Give us the “elevator pitch” on your app, “ROAR”?
ROAR is about recreating at home the social experience of watching a sports event in a stadium.
When you go to stadium, you watch with friends. We wanted to recreate that local experience. We approached this in two ways: the local experience and the audience experience on the whole.
The local experience:
- First, enabling text chat with friends who are watching the same event.
- Second, with tablets specifically, we saw the potential to recreate the visual communications at sports events—drawing signs on the tablet device and sharing those signs with friends and other people in your “section.” If your sign is really popular, it will spread and more people will see it.
- Third, we wanted to implement polls. So a sports producer can push a poll out to the viewers, such as “bases loaded, one out, are they going to score?” and the results get aggregated back to producer. This closes the feedback loop between the producers and the audience.
The audience as a whole:
- Our iPad interface is two columns, bubbling up two kinds of content:
- One is specific messages from your “section” and the sections around you. In the same way you might hear someone at a stadium shouting from the stands behind you, we wanted vocal users to gain visibility across sections.
- The second is the representation of the crowd overall, allowing the app to display what is trending across the audience, mash it together, and put it in a stream
You mentioned “closing the feedback loop” between audience and sports producers. Would you see producers using ROAR as a tool for broadcast?
Definitely. We see a whole suite of feedback tools that producers can use to interact with audiences and see what their interests are. The polling is one example. Another that I’m excited about is the fan-created signs you can create on the tablets. Just like fans bring homemade sings to games and they get thrown onto the jumbotron, or even make it onto the telecast, we imagine a fan at home creating a sign on his tablet and having that sign appear on the jumbotron at Fenway Park. Producers can monitor the popularity of content and pull content out of the “remote stands,” just as they do the real stands, and use this content in the telecast.
What drew you to sports programming for the application of Automatic Content Recognition and second screen synchronization?
It was partially pragmatic—there is something about the sports experience that lends itself to liveliness and authentic reaction. Scripted shows use a number of production elements that set you up to feel a certain way, whereas reactions in sports are much less predictable and thus much more surprising—you never know what is going to happen. I also came at this from a personal angle, though an indirect one. A number of my friends and I follow live videogame competitions, and during the Starcraft finals there would be thousands of people watching these competitions, but no single place for everyone to exchange in dialogue about what they were watching. I asked myself: how do you do large online group interaction based on what they are watching? In my work toward my Ph.D, I’ve been exploring the mechanics required to do something like that at large scale.
How functional was the app by the end of the hack-a-thon?
It’s a long way from being functional. Our concept was pretty ambitious. The chat works, the drawing works, but the crowds and business aspects are mostly dummy information. This is often the challenge with demos of social technology: if you build the framework with no dummy content it looks empty. It behooves you to fake interaction, so it looks like there is activity, and people can understand what the full experience would look like. It’s really a conceptual prototype—doing it correctly would take tons more engineering.
Do you have a vision of how content recognition will play into the social experiences of the future?
Content recognition is critical to the social experience. When I started exploring this issue of social interaction around content, I was working with streaming media where recognition was not an issue. But now that people are using multiple screens to consume content and interact around that content, recognition becomes a real problem to solve. There is no way to build an experience without it right now. As TV cable boxes publish API, it might be possible to do this without audio recognition across devices—it will be easier to have a centralized solution. But that is far off in the future.
YouTube link to the presentation video: http://www.youtube.com/watch?v=zMtVzbbFLEg&feature=g-all-u
ROAR - A super scale, social TV chat system that makes it easy for users to find friends watching live events, invite friends to come watch with you and keep in touch with the mood of the crowd -- all at the same time.
This dynamic duo of a PhD student and a computer science freshman at MIT are a force to be reckoned with. They are well versed in designing synchronous social experiences and building the technical infrastructure to make it all work. The team Captain has been writing and speaking about this space for both academic and corporate audiences for the past six years.
Drew Harry - Captain
For a PDF of the interview, click here>>