Linköpings universitet SE–581 83 Linköping
Linköping University | Department of Computer and Information Science
Bachelor’s thesis, 16 ECTS | Information Technology
2019 | LIU-IDA/LITH-EX-G--19/059--SE
A three-phase user study
evalu-ating the integration of a
general-ized playback bar for a branched
video player
En användarstudie i tre delar som evaluerar integreringen av en
generaliserad uppspelningsvisare för förgrenade videor
Madeleine Bäckström
Linn Hallonqvist
Supervisor : Niklas Carlsson Examiner : Marcus Bendtsen
Upphovsrätt
Detta dokument hålls tillgängligt på Internet - eller dess framtida ersättare - under 25 år från publiceringsdatum under förutsättning att inga extraordinära omständigheter uppstår.
Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut en-staka kopior för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning. Överföring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten, säkerheten och tillgängligheten finns lösningar av teknisk och administrativ art.
Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfat-tning som god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan form eller i sådant sammanhang som är kränkande för upphovsmannens litterära eller konstnärliga anseende eller egenart.
För ytterligare information om Linköping University Electronic Press se förlagets hemsida http://www.ep.liu.se/.
Copyright
The publishers will keep this document online on the Internet - or its possible replacement - for a period of 25 years starting from the date of publication barring exceptional circum-stances.
The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/hers own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility.
According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement.
For additional information about the Linköping University Electronic Press and its proce-dures for publication and for assurance of document integrity, please refer to its www home page: http://www.ep.liu.se/.
©Madeleine Bäckström Linn Hallonqvist
Students in the 5 year Information Technology program complete a semester long software development project during their sixth semester (third year). The project is completed in mid-sized groups, and the students implement a mobile application intended to be used in a multi-actor setting, currently a search and rescue scenario. In parallel they study several topics relevant to the technical and ethical considerations in the project. The project culminates by demonstrating a working product and a written report documenting the results of the practical development process including requirements elicitation. During the final stage of the semester, students form small groups and specialise in one topic, resulting in a bachelor thesis. The current report represents the results obtained during this specialization work. Hence, the thesis should be viewed as part of a larger body of work required to pass the semester, including the conditions and requirements for a bachelor thesis.
Abstract
When watching interactive branched video, the viewer is given the opportunity to tai-lor the storyline of the video playback. This type of video puts the users in control of their viewing experiences and provides content creators with great flexibility how to personalize the viewing experience for individual viewers. When it comes to regular linear videos, the concept of having a playback bar visually presenting the playback is a well established im-plementation used for most (if not all) video players, but for interactive branched videos, that is not the case. Instead, most branched video implementations are typically custom-made on a per-video basis (e.g., see custom-custom-made Netflix and BBC movies) and do not use a playback bar.
With the goal to fill this void, a branched video player with a generalized playback bar that visualizes the tree-like video structure was developed in 2018 by students of the Information Technology program at Linköping University and will be made public with a publication in the near future. Within the preparatory work prior to making this branched video player public, this project included a three-phase user study, where we evaluated the playback bar and its implemented features and compared the video player with alternative designs.
With this thesis, we highlight the value of a branched video playback bar and provide interesting insights into how it, and other design customization features, may best be in-tegrated into a branched video player. Furthermore, we present an improved version of the video player in which the new implementations are based on, and motivated by, the results from the user study. Finally, we describe how further investigations may be done to evaluate the improved version of the video player.
Acknowledgments
We would like to thank our supervisor Niklas Carlsson for all of his support throughout this work. Also, we would like to thank Eric Lindskog and Jesper Wrang for all their valuable help and input. We also want to thank all the participants taking part of our study, providing us with useful observations and feedback. Lastly, we thank Mimmi Cromsjö and Martin Christensson for giving us valuable comments and ideas for improving the thesis.
Contents
Abstract iv
Acknowledgments v
Contents vi
List of Figures viii
List of Tables x 1 Introduction 1 1.1 Motivation . . . 1 1.2 Aim . . . 2 1.3 Research questions . . . 2 1.4 Delimitations . . . 2 1.5 Contributions . . . 2 1.6 Publication list . . . 3 2 Background 4 2.1 Branched video streaming . . . 4
2.2 Usability testing . . . 4
2.3 Related work . . . 5
2.4 Existing systems . . . 6
3 System design 8 3.1 The video player . . . 8
3.2 The playback bar . . . 9
3.3 Features: Visual appearance of the playback bar . . . 10
3.4 Features: Couple story choices to the playback bar . . . 12
3.5 Additional design options . . . 14
4 User study 16 4.1 Part 1: Branched video player implementations . . . 16
4.2 Part 2: Comparison questions . . . 17
4.3 Part 3: Features and additional design options . . . 18
4.4 Setup . . . 18
4.5 Potential influence during the study . . . 18
5 User study results 20 5.1 Part 1: Branched video player implementations . . . 20
5.2 Part 2: Comparison questions . . . 20
5.3 Part 3: Visual appearance of the playback bar . . . 26
5.5 Part 3: Additional design options . . . 31
6 User study analysis and discussion 34
6.1 Part 1: Branched video player implementations . . . 34 6.2 Part 2: Comparison questions . . . 34 6.3 Part 3: Features and additional design options . . . 35
7 Post user study work 37
7.1 Applied improvements based on user study . . . 37 7.2 Ideas for a future user study . . . 37
8 Discussion 41
8.1 Method . . . 41 8.2 The work in a wider context . . . 42
9 Conclusion 43
List of Figures
3.1 The branched video player; without a playback bar . . . 9
3.2 The branched video player; including the playback bar . . . 10
3.3 The zoom and follow feature; before reaching first branch point . . . 10
3.4 The zoom and follow feature; upper branch selected at first branch point . . . 11
3.5 The prune non-selected paths feature; no story choices have yet been made . . . . 11
3.6 The prune non-selected paths feature; lower branch selected at first branch point . 11 3.7 The fish-eye effect feature; closing in on first branch point . . . 12
3.8 The fish-eye effect feature; leaving last branch point . . . 12
3.9 The mouse-eye effect feature; hovering close to second branch point pair . . . 12
3.10 The highlight path when hovering over button feature; hovering left button . . . . 13
3.11 The clickable playback bar feature; hovering upper branch . . . 13
3.12 The buttons on playback bar feature . . . 14
3.13 The matching branch labeling feature . . . 14
3.14 Tree structure when using arc tangent . . . 15
3.15 Tree structure when using inverse hyperbolic cosine . . . 15
3.16 Tree structure when using absolute value . . . 15
4.1 The different parts of the user study . . . 16
5.1 Average NASA-TLX effort scores; independent of implementation order . . . 21
5.2 Average NASA-TLX effort scores; seeing implementation with playback bar first . 21 5.3 Average NASA-TLX effort scores; seeing implementation without playback bar first 22 5.4 Average SEQ; independent of implementation order . . . 22
