- _..... - ..---..---. .. o - o o.-.--.-' o»-ro . u»-on"pwfl0v.h= :—

EFFECTS OF SUSITISIONAI. INTERMITTENOY
ON THE APPEARANCE or TARGETS
0F COMPLEX SPECTRAL COMPOSITION .

Thesis for theDegree of M. A.
MICHIGAN STATE UNIVERSITY '
ALLEN LOUIS NAGY
1971

 

. o I
......

ABSTRACT
EFFECTS OF SUB-FUSIONAL INTERMITTENCY

ON THE APPEARANCE OF TARGETS
OF COMPLEX SPECTRAL COMPOSITION

By
Allen Louis Nagy

The effects of sub-fusional intermittency on

the appearance of targets of complex spectral composi-
tion were observed in this study. Achromatic targets
of varying spectral composition, purple targets, and
metameric pairs were examined. A dual—beam optical
arrangement was used in conjunction with narrow-band
interference filters to produce complex targets with
known spectral composition. Brightness matches and
judgments of hue and saturation changes were made with
four conditions of intermittency. Temporal effects

on color and brightness do not appear to be dependent
on the wavelength composition of the target in a simple

way.

\
Aomammfi

Date: VNww Ti L177!

 

EFFECTS OF SUB-FUSIONAL INTERMITTENCY
ON THE AFFEARANCE OF TARGETS

OF COMPLEX SPECTRAL COMPOSITION

by

Allen Louis Nagy

A TIESIS

Submitted to
Nichigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Psychology

1971

Dedicated to my wife, Linda

ii

ACKNONLEDGNENTS
The writer wishes to thank Dr. S. Howard Bartley,
chairman of the guidance committee, and Dr. Richard
J. Ball for their guidance and encouragement through-
out the course of this project. Thanks are also ex-
tended to Dr. Theodore H. Forbes for his aid in cali-

bration of the apparatus and writing of the thesis.

A. L. N.

iii

TAB LE OF CLO}. TENTS

LIST OF TABLES ....... . ...... ....................... v
LIST OF FIGURES....................................vi
I INTRODUCTION...... ...... ..................... l
Brightness Phenomena...................... l
Chromatic Phenomena....................... 2
Problem................................... 3
II METHOD .......... ............................. 7
Aoparetus................................. 7
Specification of Stimuli..................lO
Observers.................................lO
Procedure.................................l3
III RESULTS............... ....... ................l6
Brightness Data...........................l6
Chromatic Data............................25
IV DISCUSSION...................................33
BIBLIOGRAPHY..................... ......... .........37

iv

r1

 

' 'I_- .. EL ,

“I'll“ .-

 

aser Ly.

 

 

 

rl'o _:-"

TMflEl

TABLE 2
TABLE 3

TABLE 4

LIST OF TABLE"

Description of Spectral Composition

of Targets............................ll
Hues Acquired by Achromatic Targets...26
Hue Changes in Purple Targets as a
Result of Intermittency...............27
Comparison of Changes in Chrometicity

Of II'Tet9.n’1eriC Hues...0.000000000000000031

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

LIST OF FIGURES

Schematic Diagram of Dual-Beam
Apparatus............................ 9
Brightness Index of Achromatic Targets
for Observers M. D. and A. N.........18
Brightness Index of Achromatic Targets
for Observers h. E. and R. B.........l9
Change in Brightness Index Over Sessions
for Observers M. E. and R. B.........20
Brightness Index of NarrowéBand Purple
Targets..............................22
Brightness Index of Wratten Purple
Targets..............................23
Desaturation Estimates for Narrow-Band
Purple Targets.......................29
Desaturation Estimates for Wratten Purple

TargetSOOOOOOOOOOOOOOOOOO0.00.00.00.030

vi

INTRODUCTION

Brightness Phenomena

A large amount of research concerning the effects
of sub-fusional intermittency on the brightness of
full-spectrum targets has been carried out by Bartley
and his colleagues. Generally they have found that
an increase in brightness occurs when full-spectrum
photic stimulation is made intermittent. The effect
was maximum when the target was pulsed at rates of
approximately ten cycles per second with the pulse
occupying 1/4 of the cycle. This phenomenon has been
termed “brightness enhancement“ or the “Bartley effect“.

Much of this research has been devoted to deline-
ating the particular stimulus characteristics important
to the production of brightness enhancement. Pulse
length, rate, target size, intensity level, and adaption
level are some of the variables examined ( Bartley,
1951a, 1951b, 1952, 1957; Valsi, Bourassa, and Bartley,
1959; Nelson, Bartley, and Jewell, 1963; Bourassa
and Bartley, 1963; and Kazsuk and Bartley, 1969 ).

