ADVANCED REAR PROJECTION VIEWER(Sanitized)PROJECT #02157)

• Document Type: CREST
• Collection: NGA Records (Formerly NIMA)
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78B04770A001900020031-7
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 82
• Document Creation Date: December 28, 2016
• Document Release Date: April 23, 2012
• Sequence Number: 31
• Publication Date: July 13, 1967
• Content Type: MF

Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
,NP.IC INTERNAL USE ONLY Nor
MEMORANDUM FOR:
NPIC/TDS/D-935-67
13
Chief, Exploitation Systems Branch,
July 1967
DS
STAT ATTENTION
SUBJECT
:
Advanced Rear Projection Viewer
Project
STAT
#02157)
REFERENCE
:
RADC Report of NOD-100 Viewer Test, dated 26 June 1967
STAT
STAT
1. Reference report was delivered to who brought it to STAT
me for info. I have reviewed it and am generally impressed.
V 7
2. I suggest the report will be of considerable value to you for
guiding monitors' attention in directing pursuit of the viewer. STAT
I think it will also provide some help in refining our development object-
ives and evaluation procedures for this type of equipment. In that regard,
I suggest you advise of the availability and applicability
of this report.
3. As a final wrap-up I'd appreciate your arranging for a feed-back
comment to RADC in cooperation with when you've had a chance
to evaluate the report--say by the middle of Au ust.
Attachment: (1)
Reference RADC report
Deputy Chie Development Staff, TDS
Distribution:
Original - Addressee
1 - NPIC/TDS/EPS
2 - NPIC/TDS/DS (1 - Chief, SSB)
NPIC/TDS/D
(13 Jul 67)
2v61
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
STAT
STAT
ILLEGIBIAT
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
) P.-- 7 Z. /7
rvi
1-1 C / Alt.)/ Y
94.75 2453..35 1.00000
12158.72_ _16.12_ 1.69100
, 203.70 35.43 1.6490U
314.0 9.66 1,00000
551.36 16.12 1.69100
1324.37 22.57 1.00000
449.88 12.74 1.69100
948,31 7.62 1.00000
512.82 27.93 1.64900
283.36 12.74 1.69100
1194,42 ?274.00 1,00000 _
_1263,30 11.38 1.72325
462,76 1.30,06 1.56873
362,66 12.44 _1.00000
.200,65 1.9.57 1.56873
946,49
I-21701
'4173.23
? -883.71
, 82.95
* 92.%
11 02
510.03 / 4
t14 9.7870
198.12 $13.72
4.17
-
?0003u
.01261
.01920
.00000
0126
,
1
.00000
. 012 6 1
00000
.01920
.012.61
00000
? 02563
.00902
.00000
.00902
3.81 1.00000
6,99 1.65160 .01113
2.89 1.00000 .00000
8.80 1.65895 .02325
7,57 1.00000
135.51
? '163.54
.00000
1.64250 -7. 01103 .
1.00000 .00000
- 1.73520 -f.01760
1.60311 1.00995
1.00000
589.36 9.83 1.52249
111.35
Ai 456.83...?36 .55_L1. 00000
147-11.25't 7.24 1.69713
79.45 I.1 3.59 _1.000019
86.32 8.85 - 1.72151
298.86___L_ 11.75 _1,00000
127.80 12.72 ? 1.69713
_ 438.79 126,0&i.0000()
1562.55-r-1-4.45 1.61272
- 209.33 .53 1.00000
387.16 30.90 1.61272
- 159.51_ _ 41.95 1.72151 _
- 1449.22 496.00 1.00000
..00000
4,00878
7 .54 1.71736 4.02431
..00000
.01239
? 00000
.02467
00000
.01239
.01(45
.00000
.01045
_ .02467
.00000
F'IL, fli
4iteeinulmiwismemespoper444.-,?,....,--
toe 11111
c.; ? ri,Vt
?-t", a
as a ,os''4 cr or in i,n00,,tict,a la?..,:iHR.113,1k711
ot s"O aa,a,1Go'?,..-..10tit 0811 iTsl.it
tni3 dtit thA axtz?ni
1ha roo!,,srl, T1111 1tt,111t,,rn 4pat-fitA
a .00varnmtrit A HAM to nit (Arum/11(1,M 1,10r
4,1111,,J a 4101 doll it It Is Obtainsi trum
Declassified in in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
.ZeZA1 LEN.5? (24 X- 70X if,4N6L---?.'i; 1./?_,,r.r.; 67. /
AT 41X" E- 7 7.- 1
?
2/1D/US TN/CA/MISS
A4 Iv;
2328.02
.< 7 De ? Classified in Parifae-rciTiZed 6-c7pY-ApWrove for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
(L. CYCLES E 70 DIVISIONs A) I U ? .
IcatIFFEL. a RIMER CO.
70X
co
cm
4.
.
1 ,
4
.
_
--4-
1_1_1
--I-
?
...1
1
4

__,
,-?-
t
1-
---,
4
?
-4
,..9
AN
,
-1,-
0
.
I
1
1
F
1
1
I
,
,
t 0
I
t
I-
I
i
?
i
-1-1
---i ,

?
i
?
[
-I
1 l'
1
1
'II I:
t
I
I
?
t
I-
, 1
-
i
t 1',4'''
a
i
I
.
t
1
-
i
i ,
I t .
1
1 {.F
; t
, _
!-, k
-, -
_,_
,--
_
.
-
..
?
1.
r
t
,
t
i.
IH
t
1,-
1
111
4
1
-...
,-,....?,,
,
'
4
1._ _
1
-1 I
?
t
I
h
1
.?"
,
zr7
1
4-
I-- -
4- -.
I
i
i
-
-
i i .
--1-
4
1
i
l','
T .1
1-
4-11H
I I.--:
[ i-1;
'
I
-
?-,
_i_
;
L 1,
I
'--- .- i
1...
-I
.I__
-
-

,
?;
tl,
1
1
HI
VP
:. 13_
, -?
-
-
.
1
-ft
.
i
,
'
,
_,
t
1
t '
-1'
.
i .
, 1
1
, 1,
-
-

1
,
1-i
-,-..---
--
'
I
1
f
4!
-
,
1
.
41
1
'

1
7
--[
-1
1+

J
1
i
I 1
if_
I
i---,-
A T
i +
I
r i
i
_
1 1
1
ji 7
'
1
,
I7-
1
-
I
I"
-
I i
I I, !..t

1 .
-I
I
-
'
1
I-
_ ,,.
i '
4
1
H.,.
i,
-3,--1
-,--- --, -, - t
-
i
'
_
._,
1
' ''
i
IH
-
-t-
-
-4
-
t
-
_
--4-
1
-_
1-
-'
-
1-1-,_11
i-1- 1 I
1 : I.
,
_,
i
, 4 :
, 4041
- -1'
r
'
t
LI
! '
' 1
I
1
,
hi
r,
'

,-1
-f;
,:.
I
I
i
,
1 I
,
-
..
1
_
_
1-11
,
1_?
...1.,_.
,
,
1
..,
T t
1
1

,
l?
i___4
r _4
,
1.
Ail
.
1
1
I
'
il
!
,_.-jti444-4
1 i
_
.1
,
1

r
r
' ]
]
' ?
L-
' II
!I;
'-'
-
{-
- T
i
,-'
.....
'
'
I 1
- -1
1 r
--I
- I-
-,

,1
1J-,
-1
,. ?
I T1
1
,
I
1
__I
,. 1
-4-
,
,

_

,.
,
i :1
-r
A
,
,, h
1
;
,,
, i
r
1
-r-
i
-i 1--
i t
+
1- .I.
,
-f-
- -I 1-=
4.
1-
-I=
.1 1
14.
11-
J,?
II
1
1r 4
'
i
4 4
-
,
'
1

1 I
-
.
1
.
,
,
1 ,
N
1
.,-
I
L
,
7
_,
-
-
I.
'
1-.-
. 1
-
II
V
, 1
%
\
;
, r(
I 1
_T__t f
_
1
.
1
.
.__.1
i_
i
:i
A
,..1111
ii
11:1_
;
I
i
'
I
1
.k_
.
i
. 1-
I
.
1
r ,
-4-t-
,
f"
'
0
I
-...
1
t.
1
1,-,1
a 1
-
1 ,
11111111
'
I
,
?
t t
lintliall11=MMI
j
_
4.
si
1-
r
t-
? 1.
'
lit
II
ii
I
'
I
I
,, ,
EMI
1
,
,
,
, 111,,
".. ,
,

I
4
, , ;
7
-
it-,
I
I
1
in
i
.
r,
i
,
r
_
, _11,
I,
I
1.1 1
I
.
i
,
1]
I
,
II
1
1
1 11
.
.1.
'Ill
I
i--
j
[hi
1
"r,
,
-
D
-.1
I
I
?
_
-
1
"
-1,

____I___
-I
1
-
1 -
-
..I''
1
dr
k
IF ;
i
1 _
:
I ,
1
;
i
.
'
I
T ,
..-
-- T,'
1 il
rr
Ftii
(k
1
1
..,: .
tf 1 1
I
I
;
4:
,
t,F
,
J
L
i
1
,
..]
,
1
1 1
*
' - 1
-
V
1
41I.r
i'
_
,
i
_
..
Fl
- 1-'ft

