The 1itPaneselburnatofItsychonomicScience 1982,Vol.1,No. L45-50
PerceivedLength
ofMoving
Figures:
The
Effect
ofLuminance
Relationship
Between
Figure
andBackground
upon
Reduction
Phenomenon
Yoshiaki
Kiznazawa
NAKAJIMA
Uitiversity,
A rectangular figuremoving horizontallyat a constant speed shews a decreasein its
perceived Iength parallelto the directionof movement. This study concerns the effect of
the luminance relationship between figure and backgro"ncl ttpon this phenomenon. The
perceived lengthina stationary state was also measured fordifferent exposure times cor-responding to the speeds used, in order to obtain the relative differencebetween the per-ceived lengthoi a figureina stationary state and that of the figurein a dynamic state. Three linearfiguresoi 1,5e,3" and 60 were used; all of them had a constant heightof
O.250. The stimulus figure was attqched to a horizontallymeving belt. Speedsof 280/s,
35M!sand 420/swere used. The pointsof subjective equality of the stimulus figureswere obtained by the method of limits.Eightstudents served as subjects. The same reduction tendencies were shown at the differentspeeds despitethedifferentluminance relation$hips. Key words: rnotion, perceived length,reduction penomenon, linear figures,Iuminance
relationships, figureand background.
The
perceived shape of a rectangularfigure
movinghorizontally
at a constant speed wasinvestigated.
The
perceivecl length of thisfigure
parallel
to thedirection
of movementbecomes
shorter as the speed increases,up toa certain speed
(Tanaka,
1943a,1943b,1944;Tanaka
&
Nakajima,
1970).Morinaga,
No-guchi andOhishi
(1963)
confirmedthis
tendency
and named this decrease in the perceived
length
of a moving figure"a reduction phe-nomenon ".
On
theother hand,Ansbacher(1944)
reportedthatan illuminatedarc-line
produced
by
trans-mitted
light
on ablack
disc
appeared shorter when rotated at a constant speed around a centralfixation
point.
Twentyyears
later,
using refiectedlight,
Marshal1
andStanley
(1964)
reversed theluminance
relationshipbetween
arc and
background
and found a reversed effect :i.e. ablack
arc on a whitebackground
tended to appear
longer
when rotating than when stationary, although a white arc on ablack
background
tended toappear shorterin
line
withAnsbacher's
results.Although
Ansbacher's
experimentdiffered
from
Tanaka's
experimentsin
type of motion(i.e.
the former used circular motion and theIatterusecl rectilinear metion), we rnay guess
that
the shrinkagephenomenon
Ansbacher
found in a retating arc is
fundamentally
thesame as the reduction phenornenon Tanaka
found
in
ahorizentally
moving rectangle. The reason isthatthere are two elements commonto
Ansbacher's
and Tanaka's experiments:(1)
the stimulus
figure
has
alonger
actuallength
parallelto the
direction
of movement than thelength perpendicular tothe
direction
of move-ment,(2)
theperceivedlength
ef the stimulusfigure
appeared shorter para!leltothedirection
of rnovement.
When
the results of Tanaka's experlment(1943a)
are compared with the results ofanother of
his
experimentsin
1944,
we seethat the reduction phenomenon can
be
found
in
both
types ofluminance
relationshipbe-tweerr
figure
andbackground,
i.
e. with ablack
figureon a white
background
or with abright
light
figure
on adark
background.
Completely・
46 The JapaneseJournalof
different
apparatuses were used in theseex-periments: the one used in 1943a was
based
on theprojection
of afigure
attached to a movingbe!t
upon a grey sQreen by an opaqueprojector exploiting surface refiection; the
other, used in 1944, was
based
onlight
ire-flected
from a rotating surface mirror which created a movinglight
figure
by
the change of angles of incidenceand refiection. Thoughthere
is
adifference
between
these systems, we may assume that theluminance
relation-shipsbetween
figureand background of these experiments are the reverse of each other.However,
in,
consideration of the results ofMarshall
and Stanley experiment concerning perceived arc-line which showed a reversedeffect
in
reversedluminance
relationshipbe-tween 'figure and
background,''I
wanted to confirm clearly 'what effect theIuMinance
re-lationship
betWeen
figure
andbackground
hasupon the redugtion' phenomenon of a
horizon-tally moving
figure,
in a closely controlledcondition so thata preci6ecomparison may
be
possiblewith the same apparatus. This
prob-lem isthe firstconcern of the present stUdy.