Ball (1964) extended this investigation to chro—
matic stimuli when he measured the brightness of inter-
mittent monochromatic targets. His results show that
brightness enhancement oecurs with only one of the
monochromatic stimuli observed. This was the waveband
centered around 510 nm. Other monochromatic targets

were observed to have a brightness quite close to
l

 

lllllll'llli‘lll'll]

2

Talbot level brightness under the same conditions.
The temporal conditions producing brightness enhance-
ment in the 510 nm target were quite similar to those
producing enhancement in the full-spectrum targets;
rates of six to thirteen Hz. and a PCP or pulse-to-
cycle fraction of approximately 1/4. Ball's findings
suggest that temporally induced brightness in inter-
mittent targets may be dependent on the spectral compos-
ition of the target, and that the inclusion of wavelengths
around 510 nm is necessary to the occurrence of bright—
ness enhancement.
Chromatic Phenomena

Bartley and Nelson's (1960 and 1961) initial
observations on intermittent chromatic targets showed
that the chromatic characteristics of a visual target
could also be very dramatically affected by the same
temporal conditions that produce brightness enhancement.
Hue shifts and large amounts of desaturation were
reported to occur. The first measurements of these
phenomena consisted of verbal judgments made in Ball's
1964 study. Results showed that hue shifts revolved
around two invariant points at approximately 510 nm and
575 nm. Hues of surrounding wavelengths shift toward
the invariant hue at 575 nm and away from the invariant
hue at 510 nm. Psychophysical measurements of these
hue shifts have recently been accomplished by Horst
and Muis (1969) and Nilsson and Nelson (1971). Their

3

data show maximum shifts occurring in the red end of
the spectrum with lesser magnitudes in other regions.
Both studies confirm the two invariant hues at approxi-
mately 510 and 575 nm and.Nilsson and Nelson suggest the
possibility of another invariant hue in the deep violet
region.

Measurement of saturation changes has not advanced
beyond the judgment technique used by Ball (1964).
His results show that large amounts of desaturation
occur with blue-green and orange hues while little
change occurs in the saturation of blue and yellow hues.
These temporally induced chromatic changes do not appear
to be directly related to the brightness changes, even
though both phenomena occur under the same conditions
(Nelson, 1971; Ball, 1964; Horst and Kuis, 1969).
Problem

 

The first section of this study deals with the
relationship between brightness and spectral composi-
tion of intermittent targets. Data were collected to
determine whether the wavelengths around 510 nm have
some special significance to the production of bright-
ness enhancement. Six targets used in this portion of

the study were all achromatic, but were composed of

different complimentary wavebands. One of these targets
included wavelengths around 510 nm while the others
did not. In addition a full-spectrum target was also

presented. All seven targets were equated in luminosity.

4

According to classical color theory, wavelength
information is lost somewhere in the visual system and
all seven targets should produce the same physiologi-
cal activity at some point since they are identical
in appearance. Thus it would be expected that temporal
effects might be identical on all targets. On the other
hand, Ball‘s results suggest that targets containing
wavelengths around 510 nm might produce greater bright-
ness than the other targets under intermittency. A
Dsychophysical brightness matching technique was used
to provide some data on this question.

In addition to the achromatic targets, temporal
effects on the brightness of seven purple targets were
examined. Purple hues are somewhat unique in that they
cannot be produced with a single narrow band of wave-
lengths. They require the blending of wavelengths from
opposite ends of the visible spectrum. Effects of tem—
poral manipulation on these targets have not yet been
examined. From examination of the results in Ball's
study it might be expected that the brightness of
intermittent purple targets should not reach enhance—
ment level. The brightness of intermittent red hues
conformed quite closely to Talbot level while that of
the blue hues was only slightly above. If temporally
induced brightness is additive, then summing the results
in the individual cases suggests that the brightness of

intermittent purple targets should not greatly exceed

5
Talbot level. The same psychophysical brightness

matching technique was used here.

In addition to the brightness matches each observer
was asked to make judgments on the chromatic quality
of the target. For the achromatic targets this was
restricted to a verbal description of any hue acquired
by the target as a result of the intermittency. It
is well-known that achromatic targets can acquire hue
under certain conditions of intermittency. There is
a large body of literature dealing with this phenomenon
often called “Subjective Colors“ (see Cohen and Gordon,
1949; Gebhard, 1943; and Festinger et al., 1971 for
reviews). It is interesting to note that the temporal
conditions necessary to this phenomenon are generally
reported to be quite similar to those producing the
brightness and chromatic phenomena discussed here.
Reasoning from classical theory again, it would be
expected that any hue induced by a particular temporal
condition should be the same for all the achromatic
targets despite differences in their spectral composi-
tions.

Judgments on chromatic changes in purple targets
were also made. The use of purple hues provides an
opportunity to determine whether temporally induced
changes in the appearance of monochromatic targets
can be used to predict changes in the appearance of a

composite target formed by their addition. It was

6
noted that long-wavelength targets become quite desat-
urated and shift toward yellow in appearance while
short-wavelength targets are only slightly desaturated
and shift toward violet in appearance. Thus when a
purple target is made intermittent, the long wavelengths
should be much less effective in adding to its chromatic
quality than the short wavelengths. Presumably then,
a purple target made intermittent should shift toward
a blue-violet hue and become somewhat desaturated.
Judgments of the direction of hue shift and amount
of desaturation in each of the purple targets were
made for each temporal condition.

The final portion of the study is concerned with
metameric hues. These are hues that have the same
appearance but different spectral compositions. Tradi-
tionally it has been supposed that metameric hues are
possible because of a loss of wavelength information
in the visual system. Classical theory supposes that
physiological activity is identical at some point in the
visual system for two metameric hues. If this is true,
temporal manipulation might result in identical changes
in the appearances of two metmeric hues. Three pairs
of metameric hues were observed to determine whether
temporally induced changes were in fact similar in two

metameric hues.

