0%)

2 1
0

'0 (JANAIR Report 680712

(¶I~~~ JOWlT ARkWY+LAVY

AIRCRAFT ISThUUENIATION RESEARCH ___
4 a

- VISUA•L RE UW6TE.2EITS STUVJY FO." 11ED-V? DISPLAYS

"- FINAL flPORT-PHASE I

Piepored by
T. Gold rýrm011
A. Hlyman ~~j
March 1970 Jll 12

t Jointly sponsoredl by
OFFICE OF NAVAL RESEARCH
NAVAL AIR SYSTEMS COMMAND
SU.S. ARMY ELECTRONICS COMMAND
Contract N00014-66-C-0114
NNR213-047
RepICd•-d by tho
CLEA RINGHOUSE
for Fo"dral Sc ,thhc & Tech,,cil

InFlormaihon Spr-agheld Va 2 151

This document haa been approved for public ielgase and sale; its ditiribution is unlimited.

SPErlRY GYP.ofltCE oIvIniDci

SPEHRY RAND CORPOPATiON
GREAT NECK, NEW YORK 11020

CS
____ ~ -791
 JANAIR Report 630712
6e

joINT AJAMY IkAVY AIFICRAR7 INSTRUMENTATION RESCAAICHi

"1S
L[ Lwiii.•2iTS SIOUY F0i1 16E,•,D P DISPLAYS
FINAL REP.T-PHAZE I

• Prepared by
T. Gold
A. Hyman

Mrch 1970

* Jointly sponsored by
OFFICE OF NAVAL RESEARCH
NAVAL AIR SYSTEMS COMMAND
U.S. ARMY ELECTRONICS COMMAND
Contract N00014-b6-C°0114
NR 213-047

This document has been approved for public release end sale; its distribution Is unlimited.

"PEpRY OYfCOC0PI rIVISION

SPEPAY RAND CORPORATION
GREAT NECK. NEW YORK 11020

0 -
I

, 0 0 00 0
 NOTICE

Change of Address

Organizations receiving JAN1AIR Reports on the initial distribution list

should confirm correct address. This list is located at the end of the report just prior
to the DDC Form 1473. Any change In address or distribution list should be conveyed
to the Office of Naval Research, Code 461, Washington, D.C. 20360, ATTN: JANAIR S
Chairman.

Disposition

* When this report i no longer needed it may be transmitted to other organi-

zations. Do not return it to the originator or the nionitoring office.

Disclaimer

The findings in this report are not to be construed as an official Department

of Defense or Military Department position unless so designated by other official
documents.

4 .: "

4
 FOREWORD

This report presents work performed under the Joint Army Navy Aircraft S
Instrumentation Research (JANAm) Program, a research and exploratory development
progra-n directed by the United States Navy, Office of Naval Research. Special guidance
is provided to the program for the Army Electronics Command, the Naval Air Systems
Command, and the Office of Naval Research through an organization known as the JANIAL'm
Working Group. The Working Group is currently composed of representatives from the
following offices:

o US Navy, Office of Naval Research, Aeronautics Programs, Code 461,

Washington, D. C.

- Aircraft Instrumentation and Control Program Area

* US Navy, Naval Air Systems Command, Washington, D. C.

- Avionics Division, Navigation Instrumentation ad Display Branch

(NAVAIR 5337)

- Crew Systems Division; Cockpit/Cabin Requirements and Standards

Branch (NAVAIR 5313)

* US Army, Army Electronics Command, Avionics Laboratory, Fort Monmouth,

New Jersey
- Instrumentation Technical Area (AMSEL-VL-I)

The Joint Army Navy Aircraft Instrumentation Research Program objective

- - is to cor ''ct applied research using analytical and experimental invwstigations for
- . identifying, defining, and validating advanced concepts which may be applied to future,
. improved naval and army aircraft h..strumentation systems. These systems include
sensirg elements, data processors, displays, controls, and man/machine interfaces for
fixed and rotary wing aircraft for all flight regimes.

Reproduction in whole or in part is permitted for any purpose of the United

States Government.
The Sperry Gyroscope Division publication number is SGD-4283-0333.

-. i

-w

, •• •• Q •
II.
V

S"SUMMARY

An experimental study was conducted to determIne critical values of the

I essential optical and electronic design parameters for head-up displays based on the
visual requirements of the pilots who will use these displays. This information Will
provide a basis for futre specifications covering this type of display, The investigation
covered two areas, size of exit pupil and binocular disparities due to optical distortions.

The exit pupil study waq conducted in a flight simulator in which a wide
field (25 degrees) head-up display with a large aerial exit pupil was installed. Four
pilots with military flight experience served as test subjects. The results Indicate that
the minimum size of exit pu'pil required it three inches in diameter, for wide-field
systems in which the pilot's head position is no more than 10 inches behind the exit pupil.
A laboratory telecen'xic viewing system was developed for the binocular

disparity study. 'Ms equipment made possible experiments in which disparate,

dynamic, head-up display Images presented t subjects who viewed these images
against a static view of a real world background. Three test subjects were employed in
: these tests. The ,esulcs show that binocular disparity tolerances are considerably lower
when head-up display images are viewed against a real world background, compared with
"1.7 a homogeneous background of moderate brightness. Maximum disparity levels are one
mil for sustained viewing with adequate comfort for vertical disparities and divergent
horizontal disparities. Convergent horizontal disparities may be as large as 2. 5 mils
for the same level of comfort.

1.
Ii

12
S

q •• •• • ... •*
 9 -.

TABLE OF CONTENTS

Section Page-
1
1INTRODUCTION
A. Background
3
B. Objectives
5
2 VISUAL REQUIREMENTS FOR HEAD-UP DISPLAYS
A. Requirements Related to Binocular Vision 5
*
B. Requirements Related to Head Movement 8
12
3 EXIT PUPIL STUDIES
12
A. Research Apparatus
B. Design and Conduct of Experiments 15 ,
*
27
C. Results and Implications
35
4 BINOCULAR DISPARITY STUDIES
35
A. Research Apparatus
B. Design and Conduct of Experiments 47
S
62
C. Results and Implications
66
5 CONCLUSIONS
67
* 6 RECOMMENDATION

REFERENCES
68
* " 7

"- 9i
Siv

0 0 0 0
0
*
 I S

IF~

i I.IST OF ILLIUSTRATIONS

* Figure Pag._.e
*I 1 Schematic Optical Arrangement of Head-Up D splay 2

2 Eliect of Horizontal Binocular DWsparity 6

J 3 Envelope of Viewlng Positions for Instantaneous Full Field 9
4 Effect of Smaller It-. on Viewing Position Envelope
for Full Field 9
5 Effects of Longitudinal Eye Posi:Jons Outside of
Envelope for Full Field 9
6 Effects of Off-Axis Eye Positions on Field of View 9
* 7 Binocular Field of View Based on Dual Overlapping
Monocular Fields 11
8 B-47 F•lght Simulator System 13
* 9 Wide Field Optica! Projection System Installed 10 B-47
Flight Simulator 14
10 Matrix of Positions for Target Image in Head-Up Display 16
11 Block Diagram of Switching Arra;,gement for Exit Pupil Tests 16
12 Experimenter's Control Station 17
13 Meatn Reaction Time for Pilot A an a Function of Exit Pupil
* Size and Head Position 23
14 Mean Reaction Time for Pilot B as a Function of Exit Pupil
Size and Head Position 24
15 Mean Reaction Time for Pilot C as a Function of Exit Pupil
Size and Head Position 25
* 16 Mean Reaction Time for Pilot D as a Function of Exit Pupil
Size and Head Position 26
17 Mean Reaction Time for All Pilots as a Function of Exit
Pupil Size and Head Position 28

"* [
12V
I:•

* 0 0 0 * 0 0 0 0 0
 I
4 S

LIST OF ILLUSTRAT!ONS (Cont'd)

18 Cross-Plot for Mean Reaction Time for All Pilots 29

19 95-Percent Confidence Limits for Mean Reaction Times 30
20 Optical Schematic of Telecontric Viewing System 36
21 Display Imagery and Fields ot View 38
22 Re. Quarter V.,ew of Telccentric Viewing Assembly 40
23 Side View o! Telecentric Viewing Assembly 41
24 Functional Block Diagrinam of Telecentric Viewing System 43
25 Block Diagram of Image Generating System 44
26 Geometry of Radial Disparity Generation 46
27 Subject Seated in Viewing Compartment 53
28 Comfort Level as a Function of Binocular Disparity for
Typical Test Session 56

i, S
_o0

V1

6 0 0 0 0 0 0 0 -4
 ii

[I

1- LIST OF TAbLES

Tab i Page

*I Number of Test Sessions Accomplished ith each

of the Pilots 18
2 Mean Reaction Times for Pilot A 19
3 Mean Reaction Times for Pilot B 20
4 Mean Reaction Times for Pilot C 21
5 Mean Reaction Times for Pilot D 22
6 Grand Mean Reaction Times for all Pilots 2"
7 Summary of Bnozular Lisparity Test Conditions 49
8 Binocular Disparity Levels Used in Test Schedule 52
9 Distribution of Responses for Typical Test Session 55
-0 S-wm-'.y ol Binoc..ular Disparity Levels for Two
Comfort Indices 58
11 Summary of Eir.ocular Disparity Levels for all Subjects
:or Reference Test Condition 62

I. 1..

ILvii
L2

- *--- -.- --- *----•---..

•.-•-
 I

SECTION 1

INTRODUCTION

A. BACK'GROITND

The head-up display is a relatively new cockpit display technique that

provides flight situation and cortrol iuorrmation to the pilot as he looks through the
windshield. One w~y of doing this is shown in figure 1. Images on the face o1 a cathode
ray tube (CRT) pass through a collimating optical system. The light ray.; are then
reflected by a combining glasc to the pilot's eye as he views the real world through
the windshield. The pilot has the visual impression that the images generated by the
CRT are located in the real world, in front of the aircraft. The images are called •

virtual images because the light rays that produce them do not pass through their
apparent locations. The hiead.-up display is an extension of thl- retfecting gun Fightt
systems, used extensively In World War I, applied to flight control problems other
than airborne fire control

Several unique advantages for flight control result with head-up displays.
Information is presented to the pilot in a form th'at requires almost no diversion from
his view of the real world from the cockpit. The pilot is not required to shift his
visual attention from the real world to the Instrument panel during critical maneuvers.
The associated changes in visua focusing (accommodation and convergcnce) from large

distance, to !he near cockpft panel are also eliminated.

o1

4 5

.1J•••• D®•• •
 II

WINDSHIELD

$ - COMBINING
GLASS
VIRTUAL IMAGE
PILO1 EYE

COLLIMATING -
SYSTEM

CRT

FIGURE 1. SCHEMATIC OPTICAL ARRANGEMENT OF HEAD-UP DISPLAY

These advantages are common to all head-up displays using optical pro-
ýection techniques. However, more sophlisticated head-up displays, such as those
.. i...ntly bel.. dv..t.-pee fno.r th Nay.all Air Systems Command bv Sperry, include
images closely related to figures ir. the real world. These projectcd Inages are de-
signed to overlay the real figures visually. Some examples of these figures are a
runway, the landing deck ort a carrier, and a ground target. Head-up displays with 4
this overlay capability provide fully compatible flight controi means for both visual
(VFR) and instrument (IFR) flight (reference l)'. For example, during landing the
pilot performs th'. approach to the real carrier deck in VFR, or in the same way to
the image representing the deck under IFR before vieual breakout. These capabilities
have led to the development of head-up display systems for general flight maneuvers,
terrain-following at high speed, field and carrier landings, and weapon delivery.

f*
*For this and other references, refer to page 68.

