Using a Digital Video Annotation Tool to Teach Dance Composition
Gina Cherry, University of Washington
Janice Fournier, University of Washington
Reed Stevens, University of Washington

Abstract
Encouraging reflective thinking is particularly challenging in disciplines that focus on learning by doing. This paper describes how a computer-based video annotation tool called Video Traces can be used to facilitate reflection and critical evaluation in dance. The Video Traces system allows users to capture video, and to annotate it by talking and gesturing. The resulting “video traces” are then saved and can be viewed, exchanged, and commented on by the creator and others, providing instructors and learners with a way to view and reflect on events which are normally ephemeral. In this paper, we present preliminary findings from a classroom-based study of our use of Video Traces in an undergraduate class in dance composition.

1. Introduction 
Encouraging reflective thinking and critical evaluation is particularly challenging in disciplines that focus on learning by doing. In dance, for example, emphasis is often on students’ performances as evidence of their understanding of the key concepts of the discipline (Warburton, 2002). Yet in central components of the dance curriculum, such as dance composition, students must learn to apply concepts and develop a critical consciousness in order to be successful (Lavender and Predock-Linnell, 2001).

Learning to create dances requires learning to coordinate different types of knowledge. It requires synthesizing intuitive, bodily knowledge and explicit, conceptual knowledge – using explicit knowledge to reflect on and revise one’s practice. It requires coordinating the experience of moving with the visual effects of moving (Arnheim, 1974; Fournier, 2003). And it requires learning to see like an expert – being able to identify the concepts and techniques of choreography in one’s own work and the work of others (Blom and Chaplin, 1982).

Developing these new ways of seeing can be difficult, particularly because learners have limited opportunities to view, reflect on, and discuss their work. In a typical dance choreography class, students make and perform short compositions with and for one another. Compositions are generally performed only once or twice. Student choreographers often dance in their own pieces and are thus unable to see their compositions from an external point of view. When choreographers are able to watch their work, their familiarity with the piece may limit their ability to understand how a viewer, seeing it for the first time, might experience it.

When communicating about a dance, students and instructors must remember the details of the dance, and must coordinate their comments with the actual elements of the piece, without a concrete representation to which to refer. Without a common external point of reference, conversation relies on the possibly faulty assumption that the participants in the conversation have a shared mental model of the dance. To apply the class’s feedback, the choreographer must remember what was said, and must be able to connect it back to the experience of the dance.

We have addressed these issues with a computer-based video annotation environment called Video Traces. The Video Traces system allows users to capture video and annotate it by talking and gesturing. The resulting “video traces” are then saved and can be viewed, exchanged, and commented on by the creator and others.

2. The Video Traces System 
The Video Traces system was conceived of a number of years ago (Stevens & Hall, 1997), but became feasible only with more recent innovations in digital video technology. The design is based on research on human interaction that has emphasized the embodied character of learning and thinking (e.g. Hall, 1996; Stevens & Hall, 1998; Goodwin, 1994; Suchman, 2000). Stevens and the PETTT Video Traces Research Group* at the University of Washington collaborated on the design of the current system. The system provides a simple computer interface that allows users to capture, save, and organize video clips. Users can capture video directly from a camera hooked up to the computer or they can input previously recorded video from digital videotape.

* The Program for Educational Transformation of Technology (PETTT) is a multidisciplinary initiative at the University of Washington that explores the interplay of technology and pedagogy in real settings. The PETTT Video Traces Research Group consists of Scott Macklin, Greg Bowman, Gina Cherry, and Janice Fournier.
An external link toPETTT Video Traces: http://depts.washington.edu/pettt/projects/videotraces.html

To annotate a video, users view the video, talking about what they see. Moving the cursor into the video frame produces a “finger tool” which allows users to point to different parts of the video as it plays. Users also have a number of playback options when annotating (Figure 1): they can change the speed of the video, freeze a frame to talk about a point in greater detail, or rewind the video in order to make a second comment on a particular section. Audio, gestures, and video playback changes are overlaid on the video and preserved in the resulting “video trace.”


Figure 1. Video annotation screen. Users annotate videos by talking as the video plays. Users can change the speed of the video or freeze the video to talk over a specific frame. These speed changes are preserved in the resulting trace. Moving the cursor into the video frame produces a “finger” tool, which can be used to point at things in the video. The movement of the finger is also preserved in the resulting trace.

Video Traces organizes video files and video traces in separate lists, which may be sorted by title, author, date and time recorded, or description. Traces can stand on their own or can be linked together in a threaded discussion—i.e., users can make a trace in response to someone else’s trace, and these responses appear visually linked to one another (Figure 2). Users can also record multiple annotations over the same video.

Figure 2. The traces screen. Users can create stand-alone traces by choosing the "New Topic" option. Users can create a threaded discussion made up of traces by choosing the "Respond to Trace" option. A thumbnail of the first video frame is shown in the lower left corner of the screen.

3. Using Video Traces in the Dance Studio
3.1 Informal Uses of Video Traces in the Classroom
We initially explored potential uses of Video Traces for teaching dance composition in spring 2001. We worked with several student volunteers who used Video Traces to make comments about one of their assignments for a choreography class at a large northwestern university. These students then shared their ideas about a how a tool like Video Traces might be useful in their work (Stevens, Cherry & Fournier, 2002).