.
fl
//7- //)- ii .2 -i- /77 .
1
2
F.- L.F., 7 - ? -
It
.G? 1,
}
0
0,./
/
a
* NUMBERS IDENTIFY BACKGROUNDS
5
Fig. 3. Pictorial representation of 9-mile wide film strip.
ccs
ml
an
rn
tio
thc
tra
ant
Litt
thi
jilt
eq
sin
tht
Pr,
itrz
sq
TA
Pc
T1
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
-ILIMAN FACTORS
3
3
1:00
o?
4
CHARLES AV. SIMON
In the moving presentation mode, the imagery
was moved steadily across the display in either
direction. This mode simulates the near-real
time situation in which a reconnaissance
vehicle flies above the terrain at a designated
rate while the sensor supplies a continuous
radar map to the observer via the display.
In the static presentation mode, the same
imagery was presented to the observer in suc-
cessive static steps, one frame at a time. The
image remained fixed on the screen for the
amount of time required for the same area to
move across the screen in the moving presenta-
tion mode. Between static presentations, while
the imagery was changing to a new frame, a
translucent screen moved between the projector
and the screen to blur the image without re-
ducing the screen brightness to zero. Blurring
the image while changing frames prevented
the observer from detecting targets during this
interval and kept the total observation time
equivalent in the two modes. The static mode
simulated a near-real time situation in which
the information from the sensor is stored until
a complete frame has accumulated and is then
presented statically to the observer while the
next frame is being stored.
Display Size. Both a 6-inch and a 12-inch
square display were used in the studies. The
TABLE 1
Perceptual characteristics of targets.
June, 1965-189
image was rear-projected onto a polacoat
screen masked to one of the two display sizes.
The screen was adjustable so that the entire
image filled either display; thus, the scale factor
on the 12-inch display was twice that for the
same imagery on the 6-inch display while
ground coverage remained constant. Average
screen brightness was adjusted so that it re-
mained constant for either display.
Observation Time. Observation time peri-
ods were limited to 10, 20, or 40 seconds. For
the moving presentation mode, these were the
times it took an object to appear on one edge
of the display, move across to the other edge
and disappear. ' For the static presentation
mode, these were the periods of time during
which a single frame was exposed.
Ground Coverage. The ground area cov-
ered by the radar imagery on the display was
either 18 or 9 miles square. For a fixed display
size, changing ground coverage from 18 to 9
miles square is equivalent to doubling the scale
factor while holding display size constant.
Target Characteristics. The eight target
types were divided into four groups with sim-
ilar perceptual characteristics. This grouping
served two purposes: (1) to combine enough
data for meaningful analysis, and (2) to allow
for broader generalizations than would be
PERCEPTUAL CHARACTERISTICS
TARGET GROUPS
2
3 4
Triangular
Airfield
Other
Airfields
(Average)
Stadium
Tank
Far ns
(Average)
0
Size
Length of Longest Dimension*
3700 ft.
4320 ft.
1080 ft.
1380 ft.
Width of Smallest Dimension*
180 ft.
130 ft.
300 ft.
300 ft.
Distinctive-
ness
Brightness Contrast
High
Low
High
Medium
Target and Background
Pattern Similarity
Low
High
High
Medium
*Dimensions expressed in feet on the ground.
8000 feet on the ground equal approximately one inch on a 12-inch square display for the 18-mile imagery.
Atra viewing distance of 12 inches, an inch subtends approximately a 5-degree visual angle from, the eye. At a
viewing distance of 18 inches, the visual angle is approximately 3 degrees.
These relationships are double for the 9-mile imagery and halved for the 6-inch square display.
I.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
190?June, 1965
possible if results were expressed in terms of
target class or the specific target patterns used
in this study.
This classification of targets by perceptual
characteristics is shown in Table 1.
Relationships Among Variables. Combina-
tions of several variables in this experiment
can be identified by other familiar classifica-
tions (Table 2). For example, combinations
of observation time and ground coverage can
. be identified with real world vehicular rates
over the terrain (Table 2A). Observation time
can also be expressed in terms of apparent
movement rates, that is, in terms of inches of
movement per second when display size is
taken into consideration (Table 2B), or in
terms of degrees of movement per second
when viewing distance is taken into considera-
tion (Table 2D). Similarly, various combina-
tions of display size and ground coverage re-
sult in different image scale factor values
(Table 2C).
It should be noted that while the observation
times expressed in Mach values appear high,
when expressed in terms Of degrees of move-
ment per second they are considerably below
the level at which a degradation in dynamic
visual acuity has been observed (Miller and
Ludwigh, 1960).
.?
HUMAN FACTORS
PROCEDURE
Observers practiced and were tested in a
semi-darkened cubical room which served to
attenuate external noise and visual distractions.
No restrictions were placed upon their seating
positions or viewing distance from the display.
The practice periods and experimental sessions
are described below. I;
Practice Period. The practice period served: ti
(1) to select only observers proficient in recog-
nizing radar targets, (2) to familiarize the ob-
servers with radar imagery and with the
specific targets used in this study, and (3) to
allow them to develop optimum search tech-
niques for both the moving and static presen-
tation modes.
A series of one-half to one-hour practice
periods was given to each observer. An ob-
server's total practice time was from two to
four hours, depending upon his previous ex-
perience and ability to reach arbitrary per-
formance criteria.
Those observers unfamiliar with radar were
informed briefly how radar operates and how
its pulses reflect from various ground objects
to produce the light and dark returns on the
display. All were given training in finding
targets in radar imagery similar to that to be
1,01
C,!
0'
St
TABLE 2
Some relationships among observation time, imagery ground coverage, and display size.
Observation
(A) (B)
Real World Rates for Different Imagery Ground Movement on Screen for Different Display Sizes and
Coverages and Observation Times Obervation Times.
Time, seconds Ground Coverage
Display Size
18 x 18
miles
9 x 9
miles
6 x 6 12 x 12
inches inches
10
Mach 10.8
Mach 5.4
0.6 inch /sec
1.2 inch /sec
20
Mach 5.4
Mach 2.7
0.3 inch /sec
0.6 inch /sec
40
Mach 2 . 7
Mach 1.35
1.15 inch/sec
0.3 inch /sec
(C) (D)
Image Scale Factors Resulting from Combinations of Apparent (Retinal) Movement with a 10-second Ob-
Display Size and Ground Coverage servation Time for Different Display Sizes and at
Different Viewing Distances
Display Size,
Inches Ground Coverage
Distance from Display
18 x 18 9 x 9
miles miles
6
inches
12
Inches
18
inches
6 x 6
12 x 12
1/216,000
1/108,000
1/108,000
1/54,000
5.3?/sec 2.8?/sec
9.0?/sec 5.3?/sec
*20 seconds (by 2);
40 seconds (by X)
1.9? /sec
3.7?/sec
it
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
iMAN FACTORS
DURE
:Ind were tested in a
-oom which served to
ind visual distractions.
-eed upon their seating
_ance from the display.
experimental sessions
]Iractice period served:
2rs proficient in recog-
to familiarize the ob-
;very and with the
ids study, and (3) to
pptimum search tech- ,
ing and static presen-
to one-hour practice
zh observer. An ()b-
ale was from two to
llon his previous ex-
reach arbitrary per-
iliar with radar were
ar operates and how
_rious ground objects
dark returns on the
training in finding
similar to that to be
e.
.11)
ifferent Display Sizes and
Nay Size
12 x 12
inches
1. 2 inch /sec
0 . 6 inch /sec
0.3 inch /sec
7)
,it with a 10-seeand Ob-
au' Display Sizes and at
_s
-am Display
2 '
_hes
18
Incites
/see 1.9? /sec
/sec 3.7?/sec
CHARLES W. SIMON
used in the experiment. Practice under both
static and moving presentation modes was
given; near the end of the practice period, the
observation times were limited to those the ob-
server would experience during the experi-
ment.
Through these experiences, observers de-
veloped their own methods of search and estab-
lished viewing distances most suitable for
themselves.
Instructions. The following points were
made during the initial instructions to the sub-
jects and were re-emphasized throughout the
experiment:
1. Simulated Mission. The following hypo-
thetical situation was described to the ob-
servers:
An intelligence report has been received
stating that enemy aircraft are about to take
off from certain specified airfields to bomb
friendly troops and installations, Also, missile
launchers, hidden near certain specified fuel
tank farms, are being prepared for immediate
launch. An enemy headquarters coordinating
these efforts is stationed in a football stadium.
It is vital that all these critical targets be
found and destroyed as quickly as possible.
Every second of delay could be critical, for our
reconnaissance plane risks destruction if it is
detected before it can detect and destroy the
enemy. Furthermore, if the plane is detected,
this would serve as a warning for enemy planes
to take off and for missiles to be fired. Under
these circumstances, it is better to bomb a
nontarget than lose a real one.
2. A Priori Information. Observers were
also shown small sections of radar imagery in
which the targets to be recognized were em-
bedded. These targets were the same size and
shape as the ones to be recognized, but were
not necessarily shown in the same orientation.
During a "hot" war, military-type targets
would be less likely to be located in logical
places and must be searched for everywhere.
Thus, observers were told that targets would
be located in practically any section of the
display and would not appear in a systematic
order or with a fixed frequency.
3. Observer's Task. The basic task before
each observer was one of target recognition,
June, 1965-191
that is, of finding targets on the display essen-
tially identical to those shown during the
briefing period. As soon as a target was recog-
nized, the observer immediately pressed a
button and put his finger on the target and
named it. The observer was instructed to "get
as many targets as quickly as possible, as soon
as they appeared on the display."
Scoring. When the observer pressed his
button, a mark was made on a graphic recorder
on which a time base was provided. Locating
and naming the target was done quickly and
did not appear to disrupt the observer's con-
tinuing search. It did enable the experimenter
to check the correctness of the response (both
target and location) from a scoring chart that
he held and marked. If the information was
correct, the experimenter also pressed a button
which put a verification mark on the graphic
record.
In addition, a mark was automatically reg-
istered on the graphic record for each target
after it appeared on the display. Time differ-
ences between the automatic target mark and
the observer's mark indicated the time required
to recognize that target once it appeared on
the display. Time was determined to the
nearest second. The presence of the observer's
mark and the absence of the experimenter's
mark enabled the number of false targets to be
determined.
Experimental Sessions. Before each ex-
perimental session, observers were given short
practice runs with the imagery, observation
time, and display size similar to those used in
the experiment.
EXPERIMENTAL DESIGN
The experimental design used in this experi-
ment is shown in Table 3. Performance data
was collected for all 24 combinations of pres-
entation mode, ground coverage, observation
time, and display size. Each observer was
tested under the four combinations of mode
and ground coverage. Different observers
were tested under each of the different observa-
tion times and display sizes. Thus, six ob-
servers,? each tested under four sets of condi-
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
192?June, 1965
TABLE 3
Experimental design
HUMAN FACTORS CIIARLE
Observers
Observation Time.
(seconds)
Display Size
(inches)
Presentation Mode
Moving
Static
Ground Coverage (miles)
9 x 9
18 x 18
9 x 9
18 x 18
Trials
1
10
6 x 6
3
2
4
1
7
10
6 x 6
4
2
3
2
10
12 x 12
3
2
4
8
10
12 x 12
1
4
2
3
' 3
20
6 x 6 ?
4
1
3
2
9
20
6 x 6
2
3
1
4
4
20
12 x 12
4
1
3
2
10
20
12 x 12
2
3
? 1
4
5
40
6 x 6
4
1
3
2
11
40
6 x 6
2
3
1
4
6
40
12 x 12
3
2
4
1
12
40
12 x 12
1
4
2
3
tions, were required for a single replication.
Two replications were run.
The order in which observers *ere tested on
the four presentation mode and ground cover-
age combinations was partially counterbal-
anced. The order for observers tested under
the same conditions of observation times and
display size was reversed. Over the entire ex-
periment, each of the four conditions run by
MOVING PRESENTATION MODE
to 16 ii to It 24 24
all subjects appeared the same number of
times during each of the four testing sessions.
Observers were tested morning and after-
noon on two consecutive days. A different
combination of mode and ground coverage
was tested at each of the four sessions. A
session consisted of two runs through the
imagery?first forward and then backward,
with a rest period in between. The actual
STATIC PRESENTATION MODE
14 II ii a 22 24 24 II SO 33 34
00
36
time mitt
mile lonj
?licular cd
ground cc
Results':.
presented
TABLE 5
Analysis
TOTAL
AETW
(MR',
Displ
Time
Error
WITHI
l'rese
.Grout
Tarp
Mod,
Mod.
Grou.
Mod.
Mod
Groc
Groc
Targ
Targ
Mod
Mod
Groi
Targ
Mod
Mod
Groi
(1 rot
Moc
Moc
Mcic
Moc
Moc
Gro;
Moc
Moc
Errc
This a
(G. W.
Two sc
related
the cor
Only F
using tl
of the
TIME (SECONDS)
TIME (SECONDS)
2
DISPLAY SIZE
OBSERVATION TIME
GROUND COVERAGE
12-0-
10 - 20 - 40 -
9---18 ?
3
INCHES SQUARE
SECONDS
MILES SQUARE
Fig. 4. Cumulative percent of radar targets recognized on moving and static display.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
HUMAN FACTORS CHARLES W. SIMON
entalioa Mode
time required to run once through the 324-
mile long imagery depended upon the par-
ticular combination of observation time and
ground coverage being studied (Table 4).
Static
TABLE 4
Actual testing time for one run.
June, 1 9 65-1 93
Coverage (miles)
9 x 9
18 x 18
Trials
4
1
2
3
4
2
3
3
2
4
3
2
1
4
3
2
1
4
4
1
2
3
the same number of
:he four testing sessions.
:Led morning and after-
itive days. A different
and ground coverage
t the four sessions. A
two runs through the
_I and then backward,
a between. The actual
RESENTATION MODE
1440-1.
40.4
.401
...411.4044
?
?BUI4UUMUMMAN
(SECONDS)
:OVERAGE
I 8 -
souAnE
ind static display.
RESULTS
Results of the experiment are graphically
9 miles 6 minutes 12 minutes 24 minutes
presented in Figure 4. The two plots show the
Ground
Coverage
(width)
Observation Time (seconds)
10 20 40
18 miles 3 minutes 6 minutes 12 minutes
TABLE 5
Analysis of variance: percentage of targets recognized correctly.
Source of Variation
Degrees of Freedom
Mean Square
F.
TOTAL
191
BETWEEN SUBJECTS
11
624.39
Observation Time
2
2445.19
10.235
Display Size
1
1725.97
7.225
Time x Display Size
2
7.73
Error uncorr.2
6
.39.42
WITHIN SUBJECTS
180
Presentation Mode
1
3.70
Ground Coverage
1
38844.50
49.785
Target Characteristics
3
13137.88
31.356
Mode x Ground Coverage
1
10.39
Mode x Target
3
39.64
Ground Coverage x Target
3
780.20
10.284
Mode x Time
2
15.96
Mode x Display Size
1
231.04
Ground Coverage x Time
1
16.83
Ground Coverage x Display Size
1
60.76
Target x Time
6
238.66
Target x Display Size
3
344.84
Mode x Ground Coverage x Target
3
20.59
Mode x Time x Display Size
2
32.58
Ground Coverage x Time x Display Size
2
76.44
Target x Time x Display Size
6
238.96
3.154
Mode x Time x Target
6
43.73
Mode x Ground Coverage x Display Size
1
101.13
Ground Coverage x Target x Time
6
133.61
Ground Coverage x Target x Display Size
3
78.72
Mode x Ground Coverage x Time
2
148.89
Mode x Target x Display Size
3
38.10
Mode x Ground Coverage x Target x Time
6
44.27
Mode x Ground Coverage x Display Size
3
141.84
Mode x Ground Coverage x Time x Display Size
6.96
.Ground Coverage x Target x Time x Display Size
,2
6
47.49
Mode x Target x Time x Display Size
6
26.41
Mode x Target x Time x Display Size x Ground
Coverage
6
105.82
Error corr.2
90
75.89
This analysis was made on the arcsin transformation of the square root of the percentage data
(G. W. Snedecor, Statistical Methods (4th Edition), 1946, Iowa State_College Press, pp. 447-448).
2 TW-6-s-oijrces of variability attributable to chance can be calculated, i.e., that which is based on the uncor-
related data obtained from conditions on which different observers were tested and that which is based on
the correlated data obtained from conditions on which the same observers were tested.
3 Only F ratios significant at the 0.05 probability level or better are shown. Tests of significance were made by
using the mean square of the appropriate error term4 or by using the mean squares (or pooled mean square6)
of the significant interactions containing the source of variation being tested.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
194-June, 1965 HUMAN FACTORS
/ CliAi
i
TABLE 6.}7 prob.t
Analysis of variance: median tinie to recognize targets.
each i
Source of Variation
Degrees of Freedom
Afean Square
Fa
TOTAL
47
18.43
BETWEEN SUBJECTS
11
Time ?2
60.58
Display Size
1
7.14
Timex Display Size
2
4.22
Error uncorr.1
6
3.73
WITI IIN SUBJECTS
36
Mode
1
231.89
10.614
Ground Coverage
1
125.14
Mode x Ground Coverage
1
59.62
Time x Ground Coverage
2
27.25
Display Size x Ground Coverage
1
.41
Mode x Time
2
43.28
Mode x Display Size
1
.03
Time x Display Size x Ground Coverage
2
3.86
Mack x Timex Ground Coverage
2
,27.26
12.173
Timex Display Size x Mode
2
16.52
7.383
Mode x Display Size x Ground Coverage
1
2.77
Mode x Time x Display Size x Ground Coverage
2
2.44
Error corr.'
18
2.24
time,
the
al
T1-.
show
'aria
Targ,
Med.
gets
by a
the;
the
aloft
Valu
etTcc
. plott
TABL
1 Two sources of variability attributable to chance can be calculated, i.e. that which is based on the uncorrelateeFalsL
data obtained from conditions on which different observers were tested and that which is based on the cor-
related data obtained from conditions on which the same observers were tested.
2 Only F ratios significant at the .05 probability level or better are shown. Tests of significance were made b)
using the mean square of the appropriate error term3 or by using the mean square (or pooled mean square;
of the significant interactions containing the sources of variation being tested.