122

0 0 • 0 9 9 *
 4

Optical projection sys~enis with large fields of view, and with qualities 4.
which permit considerable freedo. i of head movement for the pilot, are desilable.
High accuracy in positioning the limagoo, both ill relation to each otier and to the real
world, is also required. Ideally, this precision should be independent of the pilot's p
head position, and the same Iimage should be presented to each eye if the head-up
display is viewed binocularly.

Unfortunately, the visual requirements for projection systems based on

existing literature in vision have led to large, complex, heavy, costly optics. Perhaps
these visual requirements may be rela.ed. Hiowever, to date there has been no
scientific baris for doing bo.
In fact, there is evidence that some head-up displays have manifested
undesirable visual characteristi.s. In a flight evaluation of a head-up display in an
A-5A aircraft at the Naval Air Test Center (reference 2) the findings Indicated that it
was necessary for pilots to relocus their eyes when shifting their view from external
objects to the HUD symbols. In a flight program conducted for the Federai Aviation
jigency by Sperry (reference 1) oi.x pilot reported, "Optics problem of double image
at edge of fReld. Feel it made me tire visually". These reports affirm that visual •
design criteria for head-up displays must be carefully generated.

The expi'rimenta. eturies dpe9_rrt2ied herein nrovide' rL;lts which ohigepttl~vl

define the visual requirements i elating to the design of wide tield (25 degrees), virtual
image, optical projection systems. These studies were conducted with highly special-
zed viewing apparatus demanded by the research. The results permit the design of the
optical systems to be matched to the visual characteristics of the human pilot.

B. OBJECTIVES
The objectives of this study were to determine the essential optical and
electronic design parameters for head-up displays, based on the visual requirements
of the pilots who will use these displays.

3
"7 I

• • • •• • o
 The key parameters investigated were:
I Minimum permissible size of exit pupil, as a function of pilot head

position

9 Permissible binocular dliparitles produced by optical distortions,

including the effect of:

- Image brightness

- Image motion

- Real world background

- Image line thickness

- Overlapping monocular fields.

I4
[I
 SECTION 2

VISUAL REQUIREMENTS FOR HEAD-UP DISPLAYS

A. REQUIREMENTS RELATED TO BINOCULAR VISION

A well-designed optical projection system should provide a large exit pupil

to permit some freedom of head movement for the pilot. When this exit pupil is larger
than the interpupillary distance of the eyes, viowing becomes binocular, and image
distortion presents a major problem. With monocular viewing, distortion would
cause an overlay error in those cases where display symbology must superimpose or
real world elements, and tolerance for distortion at various eye positions with respect
to the exit pupil would have been determined by the permissible overlay error. When
viewing is binocular, differences in image distortion for the two eyes becomes a
critical matter. Such differences can cause eyestrain, display image doubling, and
an apparent depth displacement of disulay elements.

In the normal visual environment, the perception of the depth can arise
from a mimber of monocular and binocular cues (references 3 and 4). With a virtual
image optical projection display viewed binocularly, however, only differences in
image uistortion for the two eyes which result in retinal disparities need to be
considered. (Images from display points that do not fall on corresponding points on
the two retinas are said to be disparate, ) Perceptually, points slaving horizontal
disparate separation in the two eyes are interpreted as being localized in different
frontal planes.

S]5
4- h

-m

" "" I

. .t
4 17
() rigure 2 illustrates how a horizontal disparity occurs for the two eyes when
Stwo objects are displaced in depth in the real world. The upper part of the figure, A,
- shows a plan view of two vertical rods, F and N, separated in depth and viewed simul-
taneously by the left eye and the right eye. Rod F is fixated and hence imaged on the
foveae (f' and f") of the two eyes. This causes the image of rod N to fall left of the
4 fovea in the left eye and right of the fovea in the right eye. The resultant monocular
and binocular perceptions are shown in the lower part of figure 2. The view shown in
B illustrates what is perceived if the right eye Is occluded; D, if the left eye is occluded;
and C. if both eyes view the display simultaneously. In the last instauce, the rods are
perceived as separated in depth as well as laterally. This is a cane of steroscopic
4 depth perception, i. e., perception of depth based on retinal disparity.

F
A

* 4

LEFT EYE IGHT EYE

4 I'" 1v

UýFT 801~H C RIGHT D

F N N F N F

FIGURE 2. EFFEf:T OF HORIZONTAL BINOCULAR DISPARITY

b 11
• [_"
 * S

Steroscopic depth perception does not require real stimuli separated in

depth. Any device which would present view B to the left eye and view D to the right
eye would be interpreted as the situation represented in A. This is the principle
employed in the stereoscope where two disparate pictures (I. e., stereograms) are
* viewed, one by each eye (reference 3). It is to be noted that the magnitude of the
disparity, L, e., the difference between the angles subtended by the two eyes at N and
at F, is the same whether N is fixated, or some other point not shown is fixated.

The binocular perception of a single rod at N, does depend, hcwever, upon

the location of the fixation point. If the latter is much beyond or much in front of rod N
this rod will be seen as double (binocular diplopla) rather than sir.gle (visually fused).
The fore and aft region of single vision about a fixation noint (i.e., Panum's area) is a
function of the lateral angular separation between target and fiYation point (references 5,
6, and 7). The greater the separat'on, the greater is the permissible i-etinal disparity
* for single vision.

Horizontal retinal disparities occur naturally when real world objects

separated in depth are viewed by the two eyes. Optical systems, however, can
generate disparities in which a vertical component is also present. How are these
* * perceived? By a perceptual mechanism not fully understood, vertical disparities
causing retinal image size differences can be converted into equivalent horizontal
disparities (induced size effect, reference 5), and subjectively interpreted as depth
displacements. Thus retinal disparities whether horizontal or vertical, when small,
can give rise to perceptions of depth displacement, and when large, to image doubling.
By introducing both vertical and horizontal disparities, it is possible to either enhance
or to decrease the depth effect. What occurs depends on the direction and magnitude
of these disparities.

The vinral image optical projection display presents conditions comparable

* to the situations previously described. U, as in figure 2, F is a real world object and
N is a display system symbol which be. ause of differences in image distortion is seen
as in B by the left eye and as in D by th,. right eye, then F and N will appear to be
separated in depth. If F were a, optical infinity and N had a retinal disparity of 5 mils,
for an observer with an interpupillary separation of 2. 5 inches N would be perceptually
located about 42 feet forward of the observer. In this case, would a pilot have a vision

9 00
 4 p0

lr

4
Sproblem under conditions of prolonged viewing when 14 must be considered as an
overlay In the plane of F? On the basis of the classical literature (reference 5), one

would not expect this simple configuration to generate meaningful visual discomfort or
I - a perceptual problem. However, what happens if the display figure is not a simple rod
but a more complex figare? While such a situation may generate a perceptual conflict
Sto an observer, it is not certain that it wiU create a vision problem, for his task is not
to localize accurately the display in space but merely to superimpose it on the real
world. However, studies concerned with binocular disparity have not to date con-
sidered this problem.

Applied experimental studies specifically directed to help establish per-

missible disparities in head-up display configurations are required to resolve questions.
These studies must determine the acceptability of various binocular disparities typical
of virtual image optical projection systems. Also, the displays employed in thebe
studies must include complex configurations characteristic of the symbology developed
I for projection displays generating dynamic virtual images.

B. REQUIEIEMENTS RELATED TO HEAD MOVEMENT

In every optical system there is a diLaphragm, lens rim, or other element

which acts to limit the extent of an object, or the field of view, that will be represented
in thc image. This element is known. as the "el, sp. Thorp is also another phyRical
stop, called an aperture stop, which determines the size of the bundle of rays reaching
any given point in the image. These stops esttablished the region in which an eye will
see an I ,age and the field angle in which this image can occur. Detailed discussions
I of the theory of stops are given in referenceb 8, 9, and 10.

In analyzing the behavior of an optical system, it is not necessary to deal

directly with the stops and the associated optical components. One can study system
characteristics in terms of image piane locus, exit port, and exit pupil. (The exit
of the field stop and the aperture
-. port and exit pupil are, respectively, the images
stop in image space.) Figures 3, 4, 5, and 6 represent some characteristic situations
[ " that can occur with wide field, virtual image, optical protection systems. The rule to
be followed is that only rays which are unobstructed by the exit port and pass through
j. the exit pupil are physically realizable. Since exit ports and exit pupils are "imaginary"

a,)ertures, the eye can be located in front of them as well as behind them. The eye,
however, must be located beyond the last optical component of the system.

4 8
[4

* I9 00
 FIGURE 3. ENVELOPE OF VIEW)ING
POSITIONS FOR INSTANTANEOUS FULL FIELD
c
A IDD
Ep

FIGURE 4. EFFECT OF SMALLER STOP ON

VIEWING POSITION ENVELOPE FOR FULL FIELD

* ~A
~~--_
N IF

FIGURE 5. EFFECTS OF LONGITUDINAL EYE

POSITIONS OUTSIDE OF ENVELOPE FOR FULL FIELD

REGION
I ý

REGIONr

FIGURE 8. EFFECTS OF OFF-AXIS EYE POSITIONS

ON FIELD OF VIEW

9S
 r
Figure 3 illustrates the case where the image AB and the exit port C are

F- in the same plane. The exit pupil is located at EP. The cross-hatched area defines
the region where rays from all object points are present. If both eyes of the observer
4 are within this region, he will obtain a binocular view of the total field. It is to be
noted that the presence of an additional stop will have no efiect on the display if the
Image of that stop has an aperture greater than the extreme rays drawn in the figure.
Such a case is represented by aperture D. V the diameter of D were smaller, D would
have become the exit port for these eye positions where the field of view was deter-
4 mined by D. Figures -4, 5, and 6 illustrate what occurs when D acts as the exit port.

The crosv-hatched area in figure 4 again represents the region where the
full field of view can be seen. A comparison with figure 3 shows this region to have
become smaller. Also, its largest cross-section has moved closer to the Irr.age. If
4 this cross-section is smaller than the interpupillary distance for the observer's eyes,
binocular viewing of the total field is not possible for any head position. The effect of
forward and aft movement of the head beyond the cross-hatched region, when tMe
viewing eye is on axis, is shown in figure 5. In either case the instantaneous field of
4 "view, as represented by the extreme rays possible at the given eye position, is less
than the total possible field of view. However, hidden portions of the field can be
scanned by transverse (off-axis) movement of the head.

The instantaneous fields obtained at several off-axis eye positions are

shown in figure 6. When the eye is at N, region p is seen, a:nd when at F, region r
I is seen. At 0, of course, the total field is seen.

The net effect with off-axis eye positions which occur with binocular viewing
and a field stop that provides an instantaneous field less than the total available field
uf view, is shown In figure 7. R is the portion of the total Instantaneous field seen
4 only by the right eye, L is the portion seen orily by the left eye, and B is the portion
seen binocularly. In the parts of the field seen by one eye only, there would be
"wash-out" of projected imag( ry should retinal rivalry occur (references 3, 4, and 7).

4 10

U ~10
F

4
I
 R.

II

FIGURE 7. BINOCULAR FIELD OF VIEW BASED ON DUAL

OVERLAPPING MONOCULAR FIELDS

*
Must the optical designs, therefore, reduce these monocularly viewed
* regions to a minimum? It should be remembered that binocularly viewed rcgions
can present undesirable retinal disparities, a phenomenon which is not possible in
monocularly viewed regions. Also, systems which have large binocular fields ,"re
* complex, heav.y, and costly. To help establish a trade-off, one must determine the
magnitude of retinal disparity permissible at the boundary between the monocular and
binocular regions.