5.5 Added complexity compared to linear videos . . . 22
5.6 Added understanding when having a playback bar . . . 23
5.7 Added understanding when having a playback bar; depending on implementa-tion order . . . 24
5.8 Comparison of video player implementations . . . 25
5.9 Comparison of video player implementations when seeing implementation with playback bar first . . . 25
5.10 Comparison of video player implementations when seeing implementation with-out playback bar first . . . 26
5.11 Preferred implementation; independent of implementation order . . . 27
5.12 Preferred implementation; seeing implementation with playback bar first . . . 27
5.13 Preferred implementation; seeing implementation without playback bar first . . . 27
5.14 Zoom and follow result . . . 28
5.15 Prune non-selected paths result . . . 28
5.16 Fish-eye result . . . 29
5.17 Mouse-eye result . . . 29
5.18 Highlight when hovering over button result . . . 30
5.19 Clickable playback bar result . . . 30
5.21 Matching branch labeling result . . . 30
5.22 Tree structure design result . . . 32
5.23 Timing result . . . 32
List of Tables
4.1 Explanation of implementation and video order . . . 18 5.1 Ideas for alternative and additional design options that were brought up during
1
Introduction
In this paper, we evaluate the design and implementation of a novel branched video player that includes a generalized playback bar by performing and presenting the results from a three-phase user study. Additionally, we compare this player with alternative designs and analyze the integration of the playback bar and other branch video related features and lastly we implement additional features and present an outline for yet another user study to be performed in future work. Within this chapter, we give an extensive illustration of the work as a whole and explain the aim and purpose of this thesis.
1.1
Motivation
The concept of interactive videos is giving the viewer the opportunity to create a storyline of the viewers’ liking. This is done by having the video player presenting an arbitrary number of options on the screen that the user can choose from and these choices impact how the story of the video plays out. In abstract terms, the potential plot sequences available to the user can be portrayed with a tree structure. Within the tree structure, the separation of tree branches corresponds to a branch point in the video, at which the user is asked to make a path choice and the edges represent video segments that the are played out between these branch points (or the start and end of the video playback sequence). For the video player, this means that the playback path is adapted based on the selections of branches that the user makes and it results in a customized plot sequence.
The description above illustrates the flexibility and creative power that branched video streaming provides. However, despite attempts by production houses such as for example Netflix and BBC, branched video has yet to become mainstream. We believe that one reason for this is that most of these videos use per-video custom made user interfaces that gives each video a unique feel. Instead, we argue that it is important that users are presented with a generalized interface that easily can be reused for many videos and that it provides a clear visual way to extract information about upcoming branch choices, playback progress and buffer levels. In the context of traditional ”linear” videos, these are important aspects that we all expect modern video players to provide. Yet, as of today, there currently does not exist any generic playback bar for branched video that helps visualize these aspects. In addition to providing information about upcoming branch choices, playback progress, and current buffer levels, we have also found that such a generalized playback bar can help quickly increase
1.2. Aim
users’ understanding and appreciation of both the choices they have and the branch video concept itself.
1.2
Aim
In broad terms, the main goals of this study are to determine how a video player for branched videos should be designed and how a playback bar for these types of videos best is integrated and displayed. To achieve these goals, we leverage an existing implementation to perform a three-phase user study, in which each part focus on one specific part of the evaluation.
1.3
Research questions
Given the above aims, we set out to address the following research questions:
1. How is a playback bar best designed in a video player used for branched video? 2. How can a playback bar help users when watching branched video?
3. How much complexity and perceived effort does a playback bar add to the viewer experience when watching branched video?
4. How should a branched video player be designed to provide the best user experience?
1.4
Delimitations
As explained in earlier sections, we wish to determine how a branched video player is best implemented. However, it is worth pointing out that the participants taking part of the study are all university students, of ages ranging from 20 to 30 years old. Them being students can imply that they all have quite a good sense of how video players in general work and in combination with the fact that the range of age is limited we can not ensure that the results are applicable on people in general.
Also, many of the participants have experience of seeing other branched videos prior to taking part of the study, which can lead to them making comparisons to those particular videos and its corresponding video players when describing their impression of the video player implementations in this work.
The video material that is used to create the branched videos that is presented during the study originates from a linear video. This can give the participants the impression that making story choices in a branched video creates a storyline that does not completely fit and therefore impacting their overall sense of the video player. Therefore, it would have been more appropriate to use branched video material, but this was not an available option for us. This thesis mainly focuses on evaluating the usage of branched video players when used in the purpose of entertainment. We do not explore the opportunity to use this type of video player within other areas.
1.5
Contributions
In this thesis we present three-phase user study that evaluates a novel branched video player including a generalized playback bar. Based on the results from the user study we implement additional features to the video player that we believe are good additions to the video player that improves it even further. Lastly, we design an outline for yet another user study that can be used as a foundation to evaluate the new video player including the features that is to be applied due to conclusions from the user study results.
1.6. Publication list
1.6
Publication list
This thesis is based on a conference paper [14]. Here, we summarize this work.
• Eric Lindskog, Jesper Wrang, Madeleine Bäckström, Linn Hallonqvist, and Niklas Carlsson. “Generalized Playback Bar for Interactive Branched Video”. In: Proceedings of the ACM International Conference on Multimedia. 2019, pp. 2369–2377.
2
Background
2.1
Branched video streaming
Branched video streaming is a concept that allows viewers to affect the storyline of a video by providing the viewer with various choices throughout the video. This kind of interactive story telling is also referred to as nonlinear or multipath video streaming [4][12]. Depending on the choices a viewer makes, the story path will be different and the video can possibly result in different endings.
Branched video streaming enables viewers to be more interactive when watching a video and gives video creators a wider range of creative possibilities since multiple storylines can be created. The structure of branched videos differs from regular, linear videos that consists of video frames in a single sequence. It can instead be described as a tree structure composed by multiple linear sequences that are linked together in a tree or graph structure [4].
