FULL AND SHORT PAPERS

The following full and short papers were selected through a double-blind peer-review process. During the TPC meeting on 17 March 2017 in Delft, The Netherlands, 32% of all submitted papers were accepted, based on at least three reviews and one meta-review by an Associate Chair.

Full papers will get a 20-minute time slot (+Q&A) for presentation at the conference, short papers get a 10-minute time slot (+Q&A). The final program with the exact timing of the presentations will be made available in May.

Enhancing Interaction with Dual-Screen Television Through Display Commonalities

Timothy Neate — Department of Computer Science, Swansea University, Swansea, United Kingdom
Michael Evans — Research & Development, BBC, Salford, Greater Manchester, United Kingdom
Matt Jones — Swansea University, Swansea, Wales, United Kingdom

Abstract: Second screening – engaging with a mobile device while watching TV – is ubiquitous. Previous research demonstrates that this is hampered by cognitive and physical disjuncts between the simultaneous content streams. To engage effectively with more than one screen, users must manage their attention, for example, by frequently adjusting their gaze or posture. This can lead to cognitive effort, which leads to disengagement, content sacrifice, and ultimately, affects user experience (UX) negatively. In this paper, we look to improve the design of the dual-screen scenario through display commonalities; the mirroring of one content stream (e.g., TV material or second screen content) within the other. We evaluate this design space with professional broadcast practitioners, and then conduct an empirical investigation to determine the impact of the most successful methods towards understanding their impact, and designing towards positive UX with multi-device scenarios.

User Experience of Panoramic Video in CAVE-like and Head Mounted Display Viewing Conditions

Adam Philpot — School of Science and Technology, Middlesex University, London, United Kingdom
Maxine Glancy — Research & Development, BBC, Salford, Greater Manchester, United Kingdom
Peter Passmore — School of Science and Technology, Middlesex University, London, United Kingdom
Andrew Wood — Research and Development, BBC, London, United Kingdom
Bob Fields — School of science and Technology, Middlesex University, London, United Kingdom

Abstract: Panoramic 360 video is a rapidly growing part of interactive TV viewing experience due to the increase of both production by consumers and professionals and the availability of consumer headsets used to view it. Recent years have also seen proposals for the development of home systems that could ultimately approximate CAVE-like experiences. The question arises as to the nature of the user experience of viewing panoramic video in head mounted displays compared to CAVE-like systems. User preference seems hard to predict. Accordingly, this study took a qualitative approach to describing user experience of viewing a panoramic video on both platforms, using a thematic analysis. Sixteen users tried both viewing conditions and equal numbers expressed preferences for each display system. The differences in user experience by viewing condition are discussed in detail via themes emerging from the analysis.

Integrating Mid-Air Haptics into Movie Experiences

Damien ABLART — SCHI Lab, School of Engineering and Informatics, University of Sussex, Brighton, United Kingdom
Carlos Velasco — Marketing, BI Norwegian Business School, Oslo, Norway
Marianna Obrist — SCHI Lab, School of Engineering and Informatics, University of Sussex, Brighton, United Kingdom

Abstract: “Seeing is believing, but feeling is the truth”. This idiom from the seventieth century English clergyman Thomas Fuller gains new momentum in light of an increased proliferation of haptic technologies that allow people to have various kinds of ‘touch’ and ‘touchless’ interactions. Here, we report on the process of creating and integrating touchless feedback (i.e. mid-air haptic stimuli) into short movie experiences (i.e. one-minute movie format). Based on a systematic evaluation of user’s experiences of those haptically enhanced movies, we show evidence for the positive effect of haptic feedback during the first viewing experience, but also for a repeated viewing after two weeks. This opens up a promising design space for content creators and researchers interested in sensory augmentation of audiovisual content. We discuss our findings and the use of mid-air haptics technologies with respect to its effect on users’ emotions, changes in the viewing experience over time, and the effects of synchronisation.

Countering Contextual Bias in TV Watching Behavior: Introducing Social Trend as External Contextual Factor in TV Recommenders

Felix Lorenz — Technische Universität Berlin, Berlin, Germany
Jing Yuan — DAI-Labor, Fakultät IV für Elektrotechnik und Informatik, Technische Universität Berlin, Berlin, Germany
Andreas Lommatzsch — DAI-Labor, Fakultät IV für Elektrotechnik und Informatik, Technische Universität Berlin, Berlin, Germany
Mu Mu — The University of Northampton, Northampton, United Kingdom
Nicholas Race — School of Computing & Communications, Lancaster University, Lancaster, Lancashire, United Kingdom
Frank Hopfgartner — University of Glasgow, Glasgow, United Kingdom
Sahin Albayrak — DAI-Labor, Technische Universität Berlin, Berlin, Germany

Abstract: Context-awareness has become a critical factor in improving the predictions of user interest in modern online TV recommendation systems. In addition to individual user preferences, existing context-aware approaches such as tensor factorization incorporate system-level contextual bias to increase predicting accuracy. We analyzed a user interaction dataset from a WebTV platform, and identified that such contextual bias creates a skewed selection of recommended programs which ultimately locks users in a filter bubble. To address this issue, we introduce the Twitter social stream as a source of external context to extend the choice with items related to social media events. We apply two trend indicators, Trend Momentum and SigniScore, to the Twitter histories of relevant programs. The evaluation reveals that Trend Momentum outperforms SigniScore and signalizes 96% of all peaks ahead of time regarding the selected candidate program titles.

