http://www.dmst.aueb.gr/dds/pubs/jrnl/2004-CIE-VC/html/CS04b.htm
This is an HTML rendering of a working paper draft that led to a publication. The publication should always be cited in preference to this draft using the following reference:

Citation(s): 13 (selected).

This document is also available in PDF format.

The document's metadata is available in BibTeX format.

Find the publication on Google Scholar

This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.

Diomidis Spinellis Publications


© ACM, 2004. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in Computers in Entertainment (CIE), 2(3), ISSN 1544-3574, (July 2004) http://doi.acm.org/10.1145/1027154.1027177


Affective Usability Evaluation for an Interactive Music Television Channel

Konstantinos Chorianopoulos

chk@aueb.gr

k.chorian@imperial.ac.uk

Diomidis Spinellis

dds@aueb.gr

Athens University of Economics & Business

Department of Management Science and Technology

47 Evelpidon & Lefkados Str., 113 62 Athens, Greece

 


ABSTRACT

Computer mediated television raises brings up new requirements for user interface design and evaluation, since because interactive television applications are deployed in a relaxed domestic setting and aim at the gratification of entertainment needs. The dDigital video recorders, the generation of custom computer graphics on each digital set-top box, and the introduction of new advertising formats are important issues for research and practice. We explore the employment of an animated character and the dynamic insertion of advertising in the design of an intuitive user interface for interactive music video television. We found that the animated character and the skippable video-track clip feature seamlessly enhanced consumer satisfaction, as evidenced by affective usability questionnaires.

Categories and Subject Descriptors

H.1.2. [User/Machine Systems]: Human factors. H.5.2 [User Interfaces]: Interaction styles, Prototyping, Screen design, User-centered design, Input devices and strategies. H.5.1 [Multimedia Information Systems]: Video. J.7 [Computers in Other Systems]: Consumer products.

General Terms

Design, Human Factors.

Author Keywords

Interactive television, set-top box, user interface, affective usability, music video clip, animated character, TiVo.

INTRODUCTION

Computer mediated entertainment (CME) is emerging as a major economic force and is of scientific discourse subject. In the past, it was military and space exploration that drove scientific innovations that later became consumer products, like the PC and the Internet. More recently, forms of CME (e.g. video games) have been the main drivers of innovation, which is later applied to areas like scientific visualization, simulation, and education [23222122]. Besides its scientific importance, CME has emerged as a major economic factor in the media industry (cinema, video rentals, music, books, radio, television), taking-up a large portion of consumer spending and leisure time. In particular, television, which has traditionally occupied the largest share of consumer leisure time, is now undergoing a process of technological transformation.  Digital set-top boxes (STBs) and digital video recorders (DVRs) provide the platforms for new iinteractive ttelevision (ITV) applications. Is this an adequate definition for introducing the ITV acronym?

We employed the VVirtual CChannel UI user interface development toolkit [74] in the implementation of an ITV application that allows the user to ask for information (e.g. trivia, now playing, coming next, etc.) and to browse through music video clips. Since tTraditional human-computer interaction (HCI) settings, in contrast to computer-mediated leisure applications, involve a task-oriented approach where the human interacts with an application to accomplish a particular goal.  The different situational environment and goals of computer-mediated leisure applications, computer-mediated leisure applications require a fresh view of the current interaction paradigms. There is growing evidence that traditional desktop usability principles do not account for the pleasure of the user experience [1213]. Affective usability studies provide an alternative conception of user goals to consider elements like enjoyment, fun, and relaxation [712, 910]. In the light of the above findings, we employed an affective usability instrument in the evaluation of the interactive music television prototype.

The rest of this paper is organized as follows. The next section presents a critical review of the contemporary ITV application research  about ITV applications, and in particular, about focusing on animated characters, and video skipping. In section 3, tThese features are employed in a music ITV application, and its which is evaluated by consumers evaluation, as described in section 3. Section 4 presents the results and the discussion of the usability evaluation test for each one of the ITV applicationÕs features. In the last section, we draw present the implications of the studyÕs findings for the media industry.

USER INTERFACES FOR INTERACTIVE TELEVISION

Previous HCI research about ITV focused only mainly on the design of the Eelectronic Pprogram Gguide (EPG) and did not rarely considered the enhancement of the TV content. In particular, previous research approached ITV from a technological perspective [2252423]  and it did not consider the ITV user as a TV viewer [23]. While there have been numerous approaches for the design of ITV applications design, few have focused on enhancing a traditional TV program type with only a standard remote control and a TV screen in a relaxed setting. We chose to study music video television, which is a commercially successful, popular, and widely available format of TV content. Music video clips have a large lifespan Ñcompared for example with TV newsÑ that makes the experimental set-up and usability evaluation more convenient.