We also gave the instructor an opportunity to view the traces that these students created. Dance instructors rarely have a way to collect representations of students' process of creating a dance, and students' work is usually judged on the basis of what is successfully communicated in their final presentations. When the instructor viewed the traces we had collected, she discovered that she learned new things about her students, particularly about their intentions for their work.

3.2 Integrating Video Traces into the Curriculum
Our initial observations suggested that Video Traces had the potential to be a useful tool for teachers and learners of dance composition. The following year, we returned to work with dance instructors to integrate Video Traces into a three-quarter dance composition sequence. In this paper, we present preliminary findings about students' experiences with Video Traces in the final quarter of this sequence.

In the third quarter of the dance composition sequence, students are asked to apply choreography concepts that were introduced in previous quarters to three pieces that they create for dancers other than themselves. For each assignment, students have an opportunity to show a “rough draft” of their work to the class, and receive feedback about their work during a subsequent class discussion. Students then have a week to revise their compositions, after which the final piece is performed for the class.

The instructor for this class decided to incorporate Video Traces in two ways. After the rough draft showings and class discussion, the instructor used Video Traces to give students individual feedback. Students had an opportunity to view this feedback before their final showing. After the final showing, students used Video Traces to reflect on and evaluate their own pieces.

3.3 Data Collection
The findings presented in this paper are based on several sources of data. During our exploratory work, we interviewed our student volunteers about their experiences using Video Traces. We also interviewed the course instructor at the beginning and end of the quarter in which she integrated Video Traces into her curriculum. In addition, we observed class sessions, collected the traces created by the students and the instructor, and asked students to complete a final survey.

3.4 Using Video Traces for Instructor Feedback
Although students do get verbal feedback when they present their rough draft to the class, this feedback is dependent on the audience’s memory of the piece, and tends to be general in nature. Viewing the piece a second time on video gave the instructor an opportunity to embellish on comments she had made during class discussion, to notice aspects of the composition she may not have noticed in the first viewing, and to focus her comments on more specific elements of the dance. Using Video Traces allowed her to freeze the action and point out specific details of the composition with the “finger” tool – a type of feedback that would not be practical during a live performance.

The experience of making traces also gave the instructor insight into her own teaching practices. After her first experience making a trace, she realized that her comments tended to be intuitive rather than focused on explicit criteria for evaluating dances. She thought that it would be helpful to develop specific criteria for evaluating students' compositions. With the help of researchers, she developed a rubric that she used to guide her comments in subsequent traces.

The instructor believed that using Video Traces had a significant impact on the quality of the feedback she was able to give her students. In an interview at the end of the quarter, she commented, "I hear this in different ways every year - 'I've never known what the teacher has wanted' ... This year they said that they had never gotten feedback like this for projects before. Students need feedback every single time and it's got to be in a way that will be useful for them to use for the next project. I've always given them feedback after each project, but it's been less clear, less specific."

Students were asked to view the instructor's traces before their final showing, and to respond to her feedback in their final draft of the composition. In a survey at the end of the quarter, students reported that viewing the instructor’s traces about their work was the component of the course that contributed the most to their learning. Students indicated that they found it helpful to be able to watch their piece and listen to feedback multiple times, and to be able to view their work through a "different lens."

QuickTime movies show an example of the instructor feedback in a rough draft:
Longer version (~62 MB)
Shorter version (~12 MB)

3.5 Using Video Traces for Self-Evaluation
One of the ideas stressed in dance composition classes is that all aspects of a dance should reflect intentional use of the craft. However, many dancers share the belief that “the work should speak for itself” (Blom and Chaplin, 1982); as a result, students in dance composition classes normally have few opportunities to talk about their own work. During class discussions, students are generally expected to listen to how the audience perceived their piece, rather than offer their own explanations about choices they made for the piece. Students rarely practice reasoning aloud about their work.

By making traces, the students in this class did have an opportunity to explain and reflect on their work. The instructor asked students to apply the criteria from the rubric she developed when discussing their own compositions. She encouraged students to discuss their work with a critical eye, and to speak as creative artists about the motivations for their choices.

We found that students were able to apply the rubric to their own work with varying degrees of success. While many students created traces that demonstrated thoughtful reflection about their work, some students simply reported on what they saw in the video, rather than offering explanations or evaluating the success of the piece. Many student traces contained responses to comments made about their rough drafts in the instructor's trace or during class discussion.

Students told us that creating traces helped them look at their work differently. Although students found that simply viewing their piece multiple times was helpful, they also valued the annotation process itself: "With Video Traces, I wasn't just sitting back to be entertained the way I would if I watched a videotape - I was constructing changes, formulating my views in words." In contrast to previous years, when students' final drafts were videotaped for documentation purposes, making traces enabled students to continue thinking about their work even after their final showing.

Viewing the students' traces gave the instructor new insight into how students thought about their work. Reflecting on her experience after viewing the students' traces, she told us that she found that, "...they are thinking more than what is represented in their work visually, and I think that's really important. They had a point of view and a real thought process, and they were thinking about the criteria."