METHOD
Apparatus

The dual—beam apparatus diagrammed in Figure 1
consisted of two major sections. Section A allowed
blending of two narrow wavebands while section B pro-
duced a circular divided target with the bottom half
variable in luminosity and continuous in time and the
top half intermittent.

In section A the radiation from a lOO-watt con-
centrated arc-lamp is passed through a collimating
lens (L1). H1 is a half—silvered mirror which splits
the beam into two pathways that are blended together
again at H2, another half-silvered mirror. The two
pathways produced (Channel A and Channel B in the
diagram) are equal in length. Each beam passes through
one half—silvered mirror and is reflected by another.
Each is also reflected by one first surface mirror
(either M1 or M2). The division into two beams allows
placement of various chromatic interference filters
and neutral density filters in the channels.

In section B the blended beam is divided in half
spatially by the mirrored prism, P1. One half of this
blended beam is then chopped by a flag on a small pen
motor driven by a Grass S-4 medical stimulator. The
beam was focused at the point of chopping by a lens
system (L2) so as to produce short rise and fall times.

The flag arrangement was calibrated with a Hewitt-Packard
7

8
120 B cathode ray oscilloscope and a cadium photocell.
A check of rise and fall times showed that they were
both approximately two milleseconds in duration.

The other half of the beam is passed through a
circular neutral density wedge. The circular wedge is
connected by a belt drive to a knob within reach of the
observer and a circular dial with 360 degree markings.
Thus, the observer can manipulate the brightness of
the beam by turning the knob and the experimenter can
obtain a reading from the dial. This arrangement was
calibrated with a Pritchard photometer so that per cent
transmission could be derived from the dial readings.

Finally, the two halves of the beam are reflected
off another mirrored prism, P2, and into an eye piece
containing a lens arrangement, L3. The result is a
circular two degree target in Maxwellian view with a
dark line dividing the intermittent top half from the
continuous bottom half. The two halves have identical
spectral composition.

The apparatus was calibrated for the conditions of
this study with a Pritchard photometer. The interference
filters to be used were placed in channels A or B and
measurements of the luminosity at the eye piece were
recorded. In addition visual matches were made using
the cascade method of heterochromatic color matching.

The two sets of data were in substantial agreement.

Q Observer
L
3 \ e/Pe
mirror -__../______23 _____ 5%

l
I
I
I
I
L2 v—————-€P Section B

 

 

 

 

flag ————-—-—'\“$ ==Tl==
I
O
L2 ; I neutral
I I density
: I wedge
mirror ——-—\-—-———' ------- )
I
Pl l Mg
' /
I
I
Hz a e—x-—---—-------—\r
I
I
|
mounting for 'Chafinel

l
I
I
I
I
filters \+ SectionA a;
I
I
l
I
l
I

 

 

 

Channel I

B i

l

I

Ml —3V--—-—————————-)}

_______.__.————-——f I
H1 1‘1 c'>
Light ..——***"”"”flfldflfl" Q?

Source

Figure 1 Schematic Diagram of Dual-Beam Apparatus

lO

Specification 2f Stimuli

 

Three variables were manipulated in this study;
rate, PCF, and spectral composition. The rates used
were 8 and 20 Hz. with PCF's of l/4 and 3/4. A descrip-
tion of the spectral composition of the targets is pre-
sented in Table 1. Two types of filters were used in
production of the targets. The Oriel Narrow Band
Interference filters pass very narrow bands of wavelengths.
The band width at 50% peak transmission was 5 nm and at
1% peak transmission only 10 nm. Wratten gelatin
filters allow much wider bands of wavelengths to pass.
Descriptions of the passbands for these filters as mea-
sured using a spectometer are given in the table.

In the case of the purples composed using the
interference filters, the contributions of the red and
blue wavebands to the total luminosity of the resulting
purple target are shown as percentages in parentheses.
This has not been done with the Wratten filters. In
the treatment of the data spectral compositions are
indicated by the labels given in Table 1.

Observers

 

Four males served as observers. Two of them,
M. E. and M. D., were undergraduate psychology students
and had no previous experience in brightness or color
matching experiments. The two ether observers, R. B.
and A. N., both had extensive experience prior to this

study. All observers had normal color vision and hue

ll

 

 

 

 

TABLE 1 Description of Spectral Composition of Targets
KARE FILT RS MAJOR WAVEBALDS TOTAL
LURE IN A}! C 13‘
Achromatic Targets
A Interference 455—465 nm 195 foot-
565-575 nm lamberts
B Interference 475-485 nm 125 foot—
575—585 nm lamberts
C Interference 485-495 nm 125 foot- {a
605-615 nm lamberts -'
D Interference 450-470 nm
and Wratten 495-505 nm 125 foot- ‘
f 33 600—705 nm lamberts ;
E Interference 450-470 nm E
and Wratten 505-515 nm 125 foot-
# 33 600-705 nm lamberts ;
F Interference 410-460 nm
and flratten 555-565 nm 125 foot-
# 35 640-720 nm lamberts
G None Full Spectrum
Purple Targets
Red Interference 475-485 nmg25%g 30 foot-
Purple 665-675 nm 75% lamberts
Purple Interference 475-485 nm(50%) 30 foot-
665—675 nm(50%) lamberts
Blue Interference 475-485 nm(75%) 30 foot-
Purple 665-675 nm(25%) lamberts
Xylene Wratten-y 33 450-470 nm 120 foot-
Red 600-705 nm lamberts
Wratten Wratten # 34 440-480 nm 190 foot-
D light 630—7?0 nm lamberts
Wratten Wratten f 35 410-460 nm 120 foot-
D 640-720 nm lamberts
Iethyl Wratten # 36 425-450 nm 120 foot—
Violet 650-715 nm lamberts

 

12

Table 1 (cont'd.)