2.2
Usability testing
The user study that we preform in this thesis is built on some selected methods for testing the usability of a product. The overall structure of the user study is constructed using a previous study conducted by Gutwin et al. [9] as an inspirational source. The methods we use are NASA Task Load Index (NASA-TLX) and Single Ease Question (SEQ) which are described in this section.
In addition to this we construct additional, more open, questions that focused on the overall impression of the video player, letting the users explain in a more detailed manner how they perceived the video player implementations.
NASA Task Load Index
NASA-TLX is a popular multi-dimensional technique used for measuring subjective men-tal workload. It was developed in the 1980’s [7] and can be used in many different areas. Through the weighted average of ratings from six subscales, NASA-TLX can derive an av-erage workload score. The six subscales are ranged from ”very low” to ”very high” and are labeled mental, physical and temporal demand, performance (which in this thesis is called ”difficulty”), effort and frustration [2]. In the user study preformed in this thesis, a 1-20 scale
2.3. Related work
(non weighted) version of NASA-TLX is used. We choose to evaluate each of the six subscales separately by calculating the average score subscale-by-subscale, not calculating an average workload score. The reason for choosing to work with this technique in our user study is the fact that the tool has a purpose that corresponds well with the purpose of this work, i.e., evaluating workload, and is very easy to use.
Single Ease Question
The SEQ, revealed by its name, is one single question regarding the perceived difficulty of a task. Participants are asked to finish the sentence ”Overall, this task was” by filling in a scale from 1 to 7, where 1 corresponds to ”very difficult” and 7 to ”very easy”. A single question that is asked right after a task is performed is an easy way to compare what different parts of a product are perceived as most problematic to users. Since the question follows right after a task is preformed, participants are more likely to remember their exact experience of the task compared to being influenced by multiple different tasks before answering questions [8]. By using this method for measuring perceived difficulty, we are able to get an overall sense of how the branched video player is received by the participants. Since the aim of the technique corresponds well with the purpose of this work, we believe that it is an appropriate tool for us to use.
2.3
Related work
Previous work within the area of branched video streaming has been done by Meixner et al. [17][16]. They present a web and a mobile version of a video player for showing branched video. The web version displays a choice-making panel in the video when arriving at a branching point. It also has clickable objects in the video which allows the user to jump to another scene in the video [17]. The mobile version of the player has buttons for choice making beside the video instead [16].
Other work within this area has been done by students, which this thesis is strongly con-nected to. In particular, Lindskog and Wrang [13] have developed a player for branched video that includes a playback bar to display the chosen story path and the amount of data buffered. In their thesis, Lindskog and Wrang present a few features and designs for the playback bar which are tested in a smaller user study with eleven participants. Based on the results from the user study, they present various features and design options for the play-back bar which is intended to be tested in future work. Furthermore, Lindskog and Wrang developed a basic authoring tool for creation of branched videos.
Meixner et al. [17] have work related to the technique of creating branched videos as well. They have developed an authoring tool called SIVA Producer. This tool allows users to cut and arrange parts of the video by drag-and-drop.
Krishnamoorthi et al. [12][11] have explored the challenges brought by branched video during playback and how adaptive streaming can be used for playing multi-path video. They present a branched video player that is using HTTP-based Adaptive Streaming (HAS) which focuses on delivering video with seamless transition between video segments without play-back interruptions.
One area that is similar to branched video is called multi-view video streaming. This is a concept explored by Carlsson et al. [3] and it allows viewers to switch between multiple per-spectives of the same event provided by different cameras. In their work, they introduce the concept of a ”multi-video stream bundle” which consists of several parallel video streams, all providing videos of the same event filmed by different cameras and being synchronized with each other in time. The delivery is made using HAS, adapting both content and rate since viewers have the possibility to choose their viewpoint. They present a prototype implemen-tation and sample results which are shown to provide close to seamless playback switching when the bandwidth is adequate to prefetch the video streams.
2.4. Existing systems
The type of setup of multiple cameras that are used to provide different angles of objects and scenery’s is an important subject when it comes to 3D videos, which is quite similar to nonlinear videos in the aspect of showing a viewer content based on what this person chooses to see. Toni and Frossard [20] describe that this can be done by using reference camera views that then is used to synthesize additional viewpoints that could be of interest. Toni and Frossard explain the difficulties of providing a client with the best possible reference camera views and present a concept called multiview navigation segment that reduces the complexity for computation that still preserves the optimality in the majority of 3D scenes.
Other work on multi-view video streaming has been done by Ma et al. [15] where the challenges brought by this type of interactivity is analyzed. They look for the optimal design of how to represent the data by introducing new rate and distortion models and solving algorithms. The experiment results show an out-performance of the used baseline solution, providing a lower resource consumption and higher visual quality.
Xiu et al. [22] investigate the challenges of providing seamless transmission between dif-ferent viewpoints, making the view of the content appear to be completely connected.
Hamza and Hefeeda [10] have explored HAS for video delivery, focusing on free view-point video streaming. This is a concept that allows viewers to choose viewview-points that have not been capture by any camera. Instead, the viewpoint needs to be composed by multiple video streams and depth information from different viewing points.
For multi-view video streaming to be possible, video synchronization is required to align the different video segments captured by different cameras. Wang et al. [21] present a model to handle spatiotemporal alignment of an arbitrary number of videos. Within their work they bring up different example applications where the model can be useful. One of these applica-tions is Crowdsourced Video which is quite similar to multi-view video streaming. They show an example where they use their model to align a series of videos of a pool game captured by different cameras to create a single video that cuts between the different cameras.
Another aspect of synchronizing different videos is presented by Shrestha et al. [19]. To capture the moment of different events such as weddings, parties and vacations, a lot of people tend to record these using for example their mobile phones. This can result in many people recording the same event from different angles. To be able to combine the videos taken by different cameras, a synchronization of time is needed. This is something that is usually done manually, but Shrestha et al. present two methods to synchronize different videos of the same event based on the audio content of the recordings.
2.4
Existing systems
As mentioned in section 2.3, Meixner et al. [17][16] have developed a video player for branched video. Beyond this, a couple of other video players have been launched.
Netflix has recently aired the interactive video ”Black Mirror: Bandersnatch”. The video player that is used for watching this movie has buttons integrated in the video player. When-ever the viewer is presented with a choice, a black bar at the bottom of the player appears. In this player, the viewer has limited time to make a choice. The remaining time a viewer has to decide is visualized with a thin line that gets shorter as time runs out. If the viewer fails to make a choice within the given time, the player by default takes the choice to the left in the screen [18].