- °

* p

* *0 *0
 3

SF fON 3S

EXIT PUPIL STUDIES

A. RESEARCH APPARATUS
The exit pupil studies were conducted in a B-47 Flight Simulator which was
* developed from a model S6A Flight Trziner for the B-47 aircraft (figure 8). A head-up
-- display with a wide field of v-ew (25 degrees) and a large aerial exit pupil (5 inches wide
by 3 inches high) were used as the optical display medium. The configuration of this
display system is shown in figure 9.

* V The exit pupil in the display could be reduced by inserting aperture stops
S. in the optical system. This permitted the study to encon.pass smaller exit pupils. 0
. I' aýddUor to the full-Sh-b r- inch size, circular exit pupils of 3-, 2-, 1-, and 1/2-inch
diameters were used for a total of five exit pupil sizes. The location of the pilot's
Seye could be adjusted relative to the exit pupil by fore and aft change in seat position
in the cockpit. The study covered three head positions for the pilot, for which his eye!s
I "were in the plane of the exit pupiL, five inches aft of the exil pupil, and ten inches aft
of the exit pupil.

0 The exit pupil studies were designed to measure pilot's reaction time or
latency in locating an image in the field of view of the head-up display, starting with a
head position laterally displaced from the limits of the exit pupil.

The laten(y criterion was selected because normal operations in the cockpit
frequently require body motions by the pilot which remove his eyes from the exit pupil.
* It is operationally necessary that the pilot recover the exit pupti rapidly.

* ,12

0
 4 1 6"
4 i
9D

41g • " "••" c:"

S€~~~. ...

• \ - ,':
- . . . .. •,' \ \!

S "; :1'

:\ 1 "
J>
Ig

1 1
'1 I

S@ @ • • • •
 I- i

i I

WIDE-FIELD OPTICAL PROJECTION SYSTEM

INSTALLED IN CCCKPIT OF B-47 SIMULATOR

r--,• a,,m,,.• ,-.-.-• . .-.- i .. ,-.,

\ I

FIUR WD
. IL PIA RJCINSSE INTLE
i e

* IfAR VIE.W OF WIDE-FIELD OPTICAL PAQJECTION~ SYSTEM

IN B-47 SIMULATOR

S~FIGURE 9. WIDE FIELD OPTICAL PROJECTION SYSTEM INSTALLED

l IN B-47 FLIGHT SIMULATOR

1. 14
ii.
I

I
 Q91

The pilots normal loading Wask was to maintain prescribed heading and
altitude using a simple head-up display ard panel Instruments as required. The
experimenter was to remove the head-up display images at some point in time to
simulate a display failure; the pilot is advised of this situation by a warning light.
The pilot then must reach to his low left to throw a switch that returns the display
and also energizes an image (circle) in one of sixteen possible positions in the field
of view (figure 10). During the physical motion required to actuate the switch the
pilots eyes are removed from the exit pupil. At the same time, a timer is started to
measure latency. When the pilot locates the display image he presses a switch button
on his control wheel, which removes the circle image and stops the timer. The pilot
advises the experimenter regarding the position of the image in the 4 by 4 matrix,
e.g., A2, to assure that there has been no guessing. The timer ia stopped by the
pilot when he located the Image in the display, so the time contumed by the pilot in p
responding verbally to the exper imenter is not included In the measured reaction time.
Furthermore, the absence of the circle does not permit the pilot to reassess the
situation after the timer has been btopped.

The switching scheme to permit the routine dcscribed is present - in the O

block diagram in figure 11. The experimenter controls the presence and loc- ion of the
test symbol in the display, and turns the normal head-up display symbology off at the
desired time. The test subject turns the test symbol on and starts the timer. He also
simultaneously turns the test symbol off and stops the timer when he locates the symbol
in the display. A photograph of the experimenter's control station Is shown in figure 12.

B. DESIGN AND CONDUCT Or EXPERIMENTS

1. Test Procedures

Four p'lots with military flight experience served as test subjects in the
exit pupil studies. One of the pilots was currently in professional service in executive
air transport operations. The visual capabilities of the subjects were not critical for
the conduct of the exit pupil studies.

The pilots were given a few hours in the simulator to familiarize them with p
the cockpit layout, the head-up display, and the handling qualities of the B-47 aircraft
before the start of any formal tests.

- p
15

-o

4 p

0 0 0 0 0 0 0 0
 F0 A a

0
c 0

0 0'
F0 0 0 02
CENTER OFFIL

0 0 0 0 3

0 0 0 0

FIGUR~E 10. MATRIX OF POSITIONS FOR TAnGET MIAGE INJ HEAD-UP DISPLAY I * 4

1i [ " SELE.CTION HEMAD-

UP
I

ONO
OFF-lCONTROL
• • FIGUR~E 1 0.L CKDA GRAM
OFP S WIT CIO
N GRAO RTANGE MENT
AGE I?1
HE XITUP I TES TSA

S/"N OTIMER
16
S

• @@
• • TEST

0 0 0 0 .
* 4I
 p

FIGURE 12. EXPERIMENTER'S CONTROL STATION S

Each pilot was tested in each of the three head

.//"/ positions, using five exit k,•,,S
pupil szsaechposition. An experimnental session was con"'ucted with a single head
position. Exit pupils were changed after every six presentations (trials) for which
*reaction times were0 measured.
The
00 location of the test symbol in* the display field
09 0"
was varied randomly from trial to trial. The first of each set of six trials was con-
sidered a practice run, and the results were not Liciuded In the analyses of the data.

Each set of tests was replicated four ti.-nes in a session. Therefore, a total
of 100 data points were obtained in each experimental session, consisting of 5 exit pupil
sizes x 5 points per exit pupil x 4 replications per exit pupil. Each pilot participated in
at least three sessions, covering the three head positions studied. Additional replicated
sessions were accomplished with some of the pilots as permitted by their availability.
The number of sessions obtained with each pilot for each head position is summarized
in table 1.

17 0
 TABIE 1. NUMBER OF TEST SESSIONS ACCOMPLISHED
F WITH EACH OF THE PILOTS

4I Ele Position
5 in Behind 10 in Behind
Pilut At Exit Pupil Exit Pupil Exit Pupil

A 2 2 2

B 1

C 2 1 1

V) 2 1 1

The pthots were directed to maintala ,pecified altitudes and headi-gs as

a normal cockpit work load. They were requested to make changes in altitude of
500 feet, or 30-decgree changes in heading, after each set of six presentationf for a
particular exit pupil. The experimenter changed the bize of the pupil while the pilot
was occuped with this flight maneuver.

2. Data Reduction

Mean reaction times were determined for each set of five data trials in all
of the experiment~d sessions. Therefore, four mean reaction times were obtained for
every exit pupil size used in each session. These dat2 are given in tables 2 through 5
for each of the four pilots. Each table includes the results obtained for each of the
three head positions. The number of data points in each table is a function of the
number of test sessions obtained with each pilot (table 1). G,-wid mean reaction times
for all combinations of exit pupil size and head position are also shown in tables 2
through 5.

The grand mean reaction times for each pilot were plotted as a function of
exit papil size, for each of the three head positions in figures 13 through 16. The re-
sults axe a consistent family of J-curves, with increasing reaction time as the exit
pupil is reduced. The curves are also displaced vertically to increasing reaction time
as the pilot's head position ib displaced aft of the exit pupil.

18
aS)
TABLE 2. MEAN REACTION T- TES* FOR PILOT A

Size of Exit Pupil

Eye Position Session Full 3 in. 2 in. I in. 1/2 in.

At exlt pupil 1 2.12 1.64 2.68 2.72 3.64

2.38 1.76 2.16 2.42 2.60
1.80 1.87 2.09 2.30 2.44
1.52 1.60 2.25 2.09 2.53

2 1.69 2.27 1.94 2.58 2.27

1.88 1.61 2.16 2.17 2.84
1.77 1.59 1.99 1.81 2.62
1.38 1.34 1.65 2.29 2A13

Mean 1.82 1.71 2.10 2.30 2.76

5 inches 1 1.U6 1.86 2.60 2.48 5.07

behind exit 2.00 2.09 2.41 5.29 8.27
pupil
4 1.97 2.18 2.18 3.29 3.56
1.42 1.63 2.28 2.83 5.74

2 1.63 1.94 2.65 6.36 5.92

1.69 2.18 2.13 2.27 2.85
1.33 2.36 2.61 4.20 5.26
1.80 1.78 1,77 2.68 5.09 S
Mean 1.72 2.00 2.33 3.61 5.14

10 inches 1 1.35 7.21 3.45 13.27 26.88

behind exit 1.71 2.31 2.10 8.56 5.62
pupil
1.40 1.59 2.57 3.56 12.12
1.55 1.50 2.49 4.63 14.47

2 2.14 1.67 4.05 4.92 13.36

2.19 2.21 3.59 8.40 16.12
1.85 1.74 3.03 4.75 8.03
1.71 2.20 3.47 4.29 6.20
* Mean 1.74 2.55 3.10 6.32 11.91

"*Reactiontimes shown are in seconds.

19
®I

4
TABLE 3. MEAN REACTION TIMES* FOR PILOT B

Size of Exit Pupil

"Eye Position Session Full 3 in. 2 In. I In. 1/2 In.

At exit pupil 1 1.40 1.33 1.71 2.55 3.06

1 1.23 0.98 1.25 1.65 3.78

1.32 1.29 1.31 1.54 1.93

1.25 0.92 1.14 1.20 3.38

Mean 1.30 1.13 1.35 1.74 3.04

5 inches 1 1.27 0.93 1.22 1.78 2.41

behind exit
4 pupil 1.32 1.24 1.26 3.26 4.78

0.98 1.02 1.39 1.h7 2.88

1.08 1.13 0.96 1.96 1.69

SMean 1.16 1.09 1.21 2.14 2.94

10 Inches 1 1.22 1.03 1.73 1.39 8.80

behir.d exit
pupil 0.84 1. 11 1.73 2.18 4.41

1.02 1.07 2.09 3.41 2.06

a
1.38 1.36 1.59 3.92 3.05

Mean 1.12 1.15 1.79 2.97 4.73

*Reactilon times shown are In seconds.

4 1-
I-
20

IE
 TABLE 4. MEAN REACTION TIMtrS* FMR PILOT C

Size of Exit Pupil

0 Eye Position Session Full 3 in. 2 in. I in,. 1/2 In.

At exit pupil 1 0.97 0.82 0.88 1.20 1.48

0.85 0.87 0.87 1.98 1.38

0.93 0.89 0.96 0.97 1.09
0.93 0.98 1.04 1.62 1.42

2 0.72 0.69 0.76 0.82 0.96

0.70 0.80 0.67 0.94 1.17
0.82 0.78 0.82 0.93 0.87 S
0.79 0.81 0.94 0.98 1.17

Mean 0.84 0.83 0.87 1.18 1.19

* 6
5 inches 1 0.72 0.88 1. 49 2,73 4.38
behind exit
pupil 0.93 0.85 1.03 1.90 4.91
0.95 1.00 1.50 2.71 1.85
* 1.13 0.94 1.24 2.23 4.75

Mean 0.93 0.92 1.31 2.39 3.97

10 inches 1 1.12 1.08 1.67 3.85 5.50

behind exit
* pupil 0.95 0.94 2.04 4.20 5.78
0.89 0.96 1.34 3.29 6.73
0.86 1.12 0.99 3.74 9.91

* Mean 0.96 1.03 1.51 3.77 6.99

*Reaction times shown are in seconds.

