ATtythsceeenpsmc
AE
AHM
ITsmcpsamcbae‘du
PodaoptofrD©h
Yo
ABSTRACT This study aiypography elehe conveyance
studies were coconducted throuextracted to fexperiment. Seexplore the enamely animatiparticular emotstepping stonmessengers as communication
Author KeywoEmotive text; in
ACM ClassificH.5.2. InformaMiscellaneous
NTRODUCTIOThe maturing osignificant parmessaging [9]. communicationphones to smsmartwatches. advent of new messaging hcommunicationbutton messeapplication (apexchange limitYo’ app is a
does nothing busers, as show
Permission to make or classroom use is distributed for profiand the full citation owned by others tpermitted. To copy o lists, requires pri
from permissions@aDIS 2016, June 4–8,© 2016 ACM ISBNhttp://dx.doi.org/10.
o!: EnricMess
ims to investements in single of emotionalonducted: Firstugh which eighfacilitate the econd, an empeffects of kinion type, speedtions. The stud
ne for furthean alternative
nal aspects.
ords nstant messagi
ation Keywordation interfaces
ON of technology rt of interperIt is worthy o
n tools have shmart devices, Such environmways of com
has expanden by drasticallenger is anpp) for mobileted texts or sirepresentative
but send a one-n in Figure 1.
digital or hard copigranted without fe
t or commercial adon the first page. C
han ACM must botherwise, or repubior specific permisacm.org. , 2016, Brisbane, A
N 978-1-4503-4031-1145/2901790.2901
ching Esengers
MinhwanDepa
{minhw
tigate the effle-button messl quality. Twot, a paired useht macro animevaluation in
pirical study wnetic typograpd, and dynamicdy is expected er developinge messaging to
ing; kinetic typ
ds s and presentat
has led to thersonal interac
of close attentiohifted from com
such as smmental change
mmunication, aned the poy reducing th
n emerging e devices that ignals with a
e single-button-word message. The single-bu
ies of all or part of tee provided that codvantage and that coCopyrights for combe honored. Abstrablish, to post on servssion and/or a fee.
ustralia. 1/16/06...$15.00. 1835
Emotions throu Kim, Kyunartment of In
Daejeon, Rwan.kim, ck
fects of kinetsengers to enri complementa
er workshop wmation types we
n a subsequewas conducted phical elemencs, on conveyinto act as a goo
g single-buttoool with uniqu
pography
tion (e.g., HC
e relocation ofction as instaon that mediatmputers and cemartphones an
has allowed thnd single-butto
ossibilities fem. The singlcommunicatioallows users single tap. Th
n messenger the, ‘Yo,’ to othutton messeng
this work for personopies are not made opies bear this noti
mponents of this woacting with credit vers or to redistribuRequest permissio
nal Quagh Kine
ngah Choi, Hndustrial DeRepublic of
ka0116, colo
tic ch
ary was ere ent to ts, ng od on ue
I):
f a ant ed ell nd he on for le-on to he
hat her ger
has rapwe henumbereleaseforth.
HowevlackingfunctioSuch qlimitedNevertmessagexploremessagtypogra
KineticKineticsize, oattentioIt has emotiohave binstantdesignMore rcan adand bofocusedanimateffectsconducthe aniregard,studies
Fig
nalor
iceork
isuteons
ality of etic TypHyeon-Jeonsign, KAISTKorea
or}@kaist.ac
pidly become ad into the fu
er of similar ed, such as ‘H
ver, the singg in a commonality. What questions haved purpose, sutheless, we bging service, e new ways oging. In this raphy in single-
c Typographyc typography i
or position oveon and emphas
been used aonal quality inbeen many attet messengers [ed kinetic typorecently, Gaylodd more nuancody motion. Hd on empiricaltion types, whs of other elemcted in compuimation stimul, we conducts on the influe
gure 1. Screensh
Single-pograpng Suk T
c.kr
useful as an future of wear
applications hHodor,’ ‘Aiyo,’
gle-button memunicational as
if users want e led users to uuch as an eybelieve in theand studies nof communicastudy, we att
-button messen
y is defined as teer time [3]. It sizes the originas an alternat
n text-based coempts to embe1,4,6]. Lee et ography can coord et al. [5] shced communicHowever, thol evidence of thile they hav
ments. Furthermuter-based simli were pre-dested a series ence of kinetic
hots of a single-b
-Buttonphy
alternative merable devices have been su ‘Dingbel,’ ‘L
essaging applspect due to
to say sometuse this messeye-catching n
e potential of need to be conation using sintempt to embngers.
