transparencies and artistsillustrations. When the
original image is convertedto a bitmap during the scanning process,
certaincolours in the original cannot be matched exactlyusing RGB
colours, so pixels are assignedwhichever RGB colour is the closest
match to thecolour found in the original image.
Bit depth is a measurement of how many tones orcolours each
pixel can reproduce. It follows, then,that the greater the bit
depth, the more accuratelythe colours of the original image can be
mappedto the digital image.
At its simplest level, a pixel can be turned on(white) or off
(black). This level, or depth, ofinformation is known as 1-bit data
(2 to the power1).
By increasing the bit depth, the pixel can showintermediate
levels of gray (levels of brightnessbetween light and dark). For
example, pixelscontaining 2-bit data (2 to the power 2) can
showfour levels of information: black, white, light grayor dark
gray. 8-bit data (2 to the power 8) allowsthe pixel to show 256
levels of gray includingblack and white.
3.5 3.6
SCAN RESOLUTION AND SIZING
CHAPTER 3
Down-sampling
Down-sampling is the opposite to resampling itdescribes the
process whereby a computerprogram is used to remove pixels from an
image,to reduce its dimensions. This method has fewerdrawbacks
compared with resampling, but it canlead to staircasing if it is
used excessively.
Scaling percentages
Some scanning programs require that you set theimage size by
specifying a scaling percentage (orgive you the option to do so).
This is calculated asfollows:
required sizeoriginal size
Determining the correct scan resolution is important for high
quality results. It is calculatedusing a combination of the
following:
THE HALFTONE SCREEN RULING THE SCALING FACTOR THE QUALITY
FACTOR
The quality factor describes the fact that a scan resolution of
approximately twice the screenruling gives the best quality printed
image. Thus, if an image is to be printed at 55 lpi, the
scanresolution should be 110 ppi. This is not a hard and fast rule,
but rather a general guide whichhas been agreed upon across the
graphic arts world.
Increasing the quality factor makes no difference to the
appearance of the final print. However,at halftone screen rulings
of 133 lpi and above, it is possible to reduce the quality factor
to 1.5,to keep file sizes manageable, provided the image contains
no obvious geometrical patterns orstraight diagonal lines.
output sizeactual size
36036
The scaling factor is a measurementof the amount the original
imagemust be enlarged or reduced to fitthe dimensions of the final
printedimage. This is calculated as follows:
Suppose that a transparency measuring36 x 24 mm is to be printed
at a finalsize of 360 x 240 mm. Measuring thelonger dimension the
scaling factorwould equal 10. Thus:
The scan resolution isdetermined by a simplecalculation:
halftone screen ruling x scaling factor x quality factor
Suppose that the imagementioned above were to bereproduced using
an 85 lpihalftone screen. The correct scanresolution would be:
85 x 10 x 2 = 1700(lpi) x (scaling factor) x (quality factor) =
(ppi)
Suppose, then, that your clientsupplied a large print
measuring720 x 480 mm instead of a 35mm transparency. The size of
theprint would need to be halved tomatch the size of the final
print,so the new scan resolutionwould be:
85 x 0.5 x 2 = 85(lpi) x (scaling factor) x (quality factor) =
(ppi)
You can see from the above examples that the greater the scaling
factor (magnification), thehigher the scan resolution all other
factors being equal. This is why repro houses use drumscanners with
very high optical resolutions the higher the optical resolution of
the scanner,the higher the possible scan resolution. This allows
them to work with small originals whichneed to be greatly enlarged
for output. On flatbed scanners with lower optical resolutions
youwould have to resort to using an interpolation program to add
extra pixels, which would meancompromising the sharpness of the
image (see, Resampling).
BRING IT DOWN
If the output size of the digital image isnot known at the
scanning stage, make alarge scan (by setting a high
scalingpercentage) at a high scan resolution itis better to
down-sample than resample.