1
21
-T

S
 ' _

mF i
TABLE 5. MEAN REACTION ThIMES* FOR PILOT D

Size of Exit Pupil

I Eye Position Session Full 3 in. 2 in. 1 in. 1/2 in.

At exit pupil 1 1.49 1.45 1.78 2.02 4.25

1.72 1.90 1.56 3.54 3.68

1.36 1.62 1.69 2.15 3.66

1.17 1.37 1,43 1.51 3.19

I2 0.92 0.81 0.99 1.44 2.40

¶ 0.81 1.11 0.99 1.22 1.84

I 1.12 1.22 1.01 1.04 1.25

0.84 1.30 0.76 1.67 2.82

Mean 1.18 1.35 1.28 1.82 21 88

05 inches 1 1.81 1.75 2.60 3.24 6.72 .

pupil 1.74 1.34 2.54 4.11 4.08

1.54 1.98 2.23 2.59 2.52

1.61 1.60 1.57 3.54 4.51

Mean 1.68 1.66 2.24 3.37 4.34

10 inches 1 -4 1.52 6.31 6.10 6.36

behind exit
pupil 1.4U 1.79 4.15 8.67 4.87
4
1.88 1.27 2.80 5.81 9.82

1.13 1.72 5.18 3.38 11.86

Mean 1.49 1.58 4.61 5.99 7.71

4

Reaction times shown axe in aecondb.

L2 22

4 5
 tHCHES BHIND PUPIL

6 INCHES BHIN•O PUPIL

/ AT EXIT PUPIL 5

I-I

• " II

0OS I 3 FUL.L,(3x 51
SIZE OF EXIT PUPIL (INCHES)

"FIGURE 13. MEAN REACTION TIME FOR PILOT A AS A FUNCTION

S. ,OF EXIT PUPIL SIZE AND HEAD POSITION

-,

"" 23
4 *- 6- .-

I. I

77

'LIl aw

S4
"J "10 nCHES BEHiNo PUPIL

8 INCHES BEHIND PUPIL

£AT EXIT PUPIL. I

- L

' I
I|

0.8 1 ' ULL (3X 51

,

SIZE OF EXIT PUPIL (INCHES)

FIGURE 14. MEAN REACTICN TIME FOR PILOT B AS A FUNCTION

I. OF EXIT PUPIL SIZE AND HEAD POSITION

I" 24
 * S•

• -- 5i
II

* -'
\ -I4 U
10 INCHES BEHIND PUPIL

5 INCHES BEHIND PUPIL

AT EXIT PUPIL

I --

SO.Sa i FULL. (3 x 5)
. BI.laE OF EXIT PtaUL. (INCHES)

* FIGURE 15. Mr.AN REACTION TIME FOR PILOT C AS A FUNCTION

OF Exrr PUPIL SIZE AND HEAD POSITION

j- 25

.1

*
 ,ii i?

I " I
10 INCHES BEHIND PUPIL

5 INC•ES BEHIND OUPIL

i AT EXIT PUPIL

5/

"FGR 16 MEA RECINTM•O

. IO SAFNTO

I 1
SIEOF
VaxT PUPIL
SIE INDHEADPS) rO

12 26
 I'cd
The results of the tests for each pilot indicate consistent behavior patterna
aE, exit pupil size and head position are varied. Therefore, the data for all four pilots
w re combined; the results are presented in table 6 and plotted in figure 17. The corn-
bined results parallel those obtained for the individual pilots, with smoother variation
with exit pupil bizes and head positions as shown in figure 17.

TABLE 6. GRAND MEAN REACTION TIMES* FOR ALL PILOTS

Size of Exit Pupil

Eye Position Full 3 in. 2 in. 1 in. 1/2 in.

At exit pupil 1.28 1.26 1.40 1.76 2.47

5 inches behind exit pupil 1.37 1.42 1.77 2.88 4.10

10 inches behind exit pupil 1.33 1. 5e 2.75 4.76 7.85

*Reaction times shown are in seconds.

C. RESULTS AND IMPLICATIONS

The results of the exit pupil studies are best summarized in the plots in
figures 17 and 18. The mean time to locate an image in the head-up display (reaction
time) is plotted as a function of size of exit pupil, for three longitudinal head positions,
in figure 17. Figure 19 is a replot of flg're 17, with the addition of the 95-percent con-
fidence limits for the mean values of reaction time for each of the three head positions
presented. The statistical significance of the differences in mean reaction time for
the various head positions, for all but the largest exit pupils, are demonstrated by
these data. A cross-plot of these data, showing reaction time as a function of head
position for the five exit pupil sizes studies, is presented in figure 18.

Figure 18 indicates that for large exit pupils, 3-inch diameter and larger,
there are only small decrements in performance, i.e., higher reaction times, with
a shift in head position aft of the exit pupil.
The smaller exit pupils lead to increases
in reaction time, which become more marked as head position moves aft. The rates of
cha;nge of reaction time with aft movement (,I head position increases monotonically as
the size of the exit pupil is reduced (figure 18).

27

-4
]-
70
7 •- ___
10 INCHtL SCEIIND PUPIL

5 INCHES SOE14IND

- /AT EXIT PUPIL

I / '

228

1.

Ii SIZE OF XI•XT PUPIL (INCHCS$)

SFIGU.RE 17. MEAN REACTION TIME FOR ALL PILOTS AS A FUNCTIONb

I.OF EXIT PUPIL SIZE AND HEAD PO).%-I. )N

"L 28

1S

I
 S

SIZE OF EXIT PUPIL

I/Z INCH

SI9 INCH

S I

01

0 10

=' -L ( INCH x a INCH)

LONOITUDINAL HEAD POSITION

(INCHES$ AFT OF EXIT PUPIL)

5 INCH
FIGURE 18. CR1OSS-PL.OT FOR MEAN REACTIOD'" TIME FOlR ALL PILOTS

29
 I)

I: •
r7

I.
HEAD POSITION AT EXIT PUPIL

f; I
S _

COFDEC LIMIT
Ya
i i. o,,,. <.T

I
4U4--l

,*S%• .OFI.N...MT

II

0-5 I FULL (3 X 6)
SIZE OF EXIT PUPIL (NCHES)

TIMES
FIGURE 19A. 95-PERCENT CONFIDENCE LIMITS FOR MEAN REACTION

IL- 30
 HEAD OSITION . INCHES BEHIND EXIT PUPIL

In-

9s% CONFIDINCE LIMITS S

08~ $IULL F (SX 51

SIZE[ OF E[XIT PUPIL (INCHES)

FIGUR•E 19B. 95-PERCENT CONFIDENCE LIMITS FOR MEAN REAcTION TIMES

31-

AI m , S
Ii
HEAD POSITION 10 INCHES SEHIPEO EXIT PUPIL

. ,

95% CONFIDENCE LIMITS

0. 3- F (\
I- -\_____

0- _ \_ _ _ _ _ _ _ _

a-- OFEIPPL__CI3

FIGURE 19C. 95-P\RCENT C0\4YMENCE LIMITS FOR MEAN REAION TIMES

32
 uS

4 The signif'cance of these results may best be appreciated by the optical

diagram in figure 3. The presence of either eye within the cross-hatched area will
permit the pilot to view the entire field, extending f-om A to B, instantaneously. In
the present study head positions were at the exit pupil (EP), and both 5 inches and
10 inches behind, i.e., to the right, of this position. The four locations at the corners
of the matrix delfi.eating the search field (figure 10) a-e at 10. 4 degrees from the
center of the visual field in the display. The aft :imit-X of the full field cone repre-
sented by the cross-hatched area in figure 3 is a function of the size of the exit pupil.
For a 10. 4-degree half-field, the locations of point X behind the exit pupil EP are as
follows:
Size of Exit Pupil Posltion-X Behind Exit I uptil
(inches) (inches)

5.0 13.6

3.0 8.2

2.0 5.4

1.0 2.7

0.5 1.4

If the pilot is aft of the exit pupil anI he cannot position one eye within the
full field cone, he must search the field by vertical and lateral head motion. His %wu
eyes will provide him with two search "beams", whose angular subteknse is a function
of the size of the exit pupil and his head position aft of the pupil.

When the pilot's head position is in the plane of the exit pupil, his search
problem is one of locating the actual exit pupil, which of course becomes difilcult
with small pupils say of 1. 0 and 0. 5 inches in diameter. This is manifested by the
bottom curve in figure 17 with its more rapid rise ir reaction time for these small
exit pupils.
When the pilot's head is five Inch"s aft of the exit pupil he requires virtually
an on-axis eye position to see the extremities of the field at once with a two-inch exit
pupil. With the smaller eirit pupils, the field must be scanned in piecemeal fashion to
located an image. With a head position 10 inches behind the exit pupil, the field sc in-
"ningtechnique Is required with an exit pupil as large as three Inches. However, the

33
4 p
 I

scanning beams ai, large with the three-inch pupil, so the task is not difficult. With
* Ithe small exit pupils, the search beams are small and the search task becomes in-
creasingly difficult as head position Is moved aft. These considerations explain the
T orientation and shapes of the curves in figure 17.

These studies were conducted without simulating the cockpit motion

J produced by turbulent air in actual flight. Head motion produced by turbulence will
make the pilot's search problem somewhat more difficult. Search of the technical
literz,-tre rcvcaled !+t)e rclevant data c.vering pilots head motion in air turbulence.
4 4owever, in heavy turbulence the pilot will be physically constrained by his seat
V• harness, which will limit head motion considerably. Therefore, en the basis of the data
presented in figure 17, a three-inch diameter minimum exit pupil size is recommended
for wide-field head-up displays (25 degrees), with pilot head positions as far back as
10 inches aft of the exit pupil.

The interpupillary distance between the eyes of an adult is about 64 mm

or 2.5 inches. Although this figure varies among adults, an exit pupil as large as
three inches virtua311y assures that binocular vision is possible at the exit pupil for all
I.
adults. In the exit pupil tests conducted in this study, the sub'ects may have detected
* I the target image binocularly or monocularly, for the three-inch aid fuli exit-pupil sizes.
T-ds factor may have contributed to aorue variability in performance for these larger
- pupil sizes. On the other hand, the images were certainly viewed monocularly with the
- exit pupils two inches and smaller. With monocular viewing, there Is of course no
possibility of binocular disparity effects produced by optical distortions.

(. 3

,U
 27
4

4"

Ao

SECTION 4

BINOCULAR DISPARITY STUDIES

6

A. RESEARCh APPARATUS

1. Optical System
6
a. Description

A telecentric viewing aystem was designed and fab, icated to permit the
binocular disparity studies to be acct, nplished with dynamic head-up display imagery
**viewed against a real world background.

The vi•'ing device Is an optical system designed on the principles of on-axis

viewing tbrough telecentric systems. Each eye is provided w'th an identical viewing
system so that even an extremely small distortion, if it should occ-=r, is identical for
both eyes and, hence, wil! generate no binocular disparity. Further precision is
*I obtained by having both systems view a common object, the cathode ray tube. There-
fore, there is no problem conceruing the accuracy of object replication for the two
viewing systems. The optical schematic of the telecentric viewing system is shown
2n figaire 20.
* This telecentric viewing system functions in the following manner. For
right-eye viewing, the light from a display symbol presented on the CRT face is trans-
mitted through a prism beam splitter P, totally reflected by mirror MR1, collected by
* "lens LR1, and totally reflected by mirror MR2. The light is then converged by lens
• " L~p.2, paases through aperture stop ASR and shutter SR, d.verges, and is collected by
lens LR3. After total reflection by mirror MR3, an image is fo.'med in the plane of
* . field stop FSR. This Image is viewed by the right eye after it 's collimated by lens
LR4, and reflected by the half-silvered mirror MR 4 .