TABLE 7 -
Analysis of variance: range of es required by an observer to recognize 90 percent of recognizable targets.
Source of Variation
Degkes of Freedom
Mean Square
TOTAL
BETWEEN SUBJECTS
47
II
Time
2
846.33
40.594
Display Size
1
.34
Time x Display Size
2
.58
Error uncorr.1
6
11.58
WITHIN SUBJECTS
36
Mode
1
1365.34
65.494
Ground Coverage
1
60.75
Mode x Ground Coverage
1
33.33
Timex Ground Coverage
2
16.75
Display Size x Ground Coverage
1
3.00
Mode x Time
2
286.33
Mode x Display Size
1
10.07
Time x Display Size x Ground Cover
2
3.25
Mode x Time x Ground Coverage
2
20.85
4.62
Time x Display Size x Mode
2
8.09
Mode x Display Size x Ground Coverage
1
3.25
Mode x Timex Display Size x Ground Coverage
2
2.24
Error corr.'
18
4.51
1 Two sources of variability attributable to chance can be calculated, i.e., that which is based on the uncor?
related data obtained from conditions on which different observers were tested and that which is based on the
correlated data obtained from conditions on which the same observers were tested.
2 Only F ratios significant at the 0.05 probability level or better are shown. Tests of significance were made by
using the mean square of the appropriate error term3 or by using the mean square (or pooled mean square)
of the significant interactions4 containing the sources of variation being tested.
lo
GRO
PR
DI:
CR
LOE
7-- ?
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
-.IAN FACTORS citAitLEs w. straw-sr
''quare
FZ
_43
58
14
22
73
89 10.614
14
62
25
-11
C3
26 V6
12.173
? 7.383
77
44
24
=,ed on the uncorrelated.
2.11 is based on thd cor-
ificance were made by
pooled mean square4)
1
recognizable targets.
uare F2
33 40.594
34
58
58
34 65.494
75
33
75
00
33
137
25
.85 4.623
1)9
25
24
ii
; based on the uncor-1
which is based on the
:ficance were made by;
pooled mean square)i
1
probabilities of target recognition over time for
each of the 12 combinations of observation
time, display size, and ground coverage for
the moving and static presentation modes.
The statistical treatment of the data is
shown in Tables 5 through. 9. Analyses of
variances were calculated on the percentage of
Targets ?Recognized Correctly (Table. 5),
Median Time for Recognizing Acquired Tar-
gets (Table 6), and the Time Ranges Required
by an Observer to Recognize 90 Percent of
the Targets Acquired (Table 7). Table 8 shows
the Frequency of Recognizing False Targets,
along with the results of tests of differences.
Values for individual means of significant main
effects and interactions in these analyses were
plotted in the graphs of Figures 5 through 7.
TABLE 8
False targets.
June, 1965-195
The results of the analyses and data shown
on the tables and figures are summarized in
Table 9. The effect of each variable on each
performance measure is given, and where the
interactions among variables were found to be
significant, the data has been summarized
across variables.
In general, it would appear that within the
range of conditions in this study, the same
number of targets can be found with either the
moving or static presentation mode, but that as
observation times grow longer and target visi-
bility becomes poorer it takes significantly
longer and speed of performance is more
variable with the static mode.
One result of interest is the comparative'
effects of the two combinations of display size
FREQUENCY OF ACQUIRING FALSE TARGETS FOR DIFFERENT
TARGET
GROUP
PATTERN
BACKGROUNDS
1
2
3
4
5
A
A
i
o
o
I
o
o
I
o
i?i
o o
0
1
0
1
0
:0:
0
0
0
0
6
(?)
QD
6
5
6
5
4
X
5
5
13
2
4
It-
A
3
5
3
5
0
=-4
3
1
9
7
3
TARGET-DISPLAY-SYSTEM CONDITIONS
PRESENTATION MODE : MOVING
STATIC
DISPLAY SIZE 6
12
6
12
GROUND COVERAGE . 9
18
9
18
9
le
9
le
,i
_
TARGET PATTERN
[11 k Id@INk.':dHd[):
OBSERVATION TIMES
r851no1`861`886-LIT,61f6"?.36-1`6"8-alno
I
1
0
'a
6
1
I
1
I
I
I
11
2
I
ig ?
cA _
1,
I
1
3
I
Z
3
I
1
1
6
I
3
2
13
2
12
1
3
2
I
4
1
2
1
2
1
2
I
1
2
1
2
1
I
1
I
1
I
I
1
I
1
2
3.3
2
1
5
3
1
1
1
X2 FOR FALSE AR F
SOURCE
f
x2
d. f.
SIGNIFICANCE
p.5.0.05
SOURCES
f
x2
SIGNIFICANCE
d.f 1350.05
PRESENTATION MODE
MOVING
STATIC
49
58
0.75
I
NO
TARGET GROUP.
A
g)
i?
2
II
26
68
62.7
3
YES
DISPLAY SIZE
6 INCHES
I2INCHE4e
58
49
0.75
I
NO
GROUND COVERAGE ?
SMILES
!SMILES
70
37
0.23
I
NO
BACKGROUNDS ?
I
la
la
1St
X
le
18
31
22
18
3. 73
4
NO
OBSERVATION TIME
1 0 SECONDS
20 SECONDS
40SECONDS
35
21
51
12.6
2
-
YES .
?",
Declassified in Part- Sanitized Copy Approved for Release 2012/04/23.: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
196?June, 1965
HUMAN FACTORS
C-GROUND COVERAGE
B-DISPLAY SIZE
A-OBSERVATION TIME
g g E.7 8 2
00000000 0o
o 0 co P- IDVI If to ca
O0 0 0 0 0 0 0 9
O 01 CD N W o N
03ZIN0033a S.1.306Y1 O IN3083ci
10
01
0
ID
0
0
NAUTICAL MILES,SQUARE
INCHES SQUARE
?
1 I
1 1
I 1
1
1
0
-
O 0 0 0 0 0 0 0 2 0
O co c4- co co PI
0-TARGET GROUPS
O 0 0 0 00 0 0 0 o
O o.) 10 1 ID VI 4 rn N ?
0
111111111
O 0 0 0 0 0 0 0 0 0 0
O 0, a) tO nt N
02Z NO033N S.1.30e1V4, JO .IN 33a
Fig. 5. Percent* of targets recognized correctly (*transformed back from arcsin transformation used
in analysis).
CilAi
TAM
Sum
ME
NI
Ti
COV
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
:AN FACTORS
_
7,-
1 1
0 0 0
N
0
0
1 I I
00
N
20
D LN3d
xansformation used
CHARLES W. SIMON
TABLE 9
Summary of results showing effects of variables on performance.
June, 1965-197
PERFORMANCE
MEASURES
VARIABLES'
STATIC VERSUS
MOVING
PRESENTATION
MODES
GROUND COVERAGE
(9 x9 OR 18 x 18
OBSERVATION MILES AND
(10, 20, 40 SECONDS) DISPLAY SIZE
(6' x 6" OR 12' x 12")
TARGET
CHARACTERISTICS
PERCENTAGE
OF TARGETS
CORRECTLY
RECOGNIZED
(TABLE 5)
No Effect
(Table 5)
More targets re-
cognized as time
increased.
(Figure 5A)
More targets re-
cognized on larger
display or with
smaller ground
area.
(Figure 6B and C)
Larger and more dis-
tinctive targets were
more often recognized
(Figure 5D)
Reducing ground coverage increased re-
cognition of targets. Less distinctive
targets showed greatest increase.
(Figure 5E)
Targets with smallest widths were poorer with short observation
times and small displays. Targets with longest dimensions im-
proved with longer observation times and larger displays.
(Figure 6F)
MEDIUM TIME
TO RECOGNIZE
(TABLE 6)
Static took longer
(Figure 6A)
Recognition took longer with smaller displays and greater
ground coverage. Differences were significant only when
imagery was presented statically and for the longer observa-
tion times.
(Figure 6 B and C)
(No Analysis)
MEAN VARIABILITY
IN OBSERVER'S
RECOGNITION
TIMES (TABLE 7)
Static more vari-
able
(Figure 7A)
Variability in-
creased with
longer observation
(Figure 7BE)
With a static presentation and with a longer observation
time, variability was greatest when ground coverage was
large.
(Figure 7C)
No effect due to display size.
(Table 6)
(No Analysis)
No Effect
NUMBER OF FALSE (Table 8)
TARGETS CALLED
(TABLE 8)
Significantly fewer
false targets with
20-second obser-
vation time.
(Table 8)
Twice as many
false targets were
called when film
coverage was twice
as much.
(Table 8)
Seven times as many
false targets were cal-
led when target pat-
terns were judged
more similar to back-
ground patterns.
(Table 8)
Where information extends across more than one variable, it refers to the interaction among the variables
covered.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
198?June, 1965
and ground coverage yielding the same scale
factor. Both the 9-mile square ground cover-
age on the 6-inch display and the 18-mile
square coverage on a 12-inch square display
represented a scale factor of 1/108,000. On
both displays, therefore, targets were of equal
size. However, performance was not identical.
The results in Table 10 illustrate the difference.
Variations in ground coverage had a much
greater effect on performance than variations
in display size.
None of the previously mentioned analyses
included an examination of changes in per-
formance over the four experimental sessions.
The number of targets recognized for each of
the four sessions for all subjects under all
conditions were:
I. 455 II. 453 III. 459 IV. 494
The average times to find and recognize tar-
gets for each of the four sessions were:
HUMAN
FACTORS
I. 6.2 secs. II. 6.6 secs. III. 5.9 secs.
IV. 6.4 secs.
None of the values for either measure differs
significantly from what could be expected by
chance at the 0.05 probability level.
DISCUSSION
The present study clearly indicated that for
the conditions under investigation, target ac-
quisition performance was never better and
often poorer when the imagery was presented
in a series of static steps than when the
imagery was presented in a continuous move-
ment across the display. There was no differ-
ence between the two presentation modes in
the probability of finding targets. However,
under many conditions, those targets that were
detected were found sooner on the moving
MEDIAN TINE TO RECOGNIZ
A- PRESENTATION MODE
20
B-
20
la
INTERACTION- GROUND COVERAGE,
PRESENTATION MODE, AND
OBSERVATION TIME
20
18
ie
, X STATIC
0 MOVING eX
16
?
16
14 -
14
9 N. pm,
14
12 -
12
10 -
-
10
xi
? ?
-
4 -
2 -
MOVING STATIC
10 20 30
SECONDS
C- INTERACTION- DISPLAY SIZE,
PRESENTATION MODE, AND
OBSERVATION TIME
STATIC
0 MOVING
6 IN,
/ X
z
50 0 10 20 30
SECONDS
40
50
Fig. 6. Median time to recognize targets.
RANGE OF TRIES TO RECOGNIZE, SECS
A -PRESENTATION MODE
40
40
36
-
36
32
-
32
28
-
26
24
-
24
20
-
20
ID
-
16
12
-
12
e
-
4
-
4
MOVING
STATIC
B- OBSERVATION TIME
10
20 30
C -INTERACTION-GROUND COVERAGE,
PRESENTATION MODE, AND
OBSERVATION TIME
40
36
32
28
24
20
16
12
4
X STATIC
0 MOVING
IS N MI
9 N. MI.
40 50 0 10 20 30 40
SECONDS SECONDS
Fig. 7. Range of times to recognize targets.
TAM
Peri
nnd
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Pere
Rec(
Med
Reck
Ran
Rec
Nut
Tar;
*Si
"C
clisi
.per
?stu
col
(5,
prc
me
fro
Wa
ma
the
prc
oh!
TAI
SOI
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
1-IUMAN FACTORS CHARLES W. SIMON
6 secs. III. 5.9 secs.
-or either measure differs
at could be expected by
obability level.
-USSION
21early indicated that for
investigation, target ac-
was never better and
2. imagery was presented
steps than when the
a in a continuous move-
y. There was no differ-
. presentation modes in
ding targets. However,
those targets that were
sooner on the moving
1:114CTION-DISRLAY SIZE
EGENTATION MODE, AND
SERVATION TIME
STATIC
MOVING
iZ IN.
6 IN,
20 30 40 50
SECONDS
.TION-GROUND COVERAGE,
-0-ATION MODE, AND
-/ATiON TIME
TATIC
mACIWING
8 N. MI,
9 N. MI.
1
20 30 40
SECONOS
30
TABLE 10
Performance with combinations of ground coverage
and display size yielding equal scale factors.
Ground Coverage
Scale Factor 11108.000
9 mi. sq. 18 mi. sq.
Display Size
6? sq. 12" sq.
Percentage of Target
Recognized ( %)
Median Time to
Recognize (secs)
Range of Time to
Recognize (secs)
Number of False
Targets Acquired**
Moving 89.8
Static 80.8
Av. 85.3
52.2*
53.8*
53.0*
Moving 22.0 23.0
Static 37.5 66.0*
Av. 29.8 44.5*
Moving 7.3
Static 15.2
Av. 11.2
Moving 7.0
Static 12.0
6.8
19.8
13.3
8.0
9.0
T19.0 17.0
*Significant at .05 probability level.
**Corrected for twice as much 9-mile imagery.
display and with significantly less variability in
performance.
Only one other published experiment, a
study by Erickson (1964), was found that
compared target acquisition with a moving
(5, 7, and 10 degree/second) and with a static
presentation mode. Although his method for
measuring time with the static display differed
from that of the present study and his imagery
was not as representative of the real world,
many of the relationships he discovered among
the variables studied support those found in the
present experiment. Erickson required 16 male
observers to search for an incomplete ring
(Landolt C) among a number of solid rings
TABLE 11
Some results of Erickson's study.
June, 1965-199
arranged systematically throughout a square
field. Along with a number of other experi-
mental conditions, he compared search per-
formance in a moving and static field. Using
percent of targets detected as the criterion, he
found that performance deteriorated with a
decrease in search time whether the image was
moving or static. Erickson concluded that
"target movement, per se, does not necessarily
(sic) have a detrimental effect upon search
performance within the 0 to 10 degree/second
range" and his results suggest that, in fact, the
percent of targets detected actually was greater
with the moving imagery than with the static
imagery when the observation time was short-
est. The results comparing the two modes
were shown in Figure 5 of his paper and can
be summarized as shown in Table 11.
Although no statistics were applied, the ex-
tent of the difference between the percent ac-
quired with a moving and static mode up to
the 70 per second condition was almost negli-
gible and possibly due to chance. At 10?/
second, or the shortest observation time, the
difference between the two modes had in-
creased considerably, favoring the moving pres-
entation mode.
In the experiments reported in this paper,
the average visual angular rate was probably
considerably less than 70 /second3 and the lack
of a significant difference in the percent of
targets acquired agreed with the apparent re-
sults of Erickson's study. Erickson did not re-
port any time-to-detection, so no comparison
between the two tasks could be made on this
Since subjects were free to position their heads
at any distance from the screen, the visual angular
rate varied. This estimate of angular velocity was
based on head positions measured in a subsequent,
but similar study.
Visual Angular
Velocity in
Moving Mode
Available
Search Time in
Static Mode
Percent of Targets Detected
Moving Static
Percentage
Difference
Ratio of
Percentage
Difference
to Smallest
Percentage
5? /second
7?/second
10? /second
2.9 second
2.0 second
1.4 second
77
81
4
61
64
3
50
40
10
.05
.05
.25
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
200?June, 1965
measurement. However, when the results of
Erickson's study are combined with those of
the present study an interesting relationship is
suggested among relative performance effective-
ness of the moving and static presentation
modes, observation time, and the probability
and speed of detection. The relationship sug-
gested is as follows:
At the slower speed there is little difference
between the modes in terms of the number
of targets acquired although those targets
which are found are detected more quickly
with the moving presentation mode; at
faster speeds, there is little difference be-
tweeen the two modes in the time to acquire
a target, but the probability of acquiring tar-
gets is higher with the moving presentation
mode.
In both modes, performance deteriorates as
observation times are shortened. In general,
the moving presentation mode would seem to
offer the greater advantage for target acquisi-
tion in near-real time. In the sections to fol-
low, a rationale for the above statement will
be developed and the implications for equip-
ment design will be discussed.
An Explanation Of The Results Of The
Present Experiment
The conditions under which the results of the
present experiment were obtained should be
carefully delineated. Although it is believed
that these conditions are representative of those
found in a great many actual reconnaissance
missions, they are specified to avoid overgen-
eralizing to situations in which they do not
exist. However, as it will be shown, the re-
sults obtained appear to have generality beyond
the limits of the present study.
The conditions which existed in the present
experiments were: (1) targets could essentially
appear anywhere on the display; (2) the ob-
servation times were so short that a thorough
search of the display was not possible, al-
though at the slower observation times, several
rough scans could be made; (3) some targets
tended to be marginal in size, near the per-
ceptual threshold; and (4) the observers were
thoroughly briefed and made familiar with the
HUMAN FACTOR;
targets' relatively well-defined patterns. Thes{
four conditions meant that a very fine, system.
atic scanning pattern was necessary to search
the display and that the use of periphera:
vision was limited. The observation time was
spent primarily in searching for the target
which, when seen foveally, would generally be
recognized immediately. Erickson, too, noted
,that with moving imagery, particularly at the
faster rates, use of peripheral vision was lim-
ited.
In the present study, the observers were in-
structed to "find the targets as soon as possible
after they appeared on the display." This
meant that with the static presentation mode,
a thorough and careful search in two dimen-
sions over the entire display was required. With
the moving presentation mode, however, the
optimum strategy was to search the display in
a single dimension along the edge at which the
targets first appear while the imagery moves
across the display. This searching procedure
minimized the time to find the target and
maximized the time available to study it be-
fore it disappeared (if further study were
needed). The two search modes are illustrated
in Figure 8.
Assuming no difference in scan rate or the
effective cone of vision, the observer could
scan the same area of the imagery in a speci-
fied period of time in both modes. This would
suggest that the probability of finding a target
would be the same with both presentation
modes, and this theoretical equality was borne
out by the empirical results. As observation
times increased, a greater area could be
scanned, and theoretically and empirically,
more targets could be found.
The difference between modes in the time
required to find a target after it appeared on
the display is the result of the difference in the
search techniques used. With the static pres-
entation mode, the length of the search path
between the initial fixation point and the target
position was proportional to the time to find a
target. Since these two end points could be
located anywhere on 'the display, the time in
which a target could be found could vary
from almost instantaneously (if the two end
CHARLES
points should
entire obscrva
cient in this ;?
of the display
which a targc
was a pprox i rr
period. As o
fore, so did t
and the varia
These relat
ing prcsentati
optimally alot
the maximun-
by the time ?
covered by ti
leading edge.
erally half o
being considc
entire display
with the me
shorter than
more, the wit
effective visu
stant even v,
time was vai
variability of
Casual obser
the effective
DIRECTIC
Fig. 8. Di.
for two rates.
imagery scann
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
JUMAN FACTORS CHARLES W. SIMON June, 1965L-20I
-defined patterns. These
that a very fine, system.
-was necessary to search
the use of peripheral