ext that changeeasily catchesnal meaning oftive solution ommunication,ed kinetic typoal. [7] asserteonvey specifichowed that anication cues of ose studies hathe influence o
ve not yet spemore, earlier stmulated messensigned by expeof two comp
c typography e
button messeng
n
essenger as [2], and a
ubsequently Lo,’ and so
lication is its simple
thing else? enger for a
notification. f this new nducted to ngle-button bed kinetic
es in color, the user’s f the word. to deliver
, and there ography in
ed that pre-c emotions. imated text
f intonation ave mainly of different ecified the tudies were ngers, and erts. In this plementary elements in
ger ‘Yo.’
Personalisation DIS 2016, June 4–8, 2016, Brisbane, Australia
276
samthosLm
P
PTaumAaoeainTeewfdsthsw
RFwaAaETo(aa(
single-button ma user-participamacro animatiohe following s
of kinetic typogspeed, and dyLastly, we sumessengers wit
PRELIMINARY
Procedure The preliminaranimations for users for the wmade up of fAmong the paand the rest weopposite each emotions, suchand disappointmn the dummy s
Two letter sexpressing elaexpressing rigiwere allowed feedback to tdesigned for easketched form.he workshop,
six emotions. Iworkshop took
Results and DFifty animationworkshop. Hoalphabet that wAs a result, eigand were givenEight animationThe first group of the original (see Figure 3(aanimation (see animations with(see Figure 3(c
Figure 2. S
messengers for ated workshopon for mobile tudy was condgraphical elemnamics, on coggested new th design impli
Y STUDY
ry study aims mobile device
workshop. A tfive males anrticipants, twoere intimate wother and wer
h as happiness,ment [10], usinsmartphone screts had diffastic motion d motion. Dur
to talk freetheir partner’sach emotion we Ice-breaking and five minuIncluding a tenapproximately
iscussion ns were designowever, animawas specific toght types of man appropriate nn types were this composed otext in three
a)). The seconFigure 3(b)). Th a translation
c)). Finally, the
Settings of the u
mobile devicep, we extracteddevices. Based
ducted to investents, includingonveying partiways of usin
ications.
to extract uses. We recruittotal of eight
nd three femalo pairs were inith each other.re asked to ex sadness, angeng tangible letreen, as illustra
ferent texturesand an acr
ring the workshely and encous animation. ere archived asexercises wer
utes were givenn-minute user
y 50 minutes.
ned from the uations with ao ‘Y’ and ‘o’ wacro animationnames, as shohen classified iof animations wdifferent stret
nd group consThe third groupn in two differee last group con
ser-participated
es. First, througd eight types d on the resultigate the effec
g animation typicular emotionng single-butto
er-driven macted four pairs college studenles participaten a relationshi. Participants sxpress six baser, fear, surpristters ‘Y’ and ‘ated in Figure s: silicone frylic plate fhop, participanuraged to givThe animatios thumbnails inre given prior n to each of thinterview, ea
user-participata morphologicwere eliminate
ns were extractown in Figure into four groupwith deformatiotching directioists of vibratinp is composedent motion patnsists of rotatin
d workshop.
gh of ts, cts pe, ns. on
cro of
nts ed. ip, sat sic se, o’ 2.
for for nts ve ns
n a to he ch
ed cal ed. ed 3.
ps: on ns ng of
ths ng
animat
Despitseveralanimatdisappofrequenthere aspecifiwere csome pangry ejoyoustouch investithan on
MAIN SThe mtypogradynamwe atinteracwe inteof kine‘emotio
ProcedThree empiriceight mstudy wthree le‘fast,’‘slow’level, intervadiffereanimat
Figure
tions in two dif
te the differenl animations wtion type wointment. Thntly used to are stereotypesic emotions. Oncontroversial inparticipants stremotion, while
s feeling. For was added.
igate the interne kinetic attrib
STUDY main study aimraphical elem
mics—on enricttempted to iction effects amended to propoetic typographyon presets’ in t
dure kinetic typogrcal study: animmacro animatwere adopted aevels were defaccording to level had fivewhere each l
al. Lastly, theent levels to extion types w
e 3. Eight types
fferent rotation
nt participants were replicated.was repeatedhe ‘jumping’ express happi
s of animation n the other hann use among pretched the texe other particip
a joyous feeThis trigger
raction effect bute.