12
3
4
5 6
FIG 3.3. It is very simple to resample and down-sampleimages
using Adobe Photoshop (though try to avoidresampling whenever
possible). First check that BicubicInterpolation is selected in the
General Preferences box(File/Preferences/General Preferences). Once
selected,there is no need to alter it on subsequent occasions.Next,
open the Image Size dialogue box (Image/ImageSize) and make sure
that resample is selected (1).Select Constrain Proportions (2) if
you want to maintainthe same height/width proportions and enter
newvalues for the width and height in the Pixel Dimensionsboxes
(3). (If you have clicked Constrain Proportions,the width will
automatically be changed when you entera value into the height box
and vice versa). If you wishto enter values as a percentage of the
currentdimensions, select Percent as the unit of measurement(4).
The new file size for the image appears at the top ofthe Image Size
box with the original file size shown inparentheses (5). Click OK
(6) or press the return keyon your keyboard to resample the image
or down-sample it if you have entered smaller image
dimensions.Finally, apply the Unsharp Mask
filter(Filter/Sharpen/Unsharp Mask).
x 100= 10
-
1-bit 2 shades(levels of brightness)
2-bit 4 (levels ofbrightness)
8-bit 256shades (levelsof brightness)
Fig 3.4
Black and white 2 shades(1-bit)
Grayscale 256 shades (8-bit)
RGB colour 16.7 millionshades (24-bit)
Fig 3.5
Which colour model?
A scanner can record a full-colour original imagein RGB or CMYK
colour.
Specialist repro houses use powerful colourseparation software
for converting the RGB datacaptured by the scanner to a CMYK file
and havethe necessary experience to make high qualityseparations.
If you are using a repro house, thegeneral advice is to leave the
separations to them,rather than handling them yourself in an
imageediting program.
However, you should bear in mind that reprohouses are generally
set up to service lithoprinters, so the scans they supply may be
far fromideal for the screen printing process. Unless youare
scanning the images yourself, or the reprohouse is willing to
change the settings on itsscanner to accommodate your particular
screenprinting parameters, some image editing will benecessary. In
this case, it may pay to have scanssupplied as RGB files and handle
the colourseparation yourself. The bonus of working withRGB files
is that they are smaller than CMYKequivalents, so processing times
are faster andcomputer memory overheads lower.
THE PROCESSING STAGEOnce the original image has been scanned,
thedigital image file is transferred to a desktopcomputer or
workstation (see, File formats andtransfer) for processing.
Computer processing may involve the use of animage editing
program to alter, say, colourbalance or repair scratches. An image
editingprogram may also be used to separate the imageinto CMYK
colours and set relevant parameters for example, dot gain and black
generation (howto control these variables is explained later.)
The amount of image editing already carried outat the scanning
stage will determine what isnecessary at the processing stage.
Experiencedscanner operators using high end equipment andsoftware
can handle most of the essential colourseparation work, provided
they are supplied withall the relevant information regarding
theseparation parameters required for the printingpress and
substrate in question.
FILE FORMATS AND TRANSFER
Digital image files can be saved innumerous different formats.
Someformats are more open than others they can be read by a wide
rangeof computer programs on differentcomputer platforms (such
asMacintosh, Windows, Windows NTand so on). Suitable formats
include:
TIFF (Tagged Image File Format)
EPS (Encapsulated PostScript)
DCS (Desktop Colour Separation)
Avoid the use of lossy compressionformats, which discard digital
data inorder to reduce file size. Non-lossycompression formats,
such as TIFF LZWand non-lossy JPEG (Joint PhotographicExperts
Group) files are fine, as they donot degrade image quality
noticeablyand help to keep file size small.
If you are using a specialist reprohouse, ask which file formats
theyprefer to work with usually, eitherEPS or DCS. Similarly, if
you arehandling any of the repro process in-house, you will need to
specify whichfile formats you wish your customers touse. This will
tend to be dictated by thesoftware you are using, though thethree
file formats mentioned above arecompatible with all of the
mostcommonly used programs.
3.7 3.8
8-bit data is needed to produce smooth tonaltransitions in
digital images. This is because theeye can differentiate between at
least 150 levelsof gray (brightness levels). 7-bit data enables
apixel to show only 128 levels of gray, meaningthat 8-bit data
contains the least amount ofinformation necessary to reproduce all
the graylevels the eye can see.