There are very few research works that concern the consumer UI for digital video libraries, although vVideo navigation for digitally stored video has been an active research area from the beginning of the 1990Õs. However, few researchers have studied the consumer UI for digital video libraries. The latter The corresponding research was is typically performed from the perspective of the TV author and producer [25]. It is usually assumed  Researchers, for example, assume that the users of digital video libraries are in front of a desktop computer that may display tens of thumbnails and video sources at the same time, by employing an advanced  typical PC GUI [24]. These assumptions break down in the living room, in which the user is far away from a low resolution TV and has interacts with a remote control. A traditional comparative usability evaluation test revealed that consumers preferred a UI that was perceived as fun and relaxing to use, although it had the worst performance in terms of efficiency and effectiveness [1011]. Therefore, a consumer-level UI for video skipping should be evaluated with affective usability methods that do take into account attributes like enjoyment, fun, and relaxation. Furthermore, there is no previous existing research that studies has studied digital video navigation that is in close relationship with a specific type of TV content, such as music TV.

Animated characters used in desktop computing have been widely researched, but the respective commercial implementations (most notably the infamous Microsoft Office Clip) are reported to be annoying to end-users [34, ][24232223]. An explanation might be that the attention grabbing and interrupting nature of animated characters is inappropriate for productivity computing. On the other hand, television content has traditionally been about stories and character development [1617]. Therefore, animated characters might be viable in a domestic environment for leisure activities, like television watching. In fact, there is evidence that animated characters are suitable in the entertainment domain: , because users liked more and were more engaged with the UI version of an interactive game that displayed a face to depict the opponent player [1415]. Animated characters have been also applied in a consumer electronics application to set the right level of expectation and to make errors from the a recommendation engine seem more acceptable [89]. Yet, there are there is no available affective usability evaluations of animated characters in comparison with traditional information presentation for TVtelevision applications.

Leaving aside television content quality questions, the major some current interesting research questions about interactive television (ITV) user interface design concern: (a) EPG navigation, (b) local storage navigation and (c) presentation of related information with an animated character. In this paper, we address examine the latter last two issues in by means of an affective usability evaluation of an ITV prototype.

A Prototype INTERACTIVE MUSIC TELEVISION CHANNEL

We designed and implemented an application that uses local storage and to allows a television viewer to skip a music video clip. Commercial Digital Video Recorders (DVRs) offer a 30 second skip button, but content and network providers are reluctant to adopt a Set-Top Box n( STB) technology that neglects their main revenue source (advertising) and are interested in alternative forms of television advertising. For example, TiVo is offering a special space (named ÔTiVo ShowcaseÕ), whereby advertisers may can store their advertisements (ads) and users can watch them on-demand. Alternatively, short ads (5 to 10 seconds) or video clip-like ads might be a choice when users choose to skip recorded content, although a subscription service may still be ad-free. ThusSynthesizing these options, we designed a video trackclip- skipping feature to that inserts an ad, before the next video clip begins (Figure 1Figure 1Figure 1Figure 1Figure 1Figure 1). Unless the trackclip-skip button is pressed, there is a continuous flow of music video clips, just like a normal music video television channel.

Figure 1111 Interactive music video television channel with dynamic advertisement insertion

In addition to the track video clip skipping, we also used the dynamic video overlay property of the our VVirtual CChannel programming library [7] to superimpose information over the music video. The MTV channel typically displays which music video comes next, a feature that we made interactive by allowing the user to ask for Ôwhat is playing now and what comes nextÕ on-demand. The dynamic insertion of related information to the music video raises the issue of the presentation style. We implemented and tested two alternative presentation styles: (a) The traditional MTV information box and (b) an animated character with a balloon dialog box (Figure 3Figure 3Figure 3Figure 3Figure 3Figure 3).

