Density variation and spread of the printed halftone dot in offset lithography

Rochester Institute of Technology

RIT Scholar Works

Theses

Thesis/Dissertation Collections

1-1965

Density variation and spread of the printed halftone

dot in offset lithography

Walter L. Covington

William L. Heaps

Follow this and additional works at:

http://scholarworks.rit.edu/theses

This Senior Project 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 [email protected].

Recommended Citation

DENSITY VARIATION AND SPREAD

OF THE

PRINTED^HALETONE DOT IN OFFSET LITHOGRAPHY

Senior Research Project

Presented to

The Senior Research Committee

Rochester Institute of

Technology

In Partial Fulfillment

of the Requirements for the B.S. Degree

In Photographic Science

by

Walter L. Covington and William L.Heaps

ABSTRACT

No detectable functional relationship was found

between dot area and dot

density

within a solid ink

density

and paper combination over a 65 to 300 line

screen _range.

Dot _spread, measured as per cent difference from the

per cent dot value of the half tone negative, was found to

be more significant in the midtones"and the 300 line

screen w5s observed to increase spread _appreciably as well

INTRODUCTION

In _making offset lithographic halftone reproductions

from originals there are a great _many variables which

contribute to the final product. Tfte most significant

ones are

ink,

paper _stock, screen _ruling, and press con

ditions. It is the

objective-of this paper to show the

significance of dot

density

as a ftmction of dot area and the variation of dot size (the dot spread) among a par

ticular set of tints as a function of change in paper

stock, ink tackiness and screen ruling. These variations

affect the _accuracy of tone reproduction studies. In

light of previous

investigations

_suggesting that such

functional relationships _may exist and the need for an

easier, sore reliable method of _predicting tint

density,

these areas

justify

this report.

The features of tone reproduction _study in offset

lithography

can be better _{imderstood, perhaps,}

by

compar

ing

them with tone reproduction studies in photography.

Studies in the latter case -re made

by

_comparing macrc -'

density

readings of a sensitometric _step wed/,e in each

11

at the macro level of

investigation,

but consists of dots

varying in size, we must measure both the

density

of the

dot as well as its relative area. If the individual dots

vary in

density

we also have an additional factor to con

sider;

however,

presently this is assumed to be a constant.

Considering,

first,

the ideal case in a

halftone,

the follow

ing

equation has been written:

A =

aAs

which states that the total absorption

(A)

is

a product

of the absorption of the dots (a solid area As) and the

area of those dots. If we transform this equation to den

sity

terminology

we have: i

Dt

=

-log(l-a(l-antilog^Ds)).

This is a form of the Murray-_{Davies equation where}

D-f-is the

density

of the tint (defined as

having

a constant

dot area) and

Ds

the

density

of the solid.

This

formula,

as _written, does not account for light

penetration and multiple internal reflection of the paper

stock, and therefore a factor

(n)

was acded to account

for this effect: ?

w

*

Dt

=

-n

log(l-a(l-antilog

_-Ds/n))

This modification

by

Yule and N&ilsen increases the cal

culated

density

and gives a better fit with observed

actual curves. The empirical range of the n value

is

from

1.0,

for an

ideal

_case, to approximately

2.0.

Another possible variable

facilitating

a better

Ill

be a variation

in

dot

density

(solid

ink density)

with

change

in

dot area. In order to test whether a functional

relationship exists, the Yule and Neilsen equation speci

fied parameters over a range of tints and screen _rulings,

which must be

investigated.

Much work has been done in the past few years in an

attempt to explain the factors _affecting the

density

of

I

6 a tint and

being

able to predict a tint density.

With the advent of the microdensitometer and the

-Ansco Model 4

having

been made available to R.I.T. students,

there was provided a method of

investigating

the

density

fluctuations within halftone dots. Major investigations

of dot siae and dot

density

have been carried out with

L.A.R.D. (Large Area Reflection _{Density) dot-area meters,}

probably because the type of investigation needed was so

laborious. The Dietzgen planimeter (made available through

the R.I.T. Mechanical

department)

and photomicrographic

enlargements of tints provided less _variability

in

measure

ments of dot area than can be acquired

by

_estimating area

from dot diameter measurements because the _perifery or the

lithographic dot is not uniform and does not lend itself

MET^ODl

Investigation of a Functional Relationship Between

Dot Area and Dot

Density

A project design was

formulated'by

which this pos

sible variable could be investigated over a wide range

of dot areas and over three screens. If a _relationship

was found under one set of conditions it would be inves

tigated over a three paper and two ink combination. How

ever, no significant _relationship was found.

Based upon the above _conclusion, the project was

subsequently expanded, yet still within the data _obtained,

to investigate the _relationship between the 'n* _factor

for light penetration and screen _ruling and the dot _area,

tint

density,

and solid ink

density.

This

investigation

involves regression analysis of a four factor curvilinear

mathematical expression. This phase was terminated in the

last available

day

due to complications which arose in

programming a Murray-_{Davies equation to derive the} 'n'

factor

from area, tint

density

and solid ink density.