With an RGB colour image, 8-bit data (256 graylevels) is
required for each colour channel. Hencethe term, 24-bit colour (3 x
8-bits). When adigital file is separated into the CMYK colourmodel,
extra colour and tone information isrequired, as there are now four
colour channels,each requiring 8-bits of data. Hence, the
CMYKcolour model is often referred to in terms of 32-bit colour (4
x 8-bits).
Some scanners are able to capture more than 8-bits of data. The
latest desktop flatbed scannersare capable of capturing 12-bit data
and top of therange repro flatbed and drum scanners cancapture
16-bit data.
The primary reason for capturing this extra data isto allow
pixels to show a wider range of subtleshadow details an important
consideration whenworking with high density transparency films,
suchas Fuji Velvia. Image editing software is now alsocapable of
handling 16-bit data, which means thatthere is more scope for
colour correction of digitalimages and greater control when it
comes toseparating RGB images into the CMYK colourmodel.
-
FILE TRANSFER
At some point in the repro workflow you will need to transfer
files between different systems for instance, from the scanner to
the computer, or the computer to the imagesetter. You will alsoneed
to have a facility for your clients to transfer their layout files
to your system and you willneed to transfer files to and from a
repro house if you are sending out part of the work.
The most commonly used means of transferring files is via
removable disk drives. The disk is used tostore the data. They are
relatively cheap to buy and the removable disks mean that storage
capacityis, effectively, unlimited. Typical drives include:
Computer processing equipment can be roughlysplit into two
types: desktop computers and highend graphics workstations. Desktop
computers aremuch cheaper than workstations, but workstationstend
to process data faster. The right choice foryour operation will
depend on numerous factors far too many to be covered in the space
availablehere. Sericols Imaging Team or a specialistcomputer
consultant will be able to advise on theoptions.
Colour monitors Again, there are a number ofkey factors to
consider when you are choosing amonitor for repro work. The most
important are:
screen size (the bigger the better, as it reducesthe amount of
scrolling and screen redraws, whichincrease editing times)
dot pitch the size of the individual screenpixels (smaller is
better)
display resolution the number of pixels whichcan be shown on the
screen (the higher thebetter)
bit depth how much colour information amonitor can display (a
24-bit colour display canshow more than 16 million colours)
refresh rate a high refresh rate helps toprevent the display
image from flickering
calibration profiles some monitors can becalibrated so that
on-screen colours more closelymatch both the colours captured by
the scannerand the ultimate colour of the printed image
Add-on graphics cards can increase the rate atwhich the screen
redraws. Similarly, graphicscards or extra video memory can allow
you todisplay higher resolutions (more pixels) or agreater number
of colours (higher bit depth).
The monitors surroundings must also beconsidered carefully, as
reflections and glare canaffect on-screen colours and contrast, as
well asthe comfort of the operator. SericolsEnvironmental Services
team or a computerconsultant can advise on these and
otherconsiderations.
Software Most important of all is your choice ofsoftware, as
this will determine the level ofcontrol you have over the final
printed image.
Adobe Photoshop has established itself as theindustry standard
image editing program. Itcontains an extremely powerful set of
digitaltools, which afford a remarkable amount ofcontrol over the
image editing and separationprocesses. (Photoshops most useful
tools areexplained in the following section.)
The major page layout programs are currentlyQuark Xpress and
Adobe Pagemaker. Bothprograms afford precise control over
thepositioning of elements on a page. They areinvaluable for
arranging text, graphics and digitalimages within a design and can
also be used toset separation parameters, halftone screen
rulingsand so on. They also allow different scanresolutions to be
brought together on the samepage. All images are represented on the
page bylow resolution screen images which are linked tothe high
resolution file used for output. Thisresults in faster processing
times.
Page layout files often contain low resolutionpositionals to
show where the scanned imagesshould be placed and how they should
becropped. Once the scans have been made andedited, they can be
substituted for the lowresolution positional. Clients may even
scanoriginals themselves and supply the files at the
correct size and resolution for output. In this case,all that is
required is to open the scans in animage editing program to check
that the correctseparation parameters for your printing presseshave
been specified.
Digital printers Proof prints of the completedlayout file can be
output from a digital colourprinter (such as a 4-colour dye
sublimation printer)for client approval. However, this type of
proofcannot be used to check dot gain, exact colourbalance and so
on, so the proof is only useful forchecking that the design layout
is correct.