Eko Hello is an interactive storytelling platform that started with making interactive mu-sic videos and has now different entertainment videos based on branched video [5]. Their video player has integrated buttons that are graphically designed to fit the choices presented to the viewer. The player includes a subtle timeline that highlights the previous branching points where a user has been presented with a choice [5].
2.4. Existing systems
Finally, Eko Hello has an existing authoring tool to create interactive videos and lets cre-ators design their own user interface for their video. The different user interface components are React components. The user interface system is based on React and Redux [6].
3
System design
As explained in section 2.3, the design of the video player was not implemented in this work. However, we believe that it is of use to give a brief description of how the video player operates as a whole to give the reader a good understanding of the system. Moreover, we explain in a more detailed manner how the design options (i.e., the features for the video player and additional design options) that is to be tested within the user study works, looks and the different features’ corresponding purposes.
The design options that are presented below are developed with different goals, these be-ing either to help the viewer of the video with the understandbe-ing of how the video and the story choices corresponds to the tree structure in the playback bar, or enhancing the under-standing of what parts of the tree structure that is currently being played. All design options will be presented separately but it should be mentioned that it is possible to apply them to the video player not only singly, but they can be combined together in any way one might prefer. However, they are all evaluated separately in the user study to make the user focus only on one part at a time.
3.1
The video player
The branched video player that the user study within in this thesis is based on was developed (and, prior to this thesis, also improved) by Eric Lindskog and Jesper Wrang and is presented in their published thesis [13]. Here, we give a short introduction to the basics of the video player as a whole.
The video player is based on the framework by dash.js. The video segments that are used to set up the branched video are implemented with JSON and the segments are then included in a list of segment objects that corresponds to different branches in the branched video. All segments originates from a video and by setting the start time and length for each segment you pick what parts from the original video the segment should show when being played. Within each segment there are multiple fields, giving each a unique id, start time, length, branches and display name. The field called branches sets what other segments that the particular segment can be followed by in the branched video when played, and display name is the text which is shown on the buttons for choosing story plot during playback.
The video player is displayed as a normal video player window and when the video is played, the video starts at the root segment (which is set within the code) of the branched
3.2. The playback bar
Figure 3.1: The branched video player; without a playback bar
video. When the video is closing in on a branch point (i.e., a point in the video where the user is prompted to make a story choice), the story choices available are (by default) showed on large buttons on top of the video screen. An example of how this can look is shown in Figure 3.1, where two story choices are available to the viewer. When either button is clicked, the video continues by playing the plot of the story that the viewer has chosen. In the case where the viewer does not make a choice before reaching the end of the played video segment, the video comes to a halt and it will wait until the viewer makes a choice.
The number of story options available by each branch point is optional, but for this study it is limited to two story choice options and was not changed at any point of the study. Also, the number of times that the viewer is prompted to make a story choice is optional, but for this study it is limited to only two branch points and is not changed at any point of the study.
3.2
The playback bar
As an addition to the basic video player explained in the previous section where the structure of the branches within the video is not shown, the playback bar is introduced. With this, the viewer of the video is able to see how far into the video the playback has reached, how much of the video that is left to be watched and the particular branches that are available. How much of the video that has been played is distinguished using a red color. Also, the viewer has the opportunity to witness when and where the viewer will be prompted to make a story choice. This is done by visually showing a tree structure of the entire video, including all of the branches available, at the bottom of the video screen. Shown in Figure 3.2, where it can be discerned that the viewer chose to take the upper branch at the first branch point and is now prompted to make a second story choice, is an example of how the playback bar looks by default. In addition, it is possible for the viewer to see how much of the video that has been buffered. The parts of the video that have been buffered are colored in a light grey tone within the tree structure, while the parts of the tree structure that have not are colored in black.
3.3. Features: Visual appearance of the playback bar
Figure 3.2: The branched video player; including the playback bar
Figure 3.3: The zoom and follow feature; before reaching first branch point
3.3
Features: Visual appearance of the playback bar
When watching traditional linear videos the video playback is most often displayed on the screen in its entirety in the playback bar. This makes it possible for the viewer to easily understand how much of the video that has been played and how much of it that is left. However, when having the tree structure of the storyline for a branched video in a playback bar it is possible that it gets more complex to distinguish the current playback position and getting a good overview of the video as a whole. Here, we present different alternatives for displaying the tree structure during playback.
Zoom and follow
When a playback bar is to be applied to a branched video including many storyline options and endings, the size of the playback bar can be large. With this in mind, one can imagine that it might be of use only to display segments of the tree structure in the playback bar that is close to the currently played video section. This has been implemented using a feature resembling a zooming and following effect applied to the tree structure, making the playback bar display smaller portions of the playback in its entirety and therefore not displaying the tree structure as a whole. When playing a video using this feature, the playback bar only displays the most recent segment of the video that has been played and the upcoming branches for that specific video segment. Shown in Figure 3.3 and Figure 3.4 are images of how the video player displays the tree structure when the zoom and follow feature has been applied.
3.3. Features: Visual appearance of the playback bar
Figure 3.4: The zoom and follow feature; upper branch selected at first branch point
Figure 3.5: The prune non-selected paths feature; no story choices have yet been made
Figure 3.6: The prune non-selected paths feature; lower branch selected at first branch point
Prune non-selected paths
The branched video playback bar was, with the default settings, explained to visually present the entire video tree structure during the whole playback of the videos. However, it might be of interest to the viewer of the video only to be able to see the parts of the tree that is or can be a part of the video version that one can watch, due to previous story choices. Using the feature prune non-selected paths, the playback bar removes the branches that has not been chosen and the ones that can not be chosen, due to earlier selected branches. This makes the tree structure only display one step ahead, not giving information about how the closest upcoming story choice will affect the rest of the story plot. Displayed in Figure 3.5 is an example of how the playback bar looks when no story choices have been made, and Figure 3.6 shows how the playback bar has removed the upper part of the tree structure after the lower branch was chosen at the first branch point.
Focus-based visual distortion: Fish-eye effect
In some sense, this feature resembles the earlier presented feature zoom and follow, in the way that it makes the playback bar focus on the parts that is close to the current position of the playback point. This implementation uses a zoom in like effect that enlarges the point of the tree structure that is slightly placed ahead of the currently played position. In Figure 3.7 an example of this is shown, when the playback is closing in on the first branch point, making that part larger than the other parts of the tree structure. During the entire playback, this feature enlarges the closest upcoming segments of the tree structure and shown in Figure 3.8 is an example of how the fish-eye effect feature looks when the second branch point has been passed.