 

 

NALE FILTERS RAJOR HAVEBANDS TOTAL
LUKINANCE
Metameric Targets
Yellow Interference 525-535 nm 160 foot-
Orange A 595-605 nm lamberts
Yellow Interference 495-505 nm 160 foot-
Orange B 565-575 nm lamberts
Yellow Interference 495-505 nm 160 foot-
A 595-605 nm lamberts
Yellow Interference 475-485 nm 160 foot—
B 575-585 nm lamberts
Yellow Interference 505-515 nm 160 foot-
Green A 595-605 nm lamberts
Yellow Interference 475-485 nm 160 foot-
Green B 595—535 nm lamberts

 

13
discrimination as measured by the HRR Pseudoisochromatic
Plates and the Farnsworth lOO—Hue Test.

The two inexperienced observers were given six
hours of practice prior to the actual data collection.
During this period they were presented with various
monochromatic wavebands under temporal conditions iden-
tical to those used in the study. The same procedure
was followed during practice sessions as in the actual
data collection.

Procedure

 

During each experimental session only one spectral
composition was observed. Prior to the data collection,
the observer was adapted to the particular target
chromaticity. All observations were made with the domi-
nant eye, while an eye patch was worn over the other
eye. After positioning and adaptation, the observer
began by making six matches of the steady bottom to a
steady top half. For all brightness matches the method
of limits was used with three ascending and three des-
cending matches.

After the initial set of steady matches, inter-
mittency conditions were observed. Observers were
asked to match the brightness of the continuous bottom
half to the average brightness of the intermittent
top half. The observer was adapted to each intermittency
condition for one minute prior to the beginning of the

brightness matching. The adaption period tended to

l4
eliminate successive changes in the brightness and
chromaticity that occur in the early stages of stimu-
lation with intermittent targets (Ball, 1964). All
four combinations of rate and PCF were used in each
session.
After each set of brightness matches the observer

was asked for an estimation of desaturation and hue r5
":

CO

hift. For the judgment of desaturation a nine—point

scale of none, slight, moderate, extreme, total, and

 

the half stews between was used. This scale has been

used in previous studies and found to be suitable

‘F—V‘L-n
: u
9

(Ball, 1964). End points of the scale were anchored in
a sense by instructing the observers to make a judgment
of "none" if the intermittent target appeared to be
as saturated as the steady half of the target. A judg—
ment of "total" was defined as a complete loss of
chromatic quality in the intermittent target.

A description of the hue shift was given after
the desaturation estimate. The observer was asked to
use the terms redder, greener, yellower, and bluer to
describe the change in hue if any. During the period
of both chromatic judgments the bottom half of the
target was set at the brightness at which the observer
made the last brightness match. Both judgments were
made using the continuous bottom half of the target as
the reference. As a methodological point this is

important since the appearance of the top half is

15

influenced to some extent by the presence of the bottom
half.

Each spectral composition was observed under
each temporal condition is two separate exeerimental
sessions. Each of these sessions consisted of four
intermittency conditions and one steady condition,
totalling thirty brightness matches, four desaturation
estimates, and four hue shift estimates. Thus, for
each condition there were a total of twelve brightness
matches, two desaturation estimates, and two hue shift
responses. The order of presentation of the intermit—
tency conditions within a session was counter-balanced
between the two sessions so that order effects were
balanced. The order of presentation of chromatic

conditions was the same for all observers.

.2;
_‘ I.

 

RESULTS

Brightness Data

 

The brightness matching data of this study are
presented in terms of the brightness index derived by
Bartley and Ball (1965). The brightness index is based
on the Talbot-Plateau level as a standard reference
point. The Talbot-Plateau law states that the brightness
of an intermittent target above fusion frequency is
matched by the brightness of a continuous target whose
luminance is equal to the average luminance of the

intermittent target over time. In other words, the law

 

suggests that the visual system averages the incident
luminosity over time to produce an average brightness
which is then dependent only upon the PCF or portion
of the cycle occupied by the stimulus. This empirical
law has been found to apply over a very wide range of
intermittency conditions (Hyde, 1906) and, thus, is
useful as a standard reference point. At subfusional
rates of intermittency, however, brightness deviates
from the Talbot-Plateau level and becomes dependent
upon rate as well as PCP. The brightness index is a
useful means of describing this deviation. It may be
considered as a measure of the visual system's effi-
ciency in using photic energy to produce brightness.
It is calculated as per cent of the expected Talbot-

level brightness as expressed by the following equation:

16

17
L2

 

X 100% = Brightness Index
(L1) x (ICE)

where L = luminance of intermittent test target
PC' = pulse-to-cycle fraction
L2 = luminance of continuous comparison

target matched in brightness

Figures 2 end 3 present brightness matching data
for the achromatic targets of different spectral com-
position. Each figure presents the data for two obser—
vers. The six different composition are nresented
along the abscissa and labeled as described in the
section on experimental conditions. Brightness index
is presented on the ordinate and each curve on the
graph represents a different temporal condition. E is
the target containing wavelengths around 510 nm and G
is the full-spectrum target.