PostScript files Once the image editing has beencompleted and
the page layout has beenapproved, the page layout file must be
convertedfrom so-called application code (the code used bythe page
layout program) to a PostScript file. Theindustry-standard
PostScript page descriptionlanguage can be read by nearly all
output devicesand contains all the information needed for
animagesetter to output a file as 4-colourseparations.
THE OUTPUT STAGERIPs The PostScript file is transferred to what
isknown as a raster image processor (RIP). This maybe
custom-designed hardware or, more likely, asoftware emulator which
can be run on anysuitable computer. Its job is to rasterise
thePostScript file (convert the data into a matrix ofhalftone dots
the raster which the imagesettercan subsequently output to film).
The RIP can alsohandle the colour separation process if
required.
Imagesetters The rasterised file is transferred to animagesetter
for output to the four film positives one each for the CMYK inks.
The imagesetter usesa laser beam to record opaque, fixed-size
dots,known as device pixels, on to the films lightsensitive
emulsion. The minute dots (measuringmere hundredths of a millimetre
in diameter) areoutput at very high resolution a high endPostScript
imagesetter can achieve an outputresolution in excess of 3,200 dots
per inch (dpi).
Once the film has been exposed, it is processedchemically using
a carefully controlled processingunit, which maintains the correct
chemistry,temperature and processing speed for correct
3.9 3.10
-
S
e
r
i
c
o
l
L
i
m
i
t
e
d
1
9
9
8
CHAPTER 3
How much RAM?Image editing programs require a largeamount of
random access memory (RAM) towork quickly and efficiently. When
youopen a file, the computer attempts to loadthe entire file into
RAM. If insufficient RAMis available, the computer can load only
partof the file being worked on at any giventime. This means that
it has to continuallyswap data between RAM and the hard disk,which
slows down processing speedsconsiderably.
As a rule, you need to install at least threetimes the amount of
RAM as the largest fileyou are likely to be working on. This
allowsthe computer to hold the entire file in RAM,plus the last
unsaved version (allowing youto undo a manipulation if you dont
like aneffect), and still have enough availablememory to perform
editing operations.
With any new computer, check for themaximum amount of RAM you
can install, asthis will dictate the largest file size you canwork
on comfortably.
Jaz drives very widelyused, offer largeamounts of storage(1 or 2
gigabytes)and are currentlythe best option forthe transfer
ofdigital image files.
Zip drives cheap, widely used,but individual disks have limited
storagecapacity (100 megabytes a tenth of thesize of the smallest
Jaz disk).
Optical disks widely used, very robustand offer a large storage
capacity, but areslow to read and write to.
Rewritable CD-ROMs widely used, robust,inexpensive and have a
reasonablestorage capacity. However, writing data to acompact disc
is more complicated andprone to errors, compared with other
transfersystems.
-
results. The film can then be checked for problemssuch as
unwanted fringing (halos) around thehalftone dots, low dot density
and excessive dotgain.
A high end imagesetter is an expensiveinvestment. There are,
however, alternatives,including the most recent innovation
thethermal imagesetter. This type of imagesetter usesheat to record
dots on a special heat-sensitivefilm. It is considerably cheaper
than high end,photo-based versions and is very quick and simpleto
use. However, it cannot match the very highoutput resolution found
on more expensiveimagesetters.
Fig 3.6. High end imagesetters can output film positives atvery
high resolution, but are very expensive to buy andoperate. The
Aspect 600 thermal imagesetter has a lowermaximum resolution, but
represents a far more cost-effectiveoption for many in-house repro
studios.
Output resolution
An imagesetters output resolution has animportant role to play
in determining the qualityof the final print. As explained
previously, a digitalimage requires 256 levels of gray per
colourchannel to reproduce smooth tonal gradations andaccurate
colours. The total number of gray levelsthat can be reproduced on
the final print isdependent upon the halftone screen ruling andthe
output resolution of the imagesetter. It iscalculated using the
following formula:
output resolution2
halftone screen ruling2
Looking at this formula, it is clear that, providedthe halftone
screen ruling remains the same,higher output resolutions will
provide a largernumber of gray levels and, as a result, a
smoother,more accurate print. Similarly, higher outputresolutions
will be required when working withfine halftone screen rulings such
as 133 lpi or150 lpi if you are to guarantee the reproductionof 256
levels of gray.