Focus-based visual distortion: Mouse-eye effect
Similar to the fish-eye effect feature presented above, the mouse-eye effect zooms in on parts of the tree structure, but instead of doing this automatically during the entire playback, this is applied by hovering the video player with the mouse. The feature magnifies the area around
3.4. Features: Couple story choices to the playback bar
Figure 3.7: The fish-eye effect feature; closing in on first branch point
Figure 3.8: The fish-eye effect feature; leaving last branch point
Figure 3.9: The mouse-eye effect feature; hovering close to second branch point pair
where the user has placed the mouse cursor. Hence, when the mouse is held still over the video screen the magnifying effect freezes and is only applied to the parts of the tree that it was applied to when the mouse was moved to that position. This is shown in Figure 3.9, where the viewer is currently hovering the playback bar close to the second branch point pair.
3.4
Features: Couple story choices to the playback bar
As mentioned earlier, branched videos can have many storyline options and endings. When watching a linear video, the path within the storyline is clearly displayed as a straight line in a playback bar, but that is not the case when it comes to branched videos. For these types of videos, the playback route is not as unambiguous since the choices that the viewer makes affects it which raises the matter of linking the story choices to their corresponding branches. In this section, we propose features that deals with this.
Highlight path when hovering over button
To increase the level of understanding of how the story choices that a viewer of a branched video makes affects the path within the tree structure displayed in the playback bar, the fea-ture highlight path when hovering over button is introduced. With this addition to the video player, the viewer is made aware of which story choice that corresponds to which branch in the tree structure by highlighting the particular branch in a light, blue color. One example of how this feature works is shown in Figure 3.10, where the viewer of the video is hovering over the left story choice button.
3.4. Features: Couple story choices to the playback bar
Figure 3.10: The highlight path when hovering over button feature; hovering left button
Figure 3.11: The clickable playback bar feature; hovering upper branch
Clickable playback bar
When the branched video is closing in on a branch point, the options for choosing a story path is displayed as buttons on top of the video. With the feature highlight path when hovering over button (presented in section 3.4) it is possible for the viewer to get full control over which branch one prefers to take. However, this requires the viewer to investigate which story choice button that corresponds to which branch in the tree structure. With this feature the viewer can make a story choice decision by simply clicking the branch that the viewer wants the story to follow. This can be done on the branches that corresponds to the ones connected to the upcoming branch point at any point of time, i.e., the viewer does not have to wait until the story choices are made available on the screen. An example of this is shown in Figure 3.11, where the video has reached its first branch point and the viewer is hovering over the upper branch. Worth mentioning is that the large story choice buttons have been turned off in the figure, making clicking branches the only way to continue the video.
Buttons on playback bar
The default settings for the video player is that the story choice buttons are displayed over a large portion of the video screen when closing in on a branch point. This feature places the buttons on top of the playback bar instead. This means that the video screen is not covered when the viewer is closing in on a branch point, but gives the story choice buttons a less screen covering placement. The buttons are placed vertically in contrast to one another and an example of this is shown in Figure 3.12. Worth mentioning is that the figure showing the buttons on the playback bar has the big story choice buttons turned off, making the buttons on the playback bar the only way to continue the video.
3.5. Additional design options
Figure 3.12: The buttons on playback bar feature
Figure 3.13: The matching branch labeling feature
Matching branch labeling
The feature matching branch labeling is another version of visually illustrating to the viewer which branch that corresponds to which story choice button (i.e., similar to the feature high-light path when hovering over button). With this feature, the branches connected to an upcoming branch point is linked to their corresponding story choice button by matching each button and branch with a specific letter, letting the viewer know how the story choice will affect the path in the tree structure. Figure 3.13 shows an example of how this feature looks when reaching a branch point.
3.5
Additional design options
In addition to the features presented above, numerous other effects and design options can be applied and changed for the video player. Two of these aspects are investigated within the user study: how the tree structure in the playback bar should be designed and how long prior to a branch point that the story choice buttons should be made visible to the viewer.
Tree structure design
Regarding the design of the tree structure, four options was created. The design of the tree structure is calculated by using mathematical formulas in the code of the video player and can therefore be changed in numerous ways. This is done by applying mathematical functions to calculated values of the starting point and ending point of each branch segment. Since there are many formulas that can be used for this, we decided to focus on functions that we believe distinguishes the different branches by a branch point fairly quick but still have a clear difference in their appearance.
3.5. Additional design options
Figure 3.14: Tree structure when using arc tangent
Figure 3.15: Tree structure when using inverse hyperbolic cosine
Figure 3.16: Tree structure when using absolute value
The first tree structure design was implemented by using the arc tangent function and the result of using this formula is displayed in Figure 3.14 and this is the standard tree design that was used during the study.
Secondly, the tree structure design used the inverse hyperbolic cosine and this option is presented in Figure 3.15. This option makes the branches follow a more straight line than when using arc tangent.
The third options that was used for the tree structure applies the absolute value to calcu-late the shape of the branches, which is displayed in Figure 3.16. With this option one can see that the branches follow a straight line with an upward slope.
Lastly, one design sketch of the fourth tree structure design option was created. With this potential option, all branches are shown as straight lines, aside from the most close parts to a branch point. By the branch points, the branches are separated by letting each curve in a 90 degree upward or downward (depending on the placement by the branch point) angle. This design structure was not implemented, but we believe that it can be of interest for the study since it can be seen as a sort of mixture of the arc tangent and a straight line that normally is used for linear video players. This option is not shown in the video player as the other three design options for the participants but on paper. Because of the fact that this fourth option has not been implemented, all the four options are all sketched out on the same paper when the discussion of the tree design takes place in the study.
Timing
One additional aspect of design options for the video player that the study investigates is the timing of when to display the story choices to the viewer, i.e., how long before a branch point to make it possible for the viewer to decide on how the story should continue. To be able to get a good understanding of what timing that is preferred, we use three timing options; seven seconds, five seconds and three seconds.
4
User study
The performed user study consisted of three parts, each focusing on different aspects regard-ing the video player, the introducregard-ing of a branched video playback bar and the concept of branched video as a whole. Before beginning the study, we briefly explained the concept of branched video to each participant, giving all participants the same starting point. Figure 4.1 shows the outline of the user study, divided into three different parts which we here describe. In addition, we present the setup of the study and describe the potential influence that we might could have had during the study.