There
isanother purpose ofthisstudy.The
measurement method in the Marshall and
Stanley
experiment isdifferent
from thatem-ployed
by
Tanaka,In
hisexperiments, Tanaka used a method of ,reproduction which required the subject todraw
what he had seen ona sheet of whit,e paper. The results repprtedin
h.is
papers wereindicated
with the,index ・aobtained in the calculation
by
the following equation :x:=100+(lt'h)1(L/U>
(1)
where ldenotes,thelength of, the,reproducedfigure,
h
denotes
the height of the reproducedfigure,
L denotes the length of the originalfigureand' ff
denotes
theheight
of theoriginalfiginre.
As.we
can' seefrom
this equation,Tanaka investigatedthereduction phenomenon
taking
both
perceivedheight
andperceived
length intoconsideration.
On
theother hand,Marshall・and
Stanleyuseda method of/ comparison・in wh'ich the'subject
was required
to
state whether a straightline
inthe・center
(variable
stimulus) 'appeared tobe
longer
or $horter than the arc movingaround
it,・They
calculated the paint ofsixb-Psychonomic Science Vol. 1,No, 'l
/t
jective
equality of only the length of this arc,and
did
notyefer
tothe change ofheight
at'
all. ' '
We
carinot.directly c6mpare the resultstained by Tanaka and the results obtained by
Marshall
andStanley
because
of thedifferences
ef measurement method as described above.
So,
I
hope to provide results that canbe
directly
compared to the results ofMarshall
and
Stanley.
Experiment I
The purpose of this experirp. ent isto
tigate the
perceived
length
of,linear figures,, rnoving horizontallyat,constant speeds, and - the effect of the luminance relationship between
figure and
background
upon this perceivedlength..
' 'Method'and
ProcedUre
The basicarrangement of the appafatus is shgwn
in'Fig.
1:
A
.white(or
black)
stimulus'figure
was attached to the center of a
trans-parent'
belt
which wasdfiven
by
aTi electricT]-Al.-"(ss B ISCII))
I
i
i
l I'15e]n,1 ・i
'l
t
h
Sh SC3 SC2 Ti Fig,(illS)
1. Basic arrangement of the apparatus. S: subject. Sgi: screen used as a
back-ground. Sc2:$creen with a 20cm
(length)
by I6cm
(height)
rectangular aperture.Scs: screen with a viewing hole, B:
transparentbelt. M:'motor. D:,rotating
drum. Tr:Volttransformer.
St:strobo-scope used t6 regulate the,speed: ・Sh: shutter. Ti:timen /.
Y. Nakajima: PerceivedLength ofMoving Figures 47
motor, the speed of which could
be
changedby
a volt transformer, The $peed wasregu-lated
by
reference to stroboscepe-type systemusing a small neon
light
and adotted
line
onthe
drive
drum.
A
rnoying stimulus figure was observedthrough a rectangular aperture 20cm
<10V)
inlength
by
16cm
(80)
in height which was cutfrom a
black
(or
white) screen placedjust
before the moving
belt.
A black(or
white)background
screen was putjust
behind
thetranspatent belt, The height of a!1the
figures
used was O,5cm
(O.250),
and thelengths
were3cm
(150),
6cm
(30),
and12cm
(60).
Threespeeds of
56cmis
(28e/s),
70cm!s
(3501s)
and84cmfs
(4201s)
were used.The
luminance
ofthe white 'figures
(or.
backgfoiind)
was about20.lcd/m2'and
that of theblack
figures
(or
background)
was about1.0cd/m2.