The following examples illustrate the relationshipbetween screen
ruling, output resolution andlevels of gray:
1. Suppose you were to try outputting your fine150 lpi film
positives using a 600 dpi laser printer.The total number of gray
levels would be equal to:
6002 (dpi)1502 (lpi)
It is not possible to reproduce a continuous toneimage
faithfully using just 17 levels of gray theprint will have a
posterised appearance.
2. Having obtained an unsatisfactory result usingyour laser
printer, suppose you decide to send thejob to a repro house for
output on a high endimagesetter with an output resolution of 2540
lpi.The total number of gray levels would now beequal to:
25402 (dpi)1502 (lpi)
288 levels of gray is ample to produce a smoothimage which
exhibits accurate colours.
3. Finally, suppose you decide to print the job onthe laser
printer once again, but using a muchlower halftone screen ruling
for instance, 37 lpi.
The total number of gray levels will now be equalto:
6002 (dpi)372 (lpi)
As the above examples show, output devices witha low output
resolution are capable of producingquality results, provided the
halftone screen rulingis not too high. Indeed, the use of a high
endimagesetter for this type of work would besuperfluous. To
produce smooth tonal gradationsand accurate colour with higher
halftone screenrulings, however, requires the use of an
outputdevice with a high output resolution in theregion of 2,500 if
you are attempting to print atvery fine screen rulings, such as 150
lpi.
Fig 3.7 The relation of halftone screen angles toscreen mesh
stretched at 5 from base right.
Consider using a standard specificationsheet (see sample at back
of this manual)and use a symbol for the orientation of thehalftone
angles you require. On an ellipticaldot halftone screen, the angles
should bespecified in relation to the direction that thedots
join.
Fig 3.8 Use a simple symbol to indicate that all theangles are
measured from base right.
3.11 3.12
HARD AND SOFT
An important point to look out for is thetype of dot the
imagesetter records. Oldermodels tend to record soft dots,
whichhave a halo around their edge. This cancause colours to be
reproducedinaccurately at the printing stage. Newerversions record
hard dots, whichreproduce colour more accurately on thefinal
print.
OTHER OUTPUT CONSIDERATIONS
A commonly overlooked differencebetween offset litho printing
and screenprinting is the orientation of the emulsionon the film
positive (the emulsion carriesthe halftone dots). Offset requires
theemulsion to be downward when looking ata correct reading image.
Screen printingrequires the film emulsion to beuppermost, so that
this side is in contactwith the stencil emulsion when exposingthe
screen. If the film emulsion was on theunderside you would be
exposing throughthe thickness of the film base and lightwould
scatter under the image. This wouldresult in a loss of fine detail
and, in thecase of halftones, a complete loss of dotsin the
highlight areas.
So, remember to specify: emulsion up right reading.
Also, you must be consistent in how youspecify the halftone
screen angles yourequire when you are using angled mesh.Many
imagesetter software programsmeasure the angles from a base right.
Ifyou have established a set of anglessuitable for the mesh you
print through maybe, one screen has to have the meshangled to avoid
a clash between it and thepositive measuring angles from a baseleft
would produce a totally differentrelationship to the mesh.
Differential betweenhalftone and meshangles = 70
Halftone angle 75 frombase right
Differential betweenhalftone and meshangles = 80
Halftone angle 75from base left
Gray levels = + 1 = 17
Total gray levels = + 1.
Gray levels = + 1 = 288
Gray levels = + 1 = 264
-
-
S
e
r
i
c
o
l
L
i
m
i
t
e
d
1
9
9
8
Printing inks are not perfectly transparent, so the order in
which they are printed can have a marked effect on the appearance
of the image.
The gamut of reproducible colours using CMYKinks is limited
compared with that of the RGB colour model. This means that there
may be noaccurate CMYK match for some colours contained in an RGB
scan. (A Gamut Warning control allows you to preview which colours
cannot be matched precisely, before you proceed with the separation
process.)
Increases in dot percentage values do not produce a uniform
increase in colour density - that is, colour change is not uniform
across the tonal range.