4.1
Part 1: Branched video player implementations
The first part of the study focused on getting an understanding of what users felt when using two branched video players; one being the video player presented in section 3.1 including the playback bar presented in section 3.2 and the other one being the basic branched video player without the playback bar. This to distinguish if, or how much, a playback bar adds to the overall complexity and perceived effort when watching branched videos. This was done
4.2. Part 2: Comparison questions
by presenting the two different video players one after the other and letting the participants watch a video and make story choices, with the task to make the video as fun as possible. Each of these videos took approximately one minute to watch, depending a bit on how much time the user spent making their choices at the branching points. After each video had been shown, the users taking part of the study were asked to fill out a form. The form included the NASA-TLX questions and the SEQ explained in section 2 and the issued form included the exact same questions no matter of which player that the questions concerned. In addition to letting the participants answer questions, we collected measurements of the users’ actual response times when closing in on a branch point (and both videos included two branch points each, making all participants pass four branch points in total within this part of the study). This, to identify if the presence of a playback bar has any impact on how much time a user needs to make a story choice and to gain information about which timing option that should be used for a branched video player.
To reduce the risk of getting different results depending on which of the two video players that was shown to the users first, the number of studies which showed the video player with the playback bar first and the number of studies which showed the video player without the playback bar first was the same, i.e., 16 times each. In addition, we wanted to reduce the risk that the video material that the video players played had any influence on how the users perceived the video players. This was done by using two different videos, each being shown once to every user. The specific order of which the different combinations of video player and video that was shown to each user is shown in Table 4.1, showing the setup for the first four users that is then repeated in a round-robin fashion.
4.2
Part 2: Comparison questions
In the second part of the study the main focus was to compare the two different video players, making it possible for us to identify potential tradeoffs associated with the playback bar. This was done by giving the users a form containing comparison questions covering different aspects. First, the user was asked to answer a question comparing both the two branched video players with a regular, linear video player, regarding the amount of complexity that the branched video players add. To reduce the risk of getting different results from different users depending on what kind of linear video player that they thought of, all users were presented with our own, simple linear video player (having a regular playback bar showing at the bottom of the screen), prior to answering this question.
After the comparison question concerning linear video players versus our branched video players, the users were to answer questions comparing the two branched video players with each other. Every question within this questionnaire was answered using a NASA-TLX-like scale, rating both the video player including the playback bar and the one without the play-back bar. We were interested in finding out how much understanding the playplay-back bar pro-vides when watching branched video. Therefore we let the user answer questions about how much understanding they thought the playback bar contributed with regarding the branched video concept, the video structure and choices at hand, whether there would be upcoming branches, remaining playback time and amount data buffered. Furthermore, we asked the users to compare the two players for branched video regarding speed and accuracy of pick-ing the desired path and which player provided the best help to make good decision and the best user experience.
As a final part of the comparison questions the users were asked what their personal overall preference was when watching branched video, a player with or without playback bar. Following to the final question there was a box where the participants were given the opportunity to motivate their answer and give additional comments.
4.3. Part 3: Features and additional design options
Structure of user study: Part 1
Study no. Video player 1 Video 1 Video player 2 Video 2
1 With playback bar Version 1 Without playback bar Version 2 2 With playback bar Version 2 Without playback bar Version 1 3 Without playback bar Version 1 With playback bar Version 2 4 Without playback bar Version 2 With playback bar Version 1
Table 4.1: Explanation of implementation and video order
4.3
Part 3: Features and additional design options
Finally in the third part of the study, we investigated how the features and additional de-sign options, presented in sections 3.3, 3.4 and 3.5, were perceived by the users. All features and design options were presented one-by-one and each showing was followed by questions covering whether or not the user experienced that the feature or design option was a good addition to the video player or not. The time that each participant had with each feature depended on how fast they discovered it. Some of the features made a big difference of the video player and did not require a lot of time to notice for most of the participants. If we noticed that a user had discovered and understood the function of a feature, we did not in-sist on having the user watch the entire video. Instead we asked the user if they understood the feature and if they did, we moved on to the questions. When all features and design options had been presented the user study was finalized by giving the participants summa-rizing questions, with the purpose of getting an overall sense of the general impression from each participant. Within this part the participants were also given the opportunity to explain what features he or she might felt was missing and would have enjoyed to see, that had not been implemented.
4.4
Setup
Throughout the entire study, the same video material was used; Big Buck Bunny1, which is freely available.
The equipment that was used was a laptop (Asus Zenbook UX430) running Windows 10, with a 14-inch, 1920x1080 screen and an Intel processor (i7-8550U). When showing the users that took part of the study the video players, the player was set up to use up the full screen and the cursor was placed on the video player.
A total of 32 people participated in the study, 10 being female and 22 male, all being students at our local university and of ages ranging from 20 to 30 years old. Each study took approximately 40 - 50 minutes and each participant was offered a snack and a drink as a thank you for participating.
During each study, one of us had the role of being a secretary, responsible for reading the questions and taking notes of the answers that the participants gave orally. The other one of us was in charge of showing the participants the video player, managing the switches between the different implementations of the video player and ensured that the right feature was shown at the right time.
4.5
Potential influence during the study
To avoid influencing the participants’ answers, we made sure to not ask any leading ques-tions. In some cases, participants struggled to understand or observe some of the features that were presented. In these cases, instead of directly pointing out and explaining the fea-ture, we showed the participants the video again to gave him/her another chance to notice
4.5. Potential influence during the study
the feature. The few times that the participant still did not notice the feature, we provided hints on where on the screen to focus, after which everyone noticed the feature. In the cases that a participant did not understand a question (provided in writing) and asked us about any question, we tried to clarify the question without influencing the participant’s answer. For example, a few participants said that they had not noticed any information about buffer-ing when viewbuffer-ing the video and therefore did not understand our question regardbuffer-ing how much the playback bar help understand this aspect. Here, we simply encouraged them to answer the questions according to their own experience and explained that one could choose to answer with the option ”don’t know”.
One thing that we did that can be seen as ”misleading” was the fact that when we were to ask the user the final questions, we asked if the participant wanted a quick recap of the fea-tures that he or she had seen. That could have lead to that the people that was given the recap easier could explain what they liked or disliked and the people that did not want/get the re-cap maybe forgot a few features that they actually liked. Also, the people that were given the recap maybe had forgotten about how they actually felt about the features - having us explaining them maybe made the user see the feature in another way than they remembered them.