Experimental Set-up

The central element for our experimental set-up was a portable PC, running Microsoft Windows XP. We developed the prototype in VB.NET using the VVirtual CChannel programming library [74], and set the system up in a way to simulate a consumer ITV appliance. The PCÕs ATI display software was configured in the extended desktop setting (the desktop area extends beyond the PC monitor to the TV connected through the TV-out port) and the ITV prototype was set to display on the TV (second monitor). Then, the PCÕs TV-out and audio- output ports were connected to the audio-visual inputs of a traditional analog TV set. The ITV application was designed to run in full-screen and in window-less mode, so that there is was no visible portion indication of the underlying desktop environment. After starting the ITV prototype, we closed the portable PCÕs lid and placed it away from the TV.

  

Figure 2222 Low-budget set-up for interactive television usability evaluation

For supporting relaxed control with a normal TV remote control, the PCÕs serial port was connected to an infrared sensor (http://www.evation.com/irman/) that receivesd the signals from the remote control. The sensorÕs software driver and the supportiveng applications were used to map the remote controlÕs buttons to specific keyboard buttons. The whole set-up was unobtrusive and seamless to the television viewer (Figure 2Figure 2Figure 2Figure 2Figure 2Figure 2). It also allowed the experimenter to perform evaluations in consumersÕ homes using their familiar TV and remote control.

Experimental Design

Traditional human-computer interaction settings involve a task-oriented approach where the human interacts with an application to accomplish a particular goal. Accordingly, usability evaluation techniques measure successful task completion, efficiency, and error rate parameters that are supposed to correlate positively with user satisfaction. Most notable among the recent findings about ITV applications is the realization that usersÕ subjective satisfaction is at odds with performance metrics. For example, a usability test of three video skipping interfaces (two commercial and one novel) revealed that user satisfaction was higher for the an interface that required more time, more clicks, and had the highest error rate. In other words, the most usable interface was not the most preferred interface most users preferred to use. Users reasoned their choice on the basis of how fun and relaxing an interface was [1011].

Therefore, in the context of CME, there is a need to consider an alternateive conception  of user goals. Several conceptionts of affective usability have been proposed: eEnjoyment, fun, trust, engagement, motivation for using [910], hedonic quality of a user experience [1312], and consumer emotions [78] are the most relevant in our setting. We chose to use the HassenzahlÕs et al [11] affective usability measuring instrument of Hassenzahl et al [12], because it is a validated, freely available, short, and features an easy-to-understand verbal scale[*]. Furthermore, a fulfilling television experience depends on the subjective evaluations of the entertaining value of the content [1617], a characteristic that is partially captured by the construct of hedonic quality.

We used a seven point semantic differential scale and reversed the polarity of every other pair: (outstanding-second rate, standard-exclusive, impressive-nondescript, ordinary-unique, innovative-conservative, dull-exciting, interesting-boring) XXX Please explain this better.  Maybe use a table?. Scores were summed and then scaled from 0 to 10.

Figure 333333 The animated character (on the left, using the Microsoft Genie) and the traditional transparent box (on the right) for displaying dynamic video overlays

The objective of this our  study was to evaluate user preferences for an ITV application that offers track clip skipping for music video clip television and an animated character for presenting information. The experiment was designed to address two of the main issues identified for interactive television user interface design: (a) local storage navigation through simple video track clip skipping and (b) presentation of related information through alternative presentation styles. We were also interested to explore consumer preferences regarding the dynamic ad insertion, when they users choose to skip a music video clip. We formulated our objectives as the following research hypotheses.

á        Hypothesis 1:  Hedonic quality will be greater for the a trackclip-skipping TV channel user interfacecompared with a fixed one, since users will be able to seek preferred clips and avoid the disliked ones.

á        Hypothesis 2:  Hedonic quality  value will be greater for an different between the an animated character compared with a  and the a transparent information box for the presentation of related information.  XXX be honest here and take a stance.  Say what you expected (animated character leads to a higher hedonic quality value, I assume) and later admit that htis hypothesis was not proved.

á        Hypothesis 3: Consumers using an ITV application that employs a video clip-skipping user interface with dynamic advertisement insertion are will be exposed to more advertisements in an ITV application that employs a video track-skipping user interface with dynamic advertisement insertion than those in contained in a corresponding linear TV schedule.

Each participant received two experimental treatments (within groups) of the user interface for interactive music video television: 1) The animated character and 2) the transparent box, while both of them setups offered video track clip skipping with ad insertion. After the end of each session, participants evaluated separately the hedonic quality of (a) traditional music video television (all participants were selected to be frequent viewers of music TV), (b) music video television with track clip skipping, (c) information presentation with the transparent box and (d) information presentation with the animated character. We ran tests with 21 users (recruited from the post-graduate and under-graduate departments of a European business our university). Ages were between 22 and 35 (13 men and 8 women). Users were assigned with a random order to each treatment and the order of the music songs video clips was also randomized for each session. The music-video clip related information and the remote control were the same for all sessions.