le observation time was
_arching for the target,
ally, would generally be.
y. Erickson, too, noted.
;cry, particularly at the
ripheral vision was lim-:
the observers were in-
rgets as soon as possible
am the display." This
atic presentation mode,
I search in two dimen-
.play was required. With
Dn mode, however, the:
:to search the display in
g the edge at which the:
aile the imagery moves:
nis searching procedure'
.o find the target and
wailable to study it be.'
if further study were
:h modes are illustrated
-Ice in scan rate or the,
in, the observer could,
the imagery in a speciE
-oth modes. This would.
ility of finding a target
with both presentation,
ical equality was borne.
-esults. As observation'
reater area could be
zally and empirically,:
found.
!en modes in the time:
2.t after it appeared on
of the difference in the:
With the static pres-
gth of the search path:
on?point and the target:
al to the time to find
D end points could be;
le display, the time in
be found could vary}
ously (if the two end;
tI
points should coincide) to the length of the
entire observation period (which was not suffi-
cient in this study to allow a thorough search
of the display). Therefore, the average time in
which a target was found on the static display
was approximately one-half of the observation
period. As observation time increased, there-
fore, so did the average time to find a target,
and the variability of the time to find a target.
These relationships do not hold for the mov-
ing presentation mode. If the observer searched
optimally along the leading edge of the display,
the maximum time to find a target was limited
by the time the target stayed within the area
covered by the effective visual cone along the
leading edge. Average time to find was gen-
erally half of the width of this area which,
being considerably less than the width of the
entire display, would explain why time-to-find
with the moving imagery was significantly
shorter than with the static imagery. Further-
more, the width of the area subtended by the
effective visual cone remained relatively con-
stant even when display size or observation
time was varied which tended to reduce the
variability of the time-to-detect measurements.
Casual observation suggests that variations in
the effective visual cone came from a shift in
MOVING
DIRECTION OF IMAGERY
MOTION -ix-
thc distance of the observer's head from the
display and the distance the center of the scan
pattern is from the edge of the display; these
tended to be slight.
Supplemental Study. Although the empirical
results supported these analytical conclusions,
it was decided to check the possibility that they
could be accounted for by the observers using
a different scan rate for the two presentation
modes. Perhaps, it was hypothesized, the re-
sults of the original experiment had been arti-
fically produced by the implied instructions
concerning search techniques with the static
presentation mode. It may have been that
allowing the observer to practice and become
familiar with the length of the observation
time available may have encouraged him to use
all of that time to carefully scan the static
display once. If he had instead made a rapid
scan of the display first (or several rapid
scans), he might have reduced his median
time-to-find-targets by being able to pick up
the more obvious targets quicklyduring the
rapid scans.
To check this alternative, a supplemental
study was carried out using two of the ob-
servers previously tested under the 40-second
time limit. They were given two test sessions,
STATIC
Fig. 8. Diagrams illustrating search paths and scannable areas with moving and static presentations
for two rates. Broken lines represent path of eye movements: Shaded portion represents the area of
imagery scanned within area of effective vision.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
202?June, 1965
one for each of the two presentation modes.
Only the 18-mile wide strip and the 12-inch
square display were used with a 40-second
observation time limit. The subjects were in-
structed?in fact, urged?to scan the display
during the static presentations as rapidly and
as often as they reasonably could. They were
told to continue to scan the moving presenta-
tions in the same way they had done previ-
ously. The results are shown in Table 12.
Although performance in the static pres-
entation mode tended to improve slightly over
the original run, neither the overall results nor
the conclusions drawn in the original study
were changed by these results. This improve-
ment may have been due to the increased scan
rate used with the static presentation, or it
could have resulted from the practice which
the observers by this time had with the test
imagery.
In either case, under the conditions examined
here, the number of targets found with either
mode of presentation did not differ more than
could be expected by chance. The 'differences
in the time required to find and recognize tar-
gets with the two presentation modes, however,
remained relatively large.
Considerations for Equipment Design.
Three other factors, in addition to those al-
ready described, favor the moving presentation
over the static one. First, it has been found
that it is difficult for an observer to make a
precise and systematic search. Although he
may conscientiously try to scan in a particular
and regular manner and reports that he is
doing so, eye movement records reveal this is
not always the case (Townsend and Fry, 1960;
White and Ford, 1960). In the static mode,
this situation can lead to overlapping or un-
HUMAN FACTORS
scanned areas, for the observer must search in
two dimensions with no external restrictions
applied to his search pattern. In the moving
mode, however, using the leading edge of the
display to guide his scan in one dimension
while the imagery moves by in the second
dimension reduces the inefficiency of search.
Second, even if it took the same time to find
a target once it appeared on the display with
either mode, the moving mode is still favored
when time-to-find is critical. The reason for
this preference is that a time delay is "built in"
to the static mode. The delay is that required
to allow the image to build up in the near-real-
time mission. Thus, if a target lay just outside
the section of the terrain being displayed dur-
ing a single static presentation, it would have
to wait a complete observation time period
before it would appear on the next frame. On
the other hand, with the continuously moving
mode, the target would move onto the display
within seconds after it was sensed by the radar,
Third, if the observer's task is to find a pre-
defined target area rather than a target, the
possible breaking up of the terrain context
into two sequentially presented static frames
could detrimentally affect the recognizability
of the area pattern. Task effects will be ana-
lyzed later.
Scale Factor. In this study, the scale factor
of the imagery could be enlarged in two ways:
(1) by increasing the size of the display and
filling it with an enlarged image, or (2) by de-
creasing the area of the terrain being displayed
on a display of a given size. I3oth methods
yielded results supporting the generally ac-
cepted contention that performance is im-
proved with the larger scale factor. More
targets were acquired and found in less time.
TABLE 12
Results of supplemental study of the effect of scanning instructions to two observers.
Original
Supplemental Test*
Moving Presentation Mode Static Presentation Mode
Number
Recognized
Median Time
Number Median Time
Recognized
33 3.5 seconds 37.5 14.8 seconds
33.5 3.5 seconds 38.0 12.5 seconds
"18-mile wide imagery, 12-inch square display, 40-second observation time. Values are average of two observer's
performance.
CHARLES \:
However,
by the two
factor, per
considered \\!
sign decisio
target Size \\
varying the
in the full
than varyirv
holding grot
phasizes the:
techniques p!
the larger d:
presented ??';
quired to
times the 1-)1-
play, the in
twice as gr,
only in one
When groun
play size hel,
size increast
image area
a fact which
was also inc
One other
effectiveness
The apparer
the display
was partially
viewing dist;
in display s
servers. woul
reducing th.
target. Cor
observed N1
changed.
It should
under some
of ground c
could be d?
. formance. 1
amount of
which in cc
critical in lo
directly. TI
coverage re.
in the lengtl
fixed vehiel
limit beyonc
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
MAN I' ACTO
irver must search in
external restrictions
ern. In the moving
leading edge of the
in one dimension
; by in the second
ficiency of search.
he same time to find
on the display with !
-node is still favored '
:al. The reason for
iae delay is "built in"
=clay is that required !
; up in the near-real-
-trget lay just outside
being ?displayed dur- '
-ttion, it would have
rvation time period
the next frame. On
:ontinuously moving
DVC onto the display
sensed by the radar,
:ask is to find a pre-
than a target, the
the terrain context
:ented static frames -
the recognizability
effects will be ana-
the scale factor
larged in two ways:
of the display and
silage, or (2) by de-
rain being displayed
size. Both methods
the generally ac-
-erforrnance is im-
icale factor. More
found in less time.
1,sentalion Mode
Median Time
14.8 seconds
12.5 seconds
erage of two observer's
("11,0.1.ES W. SIN1ON
However, the difference in results obtained
by the two methods demonstrated that scale
factor, per se, was not the critical factor to be
considered when making optimum display de-
sign decisions. Although in both cases the
target size was changed by the same amount,
varying the amount of ground area displayed
in the full frame affected performance more
than varying the 'size of the display while
holding ground area constant. 'This feet em-
phasizes the role that search and scanning
techniques play in the recognition task. With
the larger display, although the ground area
presented was the same, the observer was re-
quired to search a static display area four
times the previous size. With the moving dis-
play, the increase in search effort was only
twice as great, since the observer searched
only in one dimension along the leading edge.
When ground coverage was reduced with dis-
play size held constant, not only was the target
size increased, but the portion of the total
image area occupied by the target was larger,
a fact which was not so when the display size
was also increased.
One other factor may have also reduced the
effectiveness of increasing the display size.
The apparent increase in target scale factor on
the display when the larger display was used
was partially offset by a shift in the observer's
viewing distance to compensate for the change
in display size. With the larger display, ob-
servers would tend to move back further, thus
reducing the visual angle subtended by the
target. Considerably less compensation was
observed when the ground coverage was
changed.
It should not be overlooked, however., that.
under some circumstances, reducing the area
of ground covered within a fixed display size
could be detrimental to reconnaissance per-
formance. For one thing, to do so reduces the
amount of contextual terrain information
which in certain reconnaissance tasks can be
critical in locating targets too small to be seen
directly. Then, too, the reduction in ground
coverage results in a corresponding decrease
in the length of the observation time (for any
fixed vehicle speed) and this may reach a
limit beyond which the observer no longer has
June, 1965-203
sufficient time to perceive the target.
Observation Time and linage Movement.
In a near-real-time mission, the observer may
have little control over the length of the ob-
servation time. Time, however, represents one
of the most severely limiting factors in this
form of reconnaissance and targets which
might have eventually been discovered had the
search time been longer, may not be detected.
When the imagery is moving, the shorter
observation time creates an effect in addition
to that of limiting the search period. In this
case, observation time and movement combine
to create an angular imagery rate which, if
fast enough, results in a blurring and reduction
of visual acuity (Miller and Ludwigh, 1960).
It is this possibility that has made many design
engineers skeptical of the use of moving
imagery for reconnaissance tasks. In practice,
however, this fear appears overemphasized
for the following reasons.
First, most of the rates encountered in the
airborne reconnaissance situation (as well as
in the faster moving reconnaissance missions
from space) are below the point at which
blurring occurs. Second, and possibly even
more important, current research results sug-
gest that whatever the degradation in per-
formance which results from a decrease in ob-
servation time with the moving imagery, a
greater degradation will occur with the static
imagery. As observation time decreases to a
minimum, observers eventually reach a point
where they stare at only one point on the dis-
play (Erickson, 1964). Under this condition,
success in finding a target on the static display
is reduced to being lucky enough to be staring
at the correct spot at the right time?a highly
unlikely occurrence. With the moving display,
however, staring at one spot does permit the
search of a greater area, namely, a line across
the display as the imagery (however blurred it
may be) moves by. For this reason, the prob-
ability of detection at the very fast speeds can
be expected to be higher for the moving pres-
entation mode, and this is exactly what Erick-
son (1964) found.
Both Erickson and the present experimenter
noted that at excessively fast rates, observers
have difficulty in searching only the lead edge.
"7-
7., 7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
204?June, 1965
Instead they begin to follow the rapidly mov-
ing image across the display. The absolute
value where this occurs is not known, but it
could be related to the velocity at which blur
begins to have a noticeable effect and the ob-
server attempts to track the image to compen-
sate for the movement.
The blurring effect of movement could have
an effect once the search phase of the task is
over and the observer must study particular
patterns in finer detail. This was not con-
sidered in the present studies. However, if this
task were required, once the target had been
found, the display could be turned off and the
imagery studied statically if necessary.
But short observation times are not the only
disturbing factor in this reconnaissance task.
Longer observation times may also result in a
performance decrement as an effect of the
difficulty of maintaining a sustained vigilance.
In this experiment, the longest observation
period was 24 minutes. Although there were
many more targets in this laboratory study
than might be expected under field conditions.
subjects reported feeling drowsy and of
"blocking" while monitoring the continuously
moving display over the longer time periods.
With marginal targets, even at speeds in which
there is sufficient time to study a display
thoroughly, targets may be missed if the over-
all situation creates suboptimum levels of
alertness in the observer.
Tasks. The task of the present study was
one of target recognition. As stated earlier,
the targets were clearly defined and readily
recognizable, once they were fixated foveally.
Detailed examination was not required.
An exploratory investigation has been made
comparing the moving and static presentation
modes when the task was one of finding a
target area. Tentative results suggest that the
static, sequential frame mode may be preferred
at the faster viewing rates. Aerial photographs
covering nearly forty thousand square miles
were used to simulate the view from an orbit-
ing spacecraft through an optical telescope.
For this situation, target area acquisiton relied
primarily upon the recognition of gross terrain
features and relatively few man-made objects.
HUMAN FACTORS
To accomplish this it was necessary that rivers
and mountain patterns, for example, be ex-
amined over areas which could not be encom-
passed in one or a few visual fixations, and
that numerous similar appearing features be
studied and rejected before finding the correct
ones. Observing these complex patterns ap-
peared more easily accomplished when the
image was static than by searching the lead
edge of a moving display.4 Furthermore, if
clearly visible cbcck-poInts of known space-
time distances from the target area were found,
the static, sequential frame presentation mode
facilitated keeping track of the time interval
more easily than did the continuously moving
mode.
Although it is true that a continuously
changing display does allow a continuity for
following prebriefed terrain features into the
target area while sequentially presented static
frames conceivably could break up this con-
tinuity and the target area pattern, this dis-
advantage of the static mode seems outweighed
by its advantages. Also, the continuity of the
static presentation mode can be improved if
there is not a one-hundred percent change in
the image from frame to frame.
Partial Static Presentation Modes. Most of
the disadvantages of the static presentation
mode and the differences in performance be-
tween it and the moving presentation mode
possibly may be removed if the frame-by-
frame change is only a partial one each time.
For example, if instead of each frame chang-
ing to a completely new image per frame the
change would occur when only twenty-five per-
cent of the image is new, several effects occur.
First of all, the time available for each new
portion is reduced proportionately from the
total time that would be available if the entire
frame had changed. Second, however, the ob-
server would have a proportionately smaller
area to study, and what is more important, he
may now essentially duplicate the scan mode
that proved optimal with the moving presenta-
tion mode, i.e., scan primarily along the lead
edge. Furthermore, if the optimum percent of
'Presumably a similar effect might be noted if
targets were complex and ill-defined.
part
rcla
viski
mtty
deli
tend!
in ti
imp
an
sub
Ano
13
Eric
iv3
. Mill
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
ilum FACTORS
a was necessary that rivers
rns, for example, be ex-
hich could not be encom-
few visual fixations, and
ir appearing features be
before finding the correct
-se complex patterns ap-
accomplished when the
-in by searching the lead
iisplay.4 Furthermore, if
-points of known space-
ne target area were found,
frame presentation mode
-ack of the time interval
the continuously moving