S -35

6 S
 F Legcnd
p ~Leiw MIrror 2
CAT
m Stop FS LA
OFS llM Il111filli :ijijflMlllll
(POLARIZED SCREEN)
a•
I REAL WORLD OVERLAY
OPTICAL PROJECTOR
,,•LLI MLI

4 PRISM MR2

4 LI 7

I
P
a2 ~~PHOTO-
ASR
ASI

ILL3 4 0SI FR Lp3

LL4 ML 4 M LR4

V ML3 /
I "BINOCULAR VIEWER

FIGURE 20. OPTICAL SCHEMATIC OF TELECENTRIC VIEWING SYSTEM

4 38

* -
i
 ýJ i
72)

L
S
In a comparable manner light from the CRT reaches the left eye through
the left hlf of the system. Interpupillary distance is adjusted by moving mirrors
ML4 and MR4 in unison, fore or alt. The subject's head position is stabilized with a
chin and head support.

The field-of-view is determined by field stops FSLR, FSR, and FSL.

FSLR is common to both eyes, while FSR and FSL affect only the field for the right
eye and left ,ye, resp.actively. Alternate presentations to the eyes are made by proper
phasing of shutters SR and SL. The two shutters are driven by a common motor through
a timing belt and geared pulley arrangemont. Luminrance matching om the two systems
is accomplished by inserting appropriate neutral density filters near lenses LRi and
LLI. Aperture and field stops can be varied in discrete steps.

Binocular overlay viewing is done by an optical projection system (UP)

which transmits tie image of a 35-mm color transparency to the two optical channels
from the right side of the prism beam splitter. An aerial viE v of real terrain Including
a few buildings was used for the static rendition of the real world background.

Binocular disparities are cbtained by presenting disparate images on the

CtR" aiternately to the two eyes at a ireq'lency above flicker fusion. Disparate CAr
images are generated by inserting X and Y displacementzi of the displayed images into
one of the sequential fields. The shutters are synchronized so that one eye views the
images in an undistorted field, while the other eye sees the distorted field only.

Dual overlapping monocular fields for the images generated on the CRT
are done with polarized screens in front of the face of the CRT at OFS, and at OFSL
and OFSR in the two viewing channels. OFS and OFSL are cross-polarized for aJl
but a circular central portion of the field, representing one rnonotular image field in
the display. The same cross-polarization is accomplished for OFSL and OFSR.
- •Suitable lateral placement of OFSL relative to OFSR will produce the desired dual
* . overlapping fields, as shown in figure 21. The circular central portions of OFSL and
OFSr are polarized in the same direction as OFS, to provide uniform lumina.nce for
the r"al world overlay across the full field.

"37

I •
rI
1-_
I-I 900 FULL BINOCULAR F!EL0

1.6

Ii

DISPLAY IMAGES

12-I/2* DUAL OVERLAPPING MONOCULAR FIELDS

I 4
FIGUR 2i. DISPLAY TMAG!UiY AND- FILDiS OF VIE•W

The optical characteristics of the system produce an overall system

magnification of unity. All lens components are identical, with a focal length of 192 mm
and a diameter of 52 mm. An assembled pair has an aperture of approximately f2. The
optical aperture of each channel is controlled by a series of Waterhouse stops on a disc.
The lens barrels are held in clamp mounts. The fittings which secure the
lens assemblies to the base structure use oversize holes and shims for M4
nal adjustment
of position and angular alignment of the lenses. Three spring/stud mounts allow adjust-
ment of the mirrors and the beam splitter.

b. Shutter/Synchronization Drive System

This equipment provides the following functions in the telecentrtk viewing
apparatus:

* Shutter operation for the alternate presentation of the common CRT display
to each eye

38

1.2.
ia
 (

6 S

e Timing for the pioper positioning of the disparate CRT display images
for each of the optical channels

* * Synchronization pulses to trigger the start of the display image patterns

generated by the electronic equipment driving the CRT.

These functions must be synchronized precisely and stably to one another

irrespective of changes in the absolute speed at which the primary shutter fvnction is
* being operated. The mechanical drive system that was selected is shown in figure 22.
A central drive motor operates the shutter and timing discs through a toothed-chain
timing belt which drives geared pulleys on all of the rotating staf ts. The primary drive
motor is a 400-Hz synchronous motor operating at 121, 000 rpm. The two shutter discs
operate at 1500 rpm, the speed reduction is by a 4X gear drive and a 2X increase in
6 pulley size from the gear drive output to the two discs shafts. Each shutter disc has
a 180-degree slot to open its optical channel for hall a cycle.

The timing disc also operates at 1500 rpm and provides two sets of pulses.
A wide 180-degree pulse provides levels with 50-percent duty cycle for image dis-
* parity switching. Two narrow pulses provide trigger signals for the start of each full 5
cycle of cperation. The timing disc has a pattern of holes passing in front of two
sources of light generated by two small lblbs in an enclosure. Photosensing is
done with two TI type H-35PN diffused silicon photo duodiodes, which are gated
by the light passing through the apertures in .he disc.

c. General Construction

The structure of the telecentric viewing system is self-supporting, and

is formed with plates and strut shapes. The main deck supports all of the equipment
* except the electronic assemblies. - a CRT is supported by a cradle in the face region
and at the deflection coils. The assembled equipment is shown in figure 23.

The head rest and chin support assembly for the test subjects is a Bausch
and Lomo unit, Catalog No. 71-91-10. The head rest is secured to the supporting
structure under the main deck.

39

. o 39

0
 r

t "' * '

* S

I ~ 14

* I
1 !1

* i.
0

01 FIGURE 22. REAR QUJARTER VIEW OF TELECENTRIC VIEWING ASSEMBLY

I " 401

"* '1*
 S

Col

F..
4 4

IIS
L

P o

J 1 5

i i

" . .. ..,. .'K. ...

• - FIGURE 23. SIDE VIEW OF TELECENTRIC VIEWING ASSEMBLY

" " 41 S
 r
r,
2. Electronic Display System

a. Description
A display generator provides four typical flight information images and
displays them on a CRT. These are a horizon line, a flight path marker, a square,
ar.d a two-digit numeral as shown in figure 21. The images may be fixed to form a
static display or continuously varied in position to form a dynamic display.

* The display generator is triggered by synchronization pulses from the

shutter assembly timing disc at a rate of 50 pulses per second. One complete set of
images is generated for each trigger pulse. The timing of the display is synchronized
with alternate openings of the left- and right-eye shutters. The viewer sees the
-- appropriate display with either eye at the rate of 25 times per second. This provides
flicker fusion at the light levels used, so the display appears continuous to both eyes.

S- Means are provided within the display generator to displace the position
of the display images for either eye. This is accomplished by changing the gains and
offsets of the CRT deflection amplifiero from one set of values to another during
Salternate display presentations. Due to the synchronization between the display and
the •hu,,tterLj the eyes see two displays with a controlled relative placement. The

"effectisthat of binocular disparity. Figure 24 is a functional block diagram of the

electronics for the telecentric viewing system.

b. Display Generator
The symbol generator employs stroke writing techniques; stroke writing
is the digital counterpart to analog Lissajous writing. The stroke generator applies
digitally-controlled integrated impulses to the CRT deflection amplifiers instead of
* sine waves. The resultant images are virtually noise-free, and of uniform brightness
and high resolution.

Pulses generated by the image and timing logic are fed to the x- and
y-integrators. Since the pulse amplitude is constant, the output of the integrator is a
linear ramp function. The amplitude of the ramp is proportional to the pulse duration.
Therefore, the writing speed is constant and the length o[ the line is dependent upon
the pulse duration. Brightness pulses are also generateid by theý image and timIng
logic, and are fed to the video amplifier to unblank the CRT when writing a symbol.

42

I1

I I I I I I|
 ý-I W

w 6,

4Z f

CL I

I 443I

C, cc.
 Figure 25 to a simplified Hlock diagram of the display generator electronics. The
system is arranged so that aItrigger sync pulse is required forn the shvtter system
to start the -mage generation sequence. Only one set o0 symbols is written for each
trigger sync pulse. The electronics can follow sync rates from zero to 00 pulses S
per second.

c. Disparity Generator

Two types of disparitles are generated by the disparity generator for both
the x- and y-axes, corresponding to horizontal and vertical dispari'ies. These are a S
constant disparity and a disparity varying as a funclion of position on the CRT. The
disparity generator is an opera.ional ampliiier whose bias signal Input and gain
characteristic may be changed electronically and in time synchronization with the
sh'utter system.

VERTICAL VERTICAL
DEFLECTION DEFLECTION
O"SIONALS CONTROL

L~ALSTO
-VFRTICAL-s -
5!UH-'w,: .. t "->PThCg
- .
LA4PLIFE DISPLAi
VEF.,ER AMPLIrIER

EN
T R IG OI (M iti
CRT
SYNC PULSE AM""LIF.ER
VIDE
PUSIROMI
FYG GENERATOR
SHUTTER

OHORITONTAL HORIZONAL

SUiM:NO
AMPLIFIER
Z DISPLAY
AMPLIFIER

DElrFLIECTION •]DEFLECTION

SIGNALS CONTROL

FIGURE 25. BLOCK DIAGRAM OF IMAGE GENERATING SYSTEM

1I 44

2-S
 Fixed disparity is accomplished by inserting constant bias signals into the
appropriate horizontal and/or vertical deflection channels. The desired bias signal
may be selecLed by the experimenter for either the left eye image or the right eye
image, by changing the switch position.

V•'rying disparities as a function of position, which produce disparity

gradients are accomplished by using different deflection amplifier gains for the right
eye and left eye displays. Referring to figure 26, let the voltage inputs to thZ x- and
y-axes be ex and ey. If kx and ky are the deflection amplifier gains relating position I

to volta-e in one of the chan- -1s, the CRT deflection in that channel will be given by:

x c k ex

Y2 a k y e S
It the changes in gain for the second channel are Ak and Aky, the deflections in the
second channel will be

x2 (kx + ak%) ex

,.y. -..
Ai y
~2 "'y "y'-y
The differences in deflection between the two channels Is then

ax x22 " x = eX
y ak
Ay= Y2 = ky ey

The slope of the line Joining the two positions (xl, yl) and (x 2 , y 2 ) is then

/t~X AX e
ex

It Ak. and aky are made proportional to kx and ky, this slope will be

ay • k x•• e

which is the same as the sloue of the line from the center of the tube to the point 1,
that is yl/xl from the first two equations. Hence the displacement of point 2 from
point 1 An figure 26 will be in the radial direction.
lS

* :45

AL dh S
 r
!•

I-

[I

FIGURE 26. GEOMETRY OF RADIAL DISPARrrY GENERATION

It the amplifier gains kx and ky are equal, the gain changes kakx anda ky
must also be equal to provide a radial displacement of point 2 in figure 26.

d. Control Panel for Experimenter

The experimenter's control panel (figure 22) contains all the necessary
controls to conduct a specific test, and some that are required only for the initial
set-up of the particular test. These controls and their functions are as follows:

e Real World On-Off (R/W On-off) - contrc'.s the power to illuminate the
real world projection system.

* Washout On-Off - controls the ambient ilurnination of the CRT.

* Meander Oscillator Vertical Select - allows selection of a dynamic

diplay moving along the vertical axis.