The study was performed in a relaxed setting, using a traditional TV set and a remote control. We used multiple usability engineering methods [ XXX add a citation here to an appropriate textbook or article][18]: (a) we observed user behavior during the testing session, (b) we kept a record of user actions in log files, (c) we had users completeed questionnaires and (d) we interviewed each user after the end of the testing session. In the beginning, the interviews were unstructured and directed by the users. Gradually, the interviews became more focused, to repeating issues, mentioned during the interviews or observed during the testing session.

The testing session contained 16 video-clips and an advertising break with three ads every 4 songs (approximately every 15 minutes), just like a commercial music video television channel. The study was in accordance with the selective-exposure paradigm [12625]. . Users were free to choose the music video clip they preferred to watch, like they would do if they were outside the experiment. In order to ensure selective-exposure the users were allowed a maximum of 1/3 of watching time, out of the total session duration [1214], that is a maximum of approximately 20 minutes out of the 1h program duration. Users could press the power-off button on the remote to end the testing session and they were told to watch as much as they liked, between 10 and 20 minutes.

RESULTS AND DISCUSSION

Users who were not involved in computer research and development asked how video skipping was possible and whether that it was a commercially available product or television station. Most of the other users were aware of an experimental system behind the TV program, but when asked whether they understood that there was actually a PC running the system, users said that it looked and felt like normal TVtelevision. The use of a normal TV and a remote control contributed to these positive evaluations, but they can also be attributed to the employment of the VVirtual CChannel UI development toolkit that delivers a familiar television experience [47]. Therefore, the combination of the Virtual Channel programming library and an appropriate experimental set-up may be used to create high-fidelity ITV prototypes.

Video Clip sSkipping

We had predicted that there would be a difference in the hedonic quality value between a traditional music video television channel and one that features track clip skipping of music videos. We found (Table 1Table 1Table 1Table 1Table 1Table 1) that the hedonic quality score (scale is from 0 to 10; scores less/more than 5 represent negative/positive attitude) for the traditional one is close to neutral (average 5.1/10),.  This finding can be explained by the fact that which is quite expected, since  music video television is a pervasive experience and feels familiar to consumers, irrespective of its delivery format. In contrast, trackvideo clip-skipping (average 7.5/10) allowed consumers our experimental subjects to watch their preferred music video clips and despite the dynamic insertion of ads the hedonic quality score was significantly higher (two tailed tÐtest, p=0.002, n=21). Therefore, it is argued we argue that simple audio CD-like track video clip skipping, similar to the track-skipping facility available in audioaudiio CD players, enhances the perceived television entertainment value, when compared with same fixed TV channel.

Hedonic Quality (p=0.002, n=21)

Average

Std Dev

Music Video TV (traditional)

5.1

2.1

TrackClip-skip

7.5

1.6

Table 1111 Mean hedonic quality scores for the trackclip-skipping music video television are significantly higher

Consistent with the selective exposure theory, users actively sought for the video clips and songs they preferred. This kind of interactive behavior may be due to the experimental setting and may not have external validity; users may have been more engaged than normal because the application is was novel to them and because they were specifically asked to use the new system. They reported that they used the skip functionality mainly to skip a music video that they disliked and to a lesser extent to get to a favorite one. Either way, the videoclip-skipping feature was a favorite, despite the ad insertion, and provided a relaxed way to control of the interactive music TV application, based on the dynamic video synthesis of music clips.

Animated Character

We had also predicted that there would be a difference in the hedonic quality value between the animated character and the transparent box for the related information presentation style. We found that the hedonic quality (scale is from 0 to 10; scores less/more than 5 represent negative/positive attitude) for a music video television channel is significantly higher (two tailed tÐtest, p=0.0002, n=21) when using an animated character (average 7.0/10) for presenting dynamic video overlays (average 7.0/10) compared with the traditional transparent information box (average 4.4/10). Again, consumers our experimental subjects were neutral toward the traditional information box, since it is a widely used and familiar presentation style for information related to music video clips (Table 2Table 2Table 2Table 2Table 2Table 2). Therefore, we argue that the animated character could be used to enhance the consumersÕ entertainment experience with television.

Hedonic Quality (p=0.0002, n=21)

Average

Std Dev

Animated Char.