The
view-ing
distance was 115cm.Monocular
vision'
was used, ' ''
The
subject satin
adark
room and wasrequired tofixupon the center of the
back-ground screen when the shutter
just
behind
the viewing
hole
was opened.After
theshut-terwas Qpened, a moving tlgure appeared on
the
background
screen and movedhorizontaliy
to the left.Immediately
after the mevingfigure
disappeared,
a eurtain above the20
cmby 16cm regtangular aperture dropped to cover
the aperture and at the same time reveal a comparison figurethat had
been
attached abovethe
frame
of thisapperture. This comparisonfigure was not visible until this time,
The
comparison figures were stationary linearfigures,
with a fixed height of O,5cm(O.250)
and variablelen'gth.
The
subject was requiredto $tate whether the stationary comparison
figureseemed to
him
tobe
longer
or shorterthan the moving
figure.
The
judgement
of equality was used. The length of thecem-parisen figure was varied by increments of
O.2cm
(O.10).
The method oflimits
was used.Eight
undergraduate students majoring. inpsycho!ogy served as subjects.
All
of themhad normal visual acuity er corrected nermal
acuity. They were trained forthe'observation
and the proeedure inadvance,
Results
The
point
of subjective equality of each subjectfor
each speed of each figure was calculatedfor
both
luminance
relationships of a whitefigure
on' ablack
background
and ablack
figure on a whitebackground.
The
means of those scores werefurther
calculated foreach condition and are shownin
Table''1.
Despite the restricted range of
figures
used,it
is'possible
te estimatefrom
thistabiethatthe pointsof subjective equality
for
both
types of theluminance
relationships show the sametendency of
length
reduction'when the speedincreases,though the amount of reduction
is
not greatin
either case,
These
resultsdon't
completly coincide with whatMarshall
andStanley
found intheirex-periment with the rotating arc: that is,a
tendency toward reduction of the per¢eived
length
of a white'arc on a black background, and increaseof theperceived
lengthef a black arc on a whitebackground.
Table 1, Mean perceived lengths
(cm)
two luminanceoEthree moving figuresat
relationships
(n
=8)three speeds fur
Luminance
relationship Whitet.figureandblackbackground Blackfigureandwhitebackgreund
Speed 156 cmls/1 Figure3cm 6cm 12cm MeanSDMeanSDMeanSD3.01 ,095.70 .2611.11 .79 70cmts 2.88 .2・15.44 .5910.72 .97 84cmfsf:56 cmfs
L
2.80 ,105.00 .64IO,54LOIl・
t 3.01・ .125.79 .2711,45 ,51 70cmts 2.96 .165.75 ,3011.21 ,72 84crnls i 2.82 .2・45.58 .3411.02 .79Luminance
relationship Whitefigureandblackbackground Blackfigureandwhitebackground
48 The
Japanese
Journal
of Psychonomic
Experiment
II
There
was a necessity toinvestigate
theperceived
length
of stationarylinear
figures,
when observed for
different
exposure timescorresponding to the speeds used
in
Experi-ment
I,
in
order toobtain the relativediffer-ence
between
the perceivedlength
of afigure
ina stationary state and that of the figurein
a dynamic state.
Method
amdProcedure
The same three stimulus figuresas used in
ExperimentI
and three exposure times cor-responding to three speedsin
Experiment
!
were used
in
Experiment
II.
Exposure
times used were .36s, .29s and.24s.
The
calculation of these times wasbased
upondividing
thelength
(20cm)
of the rectangular aperture in which thefigures
rnoved
by
their speeds. The exposuretime
was varied
by
a timer-controlledshutterbe-hind
the viewing hole. The subject viewed a stationaryfigure
within the aperture frameuntil the timed shutter closed.
Immediately
after the shutter closed and the curtain abovethe 20cm
by
16cm
rectangular aperture waslowered,
revealing the variable comparisonfigure,
theshutter was reopened.The
subject was then required tostate whether thecom-parison figurewas longer or shorter than the
figure
which waspresented
for
a restrictedtime.
Other
condition were the same as inExperiment
I.