The Social Construction of Targeted Television Advertising: The Importance of ‘Social Arrangements’ in the Development of Targeted Television Advertising in Flanders

Iris Jennes — Communication Studies, Vrije Universiteit Brussel, Brussels, Belgium
Wendy Van den Broeck — Communication Studies, Vrije Universiteit Brussel, Brussels, Belgium

Abstract: This paper focuses on the social construction of targeted TV advertising. In 2016, experiments with targeted TV commercials started in Flanders (Belgium). We apply a Social Construction of Technology (SCOT) approach to understand how targeted television advertising is being developed. We underline the importance of social arrangements in the development of this particular technology. Social arrangements can be defined as the relations between relevant social groups that work together to stabilize a technology. The development of targeted TV advertising can be seen as a moment of ‘interpretive flexibility’, implying that different relevant social groups can give a different meaning to targeted advertising as a technological artifact. To steer the development of the technology towards the most beneficial solution to their agenda, different social groups use different strategies. In our paper, we argue that in the case of targeted TV advertising, the audience should be approached as a relevant social group. Our empirical research thus incorporates both television industry and user perspectives on the development of targeted TV advertising in Flanders between 2012 and 2017. Based on expert interviews with industry representatives and focus group interviews with end-users, we provide an analysis of the different strategies, opportunities and challenges that different stakeholders (TV-industry, viewers and policy actors) are faced with. To conclude, we also formulate specific recommendations for a successful implementation of targeted TV advertising in Flanders.

Understanding Secondary Content Practices for Television Viewing

Frank Bentley — Yahoo, Sunnyvale, California, United States

Abstract: Secondary content experiences related to television viewing have been a frequent topic of study in the TVX community. While many new interfaces have been created and studied in the small scale, we are not aware of any larger quantitative work to study current practices now that many secondary content experiences are publicly available. We conducted a survey with a broad sample of the American population to explore current secondary content use. We report on our findings, including that 80% of these experiences occur before or after viewing the primary content, and not as simultaneous experiences, and that social posting about television content remains quite low, even for one’s favorite show. We conclude with implications for the design new secondary content systems based on our findings.

How Millennials and Teens Consume Mobile Video

Jennifer McNally — Verizon, San Jose, California, United States
Beth Harrington — Verizon, Waltham, Massachusetts, United States

Abstract: Technology and services available for mobile video have evolved since previous studies were conducted. A majority of teens now have mobile phones and there is an increase in younger users watching video through paid subscriptions. We set out to describe the scenarios in which Millennials and teens view mobile video, their motivations, and how they access video. Twenty-four participants completed diary entries over a five-day period. Nine of these participants also took part in individual interviews that followed. Our findings describe the scenarios and motivations in detail and highlight two main findings. 1) Several mood and emotional states, beyond boredom and killing time, lead to viewing mobile video. 2) When accessing video, choices are made based on desired level of engagement, stimulation, and length. This study provides information that can be used to inform mobile video experiences and proposes opportunities for future research.

Let’s Play My Way: Investigating Audience Influence in User-Generated Gaming Live-Streams

Pascal Lessel — German Research Center for Artificial Intelligence (DFKI), Saarland Informatics Campus, Saarbrücken, Saarland, Germany
Michael Mauderer — Department of Computing, University of Dundee, Dundee, United Kingdom
Christian Wolff — Saarland University, Saarland Informatics Campus, Saarbrücken, Saarland, Germany
Antonio Krüger — German Research Center for Artificial Intelligence (DFKI), Saarland Informatics Campus, Saarbrücken, Saarland, Germany

Abstract: We investigate how the audience of gaming live-streams can influence the content. We conducted two case studies on streams in which audience influence is central and in which the audience can directly participate: First, we review an existing format of the Rocket Beans TV channel and describe how the audience can influence its course of action. With this, we illustrate current practices for integrating the audience. Second, we report the results of our investigation of a “Twitch Plays Pokémon” (TPP)-like setting in which the audience shares the control of the main character through aggregated chat messages. We explored a wider range of techniques than the original TPP offered and found that this can help the audience to organize itself in more nuanced ways. From both case studies, we synthesize results that are of relevance for streams that want to give the audience more influence.