-ue that a continuously
2s allow a continuity for
terrain features into the
iiientially presented static
=ould break up this con-
..t area pattern, this dis-
c mode seems outweighed
iso, the continuity of the
_ode can be improved if
andred percent change in
ie to frame.
2ntation Modes. Most of
the static presentation '
tnces in performance be-
-oving presentation mode
anoved if the frame-by-
a partial one each time.
ad of each frame chang-
-icw image per frame the
-..hen only twenty-five per-
-iew, several effects occur.
= available for each new
broportionately from the
be available if the entire
Second, however, the ob-
i proportionately smaller
-tat is more important, he
duplicate the scan mode
vith the moving presenta-
primarily along the lead
the optimum percent of
itr effect might be noted if
-id ill-defined.
C11.1.1i1 ES \V. SIMON
partial frame change could be determined?
relative to the width of the effective cone of
vision?the sequential series of static changes
may serve to better pace the observer and to
delineate his scan without introducing the
tendency for the eyes to follow the imagery as
in the moving presentation mode. This might
improve performance at the faster rates.
Whether or not these speculations are true is
an empirical question to be determined in a
subsequent study.
REFERENCES
Anonymous. Dynamic Viewer Model 100-A. Cul-
ver City: Hughes Aircraft Company Report
B-40, 19 April 1963.
Erickson, R. A. Visual search performance in a
moving structured field. 1. opt. Soc. Amer., 1964,
34, 399-405.
Miller, J. W., & Ludwigh, E. Time required for
detection of stationary and moving objects as a
June, 1965-205
function of size in homogeneous and partially
structured visual fields. In A. Morris and E. P.
Home (Eds.) Visual Search Techniques. Wash-
ington, D.C.: Nat. Acad. Sci., Nat. Res. Coun.,
Publication 712, 1960.
Nygaard, J. E., Slocum, G. K., Thomas, J. 0.,
et al. The Measurement of Stimulus Complexity
in High-Resolution Sensor Imagery. Wright-
Patterson AFB, USAF AMRL TDR-64-29,
1964.
Rhodes, F. Predicting the difficulty of locating
targets from judgments of image characteristics.
Wright-Patterson AFB, USAF AMRL TDR 64-
19, 1964.
Townsend, C. A., & Fry, 0. A. Automatic scan-
ning of serial photographs. In A. Morris and
E. P. Home (Eds.) Visual Search Techniques,
Washington, D.C.: Nat. Acad. Sci., Nat. Res.
Coun., Publication 712, 1960.
White, C. T., & Ford, A. Ocular activity in visual
search. In A. Morris and E. P. Horne (Eds.)
Visual Search Techniques. Washington, D.C.:
Nat. Acad. Sci., Nat. Res. Coun., Publication
712, 1960.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
VI/71i
Zd:tor's Note
paper is one section of a
;epert entitled "Display Techniques
r Digital weapons Control Sys-
...erns" prepared by I-I. R. Luxenberg?
cf Q. L. Donness, of the Bunker-
Ramo Corp., ror the U.S. Naval Ord-
nance Test Station (NOTS), China
: Lake, Calif., under contract number
; ? i\D3z.',33-10519. Another section of
- the some report appeared in the last
issue of .NFORMATION DISPLAY
under are title -Photometric Units"
by ft R. Luxenberg.
introduction
Display device and system parameters
are of two types:
(1) Those pertaining to system inte-
gration requirements with associ-
ated environmental and logistic
considerations.
(2) Those pertaining to the interface
with the observer.
The first category consists of such
items as physical size, weight, power,
signal levels, symbol encoding, interface
requirements, temperature, humidity,
shock, vibration, cost, reliability (MTBF),
and maintainability (MTTR). '
The second category includes such fac-
tors as brightness, contrast, .color capa-
bility, resolution, viewing distance, and
vie,wing angle.
All of these parameters arc either avail-
able from the manufacturer (many must,
of course, be properly discounted) or
may be tested directly.
IL is not the intent of this paper to
7. :1 ri
'1 Li. Li
7Th TT r '",
b)-7. H. R. Luxenberg and Q. L. Bonness
duplicate detailed descriptive material
which is already easily available and well
presented in the current literature. For
example, several good articles describe,
in some detail, techniques for CRT char-
acter generation; others describe alpha-
numeric .indicators, etc. Furthermore, it
would be impractical to attempt to tabu-
late operating parameters for all cur-
rently available displays here. Several
extremely good summaries are readily
available, although even these suffer
from some incompleteness. It is, how-
ever, the purpose of this paper to present
a quantitative discussion of the physical
units and psycho-physical significance of
those parameters peculiar to display tech-
nology.
Since the function of a display unit is
to transfer information to the operator by
visual sensing, it is important to, consider
those parameters that determine the capa-
bility of the. human eye to perceive infor-
mation. Therefore, the following section
discusses those factors that are important
to the visual perception of information.
A definition of each factor is given, and
a discussion of the way in which it re-
lates to the evaluation of display devices
is included.
Brightness and Contrast
Contrast, not brightness, is the signi-
ficant factor in display legibility. Bright-
ness is generally specified. because it is
dependent only on the equipment, where-
as contrast is generally a function of
ambient lighting. Brightness is allo speci-
fied because it would seem that the high-
er the brightness, the greater the visibility
under higher ambient light. This is not
necessarily true, . since some displays of
lower intrinsic brightness have bettcr
visibility than far brighter ones. (See
Tables 1 and 2).
Furthermore, unnecessarily high bright-
ness may be a luxury where ambielo
light is subject to control, since it has
been determined experimentally that
where observers have control over ambi-
ent lighting for reading they tend to
choose lower values than are generally
considered optimum. For example, under
a controlled test', when the maximum
available illuminations were 10, 30, end
45 foot-candles, the observers selected 5,
12, and 16 foot-candles as the optimum
values.
The required brightness of a display
should be obtained by the following pro-
cedure. First, select (preferably at the
lowest acceptable level) the arnbidlt
light desired at the work station, and hy
means of a mockup measure the ambie?t
light falling on the display surface. The
source (s ) of ambient illumination shoul'd
be relocated, collimated or otherwise
shielded to reduce this to a minimum.
Acceptable contrast ratios are: for wli:e.
symbols on a black background, 5 to 1;
for line drawings or text on a white
background, 25 to 1; for pictorial scenes.
100 to 1.
From a knowledge of the reflectivity
of the display surface (unless it is glossy,
unity is a conservative - estimate), the
brightness of the background is obtained,
and multiplication by the desired coil:
trast ratio will specify the brightness n.7
quired of the symbols. For example, if
the ambient illumination on an eleetro-
luminescent alphanumeric display whose
luminance is 10 foot-lamberts can be
held below two foot-candles, the resultihg
Declassified in Part- Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A00-190-0020.61-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
_
3 1/2 optical-line pairs Detailed alphanumeric
7 TV-line representation
il-iCURE 1: Character representation.
TABLE 1
Typical Brightness (foot-lamberts)*
Surface of the sun
4.8 x
106
Surface of a 60-watt frosted incandescent bulb ("hot spot")
36,000
Surface of a 60-watt "white" incandescent bulb
9,000
Surface of a 15-watt fluorescent tube
3,000.
White paper in direct sunlight
9,000
Clear sky
2,000
Surface of moon, bright area
750
White paper on office desk
25
Pulsed electroluminescent mosaic panel
20
Television raster
20
Light valve, 10- by 10-foot diffusing screen, 2-kilowatt lamp
20
Theatre screen open gate
16
Note that pulsed electrolurninescent mosaic panels have brightness comparable with television
raster or opengate theatre screens.
*Brightness values.compiled from:
(1) D. G. Fink, TELEVISION ENGINEERING HANDBOOK, McGraw-Hill, 1959.
(2) IES LIGHTING HANDBOOK, THIRD EDITION, Illuminating Engineering Society, 1959.
(3) REFERENCE DATA FOR RADIO ENGINEERS, I. T. and T. Corp., 1949.
(4) Measurements and Calculations by the Authors.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
9
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
0
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
()
0 0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
FiGjRE 2: Full 5 x 7 dot mosaic (35 elements,
full slpnanumeric).
(..):Itrast of 5 to i will be adequate for
good legibility.
if, however, the observer is working
at a desk where the illumination is 50
foot-candies, a sheet of white paper will
have a luminance of 40 foot-lamberts
and the 4-to-1 brightness difference be-
tween paper and display may prove an-
noying. Since the display brightness
cannot be raised, the working lighting
can be reduced. If the observer is given
control over the ambient lighting, he
will find an optimum (for him) working
Size-Resolution-Legibility
Fiesolution is generally described in
terms of line pairs per millimeter (lines/
millimeer). The average observer can
resolve two lines that subtend an angle at
the observer's eye of one minute. Since
1 minute of arc 0.0003 radian, the eye
resolves at a viewing distance of. 10
inches (250 millimeters) about 13 lines/
There is some confusion in the litera-
ture between optical lines and television
lines. Optical lines are synonomous with.
line pairs (i.e., an optical line consists
of a black and white pair). To show one
line pair on a television raster requires
at least two televi,sion lines. Because the.
optical line pair may not coincide with
FIGURE 5: Numerics using. reduced 4 x 7 dot
mosaic.
FIGURE 3: Reduced 5 x 7 dot mosaic (27 ele-
ments, numeric only).
TABLE 2
Contrast Levels
Textual copy
(white on dark) 5-10 to 1*
Line drawings and black
on white text 25 to 1
Photographs 100 to 1
"Legibility of fine detail degrades with
increasing contrast if the eye is adapted
to darker background level, because of
dazzle effect.
the raster lines, more than two television
lines are required. The correction factor
of 1.4 is called the Kell facto?. Thus,
one optical line pair requires 2.8 tele-
vision lines, for full resolution.
Resolution in terms of line pairs is
significant for display purposes only
when photographs are shown, where the
observer is required to distinguish be-
tween tWo close objects. Actually, if the
existence or nonexistence of a black line
on a white background is to be deter-
'mined, the .line need subtend an angle
of only 0.5 second at the eye.
FIGURE 4: Reduced 4 x 7 aot mosaic (20 ele-
ments, numeric only).
If alphanumerics -are to be legible, the
vertical angle subtended at the eye
should be at least 10 minutes of are.
Since 10 minutes of arc arc approximate-
ly 1/360 radian, for each foot of viewing
distance the character must be at least
1/30 inch in height. While the eye can :
resolve 10 optical line pairs in 10 min-
utes of arc, this does not mean that the
detail present in an alphanumeric charac-
ter requires 10 line pairs for legible pres-
entation.
An examination of the "E" shown in
Figure 1 reveals that no more than 3.5
optical line pairs are required. The seven-
television-line representation of the "E"
on the right side of the figure may ap-
pear somewhat crude when viewed close-
up, but when viewed from a distance at
which the lines are not resolved, it is
quite legible.
Although the number of television lines.
used is just twice the number of optical
line pairs for the illustration, near per-
fect registration of the raster scan to the
figure being displayed was assumed. To
take care of the misregistration problem,
the -Kell factor (2.8) is introduced. The
product of 3.5 (the. number of optical
lines required) by 2.8 (the Kell factor)
is very nearly 10, the number of tele-
vision lines required to present a charac-
ter of good legibility.
0
000
000
0 0 0000
00
0000
00
00
CO
0
0
0
0 0 0
0
0
0 0
00
00
0
0
0
0 0 0
0
0
0 0
0 0
00
0
00
00
0000 000
000
0
00
000
00
0
0
0
0 0
00
0
0 0
0
0 0
0
0
0
0 0
0
0
0
00
0
0 0
0
0000
000
0...000
00
00
00
0 C.,'
C. "Displays, Picirs and Lighting,"
Informa: 45;yp!,,y, Vol. 1, No, 1, pp 16-26, Sep-
tcinberiOetki.)e;
2D, G. Fink, Television Engineering Ilandbo,k,
- 1c.)50.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
the legibility of the characters
,:ntis on their structure, there is no
.ivcfsal agreement on the optimum
..nariienDr she. Among the recommend-
ed shapes arc the following:
(1) Modified gothic (sans serif) char-
acter of height/width ratio of 3
to 2 for most letters, with the ex-
ception of 1, M, and W, and a
stroke width of 3; of the character
(2) The same as above, but with a
height/width ratio of 5 to 3 and
a stroke width of 1/6 to 1/8 of
i.he character height (MIL-SPEC-
33558).
Inc: authors prefer the 3 to 2 height/
width ratio with a bolder stroke, say 1/6
of the character width. The character is
then nine strokes *?:. strokes wide.
parity and clock hits. These six bits per-
mit display of a maximum of 63 charac-
ters (or 64 if the "blank" is counted).
This number can include 26 alphas, 10
numerics, and up to 27 special symbols,
including punctuation marks, etc.
While the 6-bit code can be decoded
into 64 "lines," one per possible charac-
ter, in many display applications it is
found more desirable to reduce the num-
ber of decoded lines by assembling the
characters from a smaller number of' ele-
mentary elements, as, for example, by
a dot or stroke mosaic.
Dot mosaics ?
The coarsest mosaic that is capable of
providing easily legible alphanumeric
symbols is a 5- by 7-dot mosaic, as shown
in Figure 2. Here, only 35 decoded lines.
are required. If only numerics and a
limited number of symbols arc required
depending on the manufacturer. The
bars, strokes, or segments are arrailgeti
in a pattern similar to one of those
shown in Figure 6.
The exact shape varies from one man-
ufacturer to another, with some rounding
of corners and minor variations in the
way adjacent segments join.
The segments may be electrolumincs-
cent strips, electrochemical cells or cath-
odes in a glow discharge tube, or they
may be trans-illuminated by neon or in-
candescent lamps. The power require-
ments, luminosity, and color differ in
each method of implementation.
The characters are nearly as legible as
those made from a 5- by 7-dot mosaic,
but logic (switching) requirements are
reduced from 35 inputs for the full 5-
by-7 matrix to 16, 14, 9, or 7, depending
on the type of bar matrix chosen. A deci-
(a) The double-
hon;-:, window (sev-
en elements, nu?
meric only).
FlCURE 3: Segmented character formats.
A minimum of two stroke widths should
be provided between adjacent symbols.
Special Character Shapes
While well shaped gothic characters'
are most attractive and are generally
-greed upon as providing the ultimate in
ieg.bility, certain types of display imple-
mentation preclude their use. Instead,
they present characters of a stylized
sl-itiPe made up of discrete elements.
3nce these are "learned" they become
i.early as legible as the pure gothic to
which they are an approximation. De-
vices in this class use either dot or stroke
rnesaies, as discussed below.
rj-:3 primary reasons for the adoption
hese types of characters are economy
irathc;ty oi implementation, as is-
apparent from the following con-
,:rations. Alphanumeric data are stored
a..:d processed in data processors in data
:;...ary coded form, requiring a minimum
SX ou:s no:? 0/m rnotor ?r
Declassified in Part - Sanitized Copy
(b) The double-
hung window (nine
elements).
(e.g., +, ., etc.), some of the ele-
ments are not required, and the number
of lines may be reduced to 27, as shown
in Figure 1
For numerics only, the reduced 4- by.
7-mosaic shown in Figure 4 is satisfac-
tory. This mosaic requires only 20 lines.
In general, dot mosaics present better
appearing characters than stroke mosaics,
but stroke matrices are more economical
in the sense that fewer lines are required.
However, stroke, matrices with as many
as 35 specially shaped elements have
been devised; these give extremely good.
characters, far better appearing than the
dot matrix, with no more operating com-
plexity. The elements may also be ar-
ranged in a parallelogram rather than in
a rectangle to provide sloped or "italic"
characters.
Stroke mosaics
This type of character presentation is
known variously as the bar matrix the
?
Approved for Release 2012/04/23:
(c) The starburst (14 elements, full
alphanumerics; also available is a
16-element starburst with the upper
and lower horizontal elements split).
NOTE: Segments may be electrolum-
inescent, gas tube cathodes, electro-
fluors, or illuminated by incandes-
cent or neon lamps.
mal point or underline bar may be added '
In some instances. The same height/
width/stroke ratios apply as for shaped
characters.
Screen Characteristics
The length-lambert units of the previ-
ous section are defined ? by defining the
luminance of a perfect diffuser, approxi-
mated by a fresh chalk surface, to- he
numerically equal to- the incident illumi-
nation in length-candles (e.g., a perfect.
diffuser in bright sunlight is illuminated
by 9000 foot-candles and has a lumi-
nance of 9000 foot-lamberts as seen from
any direction).
For other-than-perfect diffusers, a re-
flectivity factor is Used to obtain the
luminance.- Note that the reflectivity
need not be less than unity and it may
vary with direction. ?
For specular reflection the luminance
of the reflected image is that of ?the
source -itself. If the perfect diffuser in the'
CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
.n., the luminance of the mirror is
x ftPfoot-lamberts (same as the solar
ace) on the axis of the reflected rays
..nd aero off-axis. Thus, the reflectivity
" ?I of the mirro; in this case is a
J.y peaked function with a maximum
V.1 af (4.S x 10)/9000 or greater than
3 x 10.
13irectiona1 screen with flaked alumi-
num surfaces (to give high gains) are
.0e.:ently used to obtain greater image
'aii, tness, but at the price of a restrict-
cti viewing angle. An increase of granu-
larity in the aluminum paint provides a
greater viewing angle; or the surface
be embossed with tiny convex mir-
rors, which will not be visible at normal
viewing distances.
The
mirrors spread the incident para-
llel rays into a cone generally of rec-
tangular cross-section, to cover the audi-
ence space desited. The angles of the
cone are simply related. to the width or
height and focal length of the tiny mir-
rors.