It is aoiarent that when the ICE was 3/4, brightness
was approximately at Talbot level, a brightness index
of 100, for all targets. The visual system may be said
to be averaging the luminosity over time in this case.
The results with a PCP of 1/4 are quite different. For
the two observers in Figure 2 the brightness is clearly
greater than Talbot level across all the target com-
positions when FCF is l/4. In Figure 3, however, the
curve for each 1/4 ICF condition tends to fall off
toward Talbot level for compositions D, E, F, and G.
There is some evidence to suggest that this result is
due to a change in observer criterion over the course

of the study. It may be remembered that data for each

18

 

085 MD 0 RATE 8 PCF V4

40c” 0 RATE 8 PCF 3/4“
0 RATE 20 PCF V4

350L - RATE 20 PCF 3/4.
soot
zso~ ~
zoo. .
I50“ - ”a
:00» .

 

y
k

50

 

 

 

‘TL

' BRIGHTNESS INDEX
.5
C)
C)
r

 

 

 

D-
D
hr
D
D
h

A e c o ’ E F e
ACHROMATIC TARGETS

Figure 2 Brightness Index of Achromatic Targets
For Observers M. D. and A. N.

l9

 

400

350

300

250
200

ISO

BRIGHTNESS INDEX

IOO

5O

4OOM'
350

300

250

200

Figure 3

50

OO'

50‘

OBS ME RATE 8 PCF I/4
RATE 8 PCF 3/4'
RATE 20 PCF l/4

RATE 20 PCF 3/4.

0
I
o
E]

 

 

 

 

 

 

c o E E e
ACHROMATIC TARGETS

(DT-

{A

Brightness Index of Achromatic Targets
For Observers M. E. and R. B.

r':‘ ’3. ".5. '
.
’ ,
,1

i—Vanfiu 3“ _r Til-'1.

‘r 7917*».

20

 

 

 

 

 

TARGET
C} B A A R C D E F F D G G .
’ 0 ME ‘
500» 0 RB J
b ' ‘ Ira
6 400» J
C) o o
g b I ¢ 0 4
(D a
U) 300L J
LL] 0
Z s
E ' - ' '~
(9 o
E ZOOb .
m a o o
IOO . 2 3 °
T . E”r’ 1
’ u o q
I 2 3 4 5 6 7 8 9 IO ll I2 I3 l4
SESSION
Figure 4 Change in Brightness Index Over Sessions

For Observers M. E. and R. B.

21

condition were collected on two separate occasions.
Figure 4 shows the brightness index for rate 8, IEF 1/4
for the two observers in Figure 3 plotted against the
order of sessions for the achromatic targets. It can
be seen from Figure 4 that the brightness index of these
two observers tended to decline over the course of
the exocriment regardless of spectral composition of
the target.

In any case the results in Figures 2 and 3 do
show that the inclusion of wavelengths around 510 nm
(Target E) is not necessary to the temporal production
of brightness increments. In none of the cases Toes
the brightness index for this target greatly exceed
that for other targets when PCF is 1/4. Nor does res-
tricting the wavebands used to produce the targets apaear
to have any effect. In no case does the brightness
index for Target G (full spectrum) greatly exceed that
for the other targets.

Figure 5 presents brightness matching data for
the purple targets composed with narrow wavebands
centered around 480 nm and 670 nm. At the left end of
the abscissa is the reddest purple (.25 blue, .75 red)
and at the right end is the bluest purple (.75 blue,
.25 red). Brightness index is presented along the or-
dinate. When PCF is 3/4 it is again apparent that
brightness deviates very little from that predicted

by Talbot-Plateau level. With a PCF of 1/4, however,

INDEX

BRIGHTNESS

 

 

 

 

 

 

OBS RB OBS ME
400» - CRATE a PCF V4-
IRATE 8 PCF 3/3
, .RATE 20 PCF v .
DRATE 20 PCF 3/4
300* P '-
/\ L ‘
zoo »- : °\o » / .
IOOL mm _ M ‘
S AN
400. 08 .. oss MD q
- r
300* ' «
2()OL ’ / 'I
P * J W
'OOFI ; ., M 1
I» P +
RP SP RR In SF
HUE
Figure 5 Brightness Index of Narrow-Band

Purple Targets

23

 

 

 

 

 

 

 

 

035 ME 0 RATE 8 PCF l/4

400* It 0 RATE 8 PCF 34'

0 RATE 20 PCF v4

T y. - RATE 20 PCF 3/44

300T A .

I {p I

ZOOL J «I

I. “In I

x :00» v ‘
LU

2 I- nt 1

8 033‘ MD A A A

‘5 400* it «
F.

g I- 1} q
a:

m 300» -- -

I I. J

200* y t

r it I I u a a

ICXDF P a

. n -I

RP BP RP BP
HUE

Figure 6

Brightness Index of Wratten Purple Targets

24
brightness index ranges from 150 to 300 for the four
observers. It is clear that in every case brightness
index is higher with a PCF of 1/4 than with a PCF of
3/4 though in no case does the increment in brightness
reach enhancement level (greater than the brightness
of the target when it is continuous in time).