7.0

1.5

Box (traditional)

4.4

2.0

Table 222222 Mean hedonic quality scores for the animated character compared to the traditional overlay box

Those who have been exposed before to the Office Assistant (through the Microsoft Office suite of applications suite) recognized the similarity (due to the balloon-style dialog, which is standard for the characters developed with the Microsoft Agent system) despite the use of a different character (the genie) from the Microsoft Office default paperclip.  A few of these users reacted negatively to the concept of the animated character. Therefore, the animated character may have a carry-over effect from the desktop to the ITV environment. Those The users, who are already negatively predisposed to it, will continue to be so, at the expense of their satisfaction with the whole TV program. For Taking into account those users, we suggest that the animated character should be an option and an alternative UI should be available to select. Nevertheless, most of the users considered the character funny and less obtrusive compared with human presenters, who interrupt the flow of the video clips to present related information. Furthermore, users proposed that there should be a variety of animated characters to select from (e.g. cartoons, heroes, personalities). Users also asked for more control of the character, like changing its placement on the screen. Finally, most of the users reported that the solid balloon dialog that stands over the head of the character hides a considerable part of the TV screen. According to their suggestions, the best form for the balloon dialog would be a transparent one across the bottom of the screen.

Dynamic Advertisement Insertion

We were interested in finding out how many ads users would be willing to watch, as a consequence of using the music video clip skip feature. In a traditional music video TV channel the proportion is approximately 12 ads for each hour of TV video [Executive of a European music television channel, personal communication]. The findings regarding the dynamic advertisement insertion were very positive, despite the fact that advertising is one of the most controversial features of commercial TVtelevision. During the interviews and the observation sessions, it was we found that all the consumers users regarded were positive to the dynamic advertisement insertion in a positive light. Log file analysis revealed that consumers users actually watched the double number of advertising messages compared with a normal broadcast session. Interestingly, according to the log files, some users also tried to skip over the ads, with no effect since the application was programmed to ignore the video skip when inside an ad.  Personalised ad selection and presentation is a separate research topic [5]; an ad skipping facility could well be a worthwhile extension.

Overall, the consumers users justified the trade-off between skipping to the next video clip and watching a short advertisement positively. Yet, it is not clear which is the cause of the above result, since the system employed two innovative features that affect the number of advertising messages displayed. The dynamic advertisement insertion every time the user skipped a video clip increases the number of messages shown. At the same time, leaving aside for a moment appartapart from the dynamic advertisement insertion, the video clip skipping action also brings the user closer to the next scheduled advertising break, thus increasing the number of advertising messages displayed. Therefore, in order to draw cause and effect conclusions between the video skipping feature and the number of advertisements watched, a further another study experiment is needed:  that will using the same video-clip skipping UI the dependent variable should be the compare between having dynamic advertisement insertion and not having, for the same video-clip skipping UI.

Limitations and Further Research

Music video clips have a number of unique characteristics that facilitate further ITV development. Music content can be easily classified and filtered by employing open Internet databases and Wweb Sservices, classification schemes, and adaptation models that have been developed for the popular MP3 music format [20191820]. Additional metadata that describe the emotional content (e.g. from http://www.moodlogic.com) may also be used and combined with research about emotion in animated characters [1]. The latter research issue merits further research by the respective community, which has already put considerable effort in the investigation of animated characters for consumer electronics [89].

The most interesting suggestions for future improvements concerned the augmentation of the music video clip skip feature. Users familiar with CD players and the PC-based MP3 music players asked for more options when skipping a music video, like repeat the same song, play a song from the same artist, or play a new song within same music genre. Moreover, information about a longer list of the upcoming music videos would be welcome; and it would also allow users to organizinge their time bettermore efficiently, since they could leave the TV on and plan to return back when their favorite song is was on. Using the television as a time tool to structure household activities and organize time has been documented before in an ethnographic study of an STB trial [1619]. Providing on-demand information about the upcoming video clips supports the relaxed control of TV as a time management tool, while the ability to alter dynamically the upcoming play-list would further supports interactive behavior. For example, the user could bring up a play-list of 10 upcoming music videos and alter it dynamically along a number of parameters like genre, and artist, or automatically create play-lists [20191820]. The user could also decide whether to skip directly to a music video by pressing the corresponding button on the numeric keypad.