An analogous situation holds for rear-
projection screens. The analog of a per-
fect mirror is a clear layer. Either more
diffusion or a' lenticular structure May
be used to provide the desired spread of
light. Except for very small screens (i.e.,
those subtending a small angle at both
eye and source), gains greater than 2 to
-3 produce visible and annoying "hot-
spots."
Unhrcrmity of Luminance
A projection screen (either front ? or
rear) will not be uniformly luminous over
its entire surface, although for all prac-
tical purposes it may appear so. In this
section, the reasons for non-uniformity
and practical limits are discussed.
The two major causes for screen fall-
off are: (1) lack of uniformity in illumi-
nation, and (2) the variation of gain
with the angle between line of sight and
the reflected or transmitted projection
ray. This angle is called bend-angle.
Illumination fall-off is primarily due to
a cos 0 factor which enters when a
finite-area lamp source and conventional
optics are used in the projector. The
angle 0 is the half-angle subtended by
the screen at the projector. With small
sources and aspherical optics, the factor
may be increased, perhaps to cos" 0 or
better, but at a cost". The easiest cure
is to reduce 0 by increasing the projec-
tion distance. If space is limited, a fold-
ed optical path may be required.
Even with a uniformly illuminated
screen, the illuminance will fall off with
bend-:1:-:gle, unless the screen is a per-
fect diffuser (constant gain). The angle
at which the gain is one-half its peak
value is called the half-power, or 50%
"Wick Screcil Slick Proicction,"
Dispirol, Vol. 1, No. 1, pp 8-15, Sep-
. ?
bend-angle. Higher gains mean smaller
half-power bend-angles, unless the gain
is deliberately lowered by the addition of
light absorbing material for contrast con-
trol (see below).
While the eve is extremely sensitive to
luminance differences in adjacent areas,
it is relatively insensitive to a gradual 2-
to-1 variation over a large area. For this
reason, a 2-to-1 variation in screen lumi-
nance is generally acceptable, and even
a 3-to-1 variation will go unnoticed by
the casual observer. Therefore, the re-
striction to a .30% fall-off (often seen in
display specifications) is a luxury that
the eye does not appreciate; a 50% fall-
off is a much more reasonable specifi-
cation value.
Contrast Control
From a knowledge of the incident il-
lumination and the screen gain in the
direction of view, the luminance is cal-
culated by a simple multiplication. Un-
fortunately, ambiei it light is also reflected
back to the observer, adding to both
highlighting and shadow luminances, thus
reducing their ratio (the "contrast").
With a front-projection screen the only
effective means of control is to use a
high-gain (directive) screen, placing the
observer on the reflected projector ray
and avoiding all ambient light sources
in the neighborhood of the Projector.
Only scattered ambient light will then.
degrade the contrast. This directivity ex-
plains the effectiveness of the currently
popular lenticular. screen for home pro-
jection use. The lenticules direct the re-
flected light primarily into a sharply -
defined rectangular cone, with sharp fall-
off outside, rather than into a broader
region with gradual fall-off.
With rear-projection screen, more free-
dom is permitted in contrast control. A
high-gain screen is inherently a poor re-
flector; hence, contrast is immediately
enhanced, even with light sources on the
line of sight. As the gain is reduced by
increasing diffusion to provide ?the de-
sired viewing angle, the front reflectivity
is, unfortunately, increased.
The reflectivity may, in turn, be low-
ered by adding opaque material or a
neutral density filter. This reduces the
gain but does not change the bend-angle
as in the case of increasing diffusion. The
effect of the filter is discussed further
under the section on projection systems
which follows.
Projection System Parameters
In projection systems the amount of
light reaching the projection screen is a
function of a number of parameters, but
for well designed optical systems cer-
tain rules of thumb are applicable. Sev-
eral useful ones are:
(1) A light valve television system
? using a xenon arc lamp has an
n PI ?1
in
(2) A 35-mm slide projector using an
incandescent lamp has an output
of between 1 and 2 lumens/watt.
These are typical figures only, and the
upper or lower limits may be exceeded
by exceptionally well designed or by
poorly adjusted equipments.
For a uniformly diffusing matte screen,
the screen brightness in foot-lamberts is ?
equal to the luminous output of the pro-
jector divided by the screen area in
square feet. For example, the screen
brightness proclticed by a 2000-watt
xenon light valve operating at an output
of one lumen/watt on a 10-foot square
screen is 20 foot-lamberts.
Both front- and rear-projection screens
may appear either brighter or dimmer
than a uniform diffuser, depending on
the nature of the screen. and the line of
view. The ratio of brightness to illumi-
nance is a maximum when the line .of
ViCW extends directly back to the pro-
jector ?for a rear-projection screen, or
along the reflected ray from the projec-
tor for a front-projection screen.
This maxi,imun value is referred to as
the gain of the screen. Typical useful
screen gains line in the range of 0.5 to
2.0, although higher-gain screens are
used when the restricted viewing angles
associated with them are not objection-
able or are desirable.
The higher the screen gain the higher
the contrast, in general, for both front-
and rear-projection screens. This is true
for rear-projection screens, since the re-
flection of ambient light from the front
surface is low with high-gain screens.
For front-projection screens, the direct-
ivity of higher-gain screens is such that
off-axis ambient light is not directed into
the viewing area.
An additional degree of contrast con-
trol is available with rear-projection
screens, in that a neutral. density face- ?
plate may be incorporated. If the (one-
way) transmission is x%, the two-way
attenuation of reflected light is x",L, A
50% neutral density faceplate thus atten-
uates the projected beam by a factor of ?
two and the undesirable ambient reflec-.
tion by a factor of four.
Because of all the variables introduced
by the screen parameters and the am-
bient lighting conditions, it is not prac-
ticable to specify brightness and contrast
values for a projection system without
defining the viewing conditions. It Is for
this reason that projection systems are.
best defined in terms of lumens output.
Brightness in foot-lamberts for a
gain screen screen is obtained by dividing the
lumen output by the screen area in
square feet. Contrast is obtained by maul-
tiplying the ambient light in foot-candles
by the screen reflectivity coefficient and
computing the ratio of "light" to "dark"
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
BIBLIOGRAPHIC DATA SHEET
Author(s) Leifer, I., Spencer, C.J.D., Welford, W.T.
Technical Optics Section, Imperial College, London, S.W. 7
Affiliation and Richmond,?C.N., The Orangery, Kelvedon, Essex
Title *"Grainless Screens for Projection Microscopy"
Periodical JOSA
No. -;571. 47 Vol. ? V
Agency
? Publisher
Place of Publication
Date of Publication
Pages
1422-1423
December 1961
CONZIENTS:
Status
Classification
a. Ordered
?13'.-> In-house
c. Reviewed
9-22-66
SAIM Key
;2,1,
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
- Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
(;R.\ -;C.REENS IOR PROJECTION XliCROSCOPY
..,jeclive duration we mean the time during which the
speed is less titan the minimum for which the
htless effect occurs.
re 2 shows an al tachment. we have constructed
microscope, using this principle; the front screen
so that each point on it describes a circle of about
.;atin radius in a few seconds. The gap between the two
-,1-.2ens is adjustable and we have found that: a separa-
?idn of up to a quarter of a millimeter has no discernible
on the quality of the image. An alternative way of
:::oving one screen is to vibrate it at ac-power-line fre-
,,uencv with an amplitude of about 0.5 mm; this is just
cifective as the slowly moving rotating screen. and
;crhaps slightly simpler and cheaper.
With this system we have found that an over-all
:..ignitication of 1000 permits the finest detail resolvable
:'means of an oil-immersion objective to be seen as
..carly as by direct viewing andthe screen luminance is
::creased to about 0.002 stub. A further fourfold
ii.i.rease in luminance is obtained if the screens are
.,:ihed to increase the forward transmission, as pro-
:,oi.ed by Dyson'; the etching produces a clearly visible
-iructure on the screens which can be seen moving, lint
..ititougly this may be slightly distracting to the ob-
,csver it does not impair the resolution of detail in the.
:mge. On account of the strongly peaked polar diagram
,f these screens it is desirable to use a field lens as
liacated in Fig. 2 in order to obtain a uniformly
liltaninated field of view.
An alternative proposal for a grainless screen is to use
siagle rather rapidly moving screen6; we have tried
*is in the form of a disk-shaped Screen revolving in its
-4nplane and we have found that although an improve- ,
-,:cnt in definition of the image was obtained it was not
marked as for the double screens. The steady. move-
alt the screen through the field of view was always
aired and this was distracting, but worse still was the
that large-scale variations in scattering over the
? a.ed showed up as a flicker with the period of rotation.
,(1 this. could only be eliminated by using a speed of
J. Dyson, 3. Opt. Soc. Am. 50, 519 (1960).
'N. IL Mason, British Patent 590,931 (1947). See also E. Lau
J. Reinitz: "Optik alter Wellenlangen" p. 229 (Berlin, 1959)
--Al E. Lau and R. Schalge, Feingeractechnik 7, 121 (1958).
1423
FIG: 2. Grainless screen. Light from the microscope eyepiece
(lower right) after reaection from three mirrors forms the image
on two ground glass screeds (upper center) nearly in contact. The
image may be viewed through a field lens (upper right) if the
screens arc etched for high forward transmittance. The ground-
glass screen nearest- the observer undergoes circular translation
in its own plane at about 20 rpm from the motor drive (bottom
center): In an alternative arrangement the moving screen is
oscillated in its own plane at ac-power-line frequency.
rotation above about 30 rps, a- rather high speed for a
thin disk of glass. A single, rapidly oscillating screen
was found to be quite useless. ?
To summarize, we have found the best results by
using two screens separated by not more than 0.25 mm
and with relative speed exceeding 1 mm/sec.
A CXN OWLED GME NT
-This work is supported financially by .an extramural
. grant from the National Coal Board, but the opinions
expressed are those 'of the authors and not necessarily
those of the Board.
c?,, ...-t...- ---,--' 0'664-- "- ,
j
,, -'
1,^ ''' (
, , ..,e
.,..,..n.1. ,
"tt?1 ?.e ;
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
; 1961
Grainless Screens for Projection Microscopy
F. LF:IFF,R, C. J. D. SPENCER, AND W. T. wEr.roRD,
Technical optics hn prrial College, London, S.11' .7
AND
C. N. Ric:limos')
The Orangery, Kelvedon, Essex
(Received June 13, 1961)
/Back-projection screens for projection microscopy with high luminance and no loss of definition are
described; the screen grain is removed by the slow relative movement of two ground-glass screens placed
face to face,/
routine sizing and counting of coal dust samples at
National Coal Board Area Laboratories it is neces-
sary to count particles as small as 0.5 /I. This is at pres-
ent done with a projection microscope of conventional
design with an opaque screen for front projection. In
seeking ways to improve this technique we noticed that
the picture had to be at least 50 cm from the eye because
of physical obstruction of ? the view by the microscope,
so that the magnification used was high, usually about
3000; the picture luminance was therefore very low,
about 0.0002 stilb (candle/cm2), with a 250-w high-
pressure mercury lamp as light source.' At this low
luminance the Fechner fraction is about twice its normal
value2 and the visual acuity is halved,' so that it is
clearly desirable to increase the luminance considerably.
The obvious way to do this is to use a back-projection
screen, so that the distance from the eye to the screen
can be considerably reduced, the magnification reduced
and the luminance correspondingly increased ; but if
this is clone we find that the grain of the ,projection
screen obscures the detail in the image. All projection
screens have a more or less grainy, sparkling appearance,
and the scale of the grainy appearance is considerably
larger than that of the actual grain in the material. For
example, Fig. 1 shows part of a microphotometer trace
across a ground glass screen obtained by using a scan
FIG. 1. Microphotometer trace across fine, ground-glass screen.
Illuminating and collectinr, apertures both f/80, scanning spot
100/4 square. The horizontal line at the top corresponds to 1 mm
on the ground-glass screen.
This source has a luminance of 20 000 stubs and the theoretical
sceden luminance under the conditions of use described above is
0.006 stilb; the difference may reasonably be ascribed to reflection
and absorption losses and to the difficulty of completely filling the
condenser aperture with the image of the part of the source of
maximum luminance.
2S. Hccht, J. gen. Physiol. 7, 235 (1924).
?3 S. Hecht, Arch. Opthalmol. 57, 564 (1928).
fling spot 100 1.1 square and illuminating and collecting
beams of N. A. 0.006; the standard deviation of the
fluctuations in transmittance is ?23% and it can be
seen that the scale of the irregularities is such that detaii
several hundred microns across would be obscured
although the.glass was "smoothed" (i.e., ground with the
finest grade of emery as the last stage before polishing)
and the grain size of the emery was only about 10y.
The magnitude of the effect also depends on the numer-
ical aperture of the illuminating and collecting beams,
the values being chosen here to correspond approxi.
mately to conditions in projection microscopy.
This difficulty of graininess with small numerical
aperture of the illuminating beam is found with all kinds
of screens to a greater or less extent and it is probably
unavoidable. A screen must have irregularities several
microns in size if it is to scat ter at all and these must be
arranged in a random manner so that the screen does
not become simply a two-dimensional diffraction grat-
ing; it is presumably the linear scale of the random
variations in the screen structure which gives rise to the
seen graininess, just as the graininess in a photographic
emulsion corresponds not to individual grains of silver
but to variations in grain density and clumping. ,
In order to circumvent this difficulty we have there-
fore applied an old idea' for a grainless screen to be used
in engineering gauge projectors, etc. In this system two
ground-glass screens are placed with their ground sur-
faces almost in contact and one is moved slowly in its
own plane relative to the other.; the sparkle and graini-
ness are continually changing and are smoothed out by
persistence of vision to give a perfectly grainless, smooth
screen. The effect, is quite startling for low-contrast
objects of which the images are less than a millimeter in
size on the screen, such objects are almost invisible
when the screens arc stationary but become brilliantly
clear when the movement is started. The relative speed
of the screens needs only to be quite slow, about 1 mm:
sec, but the motion- must be such that there are no
stationary points or else if such a point does occur
effective duration must be less than, say, 1/25 sec; by
F. A. MacAdam and Taylor, Taylor Se: Hobson Limited,
British patent 592,815 (1947). See also K. J. Habell and A. Co.
Engineering Oplics (Pitman Publishing Corporation, New Tort,.
- 1948), p. 273.
1422
ca.ct lye dui..
tive spec,
cite-
Jur,, 2
for a micro-i.
iaoves so tint
5-mm radius
:-..-reens is ad j
tion of up to
effect on iilC r.
moving one sq
ouency with
eli.ective ii
prhaps slight
With this s
magnifica Lion
by means oi
c.carly as by d
increased to
increase in in
,Itched to inc.;
posed by Dysaf
rue' tire on I
.ltitoi.tgh this
.-erver it does
imago. On acco
of these scree'.
indicated in I
illuminated fiel
An alternati\
single rat her
his in the form
own planoand
mein in definiti
as marked as it.
:11c711 of the scr..
not iced and this
fact that, large-
?ercen showed 1.1
and this could (
.1. Dyson, J. OF
.? N. H. Mason,
j. Reinitz: "0
.ind E. Lau and K.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
-r? /7
j
?i
1 : ri, .???,-- r.....-7,.
......).L......'-)' L.-.i '.J '-'.."..;
r;?
,
, r7. 7.471
r",n 7
!"7":"`? , ? '4-1 ?,
./7
v*-
? ?,"?1
?????? ? ?
Abstt'ac,-:
presented in this
ars usefui, handbook type
tc.,7 enabic, the display
acaigner to de.rmine rapid-
values of basic
parsiteters dis-
so that he can de-
.:....f.r.ntion to those unique
to the design of a
-tonaie;-ing dis,plays in terms
of rniriimum resolvable,
se-
o ures of design charts
aevelopeci 'rrom which
, range vs:Lies can be
aisplay element size,
symbol size and n-iaxi-
c.,uantity for orOered or
these
parametric values can
in terms of the effect
of other system pare-
coin manner the feasi-
biiity o?r a display configuration can
varifiea in terms of fundamental
inconsistent require-
ments con ha modified in the con-
cocai phase.
DISPLAY ELEMENT SIZE (c ) IN INCHES
1,0
0.1
,01
0.001
. ?
by 'Glenn E. Whitham
? "7-
r--1 r,
a c a7
. _
7
,...11. .... ,.... ..... , . ..
-1-1-II -:, - I'?? :-:-,,,
...,'1?:,-4-, I' , '
-; : ? ' i I -,?!,.?,r7t;',PI, ?-: ?? -116_, --:-- :1;,./1. ... - ., ? ??,I,,???,,? , - :?
?
H ,,i....,,i:1-:-;
.. :
: . :i .. : '..r.:1
rffif,i i':ilj'e,',-L:,-:1
':-'1:.'ll.:::::,-::::::.7.?:::".:1::::. '.:: i 4! :t: :: -i.--.V: i '..
: i:::::::.::::: ..i.H..:::::::7-:-:---
I'' ':
.1
I:
j ?
:"-."!..
:;?, ,1,1
,'
t ? ?z,?-'
1, ?,4.
,1_,.....c,c, ,_,
, .," ,? - . ..7 -7:----, 7---'-...,-,-,77,1.1;
-i I , ? 11::4^., ,?,:::i:1:; :::. ?,-.1::11 ","?.:::',..',...-1.;'.
...,..: ,,,,.. ,,,, , L_,_ .. , .?? , 0 i li_
.. : . ..-,- ??-'
! -.;.4., , .. .
,
?;:.
-,;-
I?
,C',r,
.... :11-i4:;:lid:.i.:::...i
:: ,:: j,;.':',
'L.':
i LI'
'?......,
:n1
;:.i 1.....: : -77
il
'O'Z'.... f ip.
: iiHi,.../.:j::
id
......
.-.2 ...
:H._ -.1
i
I .
.4,
I.c.,
,,:i .? -.. i I
Y---', ',.' 't 1 .,-::: ""l',:i
1., :,:,,I,,.
' '''i
r ' " ' ti ''-
-:7,:.
1 ' ' '
.1
'' l' ''' I ' ' - .. '?
,...-P
.1
,- ???I--,., ,
- , ,,,,
:J.,:
/,:l:

t-!.
1 : ,
ti :
,I,
Q.,)
.
.1,
,',-!-;--7:1-7,-ff-'
?:,,;1',,?4",
?
''' V' I
'
.
1, ,
,,1-1-'.....a9--;;
i j,{ .,,IV.
:1,1'c6;?)...C.,1.
,...
11
',,,,I.;'/il
, . . :#?'
,./,1,;;
, .. .. r:._, ,; , .
,:? -,--t-,i,ii
'.'.; ..?,". -t.';';i'.
_.,, , ,,:..
-.--;;?J 11.
ti ', : ; Pt:?,
.., :.... ...,,,-i..,.;..,_..::,.:.
,!?H?1Ittiii-,.!;;;."
.;.,.,,.... :.!!'il.i:.:
H
Pm
?
\,(;)
,--,-;:-
4, -.: i:?:1,,il
::: I.., -
- (4.,,c,
:7, ;-,
,,,,
".-:,?,, ?,-,J.;s:c.., ,I,-i ,
,-,i.i.i.,
...1,. ,,,,i-, ,, -: ? ? --?
MUM:
l?::77-"c?
7i
; i 1!
ii -
J,.;;:i I i.F-v.>x permits use of tables of natural
logarithms or Poisson distributions for
evaluation if (B3) is put in the form
= N ? M ? (1 ? N
= N ? M [i_ (1 ? N
N M Li?e
?N/Mi
(B5)
The mean percentage of symbol over-
lap a is thus
? -I
a rz 100 {N M N/MJ
7.1
al 100
/1? N/M
(116)
which has been evaluated and plotted
in Figure 5 of the text.
Evaluation of the overlap distribution
function. leads to iterative expressions
which are impractical to evaluate except
by use of a computer. While this was
beyond the scope of the present investi-
gation, it is evident that further effort in
this area would yield useful data.
The author would like to acknowledge the assi
Ma-. st-
ance of M. nnvirl XA _
Declassified in. Part-Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
... es: Ellectrob:mineseent pan-
, -panel oT light vanes, light ampli-
lier
Tioe ? "Crystal Bali" display.
?. .?.3dg panel
Ose:.lating plan
intersTiccs
t.. 'ing, rods
Characteristics
(i)eormance parameters) ?
"',i;iiii-ations ?.Size of CRT or
projection screen, etc.
(More properly known
as b-r-'?nce)
Litnitat.'ons ? Darkened
nolarized lighting, broad band
etc.
.,::::asi)onse Time ? This could mean
wridiv-, speed, frame time, target up-
(.1L:ng.r, time.
? Ability to distinguish
two s.--rnirrne -Targets is measured in
many ,..vays.
Capability
,Zata Copy Capaiiiity
8. Syn-thol Limitations'
5. Go!or and Halftone Capability
Storage
Conclusions
V111:1;.; the foregoing attempt at dis-
play categorization and classification is
a(!?;ti-e.dly incomplete, it does serve to
saft em.)basis from the usual designer's
vicw.)3int to one which regards 'a dis-
a black box performing certain
lono. Once the desired functions
bat de-ToTmined and performance
then one can begin
tole hardware. Pleae, note
. . Terial represents the views
anLnor only, and in releasing it
the U.S. .Naval Research
7.....7.1:oraLory does not necessarily endorse
C. TALMADGE, Head
.....,:71ay Techniques Section
Naval Research Lab.
Vlashington,
1.)i.r.nne
dant cle:!ien:7::ti
at ;ca ae article The Detetmina-
tion of Display Screen Size and Resolti-
?ion Perceptual Limitations by Mr.
(Glenn E.) Whitham, (TD, Vol. 2. No.
'65) he makes the state-
"Tae equivalent 'number of line
is, Of c3urse, one-half the number
cleinen...: or raster lines.? This
stfite-
rue for c, chart or device
hieb at:T..ally has line pairs spaced
111-,on the raster. isfowever, if
-....-rnation than this rather limit-
.. T.:.;piayed, resolution is some-
.an
.
what you might expect.
-.oder: used _L.4. 1. 'Trust
more
.z..; to his intentions in tn... ..-?ture.
jouN SHAVIE.a,
The Bunl:er-I.---,aimo Corp.
Sierra Vista. Ariz.
Declassified in Part - Sanitized Copy Approved
LDOGISUAT0-2?
Err 11-2 c'Z'VAVJ 71
7.:\ZEUMiA,.
High standards of quality in the ma-
terials, construction and design of
DIGISWITCH promote reliable, trouble-
free performance, long life and excel-
lent appearance.
DIGISWITCH Thumbwheel Switches are
available in six (6) different series
offering the engineer maximum design
flexibility. All are panel space savers
? modular constructed for simple, in-
expensive mounting?human engi-
6r=
L7-Z n. LIE '11-i-V
neereci to increase operator efficiency
?and offer extensive codec: electrical
output capabilities.
Shown above (actual size), are MIN:-
SWITCH*? the smallest thurnbwheel
switch on the market; and the unique
rapid-setting 40-position Series 600
DIGISWITCH.
For a catalog on our Complete line ef
DIGISWITCH and MINISWITCH
write to:
*Trademark
C
A Division of Enclave? Corporation
655 South Arroyo Parkway, Pasadena, Caljornia
Phone: [phone]
MAN1.02ACTURND UNDER PATENTS 3.089 923 AND 0191.72a OTHER ?AtIEN211 PINDING.
for Release 2012/04/23 : CIA-RDP78B04770A001900020031-7
VI
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
:;ton Station, N.
and more tran-
,.. CT-displayed data from
col-...7,er readouts, telemetry,
capturod effectively
',zany today by prop-
-, phot3graphic tech-
r:,.coriiing, processing,
choice of t..chnic,oes that
ca be ied, singly or in combina-
. -Ave any problem of handling
,-;uantities of data that moves
tco faac for visual analysis.
-2:.o..o..lechanisms' engineers and
.13tographic scientists?now deliver-
photographic hard copy systems
:3-)ace. --"-? -tians?special-
;Taphically
...rid utilize
in any quantities.
?roblerns Are Not The Same
Gaof.pp::cation may require speed,
aambar volume, another very high
r..,-o..ntion, another economy. And, of
course, many applications require all
.nc,3 and more. Each requirement can
by a sn,!cilic photographic tech-
.Y..^ combination of techniqUes?
.e1;,y chosen, properly combined,
pro],-,er.y app:ied -by a specialist:
::;:lotolnee11-nisras is unique in the
of ?qotog:aphic data processing
b,:cause its capabilities include not just
one bat a whole range of specialized
technic:ues. indeed, its specialty may
be said -..;c be integrating specialized
1:ech:.iques in unusual photographic
syotern-.., for erlicicnt handling of greater
and greater quantities of transient
L'ata.
11CCOIN. rola Sono
04COONG PUN.
EVOSuNE (WC
EcKOunC Sioltiot
LEN37
NEFLCE MONO. -?
OSP'vo
[ANODE N. Tu.
Omorontm
VON+. 1.0O3
fonokoo
nOJEcOON
PoCCISSIN4 5mo,, So.
rirClolcsoaC POP.
NO1.1. SUP",
PAPER cunt,
EtrcraoiTAIC
CIOR091
['Anon
VINTON
COotit COSINOSteor,
COPY
Cao NICEMPI
CROIECOOn LENS
CNOAC,ON LON.
OC ESSE El PAU OFE
Shown here is a diagram of Photo-
mechanisms' DATASTAT II, a hard
generator chat combines the son-
.y of s'.-..ver halide photography
the zneed and economy of elec-
e,,.
..,nances are . your
problem needs a similar
?Or. .,c;chniclues for an opti-
o.utiGn.
17.7.r your copy of
Photo-
ocaao.:.L.
cnaracteristics of photo-
data handling systems. '
Whitham? submitted the following
reply, in response to ID's request for
his an.swer to Mr. Shaver's letter.-Ed.
Mr. Shaver is indeed correct in point-
ing out that the resolution of a line
raster along the axis normal to the lines
is degraded for arbitrary display subject
material to a value less than indicated
by consideration only of the line struc-
ture. This resolution loss factor has been
found by various experimental" 2, 4 and
theoretical3 studies to lie in the range of
0.53 to 0.85, with 0.7 a good nominal
figure. For obtaining a given value of
display resolution the number of lines
required is 1/0.7 or 1.4 times the num-
ber of desired resolution elements along
the axis normal to the lines. This cor-
rection factor is commonly known as the
Kell factor. Further discussion of this
subject was included on page 10 of the
paper by Dr. 7,;:antitative
Measures of Zji.4)lay Characteristics
which just preceeded my paper in the
July/August 1965 issue of Information
Display.
?
GLENN E. WHITHAM
Staff Engineer
Raytheon Company
Wayland, Mass.
1. R. D. Kell, A. V. Bedford, and' M. A.
Trainer, "An experimental television sys-
tem-the transmitter", Proc. I. R. E. Vol.
22, pp 1246-1265; November 1931.
2. A. V. Bedford, "Figure of merit for tele-
vision performance" R. M. A. Eng., Vol.
2, pp. 5-7; November 1937.
3. H. A. Wheeler, A. V. Loughren, "The
fine structure of television images", Proc.
I. R. E., Vol. 26, pp. 540-575; May 1938.
4. Baldwin, "Subjective Sharpness of tele-
vision images", Proc. I. R. E., Vol. 38,
pp. 458-468, 1940.
P.S. The second sentence of the last
paragraph on page 17 of my paper
should read, "Situation type data
displays usually require a resolu-
tion of about 1000 to 2000 ele-
ments for adequate symbol resolu-
tion and differential position dis-
eriminability."
Information requested
We have been doing a considerable
amount of work in the. field of electro-
luminescent displays, and in the course
of this work several product ideas have
evolved in the area of nioving pointer
and moving scale panel indicators.
In order to determine the direction
our product development work should
take we are trying to collect as much
information as possible on desirable char-
acteristics of panel indicators, user pref-
erences and requirements and the poten-
tial value of electroluminescent displays
in providing a useful product improve-
ment.
Any -information that you could sup-
ply which would be helnful in this in-
Declassified in Partz.Sanitized Copy Approved for Release 2012/04/23: CIA
vestigation would be greatly appreciated,
including suggestions of other possible
sources of information.
C. H. WArisinkw
Industrial Products Manager
Huyck Systems Co.
Huntington Station
Long Island, N.Y.
We are presently engaged in a study
of the graphic recording field for proc-
ess control and laboratory usage; and
are in erested in futiice trends in this
area, Therefore, I would appreciate in-
formation that you feel may be helpful
to us.
L. P. LANE
Arthur D. Little, Inc.
Cambridge, Mass.
I am very inter--. d rn an area of the
field in which i dad no reference con-
tained in recent issues of your magazine.
This is in the requirements, theory
and/or construction of display and status
boards posted manually from the rear.
have seen several in operation in mili-
tary installations but have never had the
forethought to look into the manufactur-
ing stage.
Could I impose upon your good offices
to look among your advertisers and con-
tributors and to furnish me some con-
tacts with technical competence in this
particular. area. I will be very apprecia-
tive of any assistance you can render in
this search.
HENRY D. BATEY
Chief, Graphics ? -
United Aircraft Corp. Systems Center
Farmington, Conn.
ID readers who can contribute de-
sired information are urged to com-
municate directly with the above cor-
respondants - Ed.
SID and Journal helpful
. Both the activities of the Society for
Information Display and the articles in
Information Display are of great inter-
est to me. Material presented in the
Journal has contributed significantly to
my knowledge and understanding of
data display technology.
? It is my responsibility to design and
implement the "Total Information Sys-
tem", culminating in display design in
the following categories: .
1. Large Area Display
2. Small Area Display
3. Desk Top Display
The project in which I am engaged is
designed to provide the company with
a system in the 1970's that will be suit-
able to the environment at that time.
JOHN P. THOMPSON
Director of Data Processing
Hoffmann-La Roche, Inc.
? Nutley_ Nr.t
-RDP78B04770A001900020031-7
?? ? ?? ? ? ? ? ? ? ? ? ? ? ?-???? ? ? ?? ? ? ? ?? ? ?-? ? ? ? ? ? ? ? ? ? ?? ? ? ? ?
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
STAND
2-128
SIT-STAND
The illustrations show recommendations
for angular mounting of visual displays such as
PPI-type CRT's. These dimensions are only ap-
proximate. They do, however, represent usable
standards for about 90 per cent of the male
population.
-4 _4?,?,,,A0?44
1-1,":01e..e.41,.e,,,,.
V
?
(
1LL."
..(*
J-3:4 e
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
r
?L.) 5
ontmand centers we find for the
e the need to read cathode-ray
thsplays, projection displays, and hard
e:py under the same level of illumina-
tan Neithe: the illumination of thea-
tees nor that of offices is suitable, the
one being too low for reading papers
t-,tttl the other to,) high for reading dis-
plays.
e. shows a method for corn-
the at which both dis-
- and ilard copy can be read with
at ease, or conversely for showing
he ativaotage given one data source over
a., other by the use of a different level
o ilittruination. Properties of displays
o copy needed in the computation
ore given, and a sample computation
is presented. fit is shown that the com-
puted value of illumination is both satis-
factory and compatible with the recom-
mendations of T.:no Illuminating Engineer-
ing Society. Means for avoiding eye strain
nod fatigue (applicable to any level of
illumination) are restated.
if a col-fp:nand or management center
is to be successful, the human decision
maker must receive all the assistance
from his data sources that the state of
tile art permits. To this end the display
equipment, the hard copy, and the illu-
mination must all be matched to the
needs of the user. There is much infor-
int.tion on the size,- shape, etc. of the
symbols that should be presented, a
great deal of information on the tech-
nicplcs for producing and presenting the
symbols, but very. little on the environ-
ment in which the displays should oper-
ate. This paper is an attempt to fill a
major part of this gap?the illumination
of the operating area.
One may ask why this should be a
serious question, since the subject of
illumination has received so much at-
tention in the last fifty years. The rea-
son is that in command centers one finds
for the first time that self luminous dis-
plays?cathode ray tubes, electrolumines-
cent panels, pn.ojection screens, etc. ?
are combined with redeetive displays ?
messages, on., operating plans, budg-
e..:,, sehem.e. etc. ? with the require-
ment that they be used in close sequence.
fellows that the illumination must be
at one time suitable for the electronic
displays and for the hard copy.
There is an extensive literature on the
illumination of rooms for self-luminous
displays, beginning with the data gath-
ered by the then Society of Motion Pic-
ture Engineers on the lighting of movie
theaters. This art was adapted during
the war for radar display rooms and other
weapons control centers, and extended
by use of narrow-band light to take ad-
vantage of a color difference. The philos-
ophy of minimum illumination is still
seen in rooms where edge-lit, grease-pen-
cil plots are kept. With the advent of
brighter displays there came a demand
that we "come up out of the caves" into
an office atmosphere.
1ES Recommandations
The recommendations of the IES (Il-
luminating Engineering Society) "reflect
.a consideration of many variables such
as visual data, . . . economic factors,
convenience, and availability." On ex-
amination it is found that their visual
data are based on reflective materials ?
the reading of papers, the performance
of shop tasks, etc. It has long been
known that when such tasks are difficult
because of poor form, extra fine lines,
low contrast, etc. a higher level of il-
lumination is helpful. It has been a
long time since electric illumination has
been expensive, and it is certainly con-
venient and available. One must con-
clude that there is no factor in the IES's
considerations which exerts a strong in-
fluence. toward lower illumination, and
since they have no control over the qual-
ity of material in view, it is quite proper
that their recommendations be generous.
Obviously, in a command center a
compromise between these extremes is
required, and the compromise must be
pleasing to persons of high rank. With
due attention to the opposite effect of
illumination on reflective and self lumin-
ous displays, a best illumination can be
achieved, and at the present state of the
art it will be a very satisfactory illumina-
tion if the conditions are met with rea-
sonable care.
The Data
It has long been known that the ease
siem33/3/331
77-11
of seeing is related to the illumination, ?
the size of the detail that must be seen,
and to the contrast of that detail to the
background; and various pairs of these
variables have been studied parametrical-
ly until the relations are well known.
During an eight year research (as of
1959) Dr. H. Richard Blackwell has re-
lated these four variables using the same
group of observers.
In these experiments the subjects faced
a hollow box which covered a wide field
of view and which was illuminated even-
ly with white light. Near a reference
mark on the back wall there was a
translucent spot of white plastic which
appeared the same brightness as the
rest of the wall under the front illumina-
tion, but which could be back illumi-
nated to a higher level, and the equip-
ment was so built that both brightnesses,
the size of the spot, and the duration
of its back illumination could be varied
over wide ranges.
The observers were thoroughly trained
in practice runs not included in the
data, to eliminate variation ascribable to
learning. They sat before the. box for
a sufficient time for their eyes to be-
come adapted to the light level to be ?
used; they were then told the limits of
a time interval in which the spot might,
or might not, have been back illuminated,
and were asked to decide whether or not
they had seen it.
Since the same color light was used
for the - background and the spot, the
contrast could be computed -using the
simple relation:
where
and
? Br,
C = (1)
B,
C is the contrast
B? is the higher of the two
brightnesses
B, is the lower of the two
brightnesses
This expression for contrast has a value
which varies from zero to infinity, and
(unlike some) it has the advantage that
the zero value conforms to the condi-
tion which has no visible contrast.
Over 80,000- observations were made.
For each set of conditions the data were
INFORMATION DISPLAY, SEPT/OCT, 1964
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
1 "-//7- Tr 7. r7r-v
r/ ;h./ /
L-4
/7/7;f2/J' ii'IJI
/77771 r s
first cc:reeted for the known percentage
of lucky guesses, then were fit to the
normal ogive so that the 50% point (the
point of maximum slope) could be de-
termined with accuracy. The curves were
Caen smoothed by the use of the known
-:i.:rainetric ? rcliti- between pairs of
of this proc-
c.is it was found ;bat an introduced vari-
ation as small as 2% would require an ob-
vious violation of one of the parametric
:elations.
This work has been reported in various
sages of progress, references 1, 2, and
beibg the more important and carry-
.; :eferences to the others. The corn-
ma data, without some of the details
3C 'now they were gotten, are given in /
rcierence 3. Because all the data came /
iTora the same basic source, because all
four variables are covered, and because
of the care used in smoothing, it is felt
Cnat the, results constitute an unusuallY
useful statement of fact.
To extend these data to the conditiont
and requirements of practical seeing, Di.
'2;1.1d:well uses the concept of field foe',
tors by which to multiply the required
contrast. A factor to compensate for the
difference between skilled observers using
fo:cod choice and the ordinary reader
facing a new problem was determined
experimentally. Factors for off-axis view-
ing, for tile lack of warning, and for the
f:equency of presentation were deter-
mined experimentally. The latter three
were checked with reasonable accuracy
by use of a task evaluator, an instru-.
mcnt whereby a task and a test spot
could both be viewed under conditions
of controlled contrast and that contrast
r,:doced to minimum visibility.
The development of the field factor
concept is described in reference 3. The
overall factor, as determined by the
process described above, is in the range
from 33 to 37.
While the basic data are felt to be
valid beyond doubt, the value of the
:actors has been questioned, some
?.corkers feeling 'Chat the proposed values
1?_iiid to to- .?w a level of illumination.
The author na:,. arbitrarily used an over-
::: value of 43- in making up his human
'ORMATIC,^, MSPLAY, SEPT/OCT, 1964
by A. C. STOCKER
performance curves; however, it will be.
shown later that the value of the field
factor has no bearing on the selected
illumination so long as the same value is
used in the computations for both hard
copy and self-luminous displays.
The author has applied a field fac-
tor (40) to ft-. :? ,.-crice 3 for
detail subtendi.:g 2, and 4
minutes of arc, and for illuminations
between 0.1 and 100 footcandles and
has interpolated for easier use. The re-
sulting curves are given as Figures 1, 2,
and 3.
100-
30
10
100
70
50
33
20 E
3.3
FIG.
HUMAN PERFORMANCE
FIELD FACTORS- 40
LOG
0
RIGHTNESS
Fl- I.
Fig. Human performance when reading sym-
bols with a stroke width (alpha) of one
minute of arc.
It must be stressed that the speeds
shown in Figures 1, 2, and 3 are not
reading speeds, but are the inverse of
the times when the spot was back
illuminated in the tests. The term "read-
ing ease" was considered for a while, but
it is hard to conceive of an ease of 100,
so the term "relative perception. speed"
was finally chosen.
The Selection of an
Illumination Level
It is obvious that both the self-lumi-
nous and the reflective displays must
be readable; lacking other indication it
will here be assumed that they should
be equally readable. The optimum illu-
mination level is, then, that which gives
each the same relative perception speed.
This may easily be found once the data
are put in :the pTciiT r orm.
The preparation oi data for reflective
displays (hard copy) is simple.
The steps are these:
a) Determine the minimum size of
detail that must be easily read-
able, compute the angle this will
subtend at the observer's eyes, and
select the curve with this value of
alpha.
b) Determine the contrast of the ma-
terial, and lay 'a straightedge across .
the curve at this value.
c) Multiply selected values of illu-
mination by the reflectivity of the
paper to establish the background
brightness.
d) For each value of brightness, read
the relative perception speed.
The preparation of data for the self-
luminous type of display is a little more
,complex. The steps are these:
a) Determine the minimum size of
detail that must be easily read-
able, compute the angle this will
subtend at the observer's eyes, and
select the curve for this value of
alpha.
?b) Determine the excitation bright-
ness available at the screen (that
supplied by either the electron
beam or the projector) and apply
factors for the gain and loss of the
screen to achieve the visible bright-
ness of the symbols.
C) Apply factors for the reflectivity of
. .
the screen, attenuation of its sup-
port, the effectiveness of hoods,
etc., to selected values of ambient
illumination to determine the back-
ground brightness.
d) Compute the contrast from item'
b) and c) above. (These computa-
tions are discussed at greater length
under Properties of Displays and
Hard Copy.)
e) For each value of illumination, lay
17
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
straightedge. across die curve at
resultant contrast, and at the
istirn?spooding value of background
br:ghtliciss, read the relative per-
eepnon speed.
determine the optimum value of
illumitaction, plot relative perception
speeil vs. illumination for both types of
ilisplav, It is obvious that one will have
slope and one a negative; the
where they cross is, then, the il-
lumination at which the two types can
be read widi equal ease.
Li Displays
is UI 11.1
Before the design of a command cen-
ter enn oc started, it is necessary to
have, at least an idea of the properties
of the display materials that will be
hsed therein. This section will discuss
the properties of a representative sam-
nle o; materials, and will discuss some
et toe steps ken to assure
good log,ibility.
II:trd Copy-Hard Copy will long con-
tinuo to carry a large and important por-
t:on of the information needed in a
center. Yet the information that
avanable on the properdes of hard copy
is very limited. The size of type, for
is the size of the block on which
the face is cut, not the size of the face
and there are only very general
atements on the reflectivity of papers
,ad iezs. For that reason a series of
measurements was made.
these measurements the symbol
, et was taken as the prcpenderant
:neight-i.: the word was in lower case,
-Linen the height of -a lower case a, e, .o,
:a or similar letter was measured. Some
letters have a variable stroke width.
However, ex-ocrience has shown that the
intelligence is carried in the heavier por-
tions of the symbol; it is in fact possible
to read ?material in which a faulty re-
production process has eliminated the
thin strokes completely. So when a vari-
able stroke width was encountered, the
heavier portion of the stroke was meas-
ured. These data were taken with a
measuring microscope.
The determination of reflectivity pre-
sented a problem because of the small
area available in the symbol and the
knOW11 tendency of the human eye to
be influenced by the surroundings. A
series of paper chips varying from white
to black were gotten by selecting from
paper stocks where that was possible,
and by dyeing to fill the gaps. These
cidps were placed in a diffuse white il-
lumination measured with a 'Weston II-
:ui-nhaorner, their brightness was meas-
ured with a Spectra Spot Brightness
and their reflectivity was com-
puted Tliese chips ctid then be corn-
I directly to the paper of a display.
l-ellectivity of symbols was deter-
'.3y, placing, a chip partially across
MATERIAL
TABLE 1
PROPERTIES OF HARD COPY
HEIGHT STROKE
Angle
@ 14"
Mils Mils Mins Color R3fleot. Contrast
Maps
Army Map Service, NK 18-11
Paper white .65
Names-large 160 24 5.5 black .07 8.3
small 40 9 5.2
Highways 38 9.5 rci & blk high
Railroads 7 1.7 black
Creeks 5 1.2 blue .1 5.5
? Contours 5 '1.2 tan fair
Elevations 52 7 1.7
Sectional Air Chart, Winston Salem
Paper Color wash to indicate ground altitude, Est. Avg. 50
Names-large 110 20 5 black .1 4
small 44 7 1.7
Smallest print 40 8 2
Highways 25 5.5 gray .3 Est. .6
Scale tics 8 2
Contours 9 2 tan kir
Nautical Chart, H.O. 1290
Paper
Names-large 144
small 40
Smallest print 24
Depth contours
Soundings 60
14
10
5
5
9
Geological Survey, Redmond, Washington
Paper
Green
Names-large 120 15
small 60 10
Smallest print 45 7
Highways 25
Roads dual 6
Creeks 8
Contours 5
Office Material
Good pulp paper
Average pulp paper
Yellow copy paper
Typewriter samples
Paper
Make A Electric