Fig 3.9 This RGB colour specification cannot be mapped
toCMYK
In fact, the mathematical equations used toaccurately map RGB
colours to CMYK ink valuesare extremely complex and require a great
deal ofprocessing power to calculate. To speed up theseparation
process, therefore, separationprograms store the equations for
individual colourconversions in a colour look up table
(CLUT).During separation, the colours of each pixel in theimage are
compared with the colours in the table:if an exact match is found,
the pixel is given theappropriate CMYK value; if no exact match
isfound, the nearest CMYK value in the CLUT isused.
Off-press proofs (or contract-quality proofs),such as Cromalin,
Matchprint and Agfaproof, areused for client approval before a job
goes into fullproduction. However, they have been developedto meet
the needs of the litho printing industry, sothey give a less
accurate representation of howthe image will appear when it is
screen printed.(The possibility of tailoring dot gain on a
screenprinting press so that it matches the dot gainfound in
off-press proofs is discussed in Chapter6.)
For absolute accuracy, you can run a press proof.This is the
only way to determine exactly how theparticular substrate and ink
system will affect dotgain under the specific production conditions
usedto print the job. By running a series of tests onvarious
substrates using the appropriate inksystems, you will have the
information necessaryto predict fairly accurately the amount of dot
gain,without having to proof each job.
Any problems that might arise during proofingcan be corrected by
further editing of the imagefiles. For example, suppose a models
faceappears slightly too red. A densitometer testshows that the
problem is caused by a higher thanexpected dot gain reading for the
magenta screenin the fleshtones. Instead of experimenting withink
formulations to try to correct the colourbalance on-press, it is a
simple matter to return tothe processing stage, alter the image
file tocompensate for the dot gain and re-run the films.Provided
the press set-up remains unchanged,you can be certain of an
accurate result. In thisway, colour can be controlled without the
need toconstantly modify the inks. Careful control andmodification
of the halftone film positives allowyou to reproduce any image on a
wide variety ofdifferent substrates, using standard 4-colourprocess
ink sets. It goes without saying that thisreduces the number of
different halftone inks theprinter needs to carry, introduces a
strongelement of standardisation and ultimately speedsthe print
process considerably by removing thetrial and error approach to
getting colours righton press.
The next step, then, is to look at how computersoftware can be
used to achieve accurate colourson the finished print. (Adobe
Photoshop 5.0 willbe used throughout to demonstrate the
varioustechniques involved.)
CONTROLLING THE KEYHALFTONE PRINTING VARIABLESThe most powerful,
and important, feature ofcustom scanning software and image
editingprograms is that they allow fast and accuratecontrol over
the all-important colour separationprocess.
Colour separation
In order for an RGB image to be printed, it mustfirst be
converted to the correct CMYK values, sothat film positives can be
produced and screensexposed. This process is known as
colourseparation and is carried out automatically byscanning or
image editing software all you needdo is select the correct colour
model from a pulldown menu and click the mouse. However,
theuser-friendly manner in which colour separation isselected
belies the very sophisticated processresponsible for the actual
conversion from onecolour model to the other. While it is not
essentialto master the maths, it is useful to have anunderstanding
of how the separation works,especially if you wish, at some stage,
to bring thedigital repro process in-house.
Mapping colours During the separation process,the computer maps
the RGB values for each pixelin the image to the appropriate CMY
values andgenerates a value for the black ink. (Blackgeneration is
covered later.)
Mapping RGB values directly to CMY values,however, does not
produce an accurate result.There are a number of reasons for
this:
Even the best printing inks have inherent impurities, which mean
that their spectral properties are not perfect. In effect, this
means
that the different coloured inks absorb some of the light they
are supposed to reflect and reflect some of the light they are
meant to absorb. Cyan ink is especially problematic as it tends to
absorb a high amount of the blue and green light falling on it.
This reduces the amount of blue/green light reaching the eye, which
leads to a warmer image on the final print. (For this reason,
separation software will always reduce the level of magenta and
yellow
ink when they are printed together with cyan.)