QuickTime movies show an example of the student self-evaluation of a final draft:
Longer version (~54 MB)
Shorter version (~9 MB)

4. Technology Issues
One of the challenges associated with projects such as Video Traces is addressing issues related to the technology itself. Technology is not yet widely used to teach disciplines such as dance, and running Video Traces required a somewhat elaborate setup, including a digital video camera attached via firewire to an iBook laptop computer, an external microphone, and headphones. The fourteen students in this class had access to two complete Video Traces setups. In addition to training students to use the equipment, we had to address logistical issues such as locating a quiet space for students to make annotations and finding secure storage for the equipment. We eventually developed a system that allowed students to borrow the equipment for two-hour periods or overnight.

We found that students were able to use the equipment on their own after a single training session. In a survey at the end of the quarter, students told us that the software was easy to use; however, they cited access to equipment as the biggest obstacle to using Video Traces. Students did not always plan ahead to have access to the equipment when they needed it, and some students were concerned about being responsible for expensive equipment.

At the beginning of the term, the instructor had been concerned about how students would respond to using Video Traces, and about how much class time it would take to set up the equipment. From the start, the instructor made students responsible for setting up the equipment and for videotaping each other’s work. She later expressed surprise at how readily her students adopted the technology: “I think I'm surprised, every time I see a student with the laptop, that they know even how to plug it in. But they don't seem to bat an eye when I say you've got to have this annotated by this day.”

The instructor had also been concerned about her own ability to use the equipment, since she made her traces at home without access to technical support. Despite this concern she learned to use the equipment easily, and was pleased with the results: "I would never have thought that technology would have been part of my pedagogy. But I am learning that certain kinds really can enhance teaching, beyond the bells and whistles of Power Point."

5. Conclusions and Future Directions
In addition to performing, students of dance need to learn to perceive, distinguish, and apply concepts of composition in their practice. Our pilot work suggests that using Video Traces has the potential to enhance these critical evaluation skills by providing a new mechanism for dialogue among participants in dance composition classes. We saw evidence of critical evaluation not only in the traces created by students and the instructor, but also in subsequent class discussions - the instructor told us that using Video Traces seemed to encourage students to talk more critically about each other’s work.

We are interested in how this learning community might appropriate Video Traces in the future. Both the students and the instructor felt that Video Traces should be introduced as an integral component of student work earlier in the course sequence to increase the likelihood that students would more readily use it for multiple purposes. In our exploratory work, students expressed an interest in using the tool in a variety of ways, such as preparing for rehearsals or communicating with their dancers (Stevens, Cherry & Fournier, 2002); however, their ability to do so in our pilot study was limited because of equipment access issues. We plan to address this issue by creating a web-based version of the tool so that users can view traces and annotate videos from any computer that is connected to the Internet.

6. Acknowledgements
We would like to thank the Department of Dance at the University of Washington for working with us on this project. We would also like to thank Scott Macklin, Greg Bowman, and the rest of the PETTT team for their invaluable assistance.

7. References
Arnheim, R. (1974). Art and Visual Perception: A Psychology of the Creative Eye. Berkeley: University of California Press.

Blom, L.A. and Chaplin, L.T. (1982). The Intimate Act of Choreography. Pittsburgh, PA: University of Pittsburgh Press.

Fournier, J. Composing in Dance: Thinking with Minds and Bodies. Unpublished doctoral dissertation, University of Washington, 2003.

Goodwin, C. (1994). Professional Vision. American Anthropologist, 96(3), 606-633.

Hall, R. (1996). Representation as Shared Activity: Situation Cognition and Dewey's Cartography of Experience. Journal of the Learning Sciences, 5, 209-238.

Lavender, L. and Predock-Linnell, J. (2001). From Improvisation to Choreography: the critical bridge. Research in Dance Education, 2(2), 195-209.

Stevens, R., Cherry, G., & Fournier, J. (2002). Video Traces: Rich Media Annotations for Teaching and Learning. Proceedings of the CSCL 2002 Conference on Computer Supported Collaborative Learning, Boulder, CO.

Stevens, R., & Hall, R. (1998). Disciplined perception: Learning to see in technoscience. In M. Lampert & M. L. Blunk (Eds.), Talking mathematics in school: Studies of teaching and learning. New York: Cambridge University Press.

Stevens, R. and Hall, R (1997). Seeing Tornado: How Video Traces Mediate Visitor Understandings of (Natural?) Phenomena in a Science Museum. Science Education, 81(6), 735-748.

Suchman, L. (2000). Embodied Practices of Engineering Work. Mind, Culture & Activity, 7(1 & 2), 4-18.

Warburton, E.C. (2002). From Talent Identification to Multidimensional Assessment: toward new models of evaluation in dance education. Research in Dance Education, 3(2), 103-121.

8. Note
Inquiries about this study can be addressed to Gina Cherry, gcherry@u.washington.edu or Janice Fournier, fournier@u.washington.edu

Inquiries about the use of Video Traces can be addressed to Reed Stevens, reedstev@u.washington.edu