Although we used complementary qualitative and quantitative Despite the alternative usability evaluation methods (qualitative, quantitative) that were used in the course of the present our study, all of them all instruments were employed during the limited duration span of an controlled experiment, instead of unobtrusively spanning taking place through the everyday living activities in consumersÕ homes. Previous findings regarding on the usability of everyday technology demonstrate that the consumersÕ perceptions and especially the mental models they form about new domestic technologies are very elastic and prone to change in with the passage of time [21201921]. Therefore, a longitudinal study would have provided additional information about the our research issues. A longitudinal study should employ an appropriate research framework, like such as the experience sampling method [1516]. In the context of ITV prototyping, adding a TV tuner would extend the functionality to include synchronization between broadcast and local storage and provide the ability to perform more realistic TV experiments in consumersÕ homes, over longer periods of time.

CONCLUSION

In the context of the commercial implementation of the video-skipping feature, a major open research question is who is going to control the rules for the dynamic synthesis of video and for the dynamic overlay informationof data. For example, on the one hand, the media industry may choose to subsidize consumers DVR STBs to consumers, in exchange of for having increased control of what is stored and how it is played (e.g. targeted advertising and subscription services for personalized music/news channels for each STB). On the other hand, the consumer electronics industry may offer advanced general purpose DVRs with DVD-Rrecording functionality and Internet connectivity for downloading meta-data and related information. The latter are likely to could be more complex to use, but will offer increased consumer control, while the former will be introduced by established media brands and broadcasting services. Finally, bBetween these two extremes there is a continuum of alternative product and service offerings.  As is the case with other consumer products, the affective usability of each offeringÕs ITV applications, and is applicability for the relaxed domestic setting is likely to be a decisive factor for the success of the corresponding products, services, and the underlying business models.

Appendix Ð Virtual Channel Model

The notion of the Virtual Channel refers to the television channel not being a fixed video signal that is shared by all TV viewers in the same way, but a dynamic synthesis of discrete video, graphics, and data controlled by a computer program, which is running at each digital STB [5]. The traditional television experience consists of video and overlaid graphics-text and it is created at the media source (the TV broadcast station or the TV production studio), thus it is fixed for all TV viewers. The Virtual Channel model shifts the decision-making about TV programming from the media source to the STB. The television experience is now created and controlled at the STB from a combination of locally stored material, real time broadcast transmissions, and Internet resources (Figure 4Figure 4).

Figure 44 generic model of a system employing the virtual channel metaphor, in contrast with the traditional broadcasting scheme

The Virtual Channel is a conceptual model for user interface design, which augments the familiar access method to broadcast programming (i.e. the simple notion of a channel), to an integrated model for accessing multimedia content from diverse sources. Firstly, the organization of digital media content into a small number of spatiotemporally personalized virtual channels simplifies the choice from a vast array of available broadcasts, stored programs and Internet resources. Secondly, the presentation of media programming from virtual channels gives the potential for more control to the user, who can actively shape the televised content. The Virtual Channel model suggests only a minimal shift from the current patterns of media use, while it focuses further research on the design of a content specific UI (e.g. music TV).

In brief, the main proposition is that neither the vision of five hundred channels, nor the vision of a single personalized channel is suitable for giving consumer access to the digital STB. Instead, it is proposed that a small number of dynamic virtual channels may offer enough choices to cater for serendipity in media experiences, while simplifying the access to vast and diversified sources of television content.

ACKNOWLEDGMENTS

We would like to thank the university students who participated to the experiment, and Ronald Boring and the anonymous reviewers for providing valuable comments on the research and for reading through early paper drafts. Parts of this work were supported by the CONTESSA (IST-2000-28567) and MUSICAL (EDC-22131) projects, partially funded by the European Commission.

REFERENCES

1.    Bates, J. The role of emotion in believable agents. Communications of the ACM, 37(7):122Ð125, 1994.

2.    Carey, J. Interactive television trials and marketplace experiences. Multimedia Tools and Applications, 5(2):207Ð216, 1997.

3.    Carey, J. Content and services for the new digital TV environment. In Gerbarg, D., editor, The Economics, Technology and Content of Digital  TV, pages 88Ð102. Kluwer Academic Publishers, 1999.

4.    Catrambone, R., Stasko, J., and Xiao, J. Anthropomorphic agents as a user interface paradigm: Experimental findings and  a framework for research. In Proceedings of CogSci 2002, pages 166Ð171, 2002.

5.    Chorianopoulos, K. The digital set-top box as a virtual channel provider. In Proceedings of the SIGCHI conference on Human factors in computing systems, pages 666Ð667. ACM Press, 2003.