During the first part of the user study, the setup was to hand out the NASA-TLX question-naire for each implementation immediately after presenting them each. However, it is pos-sible that this outline could have influenced some of the answers that we received. During some of the user study sessions, we noticed that participants, when receiving to the NASA-TLX questionnaire for the second implementation, had forgotten how they had rated the first implementation, and therefore had a hard time comparing them both to each other. This could therefore be considered as potential influence, since we can not be sure that the results would have been the same if both of the NASA-TLX questionnaires would have been handed out after both of the implementation had been shown.
5
User study results
In the following section, all the results that have been collected from the user study will be presented in the way that conclusions can be drawn in an appropriate and interesting manner. The results will both be of quantitative and qualitative character and will be presented using graphs, tables and text; depending on how the results are best interpreted.
5.1
Part 1: Branched video player implementations
Here, the results from the first part of the user study are demonstrated. As mentioned in section 4.1, this part focused on comparing the perceived effort using NASA-TLX and SEQ when adding a playback bar to a branched video player.
Figure 5.1 shows a summary of the mean score regarding the NASA-TLX questionnaire (scale being given 1-20; low-to-high), independently of the order of which video player and video that was presented to the users first. Included in each mean score bar is a two sided 95 % confidence interval.
In addition, we present the results from the same NASA-TLX questionnaire depending on which of the video player implementations that was shown first to the participants (i.e., either being the one with the playback bar or the one without the playback bar). The results of the mean scores are shown in Figures 5.2 and 5.3 and these results are calculated on 16 participants each (half of the total number of participants).
In Figure 5.4 the mean score from the SEQ question has been summarized, independently of the order of which video player and/or video that was presented to the users first, where a score of 7 indicates the most positive answer possible.
5.2
Part 2: Comparison questions
In the second part of the user study, as explained in section 4.2, the users got to compare the two different branched video implementations. First, the two branched video implementa-tions were compared to a regular linear video player to measure how much complexity the users experienced was added by the branched video concept itself. The other comparison was between the two different branched video implementations, with or without playback bar. Lastly, an overall preference questions was asked as described in section 4.2. In this section, we present the results from this part.
5.2. Part 2: Comparison questions
Figure 5.1: Average NASA-TLX effort scores; independent of implementation order
5.2. Part 2: Comparison questions
Figure 5.3: Average NASA-TLX effort scores; seeing implementation without playback bar first
Figure 5.4: Average SEQ; independent of im-plementation order
Figure 5.5: Added complexity compared to linear videos
5.2. Part 2: Comparison questions
Figure 5.6: Added understanding when having a playback bar
Added complexity compared to linear video
The results from the comparison between the two branched video implementations and the linear video regarding complexity is shown in Figure 5.5. The users answered a question about how much complexity they thought was added by the different implementations by filling in a scale from 1 to 20, where 1 referred to ”no complexity at all” and 20 referred to ”a lot more complexity”. The average score was 8.06 for the implementation with playback bar and 5.91 for the implementation without playback bar.
In addition, we present the results regarding added complexity when the order of which implementation was shown first to the users, with or without playback bar, is taken into con-sideration. The users that were introduced to the playback bar first had average scores 8 and 5 out of 20 for the implementations with playback bar and without playback bar respectively. The users that saw the playback bar second, after first having seen the implementation with-out the playback bar had average scores 8.5 and 6.5 for the two implementations respectively.
Comparing watching with or without playback bar
As mentioned in section 4.2, the users were asked to answer questions about how much understanding they felt the playback bar contributed with regarding different aspects. The different aspects were: the understanding of the branched video concept, the video structure and choices at hand, whether there will be upcoming branches, remaining playback time and amount data buffered. The results are shown in Figure 5.6. For the question regarding ”the branched video concept”, there was one participant that answered ”don’t know” and therefore the average score for this question was calculated with respect to 31 users. The same applies to the question regarding added understanding of the amount of data buffered, where 9 users answered ”don’t know” and therefore the average score was calculated with respect to 23 users.
In Figure 5.7, the result of the perceived understanding added by the playback bar is presented with the order of which implementation that was shown first taken into
consid-5.2. Part 2: Comparison questions
Figure 5.7: Added understanding when having a playback bar; depending on implementa-tion order
eration. The average scores are higher regarding all aspects except ”amount data buffered” for the users that saw the implementation without the playback bar first. The average score for ”the branched video concept” for the users seeing with playback bar first was calculated on 15 users, because of one ”don’t know” answer. Regarding ”amount of data buffered”, the average score was calculated on 11 users for the users seeing with playback bar first and 12 for the users seeing without playback bar first. This was due to the total of 9 ”don’t know” answers for this question.
To gain understanding about how the playback bar could be helpful for the user, the participants answered four comparison questions were they got to rank the two implementa-tions regarding speed, accuracy, help to make a good decision and user interface. The mean score results are shown in Figure 5.8. Regarding speed of picking desired path, participants ranked the implementation without playback bar slightly higher and regarding user interface for branched video, a lot higher. For accuracy of picking desired path and best help making a good decision the implementation with playback bar was ranked higher. The mean score regarding accuracy was calculated on 28 participants since 4 answered "don’t know" on this question, for both with and without playback bar.
Figures 5.9 and 5.10 show the responses to the question that compared the two implemen-tations (with or without playback bar) with the order in which the implemenimplemen-tations were shown taken into consideration. Figure 5.9 show the results for the participants who saw the implementation with playback bar first and Figure 5.10 shows the ones that saw the im-plementation without playback bar first. The average scores for ”best help to make a good desicion” is calculated on 15 users for the participants that saw the implementation with play-back bar first an on 13 users for the ones that saw without playplay-back bar first. This is because of the total of 4 ”don’t know” answers regarding this question.
5.2. Part 2: Comparison questions
Figure 5.8: Comparison of video player implementations
Figure 5.9: Comparison of video player implementations when seeing implementation with playback bar first
5.3. Part 3: Visual appearance of the playback bar
Figure 5.10: Comparison of video player implementations when seeing implementation with-out playback bar first
Overall preference
The final part of the second phase of the user study consisted of questions regarding the users personal overall preferences, as explained in section 4.2. The question regarding whether the users preferred watching branched video with or without playback bar is shown in Figure 5.11. Out of 32 participants, 25 would prefer without playback bar and 7 would prefer with playback bar. Performing binomial hypothesis testing (under the null hypothesis that there is no bias in the user preference between the players), we obtain a p-value p below 0.05 (in favor of the alternative hypothesis that there is a bias towards selecting the implementation without a playback bar). These results hence shows (at a 95% confidence level) that there is a bias towards selecting the implementation without the playback bar.