The
sarne eight students served as subjects inExperiment II.Table 2. Mean perceived lengths
(cm)
of three corresponding to three speeds for two.24 Exposure timet .36s .29s Figure Mean 3cm SD Mean 6cm SD Mean 12cm SD 3,05 .115.91 .3111.77 .42 3.02 .095.93 .2311.85 .43
Science Vol. 1,No. 1
Results
The
results obtained are shownin
Table
2.
This
table obviously shows thatthe pointofsubjective equality of the length of each
sta-tionary figure presented for each exposure
time
is
almost
equal to the originallength
in
both
luminance relationshipsbetween
figure
andbackground.
Discussion
Though
results obtainedfrom
ExperimentII
do
not appear toindicate
a clear change inperceived
length
of a stationaryfigure
pre-sented
for
a short exposure time, itseemedbetter
to use anindex
ip
which expresses the relativedifference
between
theperceived
length of a figurein a stationary state and that ofthis figure in a dynamic state. Therefore,
from
both
results of Experiment Iand Experi-ment II,the index ¢ was calculated by thefollowingequation :
ip-(A-B)fA
(2)
whereA
denotes
the perceived length of afigure
in
a stationary state andB
denotes
theperceived
length
of this figureinadynamic
state.
The
re$ults of thiscalculation foreachspeed are shown
in
Fig. 2. The statisticalP-values
of each figurefor
each luminancerelationship fortesting the hypothesisof
di=O
(i.
e.A=:B)
were calculated foreach speed andthose for testingthe
difference
between
thetwo luminance relationships were also
calcu-lated.
The results obtained are shownin
stationary figuresfor three exposure times
luminance relationships
(n=8)
s .36s .29s .24s 3 5 11 .03.11.91.17.79.42 3.06 .09 6.02 .1311.78 .51 3.03 .10 6.05 .0711.88 .40 3.06 .09 6.05 .0911.86 .38t
The calculation of exposure times was basedaperture inwhich the figr2resmoved by theirupon
dividingthe speeds.
Y.Nakajima: Perceived Length of Moving Figures 49
('PX
IIi
O).--I i'{
gL/
iileh.tsiiifillts
] ,,:.;l,,,,,ild '5Xilli
`')Fi""]'`''ti`'-M ,/ " (`'`'X:/l
,iF O) Figurc:i2cm:'
H{Rlr,fl,C::ttl:r()und
ii
/
il
' t .n10 1'O
/
lo,n p.-..f'i 9 g / [) J .:・,
..,・/lili
,/"
l
ily .tiX' ,,/ 5 JJt n ' f xt
L-
/,,.
.4 ,,:- ・ 4i・l
,・"'.・'
;,
;
deUgt 70
",l
g56
t
tt
84 056 7C)84Speecl(cm,,'s) Speed
(cmi's)
Speed(cm!s)Fig.2. The disof three figuresat three speeds for two luminancerelationships. The
di
expresses the relative differencebetween A, i.e.the perceived length of a figureinastationary state, and B, i,e,that of this figure ina dynamic state:
e=(A-B)IA.
Table
3
andTable
4.
From
thep-values
in
the question arises as to why the reductionTable
4,
we may say that the clear effects of phenomenonis
found
without referenceto
the
luminance
relationshipbetween
figureandluminance
relationshipbetween
figure
andbackground do not exist,
but
as is obviousbackground
in horizontallymoving figures,infrom
Table
3,
there are significantdifferences
spite of the existence of thefact
that,in
between
the perceivedlengths
offigures
ina rotating arcs, thisreduction phenomenonis
dynamic
state and those of these figuresina found only under the condition of a whitestationary state for
both
luminance
relation-figure
on ablackbackground. Themosteasilyships.
Comparing
Tables
1
and 2, we may proposed answer maybe
that the reductioninterpretthese significant
differences
asbeing
phenomena seen in rectilinear motion and incaused by the reduction phenomenon. Then, circular motion have no relation to each other:
'
Table 3. The
p-valuest
of three figuresat three speeds for two luminance relationships ・ fortesting the hypothesisofe=O
(i,e.