The project, to

date,

has involved four major divi

sions of operations: photomicrograph enlargements of dots

within the tints_; micro densitometer scans of the dots

within the

tint;

'planimiterizing' _{the dots}

on the photo

L.A.R.D. measurements of the selected tints along with

conversion of microdensitometer densities to L.A.R.D.

based values.

Preliminary

work involved all procedures needed to

produce the lithographic halftone _step wedges from which

all data was obtained. For these

tints,

we owe our appre

ciation to the Graphic Arts Research Department at R.I.T.

This work included _platemaking and press runs. The half

tone negatives were made from halftone positive original

screen scales for use in platemaking. The plate was of

the _{presensitized, negative-working} type.

The pressrun was made with a three paper and two ink

set of combinations. The papers were of three general types:

1. Coated stock - _white

2. Uncoated stock - _white

3. Medium-hard surface newsprint - buff

The two inks used were :

1. 1# Quickset offset semi-gloss

4/D

process black BOPG - 1833 - I lot

#38-171 8/27/63

3y

Pope &

Grey

Ink

2. 5# Web offset book black x-19880

OA .7073 _10/12/64

Inter chemical Corp. _{printing ink division}

All three screen rulings were reproduced well with

the Quickset

ink; however,

_{the 300 line screen tints}

a-bove 0.40 dot area did not resolve the white areas be

tween dots with the web offset

ink.

Two test objects (see ill.

2,3,4)

were selected for

_- J*

a

h

4

^

*

^i&ss^ftsftsss

ili::;:

, i&m

IMM

SEE:''':'XX;:' .

:-fsSWSpl*

!//*

__ o -*-CP

._ _ a

M

___ -^

(7/6_^9S

^

Z

/ ^ /, _{^ ? /_}

e

IEf&M

fliEX

zmzii zzzzz*

newsprint

Sk>_c

>l

V

si

N

1^-Ko

-

A?

IV

to

N

A. Photomicrograph Enlargements

The purpose of this phase of the lab work was to

obtain a value in per cent dot area for the halftone tints.

Since the

information

desired is a fractional area rel ative to a unit area this eliminated the _necessity for

any absolute area measurements of the dots.

The facilities of the P.. I.T.Z Graphic Arts. Research

Department were used for photograph

ing

the dots through

a Bausch & Lomb microscope. The standard B&L photomicro

graphy apparatus was used _containing a bellows with a

Graflock 4X5

back.

This allowed convenient use of standard

4X5 film holders for the work. The film used was Super

Panchro Press Type B.

Preliminary

experiments were executed to determine the proper exposure and magnification. An exposure series

with a range of 4X showed no discernable image spread

due to the photographic _emulsion,

indicating

that exposure

time isn't too critical* A

limiting

factor of dot mag

nification is the number of dots that should be present

in one exposure. A minimum of four dots is needed and a

maximum of sixteen dots is

desirable.

This _reasoning determined the use of a

4X5

format for _recording the

dots,

allowing sisteen per negative. *(see 111.

5)

-t-ecuwr^

4wr

A min:;um

of four dots is,r4ueXo"Xhe fact that

a unit area which is _{easily defined} when the number of

dots chosen form a square.

It was decided to mount the tints to be photographed

on glass slides

by

_using rubber cement as the adhesive.

An appropriate

labeling

system was used from the original

identification

of each tint to the final enlargements.

The photomicrographs were made from the tint numbers of

each _line screen test object from which

density

measure

ments were made. A press sheet was chosen at random from

the pile of printed sheets and the test objects from the

lower left quadrant were _used, (see HI. 2 & 3 \gage

4).

The negatives were processed twelve at a time in one

gallon tanks with a water jacket control

ing

temperature to

+ 1F- The _enlarging of the negatives was quite laborious

and time _consuming since the negative densities weren't all

the same. The main variable here is the change in expos

ure due to the changes in eyepiece magnification on the

microscope (the objective was kept at 10X and the

eye-<7i^>-;^yxX piece varied from

20X

to

OX).'

The negatives were grouped

according to their

density

and dot image size and this made

more efficient use of _{printing time.} The prints were pro

cessed six at a time in D-72 1:2 for two minutes and. went

through the normal

finishing

procedures (see Senior Re

search Project notebook #1 - Walter 1- Covington for _more

detailed

information).

See page

3lA

for a typical photo

8.

B. Microdensitometer Scans

The method of _scanning dots with the microdensitom

eter

involved

_many considerations such as direction of

scan across a dot

("diagonally

or

horizontally"),

aperture

changes (can

they

_{be made?),} magnification _changes, scan

ning speed,

focus,

etc. As there are

few,

if _any, stan

dards to go

by

in this

field,

much trial

testing

of the

apparatus behavior and operation was done before _any

data could be obtained.

An "aperture series"

test was run on the smallest

dot of the 300 line screen to see whether the instrument

design would allow it to be scanned. Calculations and the

instrument agreed and we were able to extend the inves

tigation to the 13% dot of the 300 line screen.