Figures 5.12 and 5.13 show the results depending on which implementation was shown first and second. Out of the 16 users that first saw the implementation with the playback bar, only 1 preferred having the playback bar (p<0.05) over not having it. Out of the 16 users that were introduced to the playback bar after first having seen the implementation without the playback bar, 6 preferred having the playback bar (p>0.05) and 10 preferred not having it. Calculations of the probabilities show that when seeing the implementation without playback bar first we see no bias towards the implementation without playback bar at a confidence level of 95 %, but we do when the implementation with the playback bar was shown first.
5.3
Part 3: Visual appearance of the playback bar
Within this section we present the results regarding the response from the users that concerns the features that affected the visual appearance of the playback bar and calculate the binomial hypothesis distribution, not taking the ”don’t know” answers into consideration. Hence, the number of trials n is equal to the number of participants that either answered that they liked
5.3. Part 3: Visual appearance of the playback bar
Figure 5.11: Preferred implementation; independent of implementation order
Figure 5.12: Preferred implementation; seeing implementation with playback bar first
Figure 5.13: Preferred implementation; seeing implementation without playback bar first
or did not like a specific feature. This type of calculation will be done for the results presented in section 5.4 as well.
In addition to the overall impression from the participants, some extra comments that were provided during the study are presented.
Zoom and follow
Out of the 32 participants, 13 liked and 13 disliked this feature (p>0.05, n=26). Additionally, 6 answered "don’t know" and this result is presented in Figure 5.14.
The most common answer to why the feature was liked was related to the decrease of tree size. For these participants, it was perceived as more screen efficient and less stressful compared to having to see the entire tree structure all the time.
The majority of the participants that disliked the feature said they felt disturbed and stressed by having two parts moving at the same time - the video itself and the playback bar. Some participants did not like the fact not being able to see the entire tree since they
5.3. Part 3: Visual appearance of the playback bar
Figure 5.14: Zoom and follow result Figure 5.15: Prune non-selected paths result
enjoyed having an overview of the tree structure and to know what paths to expect. Some participants on the other hand, found the part about only showing parts of the tree a positive feature, although they would prefer having the tree structure still.
Prune non-selected paths
Prune non-selected paths got a result that is shown in Figure 5.15. The result show significance in favor of the feature (p<0.05, n=30). Here, the users stated that they liked that it ”removes unnecessary information”, ”reduces the remaining tree size”, ”focus on the part of interest”, ”reduces the chance regretting past choices”, and that it ”feels more realistic”. The users that did not like this feature said they preferred to see the entire tree all the time.
Focus-based visual distortion: Fish-eye effect
The feature that added a fish-eye effect to the tree structure when hovering over it was the least popular of all the features, only 1 out of 29 participants said they liked it. The result is displayed in Figure 5.16 and it shows significance not favoring the feature (p<0.05, n=29). The only positive comment we received from the user that liked the feature was that it provided a bit of understanding of the current position of the user in the tree structure since it had a zooming effect. The participants that disliked the feature said they ”did not understand the purpose of it” and that it ”was hard to follow”, ”drew attention from the video screen”, ”was stressful and distracting because it moves both left and right” and ”made it hard to comprehend how much time was left until the next branch point was reached”.
Focus-based visual distortion: Mouse-eye effect
Similar to the fish-eye effect feature, the mouse-eye effect feature was not very popular. As shown in Figure 5.17, only 3 out of 32 participants said they liked it which shows significance for not liking the feature (p<0.05, n=29). The most common reason that users had for disliking the feature was that they did not understand the purpose and that the feature did not help the user in any way. They said the feature appeared as confusing, added unnecessary complexity to the video experience and was annoying.
5.4. Part 3: Couple story choices to the playback bar
Figure 5.16: Fish-eye result Figure 5.17: Mouse-eye result
5.4
Part 3: Couple story choices to the playback bar
This section will treat the responses from the participants considering the features that affect the coupling of the story choices to the playback bar.
Highlight path when hovering over button
This feature was appreciated by 18 participants, disliked by 8 and 6 did not know whether they liked it or not. This result is summarized in Figure 5.18, showing a bias favoring the feature (p<0.05, n=26).
The users that liked this feature said that it was ”simple”, ”made it clear what path you consider choosing”, ”made the user feel more in control”, and ”connects the purpose of the playback bar to the buttons”. The users that did not like this feature felt that it ”did not con-tribute with anything useful”, ”was not that noticeable”, ”takes focus away from the video”, ”reveals information about the number of candidate paths”, or felt that ”a similar feature may be better used for the purpose of informing users what paths they have already watch” (e.g., when watching a video the second time).
Clickable playback bar
This feature was not liked by any of the participants (p<0.05, n=28), as seen in Figure 5.19. It appeared to many as unclear and confusing, some participants said that using this feature would make the purpose of branched video go away. As a viewer, many said that it should be clear what choice that is being chosen, not a ”random click”.
We asked the participants whether they would prefer having only buttons to make their choice, only clickable branches or having the opportunity to click on either one of these. The majority of the participants said they would prefer only having the buttons, and that having both could contribute to a more complex viewing experience. The fact that it would be a much smaller part that is clickable if only having clickable branches was frustrating for some participants. In general however, participants liked that the branch was highlighted when hovered over.
5.4. Part 3: Couple story choices to the playback bar
Figure 5.18: Highlight when hovering over
button result Figure 5.19: Clickable playback bar result
Figure 5.20: Buttons on playback bar result Figure 5.21: Matching branch labeling result
Buttons on playback bar
More than half (as shown in Figure 5.20) liked this feature, that explicitly placed the buttons in the tree shown on the playback bar, which does not show any bias towards liking this feature at a confidence level of 95 % (p>0.05, n=30). Some of the users that liked this feature felt that this approach was less distracting and that it was easier to follow the video plot compared to the large buttons. Some participants said that this feature made it easy to know which branch that was connected to which story choice.
The users that disliked this feature said that it felt like it blocked the playback bar, or that it drew focus from the screen. Additionally, some participants said it was hard to read the text because of its small size compared to when the text was displayed on the large buttons.