6.    Chorianopoulos, K., Lekakos, G., and Spinellis, D. Intelligent user interfaces in the living room: usability design for  personalized television applications. In Proceedings of the 2003 international conference on Intelligent user  interfaces, pages 230Ð232. ACM Press, 2003.

7.    Chorianopoulos, K. and Spinellis, D. User interface development for interactive television: Extending a commercial  DTV platform to the virtual channel API. Computers and Graphics, 28(2):157Ð166, 2004.

8.    Desmet, P. M. Measuring emotions: Development and application of an instrument to measure  emotional responses to products. In Blythe, M., Monk, A., Overbeeke, K., and Wright, P., editors, Funology:  from usability to enjoyment. Kluwer Academic Publishers, 2003.

9.    Diederiks, E. M. A. Buddies in a box: animated characters in consumer electronics. In Proceedings of the 2003 international conference on Intelligent user  interfaces, pages 34Ð38. ACM Press, 2003.

10. Draper, S. W. Analysing fun as a candidate software requirement. Personal and Ubiquitous Computing, 3(3), 1999.

11. Drucker, S. M., Glatzer, A., Mar, S. D., and Wong, C. Smartskip: consumer level browsing and skipping of digital video content. In Proceedings of the SIGCHI conference on Human factors in computing  systems, pages 219Ð226. ACM Press, 2002.

12. Hassenzahl, M., Beu, A., and Burmester, M. Engineering joy. IEEE Software, 18(1):70Ð76, January/February 2001.

13. Hassenzahl, M., Platz, A., Burmester, M., and Lehner, K. Hedonic and ergonomic quality aspects determine a software's appeal. In Proceedings of the SIGCHI conference on Human factors in computing  systems, pages 201Ð208. ACM Press, 2000.

14. Knobloch, S. and Zillmann, D. Mood management via the digital jukebox. Journal of Communication, 52(2):351Ð366, 2002.

15. Koda, T. and Maes, P. Agents with faces: The effects of personification of agents. In Proceedings of HCI'96, London, 1996.

16. Kubey, R. and Csikszentmihalyi, M. Television and the Quality of Life: How Viewing Shapes Everyday  Experiences. Lawrence Erlbaum, 1990.

17. Lee, B. and Lee, R. S. How and why people watch TV: Implications for the future of interactive television. Journal of Advertising Research, 35(6):9Ð18, 1995.

18. Nielsen, J. Usability Engineering. Morgan Kaufmann, San Francisco, 1994.

19. O'Brien, J., Rodden, T., Rouncefield, M., and Hughes, J. At home with the technology: an ethnographic study of a set-top-box trial. ACM Transactions on Computer-Human Interaction (TOCHI), 6(3):282Р 308, 1999.

20. Pachet, F. Content management for electronic music distribution. Communications of the ACM, 46(4):71Ð75, 2003.

21. Petersen, M. G., Madsen, K. H., and Kjaer, A. The usability of everyday technology: emerging and fading opportunities. ACM Transactions on Computer-Human Interaction (TOCHI), 9(2):74Р 105, 2002.

22. Rosenbloom, A. Special section: A game experience in every application. Communications of the ACM, 46(7), 2003.

23. Schaumburg, H. Computers as tools or as social actors? - the users' perspective on  anthropomorphic agents. International Journal of Cooperative Information Systems,  10(1,2):217Ð234, 2001.

24. Smyth, B. and Cotter, P. A personalized television listings service. Communications of the ACM, 43(8):107Ð111, 2000.

25. Wactlar, H. D., Christel, M. G., Gong, Y., and Hauptmann, A. G. Lessons learned from the creation and deployment of a terabyte digital video  library. IEEE Computer, 32(2):66Ð73, 1999.

1.J. Bates. The role of emotion in believable agents. Communications of the ACM, 37(7):122Ð125, 1994.

1.J. Carey. Content and services for the new digital tv TV environment. In D. Gerbarg, editor, The Economics, Technology and Content of Digital  TV, pages 88Ð102. Kluwer Academic Publishers, 1999.

1.R. Catrambone, J. Stasko, and J. Xiao. Anthropomorphic agents as a user interface paradigm: Experimental findings and  a framework for research. In Proceedings of CogSci 2002, pages 166Ð171, 2002.

1.K. Chorianopoulos and D. Spinellis. User interface development for interactive television: Extending a commercial DTV platform to the virtual channel API. Computers and Graphics, 28(2), 2004.