The apertures and magnifications used for _scanning

were as follows :

300 line screen

Dot area Magnification Aperture diameter

eye p. obj.

0.13 20x llx 0.50mm

0.42 " " "

0.64 " "

2.50mm

0.80 " " "

0.06 5x llx 0.25mm

0.20 " " "

0.51 " ti

2.50mm

0.76 " "

*2n

5x llx 0.50mm

u.^y tt n tt

-5?

" "

2.50mm

0.74 tt

2 Ox llx

tt tt

tt tt

ft tt

150 line screen

5x llx

tt tt

tt tt

tt tt

65 line screen

5x llx

tt tt

tt tt

A _reading of the base

density

was taken before and after

each four dots scanned because of excessive drift.

In addition to scans of the test objects of which

there were

12,

a scan was made of a continuous tone _step

wedge made from an ink run down

image

of the control bar

(see 111. IB). This _step wedge gave us the means of pre

paring a conversion graph from microdensitometer readings

to L.A.R.D. readings. The L.A.R.D. measurements were

made with the.

instrument

having

been calibrated and zeroed

on an _{N.B.S. plapque. This information} is reproduced in

graph

#2.

As can be observed from the _graph, when the

M.D.

(microdensitometer)

measurements are converted to

N.B.S. the straight line passes through the origin. The

"run-down wedge"

was put together from control bar _seg

ments produced on coated stock with Quick set offset

ink.

Much work was done in justifyingthe_{faet that aper}

tures and magnifications could be changed at _will, with

proper controls. As these require _many

lengthy

explana

tions and data

testing

they

are not reproduced

here,

but

may be corroborated in the notebook of _W. L.

Heaps,

Vol.

1,

pages

30-43.

A method of _calibrating or of _checking cal

ibration

by

_using the range of available circular aper

tures (instead of filters as a means of attenuation of

light reaching the photocell) was also devised and docu

mented with _{supporting data} in the pre-listed notation.

In preliminary

investigating

of the effect of focus

magnification a *

20%

turn of the "fine focus"

adjustment produced no noticeable change

in

the output

density.

However,

in

investigating

the effect in _scanning

a dot above or below the subjective center,

it

was found

that if the scan

is

made

25%

above or below the center of

the dot it alters the

density

output considerably.

Therefore,

the smallest dots scanned which do not represent

a well defined _area, exhibit a high variance even when

8-10 measurements are taken within a single dot scan.

All scans were made with considerable attention paid

to

focus,

_positioning with the _{viewing scope,} base readings

and calibration checks. Four dots were scanned in each of

four tints in each of three screen rulings on e-.cn. of 6 paper

Preliminary Microdensitometer Tests of Data Collection Methods

The method of _{obtaining the} mean

density

for a dot is shown

graphically on page _11. The white paper stock

being

used as the

base

density

value for all data normalization. The two sample scans

are taken from scans of Illus.

#f,

Newsprint Stock

(150

line screen,

step 10 and 300 line _{screen, step}

7)

page

4.

Upon _observing the M.D. scan of the 150 line - _screen

dot,

one will notice that observations were taken at _every inch

whereas with the 300 line-screen

dot,

observations were taken at _every

l/20th inch along the scan. The reason for different linear observa

tion intervals is that with different magnifications of the dot

(see page

8-9),

22QX for 300 line - _screen dots _and 55X. for 150

line - _screen

dots,

the _M.D. intigrates to different degrees over

a unit area here at a ratio of 1:4. An attempt to equalize this was

made

by

_making more observations per inch of scan where the instrument

was

doing

less

intigrating,

as in _scanning 300 line - _screen dots.

This resulted _{in approximately four times} as _many observations per

inch _along the 300 line screen scan as _{along the} 150 or

65

line

-screen scan.

The method of

determining

the initial and terminal cutoff

points in

taking

observations from the scan across each individual

dot was selected on the basis of the

following

method testing.

A

05$

dot of the 300 line screen was scanned at 220X magnifi

cation with apertures of different size. Proceeding on the _assump

tion that each scan should give the same mean

density

(see page

lib),

The affect of _changing apertures on the M.D. is similar to changing

magnifications and

is,

in a similar way, affecting the degree

of intigration

by

the instrument. The small dot was chosen for the test

because it is the extreme case of where observations are taken on the

scan of a _{dot making} a significant difference in the mean

density

value obtained for that dot. The importance of this effect, logically,

decreases as the size of the projected image on the aperture increases.

The three methods of _sampling which were tested are:

1)

Selecting

points above base where the slope of the curve

is zero and

determining,

by

calculation, the mean.

Average values obtained in this manner from the scans

page lib are :given under Method 1 below.

2)

Selecting

points above

\

average

density

at equal linear

distance _along the trace. These mean values axe given under

Method 2 below.

3)

Selecting

points at equal distance intervals inside the

extreme points above base where the slope is zero

(as

shown for

13$

_{dot, step?,} page

11,

by

the slash marks on

the scan). Values under Method

3

below.