Since luminance was quite low for the narrow-
band oursles, at the low end of the range in which
brightness enhancement has been observed, similar
data was collected using Wratten gelatin filters to
produce purple hues. These filters allow much wider
bands of wavelengths to pass and thus, higher lumino-
sities can be obtained. The data are shown in Figure 6
with red purple to blue purple presented from left to
right along the ordinate. Only three observers parti—
cipated in this portion of the study. The data here
generally agree with that collected using the narrow-
band purples. With a PCF of 3/4 brightness again devi—
ates very little from Talbot level. With a PCF of 1/4
brightness indices for two of the three subjects again
range from 150 to 300. The third subject, M. E.,
differs dramatically here. Observer M. E. is the
subject who was suspected of changing criterion as
shown in Figure 4.

In any case it aslears from Figures 5 and 6 that
increments in brightness do generally occur when a

PCF of 1/4 is used with low rates of stimulation and

25
purple hues. Little change in the size of the incre-
ment occurs with changes in the degree of redness or
blueness of the hue. One rather unexpected result
throughout this data is the effectiveness of rate 20.
Brigitness index is as high or higher with a rate of
20 Hz. as with a rate of 8 Hz. In previous investiga—
tion lower rates were found to be clearly more effec—
tive (Ball, 1964).

Chromatic Data

 

. «4.! ‘A‘ r it ‘, uufiiiij
I

Table 2 contains hue data for the various achro-

matic compositions under the four conditions of inter—

 

‘rnv07 ~, m

mittency. The observers were asked to report any hue
that might be acquired by the targets. The four tem-
poral conditions are listed across the top of the table
and the seven target compositions occur in the left—
iand column. Within each cell the responses of each

of the four observers is listed. Three of the four
observers consistently report a blue or purple hue

when the TCF is 1/4. Observer M. E. is again deviant
reporting a purer white as a result of the intermittency.
When the ICE is 3/4 several reports of yellow occur,
though observers more often report no hue acquisi-
tion here. The most important point here is that

for all of the observers the spectral composition of
the target appears to have no effect on the hue re-
ported.

Table 3 presents hue shift data for each of the

 

 

TABLE 2 Hues Acquired by Achromatic Targets
White Rate 8' Rate 8 Rate 20 Rate 20
Target Obs. PCF 1/4 PCF 3/4 PCF 1/4 PCF 3/4
ME whiter none whiter whiter
A MD blue whiter purple whiter
AN blue none purple none
RB violet none violet none
ME whiter none whiter whiter
B MD blue whiter purple whiter
AN blue none violet none
RB violet none violet creamy
ME whiter none whiter whiter
C MD purple yellow purple whiter
AN purple none violet none
RB violet none violet none
ME whiter none whiter whiter
D MD purple none purple none
AN blue yellow purple none
RB purple whiter violet whiter
ME whiter whiter whiter whiter
E MD white none purple whiter
AN violet yellow purple none
RB violet none violet none
HE whiter yellow whiter greener
F MD purple none purple purple
AN blue yellow purple yellow
RB violet none violet none
ME blue none purple none
G MD whiter green none yellow
AN blue none violet none
RB grey none violet none

 

27

 

 

TABLE 3 Hue Changes in Purnle Targets as a Result
of Intermittency
Rate 8 Rate 8 Rate 20 Rate 20
Hue Obs. PCF 1/4 PCF 3/4 PCF 1/4 PCF 3/4
ME bluer none none none
Red MD bluer bluer blue bluer
Purnle AN blue none blue bluer
RB blue violet blue violet
ME bluer none none none
MD bluer blue blue bluer
Purple AN bluer none blue bluer
RB violet violet blue violet
ME none none none none
Blue MD bluer blue blue bluer
Purple AN bluer none violet none
RB violet violet violet violet
Wratten ME none redder none none
Xylene ED bluer bluer bluer bluer
Red AN violet none violet none
ME none redder none none
Wratten AD bluer none blue none
D(light) AN bluer none violet bluer
ME none redder none none
Wratten ND blue redder blue redder
D AN violet none blue none
Wratten ME none redder none redder
Methyl MD blue none blue bluer
Violet AN blue none blue bluer

 

28
purple hues. Observers were asked to resort any chrnge
in hue that occurred in the target by indiccting the
direction of the shift. Temporal conditions are again
listed across the to» of the table and the name of the
hues occur in the column at the left. Within each
cell the responses of each observer are listed. It
is clear that with a PCF of 1/4, all of the purple
hues generally become bluer in quality. With a TOP of
3/4 a shift tovard blue is often reported but reports
of no hue shift are more frequent.

Figures 7 and 8 present the desaturation estimates
for the narrow—band purples and drrttcn purples respec-
tively. The data represent judgments made on a nine—
point scale which is presented on the ordinate. Hues
are again presented along the abscissa with each curve
representing a different temporal condition. It is clear
that greater amounts of desaturation occur with a PCF
of 1/4 than a PCF of 3/4. The difference is larger
with the Wratten purples than with the narrow-band
hues. No change in amount of desaturation appears to
occur with change in the quality of the purple. Though
the data on chromatic changes presented here are only
qualitative, they demonstrate that a shift toward blue
occurs along with a significant amount of desaturation
when purples are temporally manipulated.