COLOUR LOOK UP TABLES
Some colour look up tables (CLUTs)contain more information than
others.High end scanning software works witha very detailed CLUT
containing a largenumber of RGB/CMYK colour matches,so the
separation process is veryaccurate. Image editing programs tend
towork with scaled down CLUTs (forreasons of processing speed),
which canlead to less accurate separations.However, the
sophistication of the CLUTused with high end scanning softwaremeans
that it takes a long time tocalculate, so repro houses tend
tostandardise on a table which gives thebest results for printing
SWOP(Specifications for Web OffsetPublication) inks on to coated
paper. Theadvantage of using an image editingprogram is that the
smaller size of thetables means that the image editingsoftware is
able to contain multipleCLUTs. This means that it is possible
tocreate and store a variety of tablesgeared for screen printing
with differenttypes of inks.
3.13 3.14
CHAPTER 3
Black generation Theoretically, printing cyan,magenta and yellow
inks together should produceblack. However, this is not the case:
the threesubtractive primaries alone cannot produce a highenough
density to render a true, rich black - theyproduce a muddy brown
colour. A vital part of theseparation process, then, is the
generation of ablack ink value.
Most colours in a scanned image will containvarious amounts of
each of the subtractiveprimaries. One or two of them will be
dominantand will dictate the hue. The remainingcomplementary
colour(s) has no effect on the hue,but simply makes the colour
lighter or darker. Thismakes sense when you recall that
equalproportions of cyan, magenta and yellow combineto create gray.
The equal amounts of the threesubtractive primaries contained
within any colouris known as the gray component
Gamut Warning
-
Fig 3.10. Equal proportions of three subtractive
primariescombine to make gray a colours gray component.Removing
equal amounts of the three colours and replacingthem with black ink
reduces the total ink weight printedwithout noticeably altering the
hue.
If you remove equal levels of cyan, magenta andyellow (the
amount is dictated by the level of thecomplementary colour) and
replace them withblack, therefore, the hue will remain the
same(only the gray component will be removed), butthe total amount
of ink printed will be reducedwhile the maximum density will be
increased(thanks to the inclusion of the black).
More importantly, it is possible to reduce thetotal ink limit
(the maximum weight of inkdeposited on the substrate). Controlling
the totalink limit is especially important as a high ink buildup
can cause a number of problems:
A heavy deposit of ink is more likely to cause smearing of the
halftone dots, reducing the definition of the image and affecting
colours and tones.
A high build up of ink over successive colours may cause the
height of the ink deposit to impair the lay down of the final
colour. Simply put, all of the ink printed through the final screen
may not be able to reach the substrate.
The slightest misregister in areas where there is a large
deposit of each of the three subtractive primaries can reveal
unwanted gray tones within darker colours.
Higher ink weights can lead to increased dot gain. (By limiting
the maximum dot sizein shadow areas, there is less filling in, so
alower total ink weight will help to preserve shadow details.)
Heavy ink deposits can lead to drying problems and prolonged
drying times.
For the reasons outlined above, the generation ofa black value
involves replacing a certain amountof each of the subtractive
primaries with black.
There are two methods by which black values canbe generated:
under colour removal (UCR) andgray component replacement (GCR).
UCR reduces the amount of cyan, magenta andyellow in the darkest
neutral and near neutralcolours of an image and increases the
amount ofblack accordingly. UCR becomes active in any areaof the
image where the total ink percentage ofcyan, magenta, yellow and
black exceeds thespecified total ink limit.
To select the appropriate UCR settings for yourpress and
printing parameters, first open theSeparation Setup dialogue
box(File/Preferences/Separation Setup) and click onthe UCR radio
button. Then set the total ink limitby entering the percentage
value in theappropriate box. (If it were possible to print a100%
dot in all four colours, the total ink weightwould be 400%. In
practice, it is usually setbetween 240% and 340% depending on
theprinting conditions. You need to determine theoptimum total ink
limit for your press and thetype of ink you are using.)
An area where UCR has been shown to beadvantageous is printing
with UV curing inks.These inks have a low volatile component, so
inkbuild is high, which can cause problems, asdescribed above.
UCR improves ink transfer in the darkest shadowareas of the
image because the amount of inkpresent does not exceed the optimum
level thatthe screen printing press can produce. Thestability of
the gray balance is also improved,becaus