Make B Elite
Make B Pica
Make B Pica-old ribbon
? Teletype printer ?
Paper
With new ribbon
, With old ribbon, top
With old ribbon, bottom
Computer Printer output
Make C
Make D
Make E
As a basis for comparison-
New York Times
? Paper
Text
Office Copiers
Make F (translucent material) ,
Paper-on dark desk
Print
Paper-on white paper
Print
Make G (opaque material)
Paper
Print
Pencil
Pulp paper
Mechanical pencil, .032" lead, HB
? 2H newly sharpened
2H fairly dull
B newly sharpened.
B dull
white .65
3.5 black .09 6.2
2.5
1.2
1.2
1.2
white .7
overprint for wooded areas
3.7 black .1 6
2.5
1.7
5.5 red & blk hich
1.5
2 blue goad
1.2 tan fair
.7
.65
.5
.7
110 15 3.7 black .1 6
110 10 2.5 black .12 4.8
114 10 2.5 black .1 6
114 7 1.7 gray .15 3.7
gray .6
105 20 5 black .1 5
12 2.7 black .12 4
Indef. .2 2
95 9 2.2 black .08 7.9
100 10 2.5 black .1 6
100 17 4.2 black .08 7.1
.46
55 10 2,5 black .1 3.6
.45
.12 2.3
.65
.12 4.4
.65
.10 5.5
.65
20 5 black .2 3.3
10 2.5 black .25 2.3
15 3.7
12 2.7 black .12 4.5
40 10
INFORMATION DISPLAY, SEPT/OCT, 1964.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy A
at tvpe and examining the type
,:le edge of ,he chip through a low
.eer inietoscopo, the microscope magni-
, where
red.s,
pencil
111('::?;;11"OCI
tO
C011
as
under
the good
mate-
be
prevent
be
an
diffuse
point
made.
example,
light
Some
had
to
to
pproved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
the hood factor can be changed from tiyity must be used with care.
The quantitative relations follow, oil
0.5 to 0.25 by this means.
After these steps, the direct ambient based on the fundamental relations that:
light will be reduced to the point where
the light reflected from the equipment,
from white shirts, etc., becomes im-
portant. Such secondary sources can be
objectionable when they are specularly
reflected from the tube face and appear
as unwanted images superposed on the
data. Glossy finishes should not be used
on equipment, furniture, etc.; a medium
gray mat surface is much less objection-
able. Such light as remains can be ren-
dered innocuous by etching the tube face
to a fine mat surface. This is not truly
a non-reflecting coating?it merely spreads
the reflection over such a large area
that the brightness is reduced and the
shape of the source obscured. It is,
however, a very effective step.
Circular polarizer use
A circular polarizer can be used for
the functions of the filter glass and the
mat first surface, either for better per-
formance or when the desired proper-
ties cannot be had in the available tube.
It is not a complete cure, for it itself
has bright specular reflection .from its
front surface. However, it may be used
as a plane surface, and the range of in-
cident angles that gives objectionable
reflections from a plane surface is much
less than the .range that gives reflections
from a spherical surface.
Proction displays can also take ad-
vantafarption in the screen sup-
port, of a hood, of .directive room light,
and (in some types) of a mat front
? surface on the screen. In addition, they
can take advantage of directivity.
Directive screens reflect (or transmit)
more light in a preferred direction than
in other directions. They therefore are.
said to have "gain", this being the ratio
of the brightness in the preferred direc-
tion to the brightness of a highly diffu-
sive screen subject to the same illumina-
tion. Values between 2 and 10 are com-
mon. If the viewers can be located along
the preferred axis, the brightness of the
data symbols will be increased while that
from light sources not near the projector
will be decreased, and an increase in con-
trast will result. This effect reaches its
peak in rear projection screens, where
the steps which decrease the diffusion
of the projected light- also decrease the
reflectivity to ambient light, and where
it is possible to take advantage of 4 filter-
ing support material.
However, a fundamental characteris-
tic of gain is its sensitivity to changes in
angle. Assuming the plane of the screen
is correct for the location of the pro-
jector and the center of the audience,
there will still be a change in brightness
due to looking at different portions of the
screen or to the viewer's moving about
the room. If this change is to be held
within acceptable bounds, then direc-
soteeder rcelection.
The data gathered in these tests are
given in Table 1. Special attention is
invited to the data on office copying
nee2liihos, pencil copy, and typewriter
and teletype output. It will be noted
thatnut contrast can be raised mate-
rially, and in some cases the Width of
the s,roke iriereased by a significant fac-
tor, by the proper selection of equipment
materials and the use of fresh rib-
bons. One must conclude that the leg-
ibility of the hard copy is to that ex-
tent a function of the interest shown by
the members of the staff. ?
Sell-Luminous Displays?Self-luminous
-lays?cathode ray tubes and projec-
on screens?must combat the ambient
eu:na:ion, so much of this section will
have; to do with means for maintaining
good contrast. The light in the symbol,
the light that carries the useful informa-
tion, comes from either an electron beam
or a proector; we will consider cathode
toy tees
'Rather than get into a discussion of
?ducts, the author will here assume
that ,he tube has a beightness of 50
t-Itemberrs available at the phosphor,
ci offers assurance that this level is with-
in the state of the art using crystalline
phosphors. The reflectivity of the phos-
phor is assumed to be 85(;6, highly diffu-
sive. The problem, :then, is to achieve
as good contrast as is possible with this
device.
The first step to consider is the use
?of a gray glass for the face plate of the
tube. It is true that such glass reduces
the brightness of the symbols. by the
transmission (filter factor) of the glass.
But the ambient illumination must pass
through the glass twice, so the contrast
is. increased by the inverse of the filter
factor. This can be a significant amount;
RCA lists face plates with filter factors
in two groups, one near 0.75 and one
near 0.4.
The next step is the use of a cap-bill
hood to keep as much as possible of the
ambient illumination from falling on the
screen.. Properly designed, such a hood
can easily have a factor of 0.5 without
e,esitic,ing the viewing angle.
The effectiveness of a hood may be
:creased by limiting the angles at which
the figat appeoaches the equipment, as
can be done by hanging properly de-
gned honeycomb gratings below the
ii5bt sources. The light can still come
:rem the ......for portion of the .ceiling, so
it is fehy chifuse (or indirect); it is mere-
ly to those angles which will not
get uneer the hood. It is believed that
ION DISPLAY, SEPT/OCT, 1964
BL
B = I R
L,
I? =-
For hard copy:
B? = Re
A
B, =
Rr
substituting in (1)?
c R? ? R,
RE
For both cathode-ray tubes and pro-
jection screens the brightness of tile sym-
bol is the sum o: te..t produced by the
intended excitation and that of the back-
ground as produced by stray electrons or
the transmission of "opaque" parts of the
film.
For cathode-ray tubes:
B5, = B115 T
eee Ia Rpi, T2 + Be T
substituting in (1)
C
T ? Be
For front projection:
Bil =AG B,
= Be + Be G
Here Be is the sum of the brightnesses
due to the individual room illuminants,
each with the value of gain required by
its. angles. This is so complex that the
value is usually determined by measure-
ments on a mockup or on the final as-
sembly.
(1)
(2)
3)
(4)
substituting in (1)
L?G
C
A (Be + Be G)
For rear...projection, with the diffusing
layer on the projector side of tlie screen:
? In
A
13, = Be + Be GT
B? = I, Rs T2
A single expression for B. is made pos-
sible here by the small difference in
reflectivity for the angles of the differ-
ent room illuminants.
substituting in (1)
L?G
C =
A(I?RsT BeG)
In the above:
A = Area of the projection screen,
in square feet.
B = Brightness, in foot-lamberts.
B. = Brightness due to the ambient
illumination.
135/ = The higher of two brightnesses.
13, = The lower of two brightnesses.
Bo, = Brightness of a phosphor due
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
:19
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
to lectron impact or other ex-
cianion.
= brightness due to unintentional
excitation, as by stray electrons
or transmission through the
"opaque" areas of the film.
- Contrast
= The gain of a projection screen
for the incident and reflected
angles involved. Note - com-
mercial values for G usually
include the reflectivity, so it
is not stated separately.
= An illumination, in footcandles.
___ The ambient illumination.
- The projected illumination. ?
- The light flux delivered to the
sceen by the projector, in Lu-
mens.
= The reflectivity.
The reflectivity of the paper.
The refHctivity of the ink.
A.-c..iectivity of the phos-
mor.
- The reflectivity of the projec-
tion screen.
= The transmission (filter factor)
of the screen support where
he support is between the
screen and the observer.
in order to demonstrate the process
and to get a feeling for the level of
..arnivation the process will indicate, a
sample calculation was carried through
using the nearest value of alpha for which
there was a curve available (Figures 1,
2, and 3.)
For hard copy, an Army Map Service
:Lap, two typewriters with different
weight of type face, a teletype machine,
and a. mechanical. pencil were taken as
representative of the materials that might
be found in a military command center.
100
10
FIG. 2
HUMAN PERFORMANCE
? 2'
FIELD FACTORS - 40
100
70
SO
33
V.
20 o-
15
10
7
3.3
2
LOG ? BRIGHTNESS FT - L
Fig. 2 Human performance when reading sym-
bols with a stroke width (alpha) of two
minutes of arc.
The results of the computation are given
in Table 2 and plotted in Figure 4.
For the display, a cathode ray tube
was chosen and assumedto have a phos-
phor with a brightness of 50 foot-lam-
berts and a reflectivity of 0.85. Filter
factors of 0.4 or 0.75 were used, alpha
was taken as 2' or 4', and the hood fac-
tor was taken as 0.5 or 0.25. These con-
ditions can also be met in. a projection
display if the screen size is properly bal-
anced to the available projected light..
The results of the calculation are given
100
30
1
ria 3
HUMAN PERFORMANCE
= 4'
FIELD FACTORS - 40
8
0
70
33
20
15
LOG BRIGHTNESS FT - L
Fig. 3 Human performance when readIng sym-
bols with a stroke width (alpha) of four
minutes of arc.
EFFCCT or ILLUMINATION
PERCEPTION SPEED
HARD COPY
AF,1>
-.M?V
-
SOO
Fig. 4 The effect of illumination on human
perception speed when reading hard
copy.
Jin
AMS Map
TABLE 2
READABILITY OF HARD COPY
Typewriter Typewriter
Make A Make B
Paper Refl.
Contrast
Min. Alpha Mins.
.65
8.3
2
.7
6
4
.7
6
2
Tel
Pencil
H3 .032" lead
5
4
.6
,7
3.3
4
lllemination
ft-candle
Speed?
B Speed*
Speed*
1
.65
2.7
.7
14.5
.7
.6
2
1.3
5.8
1.4
22
1.4
3.2
1.2
4
2.6
8.3
2.8
35
2.8
5.5
2.4
8
5.2
13
5.6
45
5.6
7.7
4.8
12
7.8
16.5
8.4
55
8.4
9.7
7.2
20
13
21
14
63
14
13
12
40
26
28
28
75
28
17
24
CD
52
35
56
90
56
23
48
100
65
33
70
96
70
25
60
7C-Xt ic,r meaning.
Speed*
10.5
16
25
35
42
48
62
73
80
Speed*
.7 6.2
1.4 10
2.8 14.8
5.6 21
8.4 26
14 33.3
28 40
56 47
70 50
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
a larger symbol
M the number of SAM-
Ca: be presented on a given
e
screen. The specification writer is
66X'd Wi 6:1 a co.f.prornise, which he must
Ide with the knowledge that the sclec-
don of a too-small symbol will impose
a 'penalty in legibility, and that this pen-
66' will haVe '66) be paid day after day
Ly the commander and his staff.
The display plotted in curve "B" of
Figure 5 and the Army map were then
? selected as representing the displays that
might be considered the critical pair in
NOMC fictitious command center, and their
curves were replotted in Figure 6. It is
obvious that their crossing point speci-
fies the illumination at which the critical
"
60
71,66
St( 73667 3
G
EFFLLT or tLLL/FI,NAT iON
PcNCEFIT1ON SPEED
D:LPL,S
4 166,773,1106 r t ,66066
:6; :Aumination on human
spuc,i when reading self-
.0 dus
hard copy and the critical display can
- read with equal ease.
should be stated again that the
speed scale is not the reading speed,
but merely a way of achieving compari-
son.
Fflter Factor
Symbol B ft-L
Alpha Minutes
Hood Factor .
Bockgd.
B
ft-candle ft-L
1 .063
2 0.136
4 0.272
3 0.542,
2 3.316
20 1.36
40 2.72
80 5.44
-.03 6.8
0.4
20
4
.5
Several things can be learned from
Figure 6. For one, it is possible that
either class of display can be improved.
The brightness of the cathode ray tube
may be increased or the suppression of
the ambient illumination improved; the
map may be printed on better paper,
or its contrast may be-increased by a
change of ink. The interesting point is
that any improvement will raise the
curve for that display, and if the value
of illumination is adjusted to suit, the
ease of reading of both displays will be
improved.
Second, it must be recognized that it
will sometimes not be possible, or per-
haps desirable, to use the value of illu-
mination that the computation calls for.
70
60
10
10
SELECTION OF
ILLUN INATION LEVEL
67 -.71101.1 "
100
Fig. 6 The selection of an optimum illumina-
tion for reading both hard copy and
self-luminous displays is accomplished
when the relative perception speeds are
equal.
On such an occasion the plot offers
numerical indication of the advantage
given one type of display and the pely
alty imposed on the other by the requir6d
level of illumination.
TABLE 3
READABILITY OF DISPLAYS
Phosphor Brightness, 50 ft-L; Reflectivity, 85%
And third, the small crosses marked
25 and 60 represent the crossing points
of two other sets of curves for the
same displays but based on human per-
formance curves in which the field fac-
tors were taken as 25 and 60. The three
crossings indicate the same value of illu-
mination to within the accuracy with
which the performance curves can be
read and their plot smoothed; it must be
concluded that the Value of the field
factor has no bearing on the computed
value of illumination sb long as the same
factor is used in the computations for
both self-luminous displays and hard
copy.
Suitability of the
Computed Illumination
Since the computed level of illumina-
tion is markedly lower than the general
recommendations of the Illmninating En-
gineering Society for offices, it is proper
to question whether or not it is suit-
able. This must be asked in two parts
-is it sufficient? and is it pleasing?
Many years ago a number of observ-
ers were given control of the illumina-
tion and were asked to find the value
that was best for reading the Saturday
Evening Post. The reported values fol-
low:6
Available value of illu-
mination, foot candles 10 30 45
hosen value 5 12 16
e very human tendency to choose
middle value is apparent, and it must
be remembered that the observers were
accustomed to a lower level of illumina-
0.4
20 20
2
.5
0.4
20
2
.25
0.75
37.5
2
.5
EL-
0.75
37.5
2
.25
Backgd. Backgd, Backgd. Backgd.
B B B B
C Speed ft-L C Speed ft-L C Speed* ft-L C Speed* ft.!. C Speed
294 .068 294 .034 0.239 156 0.119 315
147 0.136 147 .068 294 0.478 78.5 51 0.239 156
73.6 101 0.272 73.6 33 0.136 147 0.956 39.2 36 0.478 78.5 51
36.8 87 0.544 36.8 25 0.272 73.6 33 1.91 19.6 27 0.956 39.2 36
24.5 76 0.816 24.5 19.5 0.408 49 28 2.87 13 19 1.44 26 31
14.7 60 1.36 14.7 13.5 0.68 29.4 23 4.78 7.85 12 2.39 15.7 23
7.36 42 2.72 7.36 7.9 1.36 14.7 13.5 9.56 3.92 4.8 4.78 7.85 12
3.68 25 5.44 3.68 3.1 2.72 7.36 7.9 19.1 1.96 9.56 3.92 4.8
2.94 19.5 6.8 2.94 1.6 3.4 5.89 6.2 23.9 1.57 11.9 3.15 3.3
C.Jrve :n
A
for rn-::c.ning.
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
tlimi are we today. However, it is
..so apparent that there is no pressing
immail need for high levels when the
contrast is good. ?
Anti in their work on the utility of
colored illumnionts, Feree and Rand'
Itssfed tile eyes of their observers both
:,,:ort' and after reading for three hours
in Table 3 and plotted in Figure 5.
Tile difference. between curve "A"
our minutes of arc) and the other
carves (two minutes) is notable. One
is iempted to say that a stroke width
oi ;our minutes should always be used.
with a colored illumination of only 0.3
ft-candle. There was a measurable loss,
hr., the significant fact is that it was
possible to read for that time with an
metlesirable light and at such a low level.
Grie must conclude, then, that the corn-
'a: ted hovel of illumination is above the
inininfien by a factor of at least forty,
marl that it is in a range that has been
salected as oe. :num in tests where the
observer could control the level of illu-
mint, don.
.s clear from the above data that
the commuted level of illumination is
suifleient, but that does not assure that
it will be pleasing. The reason why it
nary be one avid not the other is that
the eve does not see illumination?it sees
rho iirig1rt?ess of the work and its sur-
roMidings that result from that illumina-
fiGn.
Tha sources of illumination should be
distriouted, but not evenly distributed.
Completely diffuse illumination creates
a somnolent atmosphere, while the in-
clusion of some concentration points
helps seeing through the formation of
-the shadows by which form is perceived.
And of course no light source whose
brig,ifiness is materially higher than that
of the walls should be 'within the field
of views of the operating people.
The ratio of brightnesses within the
field of view should be low. This effect
is most easily achieved by using the same
finish for similar objects and keeping the
illumination reasonably even. Such vari-
ation in brightness as exists should follow
as. closely as possible the rule that the
ceiling be brightest, the walls next bright,
the furniture and equipment next, and
the floors least bright, but not dark. The
...ES recommends that the reflectivities
be walls, 0.5; furniture and equipment,
? 0.35, and floors, 0.3. It should be noted
that hard copy, with a paper reflectivity
bear 0.7 and ink reflectivity near 0.1,
will have good contrast within itself with-
out either the paper or the ink being
markedly different from its surroundings.
Wall projection screens present a prob-
lem because rue symbol brightness is
and L is necessary to achieve the
contrast by keeping the back-
.ound dark. Hence, the screen must
ha cai-c.2-olly screened from all room
Declassified in
lights. However, the area immediately
around the screen must be as bright as
the rest of the wall. This surrounding
area can be illuminated by highly direc-
tive lights, it can be back lighted, it can
be set back of the screen and illuminated
by lights behind the screen, it can be
made fluorescent and excited by ?ultra-
violet light, etc. When done properly
the symbols will be brighter than the
walls and the background darker; the
presence of the -walls will make the
screen appear to have a greater contrast
than it actually has, and at the same
time will maintain the balance of bright-
ness with the rest of the room.
A clear -and public example of the
effectiveness of balanced brightness is
available in the two reading rooms of
the Congressional Library in Washington.
The old reading room in the main build-
ing is equipped with dark fdrniture. At
night the walls are relatively dark, and
the darkness of the ceiling is relieved
only by an illuminated painting. Lights
over the desks illuminate only the writ-
ing surface. In the day the walls are
much brighter, but the windows to the
south and the spots of sunlight to the
north make severe glare spots. Study in
this room is very tiring, eye strain set-
ting in after an hour or two. The reading
room in the Annex is illuminated with
artificial sources only, using desk lights
similar to those in the old room plus in-
direct illumination. The ceiling is the
brightest part of the room, the walls
are next bright, and the furniture is
relatively light. There are no glare spots.
One can study for hours in this room
without fatigue. The comparison is strik-
ing proof of the desirability of a low
ratio of brightnesses; it is even more
striking when one realizes that the illu-
mination on the writing surfaces in the
two rooms is the same..
It is clear, then, that the process out-
lined in this paper provides equal ease
of reading for both hard copy and dis-
plays, that the resultant illumination is
entirely adequate, and that it can be
made pleasing. There remains only the
. problem of those officers who normally
work in brightly lit offices and who must
move to the command center on the oc-
casion of an alert. Luckily the computed
illumination for the command center is
less than that to be expected in offices
by a factor of only three to six. It is
felt that a smooth transition can be made
possible by setting the illumination of
the intervening halls and anteroom at a
value intermediate between those of the
offices and the command center.
As in most system studies, this work
shows that much can be done by a num-
ber of small steps taken together. . It
must be noted that one of these steps,
that of providing a limited range of
brightnesses, has long been known but
has frequently been overlooked.
There is no easy way to compute the
effectiveness of hoods, means for direct-
ing the illumination, etc. One solution
is to make the display contractor respon-
sible for the entire installation and ,to
provide time and funds for either a
mockup or extensive tests in location. An
approach that would save the cost of
repeated mockups is for the Government
to make the tests, to provide a suitable
room, to specify the design of the hoods,
and to relieve the contractor from re-
sponsibility for the contrast.
Conclusions
a) The described process readily gives
the illumination level at which hard copy
and self-luminous displays are equally
readable, or gives a quantitative state-
ment of the advantage given one data
source over the other when a different il-
lumination level used.
b) The current art for cathode-ray dis-
plays permits an adequate level of illu-
mination if known techniques for main-
taining contrast are used. The same is
true for projection displays if the screen
area is reasonable
c) This level of illumination will be
pleasing if the color of the ceiling, walls,
furniture, equipment, and floors, and
the distribution of illumination gives .. a
small range of brightness with a proper
distribution of brightness.
d) The values for the contrast of hard
copy given herein should be re-examined.
If the listed values are low, then correct
values will permit a new selection of il-
lumination with which both the hard
-copy and the displays can be read more
easily. If the listed values are correct;
then similar values can be used in future
display specifications with a financial sav-
ing for the customer.0
REFERENCES
1. Blackwell, H. Richard "Use of Perform-
ance Data to Specify Quantity and Qual-
ity of Interior Illumination." /lluininating
Engineering June 1955, p. 286-299,
2. Blackwell, H. Richard and McCready,
Donald W. Jr. Foveal Contrast Thresh-
olds for Various Durations of Single
Pulses. Prepared June 1958 for BuShips,
U.S. Navy, under contract Nobs-72038.
Copies available from Institute for Re-
search in Vision, The Ohio State Univer-
sity Research Center, 1314 Kinnear Road,
Columbus 12, Ohio.
3. Blackwell, H. Richard "Development
and Use of a Quantitative Method for the
Specification of Interior Illumination Lev-
els on the Basis of Performance Data."
Illuminating Engineering June 1959, p.
317-332.
4. IRS Handbook, 3rd Edition, 1959.
5. "Recommendations for Quality and Quan-
tity of Illumination." Report No. 1, I.E.S.
Committee on Recommendations. Illu-
minating Engineering, August, 1058, p.
423.
0.. "Value of Higher Intensities of Illumina-
tion." Jour. Amer. Institute of Electrical
Engineering 41:499, July, 1922.
7. Force, C. E. and Rand, G. "Visibility of
Objects as Affected by Color and Corn-
posion of Light. Personnel factual
10:108-124 (July, 1931).
INFORMATION DISPLAY, SEPT/OCT. OtSzt
Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
instrument panel layout ?
instrument -priority and position
INSTRUMENT PRIORITY, POSITION, VIEWING ANGLE, .,ND -DISTANCE
. _ --- - .
?
(-:7--
th
Optimum Position: In general,. optirmlm_loca.tion,fornstruments ia diFely
' jaaLore the operator froal_aya_l.eyel to about 3e-b-alow eye level dt.T7
oi-Otrer'Vent'fe-wratora,-tTe-OPti'MUM1.-Otaticen-iS'JUSt benW e windshield.
The most important.and frequently'used instruments should be placed in this
,most favorable area. Instruments used for controlling direction (such as
aircraft heading indicators) are preferably located directly ahead of the
operator. If there are two or more such direction instruments they are pre-
ferably arranged along a vertical line. .
Viewing Angle: The most favorable angle for viewing instruments is perpendicular
to the dial faces. Extremely oblique viewing angles should be avoided by angling
the ends, bottom or top of large panels. Viewing angles up to 45? are considered
.X.s.aticfac_tory, provided needed information on the dial is not obscured by the
bezel, lighting shield, or other obstruction, and if some loss of reading pre-
cision due to parallax can be tolerated.
? Viewing Distance: Viewing distance for .instrument panels need be limited only
...if the operator is required to manipulate knobs 'or switches on the panel from
' his normal seated position. This distance is normally. fixed at 28. inches from
the eyes for vertical panels. The instrument size must be increased propor-
tionally.for longer viewing distances. .
? Horizontal and Vertical Separation: The difficulty of shifting between
instruments increases with separation. distance. Vertical eye movements are
more difficult than horizontal shifts in fixation. Distance between instruments
. should be minimized. Horizontal separations are preferred to vertical separa-
tions. (Fitts and Simon, 1952.)
-7-:12. ' LIA eft
WADC TR 54-160
?Declassified in Part - Sanitized Copy Approved for Release 2012/04/23: CIA-RDP78B04770A001900020031-7
e,4