Table 4 contains judgment data on saturation and

hue for the yellow-orange, yellow, and yellow-green

I}

TOT

EXT '

MOD -

SL'

TOT

DESATURATION

EXT

MOD.

SL

Figure 7

29

 

035 ME noes R8
0 RATE 8 PCF V4
.. 0 RATE 8 POPS/4
o RATE 20 PCF l/4
- RATE 20 PCF 3/4‘

 

 

f

v

 

 

 

HUE

Desaturation Estimates for Narrow—Band
Purple Targets

 

 

 

 

   

 

 

 

 

 

TOT‘ W 0 RATE 8 PCF V4 ‘
0 RATE 8 PCF 3/4
T " 0 RATE 20 PCF V4 ‘
I RATE ZOPCF 3/4
EXTT ‘* ‘
. a .
MOD‘ 4»
. 1L 4
5L. ' 1r 4
D h ‘
2
E. N D d} 1
g i a - i - 4 . .
'2 TOTb OBS AN H 033 MD A
‘4
m . H 1
LU .
O
' EXTW " .
b J! 1
NKDD' ‘r 1
b n J
SL.r 1* J
P U_R/D\£ “ 4
N ' l V .
RP t ‘ BAP RP ' ER
HUE
Figure 8 Desaturation Estimates for Wratten Purple

Targets

31

TABLE 4 Comparison of Changes in Chromaticity of
Hetameric Hues

 

 

Rate 8 Rate 8 Rate 20 Rate 20
PCF 1/4 PCF 3/4 PCF 1/4 PCF 3/4
Observer A B A B A B A B

 

ORANGE-YELLOW

Hue Shift
ME 0 O O O ‘9 O O 0
LED ‘- 9 ‘— 0 (- ‘- 0 ‘-
AN 9 O 0 0 e— es «- O
Desaturation Level
ME 9 7 l 4 5 5 3 1
MD 5 7 l 3 3 3 3 2
AN 4 8 3 l 4 6 l l
YE L L O W
Hue Shift
ME 0 O O O O O 0
ED 0 O O 0 e- e e- e
AN 6 6 O 0 +- e O 0
Desaturation Level
KE 9 9 1 1 9 9 3 1
MD 7 9 2 2 5 5 l 1
AN 5 5 l l l l l l
GREEN-YELLOW
Hue Shift
ME 0 0 e- e O O 4—
MD 9- e e- e. e- e- e. (—
AN +- 6- O -a 6- e- 0 O
Desaturation Level
MD 5 5 l 1 7 7 l 2
AN 5 5 l l 7 6 2 l

 

FR .215).

32
metameric pairs respectively. The major emphasis in
this section of the study was to determine whether
the temporal programming of the target would differen-
tially effect the two members of a metameric pair.
Each rate and ?TF is listed again across the top of
the table. Columns A and B under each condition contain
the data for the two spectral compositions which pro—
duce the same hue. Observers are listed along the
left—hand edge of the table. The uppermost section
under each hue contains hue shift judgments and the
lower section the desaturation estimates for each tem-
poral condition and observer. Arrows pointing to the
left indicate a shift to a hue of shorter wavelength
and arrows pointing to the right indicate a shift
toward a hue of longer wavelength. N mbers indicate
the level of desaturation on a nine—point scale.
Comparison of columns A and B under each hue and condi-
tion shows that no consistent differences occur between
the two members of a pair. It appears that if there
are differential temporal effects, a more sensitive

method must be used to detect them.

33
DISCUSSION

The brightness matching data of this study suggest
that temporally induced brightness changes are not
dependent in a simple additive way upon the wavelength
composition of the stimulus. The results of Ball's
1964 study on spectral stimuli show that the only
spectral stimulus whose brightness reaches enhancement
level under temporal manipulation has a dominant wave- fink
length of approximately 510 nm. However, the results
of this study show that under the same temporal condi-

tions an achromatic target which includes wavelengths

 

wr'.‘ - _

at 510 nm does not appear brighter than achromatic
targets which do not include wavelengths at 510 nm.

Thus, the results tend to suggest that spectral com—
positions of the target has no direct effect on temporal-
ly induced brightness of achromatic targets though
chromatic appearance of the target is important. This
conclusion must be presented rather tentatively, however,
since the data of two of the subjects do show some

change with spectral composition of the achromatic
target. Earlier in this paper it was suggested that a
change in the matching criterion of these subjects
occurred during the course of the experiment. It is
well-known that variance within a single subject's

data is quite high in brightness-matching tasks under
these conditions (See Schneider, 1971, for a discussion)

and changes in criteria do occur. Evidence for the

34.
'criterion change was presented in Figure 4, but no
explanation can be offered as to why it occurs. In

any case it apwears that wavelengths around 510 nm are
not necessary to temporally induce brightness in
achromatic targets.

The brightness matches for intermittent purple
targets also show that under some conditions (PCF of
1/4) brightness is two to three times Talbot Level
Brightness. This finding was unexpected since in
Ball‘s data (1964) brightness of monochromatic stimuli
from either end of the spectrum conform quite closely
to Talbot-level predictions. Temporally induced brigh -
ness does not appear to be additive in a simple way.

The results as a whole seem to suggest that brightness
enhancement is directly related to the chromatic appear-
ance of the target but not the spectral composition

of the target. Further work on this problem may provide
information on the relationship between brightness and
hue.