Method Analysis Data Aperture size Average _Density Values

Diam. Method: 12 3

Dot scan L. to fi. as

reproduced on p. lib 5.00mm .69 .62 _.69

3.75mm .70 .61 .68

2.50mm .72 .64 .70

1.25mm ,76

.63 .71

0.5Qn_a .79 .^> .70

3.75mm .70 .62 .70

With Method

1,

as substantiated

by

the data above, the mean

density

of the dot is significantly influenced

by

the size of the aperture to

Although _{both Methods} 2 and

3

above are _expressing tolerable

deviations from the _group mean

(instrument

_variability

fo.02),

one

can see that the values taken from the scan under Method 2 would

include observations when the _scanning aperture was not

totally

within the magnified image of the dot. The

density

value would

thus be reduced . For this reason and that of the paragraph above,

Method

3

was used for

determining

cutoff points on M.D. scans of

Graph

II

11.

TANCE

T.0.#/\8

2

Do+S.

150

line

Screen-

S

+_*p 10. D

]

Scan

Sample

o-f

\LLUS.*4.

A/ea/spriri+.

2.4i-T

x

-

2.2-4

to

-

T/

X

'

mtfiW-

--W

Bzm~w

JU

obj.

26*

_L

^

-n -nam >Us

_Q ___AH

Ap

cjeiurJC.

S$lC_

Si*,

e.

d5

1.6.

SAW

SWK

1.2-A/ewspnn+ T0,#/f */ Dots 30O/iwe scrn

S+ep

7

See If/US =*<f

< z i

1

1-i 1 "__._. __

2

Fajge lib

/

h

Ki

4-

1

1

1

l 1i

1

1 !

'3?

\'

s

H

L

.... _j

m

i

|

1 ! _A1IS

N3dn|

J2

i

J

P cc i_ ^r cm

6

or;

a co _rp co H co rj

_> "5T CM

C\i CM CM

P

-oq VJ- 4_-

^

<

E_

!

!

I

I

1

|L

i

1

--

|-1

L

TT")

i

J

₉

1

i)

^_>AJ i

0^-

?|

\

>

...

!

T

hS* w^ = 1

J

1

t

sir

1

1 . _-L__. 7

[<5#

V

1

1

_1

<

1

1

i

!

\

hr

^

r*+ _S

i

j

u

J)

i

\

i

!

1

i

1

1

1

<"

i

1 ! i

X

1 _\j

j

i

J

- 11

\\

i

1

rr

[

i

i

I

1 _Vs

IdU i- i i 1 i

2

. _E !

! I₁ ,

V

1

i i_i

i

j

1

"

-i

_

t\ J>

Sr

1

i

i

i

i

i

4

_CM

>

't

v i₁

i

1

i i

)

1

i i -X*--1 1 I 1 1 i i i

^

"^

1.3

1

1

" ~

1 i <i ' -*

|

.

1

* !" w

j

"

-*3

r

.

1

1 ₁

i

i /

l^T>

|

|

w

j

1 i i C1Z -'

<^

~ i _

/

i 1

1

1 r 1 1

| 1

u*

i

I

id

v

h

s

rv <

.

| X

/

p

1

1

ZA

rlt

i

i _A1ISN3(

1 ] <D 3: 00 en o CO co cm (J c

^f CM o

c\i ___

'

to <;_{t CM}

T*-xW

4&

1

I

j

i

i

Ii

i T^ i i i i

1

~ -i

J-i i _i <

I

12.

C. Planimeter Determination of Dot Area

To

determine

dot area a Dietzgen planimeter was

i

used which has an _accuracy to * 0.001 square

inches.

Since the

instrument

has

its

own standard square area

device

for calibration purposes this accuracy was _readily

determineable.

This meter has a _referencing arm with

one stable end and a moveable arm with a stylus to trace

the perimeter of the object.

Due to such factors as dot _spread, paper

density

etc.

it becomes a problem to determine the border between dot

area and white area. There is also the problem of how to

handle dispersed white areas within dots when

trying

to

determine per cent dot with the planimeter.

To give an

idea

of the method used in

determining

area of some 'non-uniform' dots please refer to

illustra-tion #6. The blue outline of the dot would be the typical

path of the planimeter stylus. Also note the hole at one

point on each dot. This defines the starting and term

inating

point of the planimeter'

s dot encompassment. A

relatively independent

grasping

component is provided to

facilitate _waassiA^fcy movement. Since the point of the

stylus

touches,

or is almost in contact with the print

surface, a paralax problem does not exist.

In order to calculate unit area the distance is de

termined from the

top

(or center) of the first row of

row of dots in the adjacent matrix. This value squared

will give the unit area

(assuming

a well made screen

was used to make the halftone negative). See 111.6 p.13.