46
IL-
I.-.
 • Meander Oscillator Vertical Gain - controls the amplitude of thE dynamic
display movemept along the vertical axis.
0 Meander Oscillator Horizontal Select - allows selection of a dynamic
display ricving along the horizontal a~ls.
* Meander Oscillato: Horizontal Gain - controls the amplitude of tl'.e dynamic
display movement along the horizontal axis.
4 Distortion Left-Right - selects the eye to which the displaced pattern is
presented.

* Vertical Gain Disparity Select - allows the selection of either a fixed or

a variable gain control ol the CRT ve-ftical deflection amplifier.

* VertIcal Gain Disparity Control - allows manual controL of the magnitude

of the vertical gain disparity.

e Vertical Null Select - allows selection of the vertical null Control.

a Vertical Null Control - allows manual control of the magnitude of the fixed
(or offset) vertical disparity.

* Horizontal Gain, Disparity Select - per'mits the selection of either a fixed

or a variable gain control of the CRT horizontal deflection amplifier.

* Horizontal Gain Disparity Control - allows manual control of the magnitude

of the horizontal gain disparity.

a Horizontal Null Select - allows selection of the horizcntal null conLrol.

o Horizor-tal Null Control - allows manual cotb-ol of the magnitude of the

fixed (or offset) horizontal disparity.
* Focus Control - allows control of the line thIckness of the CRT diLplav.

B. DESIGN AND CONDUCT OF EXPERIMENTS

1. Tesi Procedures
The b.nocular disparity studies were con.luctcd with three test sjabjects.
Two of the subjects were pilots with milltary flight experience, one currently an airline
captain in scheduled air carrier service. The third subject was a flight test engineer
who had extensive experience in aircraft under a wide range of weather conditions.

47
 4 1r
T" Two other isubjects with military flight backgrounds also participated in
I some of the early tests. However, the extreme binccular disparities that confronted
them caused high visual stress and, after short periods, visual fatigue accompanied
by tearing of the eyes. Both these subjects felt compelled to withdraw from the
program.
All subjects were given optometric examinations, including Ortho-Rater
and duction testa. Their perfoi mance in these tests was found to be within normal
limits.
The. studies were designed to measure visual performance with various
levels of horizontal and vertical binocular disparity, as functions of the following
parameters:
4
* Brightness of images
6 Image motion
* Real world background
S 4 * Line thickness of images
• Overlapping monocular fields.
Fourteen test conditions were established based on the use of these variables
in different combinations, and these are summarized in table 7. The test condition 1 in
*t table 7 consists of a slow oscillation of the display images at a 45-degree angle with
the horizontal, with high bi lghtness images consisting of thin lines, viewed against a
static real world background. Both eyes viewed the same 2J-degree circular field in
the test condition.
* Test conditions 2 through 6 are variations of the nther classes of parameters
one at a time. Image motion in the horizontal directicn and static images were pre-
sented as test conditionr 2 and 3. Images with low brightnebs ad thin lines are involved
in test conditions 4 and 5. The normal brighlness of the image, *nvolved a luminance
equivalent to ten times the luminance of the brightest region '.L •,e real world background.
For the low brightness condition, the luminance of the images was reduced to one-tenth
of Its normdl value, so that they matched the maximum luminance in the real world.
The standard line width of the images waa 2.4 r .inutes, ;knd this was increased to six
minutes to represent th-ck lines. A complex disparity having both horizontal and

48
1•
1_
 0I 13a(qnS ~o- ca c.li~~j - '

*Z W

-jrtn~ouie IrLa O0P0.D 0FIP%,'0' 0 00

0o Sroaua5owoJ 0

6 ~PI.IOM ItE.3 0 000 0 00

xaldwoo 0 -0

aldtu;S 0 00 00 a 00 10100.

0
- ~ aula 0 000 0 0060 , 0

0CUs, -4i

* _ _ _ _ _ ~ Ef
r
0 11M ZI-49
(-p nim0 0 le 1 1 1

*E
 O F
0

ll - vvrtlcal components was presented in test condition 6. A homogeneous background

I -representing flight within a cloud layer replaced the real world background to test
condition 7. in test condition 8, dual overlapping fields of view were presented to
each eye, displaced laterally to produce six degTees of overlap in the horizontal
direction.

Test conditions 9 through 14 are derartures from test condition I by the

simultaneous change of two parametcrs; the combinations used are shown in table 7.

The disparity levels used in the tests were established by an extent ive
series of preliminary experiments, and these levels are summarized in table 8.
- Several facts are prominent in these data. The disparity ranges are much smaller
for displays viewed against a real world compared with displays viewed against a
4 homogereous background. This is true for both horizontal and vertical disparities.
For b•th horizontal cases, the ranges are larger for condition esophoria (cross-eyed)
compared with exophoria (wall-eyed, divergent). For the vertical disparities, the
Sranges are the same for hypophoria (right eye low) and hyperphoria (right eye high).

SA aOuvi pea-fovae Was eiabirihed for iiese binocular

disparity studies. Tolerance for disparities Is usually associated with the ability of
an observer to retain sirgle vision, i.e., to prevent visual doubling of objects or
displays. However, In pie!lminary experiments, the subjects complained of visual
stress and annoyance caused by disparities that are considerably smaller than those
4 that produce doubling. Therefore, a psychometric rating system was developed to
measure visual stress levels at disparity levels for which single vision exists, as well
as for doubling phenomena.

After a 15-second exposure to a particular display situation, each subject

4 was required to respond by rating his level of visual comfort in one of the following
slx response categ( ries:

L.. Response Category Visual Com'ort Level

1
.3.Comfortable Excellent
2 Comfortable, short of excellent C
i

3 Mildly uncomfortable Uncomfortable

4 Severely tncomfortable J with single vision
5 Doubling less than 50 percent of the time "
>Double vision
6 Doubling more than 50 percent of the time)

50

4t
 On this basis, categories 1 and 2 represent two levels of comfortable
vision, while categories . arid 4 provide for two levels of discomfort, all with single
vision. Image doubling is covered by categories S and 6, depending on the persistence S
of the doubling.

Each test, representing a horizontal or a vertical disparity experiment

under a particular set of display conditions from table 7, involved ten replications of
each of nine disparity levels from table 8, or a total of 90 data points. The disparity
levels were presented in a random sequence to the test subjects. Considering both
horizontal and vertical disparities for each display condition, a full test session included
180 data points. The responses of the subjects to each presentation were in one of the
six rating categories.

The details associated with the conduct of the experiments are as follows.
The Interpupillary distance of a subject was measured, and the spacing of the two exit
pupils Jn the binocular viewing apparatus was adjusted accordingly. The subject was
then seated in the viewing compartment (figure 27) and his head rest and seat were
adjusted to provide a comiortable viewing condition. The subject became adapted to M
the low level of ambient illuminadon and final adjustment of the Interpupillary setting
of the apparatus was made.

The subject was then given practice trials involving both horizontal and
vertical disparities to familiarize him with the displays, the equipment, and the follow-
ing test procedures. When the experimenter had set a particular disparity level into
the apparatus, he advised the subject by calling "ready" over the intercom. When the
subject was ready to proceed, he responded by pressing a buzzer. This procedure
permitted the subject to proceed from trial to trial at his own pace. Upon hearing the
buzzer, the experimenter started an electric timer and permitted the subject to view
the display for fifteen scconds, at which time he requested the subject to respond by
calling "read". The subject then pressed the buzzer one to six times corresponding
to the visual comfort levels I through 6. The buzzer was used to transmit the response
rather thain voice communication to avoid errors caused by loss of intelligibility at the
ambient noise levels produced by the running shutter drive system. The experimenter
then recorded the response, reset the timer to zero, and proceeded with the next
disparity level.

51
 I.I
TABLE 8. BINOCULAR DISPARITY LEVELS USED IN TEST SCHEDULE

DialHorizontal Vertical
DilDisparity Dial Disparity
I.SettUag (minutes) Setting (minutes)

With Real World Background

530 -18 Esophoria 520 +12 Hyperphoria

520 -12 (Convergent) 515 +9 (Right Eye High)

510 -6 (Near) 510

1.505 -3 10 .3
500 0 Orthophoria 500 0 Ortho

495 +3 495 -

490 +8 Exophoria 490 -6 Hypopaoria

495 +9 (Wall-Eye) 485 -9 (Right Eye Low)

480 +12S (Far) 480 -12( 1te

With Homogeneous Background

800 -180 Fsophoria 570 +42 Hyperphoria.

700 -120 (Convergent) 550 +30 (Right Eye High)

600 -60 (Near) 530 +18

550 -30 515 +9

500 0 Orthophoria 500 0 Ortho

475 +15 485 -9

Lii
450 +30 Exophoria 470 -18 Hypophoria
40 10 Fspoi 7 4 yepoi
400 +60 (Wall-Eye) 450 -30 (Right Eye Low)

350 +90 (Far) 430 -42

52
 ..

I
II

FIGUE SUJ . ETDI IEWN OPRMN

0 13
 A set of experimental trials consisting of 30 disparity presentations lasted
between 1U and 15 minutes. Alter this the subject was given a five-minute rest period;
10- to 15-minute rest periods were given after each two sets. A complete experimental
session consisted of six sets, three involving horizontal disparities and three covering
vertical disparities, thereby providing the 180 data points planned. Aa experimental
session lasted about two hours. Each subject participated in two sessions per day and
In some instances three sessions. When a third session was included, two test cindi-
• lions were interlaced so that the presentations for the two conditions to be compared

would be balanced on a day-to-day basis.

i 2. Data Reduction

Each test condition yielded 90 data points for horizontal disparities and
* 90 data points for vertical disparities. The 90 data points are distributed among the p
nine levels of disparity (table 8), with 10 points per disparity. The results for a
typical test condition are prescnted in the two-way matrices in table 9, in which the
responses are binned by category for each stimulus (disp;,rity level, In minutes of arc).

The e__er't•_d data In the matrices were reduced to graphical plots by con-
aidering the percentage of the responses that were equal to or better than response
category 2, and equal to or bctter than response category 3. This was done for each
disparity level and the results were plotted as in figure 28. The solid lines represen1t
the variation of the sum of the responses categories of 2 or better, I. e., I plus 2, as a
L function of disparity level. The dotted lines represent the sum o; the response cate-
gories of 3 or better, i.e., 1 plus 2 plus 3.

I The curves in figure 28 show optimal performance levels at small dispari-

ties, near zero. However, this performance degrades as disparity, and, theref.ire,
*, [.visual stress level rises in either direction. Smaller tolerances to disparity is '

manifest by narrow band or peaked curves in figure 28. Therefore, subjects are seen
to be more sensitive to vertical disparities compared with horizontal disparities. In
addition, for horizontal disparities, esophoric (convergent, cross-eyed) conditions
I " provide lower stress levels than exophoric (wall-eyed) conditions. These two horizontal
• .- disparities will be designated as "near" and "far". p

54L
54

L2
*
 TABLE 9. DISTRIBUTION OF RESPONSES FOR
TYPICAL TEST SESSION

Disparity Response Category

(mnte)12 3 4 56

Horizontal Disparity

-181.
-12 3 5 2
-6 1 8 1

-3 6 3 1

0 9 1

+3 4 5 1

+6 4 51
+9 2 4 4

+12 1 9

Vertical Disparity

+12 101J
+9 3 7
+6 1 5 4 -

+3 3 5 1 1
0 8 2

-3 9 1 1
-6 1 6 2 1

-9 3 7
-12 10

55

4 IS
 (0%)

I-I

r. -.

u- w

[i °i,,,,•
40 " ,
0

-18 -15 -12

-

-9
NEAR
--6 -3
I
0 3
FAR
,

6 9
HORIZONTAL DISPARITY (rninutAs of arc)

COMFORT INDEX
comrORT INDEX 2
3 ,
J' ..=•.==-•-

1 100

-•RIGHT EYE HIGH RIGHT EYELO

z so
S~vERTICAL DISPARITY (minutes of arc)

1I
I.. CrI

DISPARITY FOR. TYPICAL TEST SESSION

56
[- 60 Ai
 The families of curves in figure 28 summarize the variation in visual
performance as a function of horizontal and vertical disparity levels. To apply these
data to the optical design problem for head-up displays, it is necessary to establish
maximum permiss!ble disparities. These in turn depends on minimum acceptable
visual performance levels. Two criteria were specified for this purpose:

"* Comfortable vision or better (Categories I or 2) 5U percent of the time

was designated Comfort Index 2 •

"* Mild discomfort or better (Categories 1, 2, or 3) 80 percent of the time

was designated Comfort Index 3.