Chromatic judgment data presented in the results
section showed that all achromatic targets acquired
the same hue as a result of the temporal programming.
This result is what might be exeected on the basis of
classical theory, which supposes that physiological
activity within the visual system would be identical
for all achromatic targets of equal brightnes . It

was also found that changes in the chromatic appearance

35

of the purple targets agreed with what might be pre-
dicted from temporally induced chromatic changes in
the individual monochromatic wavebands of which the
target was composed. These results are relevant to
the question of the relationship between the hue-
saturation shifts and the Subjective Colors phenomena.

Nelson (1971) has suggested that both phenomena may

 

result from the same mechanism even though elicited
by different spectral conditions. Subjective Colors
may be the result of temporal events being selective ;

with respect to hues. That is, “Subjective Colors” ;

p91 .

may simply result as the summation of the effects on
individual hues. The fact that changes in purple hues
appear to be the result of just such a process lends
support to Nelson's hypothesis. These results suggest
that the chromatic appearance of intermittent targets
may be vector additive as it is in the color space of
classical color research. Further work in this direc-
tion may be profitable.

In the final section of this study, it was found
that temporally induced changes in chromatic appear-
ance appear to be similar for metameric hues. This
conclusion is by no means firm since the data consist
of a small number of verbal judgments. It does however
lend further support to trends exhibited in all the
data on chromatic changes. Results as a whole are

consistent with the hypothesis that wavelength information

is lost in the visual system.

BIBLIOGRAPHY

BIBLIOGRAPHY

Ball, R. J. (1964) An Investigation of Chromatic
Brightness Enhancement Tendencies. American Journal
of Ontometry and Archives oi American Academy of
Cntometry, 41(5): 333—?61.

 

 

 

 

 

Ball, R. J. and Bartley, s. H. (1965) Effects of
Temooral haninulation of Photic Stimulation on
Perceived Brightness Hue and Saturation. Ameri-
Journal 2£.9239metry and Archives 2: American
Academy 9: Optometry, 42: 573—581.

 

 

 

 

Bartley, s. H. (1951a) Brightness Enhancement in
Relation to Target Intensity. Journal 9: Ps -

chology, 32: 57-6?.

----- (1951b) Intermittent Photic Stimulation at
harginal Intensity Levels. Journal 2: Psychology,
32: 217-223.

 

----- (195?) Brightness Comparisons When One Eye is
Stimulated Intermittently and the Other Steadily.
Journal 2£_Psychology, 34: 165-167.

 

---—- (1957) Light Adaption and Brightness Enhance—
ment. Percention and Motor Skills, 7: 85-92.

 

Bartley, S. H. and Nelson, T. M. (1960) Certain
Chromatic and Brightness Changes Associated with
Rate of Intermittency of Photic Stimulation.
Journal 2: Psychology, 50: 393-33 .

Bourassa, C. M. and Bartley, S. H. (1963) A Study of
the Possible Effect of Differential Light Adaption
on Measures of Brightness Enhancement. Journal 2:
Psychology, 56: 435-440.

Cohen, J. and Gordon, D. A. (1949) The Prevost-Fechner-
Benham Subjective Color. Psychological Bulletin,
46: 97-136.

Festinger, L., Allyn, M. R., and White, C. W. The
Percention 2: Color with Achromatic Stimulation,
Vol. II. pp. 591—611.

 

 

37

3 8

Gebhard, J. J. (1943) Chromatic Phenomena Produced
by Intermittent Stipulation of the Retina.

Journal of ggserimcntal Psychology, 33: 387-406.

 

 

Horst, G. J. C. van der and Luis, H. (1969) Hue Shift
and Brightness Enhancement of Flickering light.
Vision Research, 99: 953-964.

 

1‘ ‘\ TN

Hyde, s. r. (1906) Talbot's Law as noniied to the
Rotating Sectorcd Disk. Bulletin of Bureau of
Standards, 2: 1-?2.

 

 

Kazuk, J. D. and Bartley, S. H. (1969) Brightness
Hatching Under Intermittent Stimulation.
Journrl of Esychology, 71: 281—289.

 

Nelson, T. K. (1971) Stimulus Intermittency and Color
Studies in Review. Journal o£_Gcneral Psychology,

 

Nelson, T. t. and Bartley, s. i. (1961) The Role
of ICE in Temnoral fianinulation of Color. Journal
of Psychology, 52: 457—477.

 

Nelson, T. k., Bartley, S. H., and Jewell, R. L. (1963)
Effects upon Brightness Produced by Varying the
Lengths of the Null Interval Senarating Successive
"single" Pulses: Sensory Implications of the
Alternation of Resnonse Theory I. Journal of
Psychology, 56: 99—106.

 

Nilsson, T. H. and Nelson, T. m. (1971) Hue Shifts
Produced by Intermittent Stimulation. Vision
Research.

Schneider, C. W. (1971) The Problem of Brightness .
Matching and Observer Criterion at Low Rates of
Intermittent Photic Stimulation. Journal of
General Psychology, 84: 56-59.

 

Valsi, E., Bartley, S. H., and Bourassa, C. (1959)
Further Maninulation of Brightness Enhancement.
Journal of Psychology, 48: 47755.

 

Nov 12 3971.

"l
[I
I
II'
I.
III
I
l
l'
I
l
l

iIHIWHIUHI

19 7254