ms to investigments—animatiching emotionaidentify both mong the kinetose a set of prey, which in turthe messenger
raphical elememation type, sption types deras the level of fined for speed
the play timee times longerlevel was dese dynamics wxplore the activwith deforma
s of macro animpreliminary st
n axes (see Figu
involved, int. For example,
dly used toanimation
iness. This imthat are stron
nd, some animparticipants. Foxt vertically to pants used it toeling, howeverred our motjointly caused
gate the role ion type, spal quality. In
the main etic elements. Iscriptive rules rn can be desiinterface.
nts were explopeed, and dynamrived in the pf animation typd: ‘slow,’ ‘mode of the animr playtime thansigned in an were divided veness of animtion (Figure
mations extractetudy.
ure 3(d)).
terestingly, ‘flopping’
o express type was
mplies that ngly tied to ation types
or instance, express an
o express a r, a gentle ivation to d by more
of kinetic peed, and
particular, ffects and
In addition, for the use gned as an
ored in the mics. First, preliminary pe. Second, derate,’ and
mation. The n the ‘fast’ equidistant into three
mations. For 3(a)) or
ed from the
Personalisation DIS 2016, June 4–8, 2016, Brisbane, Australia
277
translation (Figure 3(c)), the level of the dynamic is defined by the space covered by the text. On the other hand, for animation types with vibration (Figure 3(b)) or rotation (Figure 3(d)), the levels were defined with the motion range, such as its rotation angle. In order to reduce the number of stimuli, while minimizing the loss of measuring the effect, we adopted an orthogonal design. As a result, 31 profiles, including four holdout profile cards, were composed out of all 72 (8 × 3 × 3) combinations. Including a text without any effects, a total of 32 animated texts were designed and created in Adobe After Effects CS6, which are available at https://vimeo.com/152108052.
A total of 20 subjects consisting of eleven males and nine females participated, and their age ranged from 21 to 32 (M = 26.33, SD = 5.67). The animation sets were displayed on a smartphone (iPhone 6s) in order to simulate an instant messaging application in smart mobile devices. Participants were asked to assess six emotional qualities of the 31 animated texts using a seven-point Likert scale. In addition, the responses were recorded in the Affect Grid [8].
Results and Discussion We performed conjoint analysis to identify the elements that have a dominant effect, as shown in Table 1. The relative importance was extracted for each of the 20 participants to perform repeated measures ANOVA. The result indicated that ‘animation type’ plays the most important role in conveying emotions across eight dependent variables (p < .05). The importance between ‘speed’ and ‘dynamics’ was not significantly different, except for the ‘happiness’ emotion (t(19) = 2.95, p < .05). Although the importance of the ‘animation type’ was always biggest among three typographical elements, its relative importance was significantly different among eight measurements (F(7, 133) = 8.27, p < .05). The ‘animation type’ played a relatively less important role in expressing ‘surprise’ (49.01%) and ‘energy’ (45.72%). It was interesting to note that, however, the importance of the ‘dynamics’ element was biggest in expressing these emotions (F(7, 133) = 6.74, p < .05). There was no
significant difference in ‘speed’ elements among eight measurements (F(7, 133) = 2.02, p = .57).
In order to propose the best kinetic typography for each of the six basic emotions, part-worth utilities were derived from conjoint analysis, as shown in Table 2. The post hoc tests indicated that there is no single representative ‘animation type’ for each emotion. Hence, the animation types with a significantly higher utility score were extracted and underlined for each typographical element.
Next, we plotted eight animation types on the Affect Grid using the ‘Mood’ and ‘Energy’ scores, as shown in Figure 4. The result from the Affect Grid is qualitatively consistent with the best animation types extracted for six basic emotions. For instance, ‘vertical’ and ‘jumping’ animation types in the excitement (1st) quadrant are suggested for ‘happiness’ or ‘surprise’ emotion. The ‘horizontal’ and ‘expansion’ animations are located in the stress (2nd) quadrant, and they should convey emotions of ‘anger,’ ‘fear,’ or ‘surprise.’ The ‘vibration,’ ‘flopping,’ and ‘shimmy’ animation type were located in the depression (3rd) quadrant, and they should be adequate for ‘sadness,’ ‘anger,’ ‘fear,’ or ‘disappointment.’ For these seven animation types, the changes in speed and dynamics merely act as an ‘amplifier.’ This indicates that the influence of speed and dynamics on the emotional characteristics was marginal.