The unit area divided into the summation of the dot

area gives the fractional dot area. It was originally

decided to do this with a sixteen dot _matrix, but upon

running an F test _comparing it with a 9 dot matrix, the

two derived s2.'sl were found to be of the same population,

i.e.

equal. See below:

F test

sx = _{Variance of 9 dot} matrix

o

S

= Variance of 16 dot matrix

H0: s_2 = s22

Hi: s_2

/

s22

F_{8 15,0.05} = l.78 X l0-2/7.6544 X 10"^

2.1185 + _13.1

15.

D. LARD Measurements of Tint Densities

Large Area

Reflection

Densities as opposed _to

M.D.

densities discussed under _B. were made with a Mac

beth RS-100 geroed and calibrated on an NBS (National

Bureau of

Standards)

calibrated placque.

Tint densities were then measured and replicated for

each of four tints on each of three screen rulings on each

of the 12

T.O.'s.

Thus four LARD measurements were used

in

_obtaining a mean value for each _step of a _paper,

ink,

and screen _{ruling combination. These} values were recorded

in

Vol.

1,

p.

44, 49,

of the notebook of W.L.Heaps.

These mean values were used

in

an attempt to test the

relationship "n=f(a)I'

This,

as stated _earlier, was curtailed

due to

insufficient

time availgble for analysis of data.

LARD measurements of the "run-doxm wedge"

provide a

Graph 2.

16,

17.

A. THE RELATIONSHIP BbTWSdN DOT AREA AND DOT DENSITY

To test the hypothesis that "dot

density

is a

function of dot area','

i.e.

Ds

=

f(a),

was the main

objective of our _preliminary project designs. The

mean

density

values from the dot scans are as listed

oti-pl.8.

They

were measured from T.O. (test objects)

#29A and

44A

which were on coated stock with a "Quick

Set" _ink

(see

Methods,

page

8,

9 ).

Each mean value is obtained from at least five meas

urements taken from the dot scan. This observation limit

of five was _emeerically emposed

by

the projected size of

the 300 line _screen, 0.10 area dot on the circular aperture.

Eight observations were taken from images of 150 line

dots and 10 observations from 65 line dot scans to de

crease the variance of the means. As can be seen from

the sample photomicrographs, there

is

not a continuous

density

level within a dot.

The group mean of each tint set of 8 sample means

converted to N.B.S. base

(+0.08)

along with two

estimates

of sigma

is

as follows-on page

18.

These were plotted against the planimeter measure

ments of fractional are converted to relative area num

bers to enable plotting of

300,

150,

and 65 line screen

halftone tint data on one horizontal axis. This plot

from the data clearly gives evidence to support

H:u=uQ

18.

M.D. Scans

(T.O.

#29A and

44A)

Mean

density

values.

SAMPLE MEAN

Screen Step#

2

2

I II III IV

300 line 1.07 0.92

D.06

0.70 1.2)1 0.98 0.86 0.66 4 4 0.87 1.32 1.22 1.17 0.98 0.90 1.12 1.21 5 5 0.90 1.17 0.97 1.26 1.15 1.09 1.00 1.20 7 7 0.95 1.28 1.13 1.13 0.98 1.38 1.06 1.19 150 line 2 2 1.01 0.98 0.74 1.14 0.98 0.13 0.89 1.18 4 4 0.93 1.12 1.04 1.02 1.10 0.99 0.86 1.03 10 10 0.87 0.88 0.98 0.97 1.04 0.97 1.01 1.04 13 13 1.04 1.16 1.00 1.05 0.90 1.08 1.07 1.00 65 line 14 14 0.86 0.93 1.11 0.89 1.07 1.17 0.94 0.88 11 11 0.93 1.15 0.96 1.15 1.09 1.06 1.04 1.08 9 9 1.05 0.91 0.960 0.99 1.05 1.05 0.96 0.95 4 4 0.93 1.07 1.00 0.99 0.99 1.09 1.02 0.98 s=0.l69 y=1.02 y=1.18 y= _l.SZ

y = _1.22

y = _1.06

y= _1.09

y = _1.05

$=0.04

y = _1.12

y = 1.06

y = 1.14

y = _1.07

y = _1.09

19.

extremely

small dots within the 300 line screen halftone

tints.

Between 0.13 and 0.80 dot area within this tint

there appears

by

the graph to be a relationship between dot

area and dot density. In _{statistically} examining this

variability

we employed a Bartlett's test of

homogeneity

of variance within the eight dots scanned to give the mean

value for the 0.13 area. From

this,

the smallest dot

scanned, we found that the variances are.equal to the _pop

ulation variance at a significance level of 0.05.

Sig

nificant data is given below:

Bartlett's Test on Variance

S;l2

log

s-l

v s_2

1. 0.00665

2.

0.01390

3.

0.00360

4.

0.02160

5.

0.03550

6.

0.04050

7. 0.03610

8.

0.01340

T5.6989

0.7334 Total

M = _11.2

C = _1.09

M/C =10.2

X2.05,7

=

^

10.2 is less than 14.1 Cone- _Variances are equal.