It was judged that a head-up display that met these criteria based on the data for
S
sustained 15-second viewing used in these studies, would be satisfactory in real flight.

As an example of the application of these criteria to the data In figure 28,

the maximum permissible disparities are:
Maximum Disparity
(minutes) .
.Type
of Disp.rity Cnmfnr;n Tndew N1ear Far q -

Horizontal 2 8.8 4.3

3 7.8 3.6

Right Eye Low Right Eye High

Vertical 2 6.8 4.1

3 6.3 4.5

The amount of data from each of the three test subjects is summarized in
table 7, where the number of test sessions for each experimental condition is indicated. s
The horizontal and vertical disparities associated with Comfort Indices 2 and 3 for all
of the test sessions are presented in table 10. The median disparity levels derived
from replicated test conditions are also included in this table. Medians were used
rather than arithmetic means, to include the effects of larger-than values which are
beyond the disparity ranges covered, without permitting these data points undue weight-
ing of the results. When two pieces of data are involved, the median ana mean values
are of course the same.

57

.......... .......... ....... ,

 TABLE 10. SUMMIARY OF BINOCULAR DISPARITY LEVELS
FOR TWO COMFORT INDICES
Comfort Index 2 Comfort Index 3
Horizontal Vertical Horizontal Vertical
Disparity Disparity Disparity Disparity
.(miuiutes) (minuteE-) (minutes) (minutes)
Test Right Rioght Right Right
Condition Session Eye Eye Eye Eye
and subject No. Near Far Low High Near Far Low Hfigh

SUBJECT A

Condition1, 1 8.8 4.3 6.8 4.0 8.0 3.8 6.3 4.5

Reference 11 4.5 2.0 4.5 2.8 >18.0 3.8 3.8 1.8
14 11.0 4.5 5.0 5.0 >18.0 3.0 4.5 4.0

23 8.0 5.3 4.5 7.0 18.0 4.3 3.8 6.8

25 8.0 4.0 4.5 4.5 18.0 3.8 4.0 4.5
28* 8.3 8.3 7.0 2.5 7.8 6.0 6.0 4.5
Median 8.15 4.4 4.75 4.25 18.0 3.8 4.25 4.5 b

Condition 2, 21 8.3 4.8 4.0 4.0 13.8 4.0 3.5 3.5

Horizontal
Motion

Condition 3, 2 8.8 5.0 5.5 2.0 13.3 4.3 4.5 1.5

Static 13 12.0 3.5 4.3 1.8 >18.0 4.0 3.8 1.3 0
Median 10.4 4.3 4.9 1.9 - 4.15 4.2 1.4

Condition 4, 4 10.5 5.0 4.5 4.5 18.0 4.0 4.0 3.5

Low Bright- 26 9.0 5.5 4.3 3.8 >18.0 6.0 3.8 3.8
ness
Median 9.8 5..% 4.4 4.2 - 5.0 3.9 3.7

Condition 5, 3 12.0 5.0 6.5 3.5 18.0 4.8 6.0 3.5

ThickLinev 15 9.5 4.8 3.5 6.0 >18.0 4.0 4.5 4.5

24 9.8 5.0 4.0 2.3 18.0 4.5 3.5 3.5

Median 9.8 5.0 4.0 3.5 18.0 4.5 4.5 3.5

*Half Session

58
.7

vS
 S
If77 1®
S
! .. (II

TABLE 10. SUMMARY OF BINOCULAR DISPARITY' LEVELS

FOR TWO COMFORT INDICES (Cont'd)

Comfort Index 2 Comfort Index 3

Horizontal Vertical Horizontal Vertical
Disparity Disparity Disparity bisparity
(minutes) (minutes) (minutes) (minautes)
Test Right Right Right Right
Condition Session Eye Eye Eye Eye
and Subject No. Near Far Low High Neat Far Low High

Condition 6, 5 9.3 4.5 4.5 3.0 15.0 4.3 4.3 3.5

Complex 12 1.5 1.3 1.8 1.3 8.5 0.0 1.5 0.5
Dispar. ty
Median 5.4 2.9 3.2 2.2 11.8 2.2 2.9 2.0

Condltion 7, 18 53.0 27.0 15.0 8.0 72.0 35.0 21.0 10.0

Homogeneous 20 86.0 82.0 22.0 11.0 84.0 80.0 23.0 9.0
Background
27* 110.0 73.0 22.0 13.0 120.0 70.0 21.0 13.0
Median 86.0 73.0 22.0 11.0 84.0 70.0 21.0 10.0

Condition 8, 7 7.0 5.3 2.5 3.5 14.0 5.0 1.5 1.5

Overlapping 17 10.3 5.3 5.0 4.5 > 18.5 4.0 6.0 3.5
Fields
Median 8.7 5.3 3.8 4.0 - 4.5 3.8 2.5

Condition 9, 22 8.8 6.0 3.8 3.0 '18.0 6.0 3.8 1.5

Horizontal
Motion and
Low Brl' 'it-
ness

Condition 10 6 > 18.0 4.8 2.5 4.8 >18.0 3.6 1.5 3.5
Static and 3.5 0
16 13.8 6.0 3.5 4.0 '18.0 4.0 2.0
4 Overlapping
Fields Median - 5.4 3.0 4.4 >18.0 3.9 1.8 3.5

Condition 11 9 11,0 4.8 4.0 4.3 >18.0 4.5 3.5 3.5

Low Bright-
ness and
Overlapping
Flelda

*Haif Session

59

4D
 ri

TABLE 10. SUMMARY OF BINOCULAR DISPARITY LEVELS

FOR TWO COMFORT INDICES (Cont'd)

Comfort Index 2 Comfort Index 3

Horizontal Verticai Horizontal Vertical
Disparity Disparity Disparity Dispartiy
(minutes) (minutes) (minutes) (minutes)
Test Right Right Right Right
Condition Session Eye Eye Eye Lye
and Subject No. Near Var Low Hi;h Near Far Low HI.h

CondJuon 12, 8 9.0 4.5 5.0 3.8 >18.0 4.0 5.0 3.0
Thick Lines
and Over-
lapping Fields
{ Condition 13, 10 - -
Thick L'nes

and Large
Complex
Disparity

Cond!tinon 14. 19 37.0 30.0 14.0 13.0 69.0 25.0 21.0 12.0
Static ard
Homogeneous
Background

SUBJECT B

Condition 1, 8 9.3 7.8 4.3 6.8 18.0 9.0 6.3 9.0

S~Reference 5.3 18.0 7.0 6.3 7.0
R3 3.0 6.0 6.0

Medtan 6.2 6.9 5,2 6.1 18.0 8.0 6.2 8.0

Condition 3. 1 9.5 6.0 3.8 5.0 12.0 7.5 4.5 5.0

Static

Condition 4, 7 12.0 6.0 3.5 4.5 > 18.0 6.5 4.5 6.0
Low Bright-
ness

Condition 5, 2 9.0 6.8 5.0 5.3 14.5 12.0 9.0 9.5

Thick IPr~es

Condition 6, 4 8.5 5.5 4.0 4.0 >18.0 8.0 6.5 6.5

Complex 9 12.0 7.3 4.8 6.5 >18.0 8&0 e.0 7.3
Dis~parity
M.-,dian 10.3 6.4 4.4 5.3 >18.0 8.0 6.3 6.9

I
60

V
 TABLE 10. SUMMARY OF BINOCULAR DISPARITY LEVELS
FOR TWO COMFORT INDICES (Cont'dj

Comfort Index 2 Comfort Index 3

Horizontal Vertical Horizontal Vertical
Disparity Disparity Disparity Disparity
(minutes) (minutes) (minutes) (minutes)

Tast Right Right Right Right

Condition Session Eye Eye Eye Eye
and Subject No. Near Far Low High Near Far Low High

Condition 8, 5 - - 7.0 - 12.0 7.0 6.8 3.0

Overlapping
Fields

Condition 12, 6 - - 7.3 5.0 >18.0 6.8 10.0 9.0

Thick LUnes
and Over-
lapping
Fields

SUBJECT C

0 Condition 1, 12.0 1.3 2.3 3.8 > 18.0 5.0 4.2 6. 0

Reference

Condition 7, 2 60.0 34.0 23.0 32.0 127.0 04. 0 32.0 34.0

Homogeneous
Background

61
 -• 4

O r
S®

C. RESULTS AND IMPLICATIONS

The results in taole 10 cover the effecLs of the principal factcrs, oiie at a
S4 subjects, A and B.
time, for two test In general, the effects of variables changed in
*- paIrs of two were investigated with subjects A and B, while only basic test data were
obtained from subject C, 1.o corroborate some key basic results.

1. The reference tes' condition I represents a baseline situaticn representing

-aormal operation for tl-h head-up diE olay, with moving display images viewed against a
* -static rra.i world b3.ckground. The disparities for Comfort Indices 2 and 3 for all subjects
are stmniarized in tabk 'I. The results are similar for both comfort indices. For
horizontal diopariiis the tolerance for wall-eý'ed or exophoric conditions (far) is con-
sldzrab'-y less Cian for the convergent (near) or esophoric condition. Tolerances are
j-uniformly low for vertical disparities. For both horizontal wall-eyed and vertical dis-
parities, the tolerar.ces are in the I- to 2-rr~l range. Convergent horizontal tolerances
are Romewhat higher, about 2-1/2 mils. (The test data are expressed in minutes of arc,
- but the results are also conv'rted to milliradians (mils), in which optical tolerances are
* conven~tionally expressed, in table 11. One m£Iliradian is equivalent to 3. 44 minutes

FP 4 1 , of arc).
TABLE 11. SUMMARY OF BINOCULAR DISPARITY LEVELS FOR ALL
SUBJECTS FOR REFER2NCE TEST CONDITION

* Comfort Index 2 Comfort Index 3

Horizontal Vertical Horizontal Vertical
Disparity Disparity Disparity Disparity
i (minutes) (minutes) (minutes) (minutes)
Right Right Right Right
Eye Eye Eye Eye
Subject Near Far Low High Near Far Low High

A 8.15 4.4 t.75 4.25 18.0 3.8 4.25 4.5

B 6.2 8.9 3.8 5.0 18.0 8.0 6.3 8.0

C 12.0 1.3 2.3 3.8 '18.0 5.0 4.0 6.0

Mean 8.8 4.2 4.4 4.7 18,0 5.6 4.9 6.2

Mean 2.6 1.2 1.3 1.4 5.2 1.6 1.4 1.8

(mils)

62

S I-
 The tolerances for all types of disparities are high for the two subjects
tested when the display images are viewed against a homogeneous background (test
condition 7). These results are summarized in table 10. The disparities exceed the
* comparable results for the reference condition by more than a factor of 10 in almost all
cases. This represents the most significant finding in the disparity studies. Subjects
can compensate for binocular disparities without loss of visual comfort by changes in
eye positions, provided these compensatory relative eye positions can be maidtaired.
This is the viewing condition with a uniform backg-round that the subject may ignore
visually. However, with an articulated background such as the real world, from which
the pilot must obtain visual information, the situation is more demanding. Compensa-
tory eye movements to handle binocular disparities in the display images must be
relaxed when viewing the real world background. Hence, alternate changes in relative
* eye positions must be made, as attention is shifted from the display to the real world,
and back to the display. Therefore, the permissible disparities are low under these
relatively stringent conditions.