Interestingly, the ‘pendulum’ animation type was plotted on a neutral point. In fact, this received positive utility scores
Elements Levels Happiness Sadness Anger Fear Surprise Disappoint Mood Energy
Animation type
Vertical 1.19 -0.45 -0.12 -0.33 0.45 -0.64 0.81 0.65 Horizontal -0.22 -0.37 0.84 0.43 0.60 -0.49 -0.34 0.80 Expansion -0.48 -0.79 1.14 0.89 1.30 -0.92 -0.39 1.05 Vibration -0.67 0.23 0.02 0.96 -0.07 0.06 -0.43 -0.65 Jumping 1.82 -0.87 -0.71 -0.61 0.36 -1.10 1.89 0.64
Pendulum 0.32 0.34 -0.23 -0.66 -0.52 0.41 0.03 0.05 Flopping -1.18 1.06 -0.49 -0.12 -0.79 1.51 -1.14 -0.85 Shimmy -0.77 0.86 -0.46 -0.56 -1.32 1.16 -0.43 -1.70
Speed Slow -0.64 0.49 -0.30 -0.15 -0.63 0.44 -0.40 -0.95
Moderate -0.06 0.02 -0.11 -0.13 -0.19 0.10 0.05 -0.02 Fast 0.71 -0.50 0.41 0.28 0.81 -0.54 0.36 0.97
Dynamics Minimum -0.34 0.43 -0.56 -0.33 -0.72 0.41 -0.07 -1.12 Medium 0.21 -0.26 -0.06 -0.13 -0.20 -0.32 0.36 0.18
Maximum 0.13 -0.17 0.62 0.46 0.92 -0.08 -0.29 0.94
Table 2. Part worth utilities of each level for eight emotions. The underlined scores indicate a statistical difference at p < .05.
Emotions Type* Speed Dynamics* Happiness* 58.41 24.40 17.20 Sadness* 54.82 23.79 21.39 Anger* 57.40 18.90 23.71 Fear* 55.37 20.92 23.70
Surprise* 49.01 24.04 26.95 Disappointment* 59.58 22.09 18.34
Mood* 65.62 18.00 16.37 Energy* 45.72 25.62 28.66
Table 1. Relative importance (%) from conjoint analysis. An asterisk indicates significance at p < .05.
Personalisation DIS 2016, June 4–8, 2016, Brisbane, Australia
278
fWpain‘5cws
DBud
Fccidwasrh
Sin
for both ‘happiWe performedpendulum typeand with slow ndicated that happiness’ (t(
5.08, p < .05changes of speewhen the animstereotype.
DISCUSSION ABased on the eusing single-budeliver the emo
Firstly, using acommunicationcan send their ls the notificati
draw rich commwith animated aspect. People such as, “It’s rrestaurant tastehealth data mig
Secondly, notifn the ‘commu
Figure 4. Eig
Figure 5
iness’ (0.32) ad a paired e with fast sp
speed and mthe effects of19) = -2.57, p5) were signifed and dynamimation type h
AND FUTURE empirical studiutton messengeotions.
animated texts n, as shown in location on theon of the user’munication. Ontext connote acan express threally nice to bes really bad.”ght also be appl
fication with annication with
ght animation t
5. Using animat
and ‘sadness’ (samples t-teseed and maxi
minimum dynamf speed and d
< .05) and ‘sficant. This iics function as as a less sen
WORK ies, we suggesers expecting th
with other datFigure 5. For
e ‘Yo’ app. Its s presence, andn the other hana much richereir feelings wibe here,” or “T” Other types licable.
nimated texts cthings,’ as sho
types plotted on
ed texts with ot
(0.34) emotionst between thmum dynamicmics. The resudynamics on thsadness’ (t(19) mplies that th‘shunt switchesitive emotion
st a new way hat they also c
ta can enrich thinstance, peopcurrent functio
d it is difficult nd, location dar communicatioth their presenThe food at thof data such
can be embeddown in Figure
n Affect Grid.
ther data.
ns. he cs, ult he =
he es’ nal
of an
he ple on to
ata on
nce his as
ed 6.