Subsequently an Analysis of Variance test to deter

mine the effect of area and error was made. This led us

20.

significant on

density

as given here.

Next,

to test

whether the

deviation

_from

u0

was of the magnitude to

cause us uo reject H0: u =

u0, a test of the

T-distri-bution

of che

density

about the 0.13 area point

(u^l.O)

was _{performed. For} a one tailed test it was found not

to be _{statistically significant} at the 0.05 alpha level

with 7 degrees of

freedom.

It would be _significant,

however,

at an alpha of

0.10.

t-Test on the Mean

Density

HQ: u =

uQ

(blue

line on Graph

#1=1.10)

Ha: u

/

u0

(

yjsl.oo, 3001ine screen, smallest

dot)

t =

**l-Uo

₌

--10

sj/

*& .169/2.828

t=-1.65

t _qcj 7=1.89 Conclusion:

u=uq

one tail as

density

could _only

decrease.

The largest dot of the 300 line screen halftone tint

is also not _{significantly} dense to cause us to reject the

.z

1

.

·

:

•· (J

+

,

'

,

; . ... i

+ '_

I

'i

.,J

.

'·

'

- - !ij

-1·

.j

.·i::

l

I I

,~,,

'1 •

·-ti I,

ti .j_

l::r

-+

'

-t

+

I.

--r_

. . .

.

et:

..

~, ... :.o;©. 111 rr'"~ ·50

·-~

-Qh! ~ i-, l>-t' ~ •1,

\

H c+ ~

"

'i\il

,

,

I~ - . i--: j

- -t

H

r

22

.

B.

D

O

T

S

IZE

VARIATION

F

our tints s

el

ecte

d

f

ro

m ea

c

h

o

f

t

h

r

ee sere.en rul

-ings

, (

65

,

15

0

and

3

00

lin

e. s

cree

ns) were used in

t

he

do

t

area

a

n

aly

sis

.

T

he re

l

a

t

i

ve

d

ot

a

r

eas

i·

e

r

e.

o

btaine

d

b

y

u

s

i

n

g a D

ie.tzgen pla

nime

t

er

no.

10309L~

a

cc

urate.

t

o

+ •

001

s

q

uare inches

on

photomicro

g

raphs

of

th

e

t i

n

t

s

.

The sam

ple

size

p

er

tint

was

9

d

ots.

Percent

d

ot

area

;r

as calcu

-lated as the ratio

o

f u

n

i

t

are.a

t

o dot

are

a

t i

me

s

100

.

T

he

d

ot

are.a

o

f

th

e

ne

ga

t

i

ve

u

se

d

t

o

m

ak

e.

the

p

r

e

ss

pla

t

e

was de

t

e

r

mi

ned

b

y

a

c

a

l

ib

r

a

ted

transmissi

on

per

c

e.

nt

dot

mete

r (

mo

d

i f i

e

d de

ns

ito

meter

).

Th

e

per

cent

d

ot

difference of

the t

ints was then

plotted

agai

ns

t

the.

per

cent

d

ot o

f

the.

t

in

ts

i

n

the

hal

fto

ne. ne

ga

tive.

It is

s.lign

±

f

.

ica:i::l.t; to

n

ote t

ha

t

with

i

ncrease i

n

do

t

s

p

read th

e.

re

is

genera

l l

y an increase in

d

ot

s

iz

e v

a

ri

a

bilit

y w

i t

hin

a t i

nt

.

This is

mainly

due

t

o

the decrease of

do

t

d

ef-inition

wi

t

h

incr

eas

in

g

d

ot s

p

re

ad

.

Graphs

o

f p

er

cent do

t

differe

nces

wer

e

conside.1~ed

to

be

mor

e

appropria

te

t

han

indiv

idua

l

signifi

canc

e

t

e

sts

due

to

t

he number

o

f

p

i

e

ce

s

o

f

dat

a

a

n

d

t

he qu2

l i t

ative

f

ac

t

ors (

di

ffere

n

t

paper

stocks

a

nd

inksD.

T

he

six g

r

a

phs on

page

s

23

throu

gh

2

8

s

h

o

w

d

o

t

spread

as

the

differen

ce

i

n

p

er cent d

o

t of

t

he pr

inte

d repro

-du

ct

i

on

f

r

om

t

hat

o

f

the

halft

o

ne ne

ga

tiv

e used

in

mak

in

g

th

e

p

l

a

te.

.

The

curv

es

a

p

pear to be

quite

v

a

ried in shape.

,

but hm

·

1eve

r,

mu

st

be

conside

r

ed

i

n

rel

a

tion

t

o

their

c

on

' '

.

-rt~+

J:.

+:!:~

:+r _;,

~++

'Tt1 .

·-t j: :1

I

1·1 -·'

,+; l

'rt I r ' --++ -,+

-H l + H

H '-l4-l--!-+-4t t

:ti

J r , I t- I ,....

: "' I

.,.