The pzttern of tolerances is the same for both the reference condition and
* * the homogeneous background. Convergent (near) disparities are appreciably larger
than wall-eye (far) disparities, and permissible horizontal disparities are larger than
vertical disparities.

Considering image motion, disparity tolerances for static images (test

condition 3) are slightly lower than for moving Images for both subjects A and B
(table 10). The differences between test conditions are less than one mil. However,
for subject A with vertical disparities, the tolerances with static images produced the
lowest levels recorded, about 1/2-mil, and theae data were supported in two tests.

Regarding image brightness, tolerances are comparable for both low

6 brightness images (test condition 4) and the referenc, high brightness situation for
both subjects.

Thick lines (test condition 5), in which the image Line thickness Is six
minutes, compared with the standard fine line of 2. 4 minutes, yield about the same
* disparify tolerances, for both subjects.

63

0
 4
1A v complex disparity (test condition 0) is one in which both horizontal and
-vertical disparities are present concurrently. For the horizontal disparity tests, a
•.constamnt three-minute vertical disparity (right eye low) was present; for the vertical
disparity tests, a four-minute wall-eye horizontal disparity was superimposed. The
complex disparity presented a more stringent condition for subject A, lowering his
"meanvertical disparity toieran(.e im one mi. Subject B had about the same tolerances
with the complex disparity as with the simple disparity.

[I With dual overlapping fields (test condition 8), a portion of the display field
is seen with one eye, another portion is seen with the other eye and a central portion
is scen binocularly (figure 7). The fields are displaced laterally. Under these condi-
tions, both .;abjects exhibited slightly lower vertical disparity tolerances compared with
4 the reference test condition 1. Minimum tolerances are about one
mil for both subjects.

Test conditions 9 through 14 involve the simultaneous manipulation of two

parameters to eliminate possible interaction effects. Subject A was the key participant
in the e! studies, and the results are included In table 10. No unusual effectq are noted,
S I with minimal toierances still about one mil, except for one situation, horizontal motion
with low brightness (test condition 9). Here, one vertical disparity tolerance was as low
as 1.,5 minutes or about 1/2 mi:. The lowest vertical tolerances for horizontal motion
(test condition 2) or low brightness (test condition 4) are about one mil. There may be
an interaction effect pi esent in the combined test condition 9, but the data from the one
test is too limited to be conclusive.

Test condition 14 involves static imagery viewed against a homogeneous

background. The results in table 10 indicate that there is a significant reduction in
permissible disparities for the static condition compared with the moving imagery with
4 the same background (test condition 7). However, the tolerances under the static
condi-
tion are still large compared with the same images viewed against the real world.

I The fact that two prospective test subjects, both with military flight experi-
ence, withdrew from the test program after a few preliminary test sessions is significant.
Both complained of severe visual stress, accompanied by tears and visual fatigue, as a
result of the test situations. These effects were undoubtedly produced by the binocular
disparities with which they were confrontd. There have been reports of test pilots
compalning of eye fatigue after s•ustained flight with head-up displays. This condition
may well have been caused by distortions in the optical systems of these displays.

L2 64
 In summary, it may be concluded that ,inocular dispaý,ity tolerances are
considerably lower when head-up display images a-e viewed against a static real world
D
background compared with a homogeneous background. Maximum disparity levels for
viewing with adequate comfort are one mil for vertical disparities and wall-eye horizontal
disparities. Convergent hortzontal disparities may be as large as 2.5 mils for the same
levels of visual comfort.

The foregoing studies have demonstrated that the presence of a static real
world background significantly influences the permissible binocular disparities. In the

tests which have been conducted, the simulated real world background was a static aerial
view, typical of flight at higher altitude, with no appreciable rates of change of attitude
or heading. The visual backgrounds remains quasi-static under these conditions. How-

ever, since presence of a real world background in important for binocular disparity
tolerances, the effects of a dynamic background, typical of low level flight, should be
Investigated.

65
4
ri
I

SECTION 5

. CONCLUSIO.'•S

1. The minimum size of exit pupil required for wide field (25 degrees) head-up display
-*-- system is three inches when the pilot's head position is no more than 10 inches
behind the exit pupil.

- -2. The binocular disparities permissible are consid. rably lower when head-up disptay
-. images are viewed against a static real world background compared with a homoge-

3. Maximum permissible disparity levels are one mil ior sustained viewing with
- - adequate comfort, for vertical disparities and divergent horizontal disparities.
Convergent horizontal disparities may be as large as 2. 5 mils.

- 4. The effects of thicker image lines, low image brightness, or dual overlapp!ng
monocular fields do riot signtficantly change binocular disparity tolerances.

-- 5. Disparity tolerances for static display images, and for combinations of vertical
. . and horizontal disparities, are slightly lower than for dynamic imagery with either
horizontal or vertical disparities.

6. There Is a slgnificant reduction in permissible disparities when static images are

* - vlowed aga-nst a homogeneous background, compared with moving images with the
same background. However, the tolerances for static imagery with the uniform
background are large and, therefore, not critical compared with images viewed
aZa"inst a static real world.

7. Interaction effects due to simultair.eous changes in two visual parameters are smail,
and do not significantly affect binocular disparity tolerances.

48
68

4t
 II
k.]

* S

SECTION 6

RECOMM ENDATION 5

The effects on binocular disparity tolerances of a dynamic visual real world

* background against which head-up display images are viewed should be investigated.
The results obtained in this manier will provide a more complete bacis for head-up
display specifications.

S S

* S.

67
 I

- S

SECTION 7

SRE FERENCES

. *1. T. Gold, R. F. Perry, and S. IKier, Simulator and Flight Test Studies of
Windshield Projection Displays for All-Weather jLading. FAA SRDS Report No.
RF-65-112, January 1966. I

2. R.K. Johnson, and T.S. Momlyama, Flight Test and Evaluation of Spectocom
Head-Up Display Installed in an A-5A Airplane, Naval Air Test Center Technical
Report No. FT 222-•SR-64, December 1964.

* 3. S. H. Bartley, Principies of Perception, New York, Harper and Brothers, 1958, I

Chapter 10, pp. 229-232, and p. 168.

* 4. C.H. Graham, Visual Perception, Handbook of Experimental Psychology, edited

by S.S. Stevens, New York, John Wley, 1951, pp. 870-871, and p. 894.

5. K. N. Ogle, Researches in Binocular Vision, Philadelphia, W. B. Saunders, 1950. i 5

6. K. N. Ogle, On the Limits of Stereoscopsis, Journal of Experimental Psychology,

Vol. 44, pp. 253-259, 1952.

7. W.D. Zoethout, Physiological Optics, Chicago, The Professional Press, 1947,

pp. 294-295. •

I 8. D. H. Jacobs, Fundamentals of Optical Engineering, New York, McGraw-Hill,

1943, Chapter 3.

* 9. F.A. Jenkins, and H.E. White, Fundamentals of Optics, Third Edition, New York,
McGraw-Hill, 1957, Chapter 7.

10. J. 1.C. Southall, Mirrors, Prisms, and Lenses, Third Edition, New York,
McGraw-Hill, 1933, Chapter 12.

* 688 ,I

L2
0I
 S

UNCLASSIFIED
DOCUME14T CONTROL. DATA- R&D
,Y,-,,,,,ft,,I,.. (t. ,,jrono f hi.L Ad,d of ahltysc(artnl ,nd...fn, at I , .,.)
AI,ouqu i.n !tntprcd,.bet, the .,, .p.rl
P•t .I...t.•#l.Ij

Sperry Gyroscope Division Unclassified

Sperry Rand Corporation ih- Oo.
Great Neck. New York 11020 !

Visual Requirements Study for Head-Up Displays 5

t
* OiCS ,D 'VF NOPES rTýn t tp ee C t d e-).t~t.

Final Report - Phase I, April 1966 - March 1970

?-- C'} ,I, ne,¢l.l-. m~ddi.• *,t,.ll. 1(Cat SAU
namalj

Theodore Gold
Aaron Hyman S
* ."0., r..'t I.. TOY*L '.0 01 _ [3E 17b. NO OF MREF

March 1970 81 j 10

N00014-66-C-0114, N-R 213-047

....., .... SGD-4283-0333
012-04-01 •
Ob. 0 Tt-( AL POMR NOiSt (An othrt numb~t. be, miay be -1C*,-1cd
this flpot()

"d JANAIR Report 680712

0O01$ T . -U PIO L - (
ECA

This document has been approved for public release and sale; its distribution I(~ I

is unlimited.
pI L - - ____________________T-T
NjE 12__

Joint Army Navy Aircraft Aeronautics Programs, Code 461, Office

Instrumentation (JANAIR) of the Naval Research, Department of
-A the Navym WVashington, D.C. 203C0

An experimental study was conducted to determine critical values of the

essential optical and electronic design parameters for head-up displays based on
the visual requirements of the pilots who will use these displays. This information
will provide a basis for future rpecificati:ns covering this type of display. The
investigation covered two areas, size of exit pupil and binocular disparities due to
optical distort:ons.
The exit pupil study was conducted in a flight simulatcr in which a wide field
(25 degrees) head-up display with a large aerial exit pupii was installed. Four
pilots with military flight experience served as test subjects. *The results indicate
that the minimum size of exit pupil required is three inches in diameter, for wide-'
field systems in which the pilot's head position is no more than 10 inches behind the
exit pupil.
A laboratory telecentric viewing system was developed for the binocular
disparity study. This equipment made possible experiments in which dip.pzrate,
dynamic, head-up display images were presented to subjects who viewed these
images against a static view of a real world background. Three test subjects were
(Contim,-d)

DD Fo 1 4 7 3 ,(PAGE
)
S/N 010i-807-6801 Srcury Ll-ssLttcaýton
JNb PPS0 13152
 4i
S
UNCIASSIFIED
Securaty 0|419th-11ntn• •
44 1. ..
UINK C

*OL
WTKOL KOL W-

Binocular Vision
Displays
Head-Up Displays
Optics
Stereopsis
Vision

* I'

i3

bakrud copae wihahroceu fmdrt rgtes wlgo

WG
I
I
I
S

13. Abstract (Cont'd)

employed .in these tests. The results show that binocular disparity tolerances are-
considerably lower when head-up display images are viewed against a real world
background, compared with a homogeneous baclkground of moderate brightness. -
Maximum disparity levels are one mil for sust!a:ed viewing with adequate comfort
for vertical dispa~rities and divergent horizontil disparities, Convergent horizontal
4 disparities may be as large as 2. 5 mills for the same level of comfort. I-

LhD ,':ft.I,73 (ACK) UNCLASSIFIED "

(rPACE 2 SCruri! CIAaElct iofC*I"--fl