Currenfunctiofrom tplatforonly commutext cacontennotificadelivernotificadepresthe not
Howevelemenconcermessenmergincommumight b
Nevertkineticemotioit is a for fast
CONCIn thistypogracommucomplemacro evaluatstudy, importkineticwe obscould aon theexpectbutton
Figu
ntly, the ‘Yo’on, providing utheir favorite rm has simply
for ‘interpeunication withan support appnts or users’ations about red with the exation about thesed animationtification.
ver, this studynts among the krned about thenger more comng between unication. Forbe utilized as a
theless, we hc typography oonal communicrelatively simpt and powerful
LUSION s study, we exraphy onto the unicational aspementary studianimations w
tion in a maiwe found tha
tant role in conc typographicalserved that thact as an ‘ampe animation tyted to strengthmessenger and
re 6. Animated
app has beeusers a new wa
online ‘thingsoutgrown its f
ersonal’ comh ‘things.’ In propriate notif’ preferencesusers’ favorit
xcited-feeling ae stock market
n. This helps p
y is limited to kinetic typograe real benefit
mplex. We aresimple text
r instance, usan automated in
have shown thon the single-bcation tool and
mple and nevertl communicatio
xplored the infsingle-button
pects by condies. From the
were created byin subsequent
at the ‘animatinveying emotiol elements. Be
he interventionplifier’ or a ‘stype. The findhen the potentd expand its us
d texts embeddewith things
en updated way to subscribes.’ This showfunction and th
mmunication this situation,
fication depend. For instante sports teamanimation, or at can be delive
people engage
the articulatioaphy. Also, wet of making te aware of the
and rich sers’ physiolonput.
he possibilitybutton messend provided evitheless emotioon.
fluence of addmessenger to
ducting a seripreliminary sty users and at study. Fromion type’ playonal quality amsides the anim
n of speed andshunt switch,’dings in this tial value of tsage.
ed in the ‘comm.’
with a new to updates
ws that the hat it is not
but also , animated ding on its nce, news ms can be a bad news ered with a more with
on of three e should be the simple successful emotional
ogical data
y of using ngers as an idence that
on-rich tool
ing kinetic enrich the
ies of two tudy, eight
adopted for m the main
s the most mong three
mation type, d dynamics
depending paper are
the single-
munication
Personalisation DIS 2016, June 4–8, 2016, Brisbane, Australia
279
REFERENCES 1. Kerry Bodine and Mathilde Pignol. 2003. Kinetic
typography-based instant messaging. In Proceedings of the CHI'03 Extended Abstracts on Human Factors in Computing Systems, 914-915. http://doi.acm.org/10.1145/765891.766067
2. Sam Byford. 2015. Messaging apps could be the best thing about the Apple Watch. Retrieved January 15, 2015 from http://www.theverge.com/2015/7/8/8910935/apple-watch-messaging-apps-asia
3. Shannon Ford, Jodi Forlizzi, and Suguru Ishizaki. 1997. Kinetic typography: issues in time-based presentation of text. In Proceedings of the CHI'97 Extended Abstracts on Human Factors in Computing Systems, 269-270. http://doi.acm.org/10.1145/1120212.1120387
4. Jodi Forlizzi, Johnny Lee, and Scott Hudson. 2003. The kinedit system: affective messages using dynamic texts. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 377-384. http://doi.acm.org/10.1145/642611.642677
5. Weston Gaylord, Vivian Hare, and Ashley Ngu. 2015. Adding body motion and intonation to instant messaging with animation. In Proceedings of the 28th Annual ACM Symposium on User Interface Software &
Technology, 105-106. http://doi.acm.org/10.1145/2815585.2815741
6. Johnny C. Lee, Jodi Forlizzi, and Scott E. Hudson. 2002. The kinetic typography engine: an extensible system for animating expressive text. In Proceedings of the 15th annual ACM symposium on User interface software and technology, 81-90. http://doi.acm.org/10.1145/571985.571997
7. Joonhwan Lee, Soojin Jun, Jodi Forlizzi, and Scott E. Hudson. 2006. Using kinetic typography to convey emotion in text-based interpersonal communication. In Proceedings of the 6th conference on Designing Interactive systems, 41-49. http://doi.acm.org/10.1145/1142405.1142414
8. James A. Russell, Anna Weiss, and Gerald A. Mendelsohn. 1989. Affect grid: a single-item scale of pleasure and arousal. J Pers Soc Psychol 57, 3: 493-502.
9. Ori Schwarz. 2011. Who moved my conversation? Instant messaging, intertextuality and new regimes of intimacy and truth. Media Cult Soc 33, 1: 71-87.
10. Zhiquan Yeo. 2008. Emotional instant messaging with KIM. In Proceedings of the CHI'08 Extended Abstracts on Human Factors in Computing Systems, 3729-3734. http://doi.acm.org/10.1145/1358628.1358921
Personalisation DIS 2016, June 4–8, 2016, Brisbane, Australia
280