,

T H' +'t -++rH-+t;-;--i-Ttt-t-c

r.±-

_{. t- .}

-+

n.:

_{H- 1-1·.} m_

. . . ; ' t •

'

-I

r:

f+ H-H l

-

~

~

\

;-' r -i+

_t

I~

. .

-+- it'

:f

±t:

t:t 1

H+

-23

.

·++

-''

r _~

r ..!.-''

.

....

ID --Iii -~ ~

4 !~

I

~

~

--+ -t IZ

g:_

'"'

"

-Graph 4

Graph

5

25.

28.

29.

spread: 1. screen _ruling,

2.

paper

stocky

and,

3.

ink.

If you. consider the dot shape change between 0 to

100%

dot area in relating its area to its

diameter,

a

correlation S shaped curve

(y

=

l/(l+e~x)

but with a

greater slope) may be applied. Its maximum slope occurs

around a

50%

dot due to its square nature. Refer to

curve on page 30 . if we assume that the change in dot

diameter is _relatively equal throughout the scale (i.e.

the spread of the dot in terms of

diameter)

then in apply

ing

this to our S shaped curve _relating diameter to dot

area we would expect greater dot spread in the midc,le of

the range. It was observed that this empirical

theory

holds

fairly

well for the 300 line screen which was graphed

with a definite rise in dot spread in the middle area.

The effect was not quite so pronounced with the 65 and

150 line screens_, however a maximum slope was observed

in most instances between the range of a 30 and 60p dot.

The greater increase for the fine screen _may be attributed

to the large number of dots which it has in an absolute

area as compared to the coarser screens.

The _sharp

drop

in dot spread found in the 300 line

screen is due to the _running together of the dots and

their _resulting change of shape. Again a less pronounced

^affect

is

found with the coarser screens since their dots

are

larger,

don't run together as much and therefore have

Graph

10

100

80

GO

h

Q

x 40

20

/

_

20 HO GO

m

6t

Diam

ETER

wgm

=3xS

80

g^

I0O

tig

ii

7

44

31

The graphs show _less

spreading tendency

with the

coated stock and also reveal maximum dot spread in the

mid-range. Since the newsprint used in this evaluation

had a medium-hard surface

intended

for offset work the dot

spread does not

increase

as fast in the 'midtones'

as

observed with the uncoated paper stock. If data had been

collected

in

the upper quarter of the dot area range the

results _{may have} shown a more pronounced effect on dot

spread due to different paper stocks.

The two inks of different

tackiness,

one a typical

web offset ink and the other a sheet-fed press ink ("Quick

set"

ink),

were quite well differentiated as displayed

in the 300 line screen _tints. The sheetfed press ink

rendered a discernable dot structure throughout the range

covered

by

the _graphs;

however,

with the web offset ink

the dot pattern could not be detected above a

45%

dot.

(p. 32).

is

This4

attributed to the greater tackiness of the ink and

the _closer, fine dot pattern. In the 65 and 150 line

screen tints there is a noticeable slope increase as com

pared with the tints printed with the ink of less tacki

U

o

-:

-4-!

cd

Pm

-p

o

p.

r-f

o

H

ft >>

Eh

O o to

& O Q

U -P U

O CO ft

CO

p.

(D C <sj o

'H_{tw O}

U

; -P i

w a cy.

& o

^ m a) r; r^

H H Z W -.]

4->

~

z

f-3 1-rj

('i) ('i) I-'• _!-'• O" ~ ::l t-'

rn

rt t-'

O"O I-'•

Hi Ii :i:t ::l

~

s· ...

~

(I> rt

~

rt

(/} I-'• rt

::l 0

339

SUMMARY OF AN INVESTIGATION INTO THE NATURE OF N

At the due date of this _project, another area of

investigation

was _proceeding with much feresight

and

expectantions. It was an

investigation

into the factors

affecting the empirical; value of "n" which is an attempt

to account for paper

translueeney

and screen _ruling

combinations in calculations of tone reproduction studies

as stated in the _{Introduction.}

The method of

investigation

was as follows:

With the press _running at "optimum" _conditions, the ink

supely was shut off and thus the ink transferred to the

press sheet in the image areas became less and less.

This decreased the value of

D<j,

Ds,

while dot area was held nearly constant. Yule has said that with different amounts

of ink

being

laid

down,

the dot spread will vary. This is

quite

logical.

We did not planimeterize the dots of the

run _{do-m wedge,} 150 line screen _{specifically,} to determine

whether or not area did remain constant. This may need to

be done before we could proceed to investigate factors

control

ing

"ni'

In addition, Yule7 has published some tone reproduction

curves with quantities attributed to the

density

differences

obtained with light compared to

heavy

inking

for equal dot

34>

As stated under

Method,

_D.

, LARD measurements were made

of _{this wedge and in}

addition,

they

were scanned with the

M.D.

The

intent

w.-s _to

calculate^ by

an equation which

was later proved to be

invalid.