A=B)ttLuminance relationship SpeedFigure 3cm 6cm 12cm
White figureand blackbackground 56cmis .4<p<.5
(t=,8936)
.05<p<.1(t.:,1.9711)
.02<p<.05(t=2.0850)
70crnls J .2<P<.3(t=1.2632)
.Ol<p<.02(t=3.1358)
.Ol<p<.02(t=3.1800)
84cmts .OOI<p<.(t=3.7650)
.OOI<p<.(t=4.1123)
.OOI<p<.(t=4.0340)
Ol Ol OlBlackfigureand white background 56cmis .4<P<.5
(t=.7650)
.02<p<.05(t=2.6170)
.05<p<.1(t=1.9504)
70cm/s .5<p<.6(t=.6643)
.02<p<.05(t=2.5000)
.02<P<.05(t=:2.5045)
84cmls .05<p<.1(t=2.1717)
.Ol<p<.02(t==3.442O)
.02<p<.05(t==2.8612)
tttThe
P-values
See Fig. 2 forwere
obtained from the results of definitionsof
di,
A and B.50
Table 4.
speeds
The Japanese
Journal
ofThe P-valuestof three figuresat three
between two Iuminance relationships
Psychonomic Speed 56cmls7ocm/sI84cmts Figure 3cm 6cm 12cm .9<P<1.0
(t=.1161)
.8<P<.9(t=.1400)
.2<P<.3I(t=-1,276s)
P=.6(t=-.5490) .2<P<.3(t=-L1696)
.1<p<.2(t=-L6041)
.9<P<1.0(t=.0934)
.05<p<.1(t=-2.2139)
.1<P<.2(t.."1.840e)
1t
The P-valueswere obtained from the resultsof t-testsbetween the corresponding groups
(cif=8-1=7,
two-tailed tests).they
have
only externally sirnilarappearances,coming from
different
sources,However,
whenwe lookat the restricted range of
figures
andspeeds used
here
and the present general stateof the poor
knowledge
on the reduction phe-nomenon, itseems too dangerous tojump
tothe conclusion that the
discrepancy
reportedhere
inlength
changebetween
horizontally moving figures and rotating arcs suggests mutualindependence.
Summary
The
perceivedlength
ofdifferent
linear
figures
was measured under differentmoving speeds and differentluminance
relationshipsbetween
figure
andbackground.
The
method of limits was used in measurement.The
stimulus
figure
was attached toabelt
movinghorizontally
at a constant speed.The
perceivdScience VoL l,No,1
length
in
a stationary state was also measuredfor
different
exposure times corresponding tothespeeds used,
in
order toobtain the relativedifference
between
the perceived length of afigure
in a stationary state and that of thefigure in a
dynamic
state.The
same reduc-tion tendencies were shown atthe
different speedsdespite
the
differentluminance
relation-ships.References
Ansbacher, H.L. 1944 Distortionintheperception
of real rnovement. Iburnal
of
ExperimentalRsychology,34,1-23.
Marshall, A,J. & Stanley, G. 1964 The apparent length of lightand dark arcs seen peripherally in rotary motion. Australian
lburnal
of
chegQgy, 16,120-12s.Morinaga,S. Noguchi, K, & Ohishi,A, 1963On the reduction phenomenon in pereeption of moving figures.fournal
of
the Cellageof
Arts and Sciences,Chiba Uhaiversit],,4,'19-23.(In
Japanese
with English summary)Tanaka, Y, 1943a Perception of moving figures
(1),
lapanesefournai
of
Rsychology,17,333-352.(In
Japanese)
Tanaka, Y. 1943b Perceptionof moving figures
(2).
laPanese
JbunutJ
of
ts'cholaev,
17,443-458.(In
Japanese)'
Tanaka, Y, 1944 Perception of moving figures
(3).
fapanese
fournal
of
llsychogogy,19, 1-11,(In
Japanese)
Tanaka, Y. & Nakajima,Y. 1970 Perception of moving figures:1,On thereduction phenomena.
laPaneseRsychologicalResearch, 12.172-175.