1.J. Clark. A telecomputer. In Proceedings of the 19th annual conference on Computer graphics and  interactive techniques, pages 19Ð23. ACM Press, 1992.

1.D. Crockford. Integrating computers and television. In B. Laurel, editor, Art of Human-Computer Interface Design, pages  461Ð466. Addison-Wesley, 1990.

1.P. M. Desmet. Measuring emotions: Development and application of an instrument to measure  emotional responses to products. In M. Blythe, A. Monk, K. Overbeeke, and P. Wright, editors, Funology:  from usability to enjoyment. Kluwer Academic Publishers, 2003.

1.E. M. A. Diederiks. Buddies in a box: animated characters in consumer electronics. In Proceedings of the 2003 international conference on Intelligent user  interfaces, pages 34Ð38. ACM Press, 2003.

1.S. W. Draper. Analysing fun as a candidate software requirement. Personal and Ubiquitous Computing, 3(3), 1999.

1.S. M. Drucker, A. Glatzer, S. D. Mar, and C. Wong. Smartskip: consumer level browsing and skipping of digital video content. In Proceedings of the SIGCHI conference on Human factors in computing  systems, pages 219Ð226. ACM Press, 2002.

1.M. Hassenzahl, A. Beu, and M. Burmester. Engineering joy. IEEE Software, 18(1):70Ð76, January/February 2001.

1.M. Hassenzahl, A. Platz, M. Burmester, and K. Lehner. Hedonic and ergonomic quality aspects determine a software's appeal. In Proceedings of the SIGCHI conference on Human factors in computing  systems, pages 201Ð208. ACM Press, 2000.

1.S. Knobloch and D. Zillmann. Mood management via the digital jukebox. Journal of Communication, 52(2):351Ð366, 2002.

1.Koda, T. and Maes, P. Agents with faces: The effects of personification of agents. In Proceedings of HCI'96, London, 1996.

1.Kubey, R. and Csikszentmihalyi, M. Television and the Quality of Life: How Viewing Shapes Everyday Experiences. Lawrence Erlbaum, 1990.

1.Lee, B. and Lee, R. S. How and why people watch tv: Implications for the future of interactive television. Journal of Advertising Research, 35(6):9Ð18, 1995.

1.Giorgos Lekakos, Kostas Chorianopoulos, and Diomidis Spinellis. Information systems in the living room: A case study of personalized interactive TV design. In Proceedings of the 9th European Conference on Information Systems, June 2001.

1.Nielsen, J. Usability Engineering. Morgan Kaufmann, San Francisco, 1994.

1.J. O'Brien, T. Rodden, M. Rouncefield, and J. Hughes. At home with the technology: an ethnographic study of a set-top-box trial. ACM Transactions on Computer-Human Interaction (TOCHI), 6(3):282Р 308, 1999.

1.F. Pachet. Content management for electronic music distribution. Communications of the ACM, 46(4):71Ð75, 2003.

1.M. G. Petersen, K. H. Madsen, and A. Kjaer. The usability of everyday technology: emerging and fading opportunities. ACM Transactions on Computer-Human Interaction (TOCHI), 9(2):74Ð  105, 2002.

1.L. Press. Personal computing: compuvision or teleputer. Communications of the ACM, 33(9):29Ð36, 1990.

1.A. Rosenbloom. Special section: A game experience in every application. Communications of the ACM, 46(7), 2003.

1.H. Schaumburg. Computers as tools or as social actors? - the users' perspective on  anthropomorphic agents. International Journal of Cooperative Information Systems,  10(1,2):217Ð234, 2001.

1.B. Smyth and P. Cotter. A personalized television listings service. Communications of the ACM, 43(8):107Ð111, 2000.

1.Wactlar, H. D., Christel, M. G., Gong, Y., and Hauptmann, A. G. Lessons learned from the creation and deployment of a terabyte digital video  library. IEEE Computer, 32(2):66Ð73, 1999.

1.D. Zillmann and J. Bryant, editors. Selective exposure to communication. Hillsdale, NJ: Lawrence Erlbaum Associates, 1985.

 


 



[*] We used a seven point semantic differential scale and reversed the polarity of every other pair: outstanding-second rate, standard-exclusive, impressive-nondescript, ordinary-unique, innovative-conservative, dull-exciting, interesting-boring. Scores were summed and then scaled from 0 to 10.