_{A second attempt to}

calculate n

by

approximation with _the 1620 IBM Computer

provided the

termination

_{point of this}

effort. 3f the

aprropriate n values were obtained _along with measurements

of

DT>DS

and <5ot _{area, the data coud be fit}

with a mathemat

ical

model obtained _{through multiple regression}

analysis _using

the STUFF

(oixteen

Twenty

_{Universal Function}

Fitter)

Library

Program.

Progress was _{steadily made until data was applied to the}

"Approximation Program for n"

which would never arrive

at a value for n. _{After many} attempts at

determining

the

reason for this _, the effort was abandoned.

Mr. Archer assumed the problem. He reqrogrammed a

different form of the

Murray

D,pvies

Equation;

discovered the

value would take almost forever to compute

by

approximation

and notified us that he was suspicious of our area determin

ations as n values were above

2.0,

the empirical _limit.

In a careful examination of the data

(following)

it was

discovere that all

density

values made with the LARD were

calibrated to the NBS placque which read the base paper

stock as 0.07. Thus all values were higher than could have

35.

Thus there is reason for subsequent work

beginning

with an investigation _using the data presented or new data

of the same general parameters to determine whether or not

the functional _relationship

"n=f(a)

or

f(Ds)

exist.

(p/I

7/1

%

W

13

A

15ft

\\eQftSS\DVL

Uala

-Vor

l^ressi^

o"r

Y\

From

RuMOdfcUDi/i

\&0

\jng

saffiA^cgl-

t'See

pao^-)

'-Ds

=

D(5*

=p\A

W

Frac-ficndl

ntccl

from ra

#

MS

D(r)

DSe-

o.13

D=0.5H

Ds

-

(P.Sfl

0,

-

0.<_

a

=

i.-.h

05p-

LI

5

Dt

0,

tv

Or

Dr

Dr

Dr

ifcrj

vx

S-^ep

Mum

be

,C*>3 .m

c(x)

.16 .11 .11

11

(3

12-t .11 .a

Affo

.on

A.C4

n

c(x)

.ao .in #

HI

73

,7

k;

cd),

-31 .MS

M

Jo

41

oil

|\\

CoMpli\-C<$,

Yrocyyam

-jd\j

36.

CONCLUSIONS

Based upon statistical evaluation of data obtained from

lithographic reproductions of

300,150,

and 65 line screen

halftone _step wedges on coated white stock with a "Quick

Set"

ink,

there is no detectable (significance level of .05)

functional

relationship

between dot area and dot

density

within this range of dot sizes.

Dot spread in the 65 and 150 line screenstints was observed

to be quite similar. There is a gradual increase in per cent

dot difference through the mit tones of the test objects

and a retarded per-cent dot difference at either end.

The 300 line screen in general exhibited a much greater

per cent dot difference in the midtones than the coarser

screens, as observed from the _graphs,

however,

the dot size

variability was also greater in the 300 line screen.

It is hoped that more thorough and extensive future studies will be made of dot spread with different ink

characteristics, paper stocks and screen _rulings, in order

to facilitate more dependable tone reproduction studies with

ACKNOWLEDGEMENTS

We wish to thank the

following

people for their as sistance and contributions towards _making our research pro

ject

possible:

Mr. John Petrycki of the R.I.T. Graphic Arts Research

Department-Printing

of the

halftone

tints.*

Mr. Milt Nudo of GARD-

Making

the press plate.

Mr. Arpag- _{Advice given in the}

stripping of halftone negatives. (PvIT-Print

ing

Dept

.

)

Mr. Irv

Pobboravsky

of GARD-

Assisting

in photomicro

graphy setup and _supplying a computer program for calculations.

B IBL JCGRAPPY

1.

Dorst,

Paul W. "Halftone Densities,"

Journal of the

Optical

Society

of

America,

33:8,

pages

43T-41i^,

August,

1943. '

2.

Miyauchi,

T., and Tajima M. "Techniques for

Evaluating

Dot

Quality

in _Halftone Processes','

Photographic

Science and

Engineering.

8:4,

_pp.

216-220,

July-August

1964.

" ~

3#

Murray,

Alexander.

"Monochrome Reproduction In Photo

engraving,"

Journal of the Franklin

Institute,

pp.

721-744,

1936.

4.

Neilsen,

W. J. and

Yule,

_{J. A. C.} "The Penetration of

Light Into Paper and Its Effect on Halftone Repro

duction,"

TAGA

Proceedings,

3:65-76,

1951.

5.

Pattillo,

Harold 0. "Tone

Density

and Printed Dot Area

Part 1',' _GraphcArts

Progress,

pp.

1,3,

Jan.-Feb. 1959.

6.

Pattillo,

Harold 0. "Tone

Density

and Printed Dot Area

Part 2," Graphic Arts

Progress,

pp.

1,3,

March-April

1959.

"

7.

Yule,

J. A. C. "Variation of Tone Reproduction in Half

tone Processes," _{Communication #2024 from}

the Kodak Research

Laboratories,

pp.

48-66.

8.

Graphic Arts Progress. "Measurement of Dot Area," Graphic