Augmented Reality Museum Exhibit - The Resting Satyrs

Teammates: Kelly McClinton, Leif Christensen, Emily Baumgartner, Kyle Overton

My contributions: Co-team lead(with Kelly McClinton), experience design, usability testing

Publication

This work was published in Interaction Magazine in July/August of 2018.

Overview

This museum exhibit was based on the research of Kelly McClinton on Praxiteles' Resting Satyr statue. This Greek statue was copied by many Roman sculptors, and more than 100 have survived into the modern era. One of these statues rests in the Eskenazi Art Museum at Indiana University. The museum asked us to create a museum exhibit using Augmented Reality technology around this statue.

We were able to create an interactive exhibit in the Microsoft HoloLens that allowed patrons to compare the physical statue with virtual statues that are housed in different parts of the world, along with a speculative reconstruction of the Eskenazi Satyr, of which only a torso survives. Patrons use gaze interactions to activate audio, and toggle between different statues in the experience.

Assumptions

Going into the project the team goal was clear, create an augmented reality museum experience for the Eskenazi Art Museum, but I hoped to learn specific things about AR technology:

• How will users react to holograms in a public space?
• What interaction methods will be appropriate for a museum?
• Will users understand how to explore in AR?
• How will we guide users through the AR experience?

Problem

This opportunity arose out of a casual conversation with the Eskenazi Art Museum, who were interested in exploring emerging technologies in the museum. After exploring the space and selecting The Resting Satyr as a focal point and the Microsoft HoloLens as a platform, we asked the question, "What experience will support this artifact in Augmented Reality?"

Methodology

• Understanding the museum space through observation of museum patrons and exhibits
• Interview with museum curator to understand how she designed exhibits and viewed art
• Behavioral prototypes with the HoloLens to test technical feasibility and reactions
• Usability Testing - We tested and iterated for 8 weeks once we had a solid demo in place, beginning in February of 2017.

Major Insights

• Museum patrons often don't know what questions to ask about a piece.
  • Teaching them how to look is valuable
• The museum can be a sacred space for many. Any experience that disrupts that space will be problematic.
• The artifact should be central to the experience

Early Concepts

Iterations

Satyr Play

Goal - We initially wanted to give the viewer context on the Satyr as a historical character. We considered animations appearing above the virtual and physical statues depicting the Satyr, or perhaps an entire section of the exhibit that depicted Satyrs in action, as the mischievous entities that they were.

Result - As we explored the space further, we realized that technical limitations of the HoloLens would make it hard for the viewer to take in and keep track of these types of animations. The field of view was too narrow. We also began moving towards a simpler view of our experience. Focused on the artifacts, rather than the technology.

Walk through Time

Goal - Our next iteration focused on how the viewer would experience the space as a historical experience. As patrons moved through the exhibit, virtual statues would appear, leading to the physical statues. We also considered breaking up the space with virtual walls with doorways to walk through.

Result - Several things kept us from pursuing this direction. First, breaking up the space with virtual walls would obscure the patrons view, and potentially lead them to run into, or knock over cases or artifacts, which was clearly unacceptable. Secondly, we did not have complete information about the historical context or provenance of the statues, so placing them accurately in a timeline wasn't possible. Also, the other exhibits in the space were organized by culture, and placing Roman statues next to artifacts from other cultures might have been confusing.

Roman House

Goal - To add more context to how these statues might have been viewed, we considered placing them in the context of a virtual Roman house. This idea extended to the rest of the gallery, where we would add a virtual Roman temple, and call out other instances of Satyrs and Dionysus, who Satyrs were related to.

Result - Time constraints, mostly related to the creation of assets needed to complete this experience led to us not pursuing this idea initially, but we have plans to pursue it in the future.

Doorway Selection

Goal - Our final iteration before moving towards our final design involved two doorways that would allow users to select whether they wanted audio included in the experience, or wanted to view the experience silently. In this design, users walking through the audio door would have walked into audio zones that would play automatically.

Result - After considering this option and having our developer test out some code, we realized that this option privileged the technology over the artifact. In the end, we chose a more traditional museum exhibit style, placing 3D models next to a physical statue. For interactions, we included a small menu to select different virtual statues to view, and audio triggers, all activated with gaze controls.

Testing

We performed 8 weeks of usability testing, covering more than 20 participants as the app developed. This testing was performed in the museum, while it was open to other patrons, and we were bound by all the normal museum rules. Not an ideal testing environment, but it was important to test in the real environment, because the AR experience is embodied. We went into the testing with important questions to answer:

• How will users react to the holograms?
• Is the HoloLens technology sufficient for the experience we want to convey?
• Will the technology be too distracting for the user to experience the artwork?
• Will the user be able to navigate the museum space while wearing the HoloLens?

In most cases, we answered these questions to our satisfaction, in others, our answers were incomplete. For example, we found that users treated the holograms as real objects, rather than trying to walk through them. This required us to redesign the exhibit in order to allow walking paths around the edges. We also found that users were able to navigate the space, because the HoloLens doesn't cover the users entire vision. We did notice that the virtual statues had a tendency to pull focus from the physical statues, but that issue could be solved. 

Our testing also resulted in many efficiencies and protections for setting up the experience, including a user mode and admin mode activated by voice commands (which were not included as part of the normal experience), and ways to set up the experience in under 5 minutes.

Reflection

This was the first design project I completed as a team lead that involved a functional deliverable, not just a design concept. I learned quickly that while we were able to come up with a strong design, the realities of the space, time constraints, and the technology would inherently limit our options. Our initial ideas involving lots of animations and dynamic interactions were pulled back to something that could be completed within the timeframe. 

We learned these lessons over 8 weeks of usability testing, taking place in the museum with the physical artifact. Our initial prototype just had virtual models that we could place in space. With that, we were able to begin testing expectations of users, and understand how they would react to the holographic images. For example, we learned early on that patrons would generally not walk through the holograms, but would always walk around them, unless they were specifically testing boundaries. In our initial design, we didn't leave enough space around the outside of the virtual statues, and patrons were hesitant to move behind them. We were able to redesign the placement for the next week, and saw immediate changes in behavior. This testing/iteration pattern continued throughout the testing process.

Of the other questions I wanted to answer, I learned that users don't really understand how to explore in AR, and it was very difficult to guide them through the experience, considering the narrow field of view of the HoloLens. Often, users would miss things that we intended them to see because they weren't looking directly at it. As far as interaction methods, hand gestures would not have been appropriate, and voice commands were questionable. So we designed everything around sustained gaze. Users initially had trouble with this, but those who stayed in the experience longer understood it quickly. 

The other major challenge was the constraint of the museum. We were working in an existing space, with a very limited ability to move anything, and strict requirements for what interactions we could design. We had to create interaction, but also discourage patrons from reaching out. Many of our decisions in the final design were thought of as ways to train patrons that they didn't need to reach out.

In addition, working directly with a stakeholder on a project in their space was challenging. For this project, we weren't creating a deliverable that we would hand off. Instead, we were creating an experience for an event which we would run, which was in an active museum. Planning this event and finding the right timing was a major challenge. The museum was incredibly gracious and supportive, but getting quick answers was not always feasible. We had to work around these constraints and set our own deadlines and milestones in many cases.