On time or not on time: A user study on delays in a synchronised companion-screen experience

Christoph Ziegler — IRT, Munich, Germany
Christian Keimel — IRT, Munich, Germany
Rajiv Ramdhany — British Broadcasting Corporation, London, Greater London, United Kingdom
Vinoba Vinayagamoorthy — British Broadcasting Corporation, London, Greater London, United Kingdom

Abstract: One major challenge in creation of compelling companion screen experiences, are the time delays between the presentation of content on the TV compared to the presentation of content on the companion screen. Through the use of a synchronised, interactive textbook application, we conducted a user study to evaluate the potential influence of different delays, between the TV and the companion screen, on how users experience watching a Shakespearean play on the TV. Our results indicate that although users do not notice delays of up to 1000 ms, for the kind of experience tested, they feel significantly more distracted by the tablet content for increasingly higher delays. We discuss the implications of our findings with regards to the time delay tolerances users might have when using a synchronised text accompaniment to these kinds of TV programmes.

How People Multitask While Watching TV

Auriana Shokrpour — Samsung Research America, Samsung , Mountain View, California, United States
Michael Darnell — Samsung Research America, Samsung, Mountain View, California, United States

Abstract: We often think of TV watching as the activity where people are fully engaged and immersed in the TV program. However, research has shown that there is a continuum of levels of attention while watching TV. We set out to understand multitasking behaviors as well as users’ motivation and intention behind simultaneous tasks performed in front of the television. We conducted an in-home qualitative research methods study inside ten households across the San Francisco Bay Area and used a quantitative method for analysis of the large amount of behavioral data we gathered. We recorded participants’ television watching behaviors using cameras that were placed in their homes and used retrospective interviews to gather purpose behind events that were observed in the video recordings. We defined eye gaze elsewhere than on the TV as accounting for a multitasking event. It was found that multitasking occurred almost 40% of the time when people were seated in front of the television. Most multitasking occurred during TV programs – not during the interval between TV programs. Of the time people spent multitasking, 36% was spent on a device, mostly a smartphone. However, only 10% of device-related multitasking was related to the content being played on the TV. With our study, we contribute to the greater body of foundational knowledge around common multitasking behaviors that are conducted in front of the television.

Project Orpheus A Research Study into 360° Cinematic VR

Mirjam Vosmeer — Amsterdam University of Applied Sciences, Amsterdam, Netherlands
Ben Schouten — Amsterdam University of Applied Sciences, Amsterdam University of Applied Sciences, Amsterdam, Netherlands

Abstract: When creating content for virtual reality, filmmakers find that they need to re-evaluate the tools they have traditionally used to tell their stories, and explore the new possibilities that this particular medium has to offer. To determine how storytelling- and filmmaking tools function in VR, the concept of presence is currently being re-evaluated for its possibilities to be used as a measurement of the relative effectiveness of these tools. The research project Project Orpheus is presented as a case study into trans-medial storytelling, exploring how the impact of a traditional television show may be reinforced by an immersive VR experience. The movie was subsequently used to conduct a small qualitative study into the use of 3D sound to guide the viewers attention in VR.

Media Multitasking at Home: A Video Observation Study of Concurrent TV and Mobile Device Usage

Jacob Rigby — UCL Interaction Centre, University College London, London, United Kingdom
Duncan Brumby — UCL Interaction Centre, University College London, London, United Kingdom
Sandy Gould — School of Computer Science, University of Birmingham, Birmingham, United Kingdom
Anna Cox — UCL Interaction Centre, University College London, London, United Kingdom

Abstract: Increasingly people interact with their mobile devices while watching television. We evolve an understanding of this kind of everyday media multitasking behaviour through an analysis of video data. In our study, four households were recorded watching television over three evenings. We analysed 55 hours of footage in which participants were watching the TV. We found that mobile device habits were highly variable between participants during this time, ranging from 0% to 23% of the time that the TV was on. To help us understand this variability, participants completed the Media Multitasking Index (MMI) questionnaire. Results showed that participants with a higher MMI score used their mobile device more while watching TV at home. We also saw evidence that the TV was being used as a hub in the home: multiple people were often present when the time the TV was on, providing a background for other household activities. We argue that video analysis can give valuable insights into media multitasking in the home.

Don’t Leave — Combining Sensing Technology and Second Screens to Enhance the User Experience with TV Content

Daniela Huber — LMU Munich, Munich, Germany
Daniel Buschek — LMU Munich, Munich, Bavaria, Germany
Florian Alt — LMU Munich, Munich, Germany

Abstract: In this paper we explore how the use of sensing technologies can enhance people’s experience during perceiving TV content. The work is motivated by an increasing number of sensors (such as Kinect) that find their way into living rooms. Such sensors allow the behavior of viewers to be analyzed, hence providing the opportunity to instantly react to this behavior. The particular idea we explore in our work is how a second screen app triggered by the viewer’s behavior can be designed to make them re-engage with the TV content. At the outset of our work we conducted a survey (N=411) to assess viewers’ activities while watching TV. Based on the findings we implemented a Kinect-based system to detect these activities and connected it with a playful second screen app. We then conducted a field evaluation (N=20) where we compared (a) four hints to direct users’ attention to the second screen app and (b) four types of second screen content requiring different levels of engagement. We conclude with implications for both practitioners and researchers concerned with interactive TV.