Authors: Maribeth Gandy, Blair Macintyre, Peter Presti, Steven Dow, Jay Bolter

The paper is about an emerging idea from Augmented Reality research (AR is like VR, except that it uses head-mounted displays to place the virtual world on top of the real world — instead of a completely virutal world) to build AR Karaoke / Movieoke experiences. Movieoke is a NYC bar game, where, in a movie analog to Karaoke, one can act out scenes from a moive. Another motivation is the rise of “rhythm games” like DDR and Donkey Konga, in which a player doesn’t just keep the beat, but also re-invents the moves to creatively interpret the dance and to impress the audience. One piece of related work is ALIVE (by P. Maes et al).

Design goals for AR Karaoke are “to give the participant the directions they need to succeed.”

• Improvisation: ability to creatively interpret the scene
• Replayability: It’s going to take a few tries to get the lines, the blocking, etc. correct. People will want to play again to “get it righter.”
• Simple Interfaces for Artistic Expression: Take a complex skill (acting) and create a simplified version of it that is doable by novices.
• Feedback: it’s not going to be points, or not dieing, but instead positive feedback from the audience.

The sytem overlays the virutal actors into the world (on the stage). The system also tracks the participant, so we can track her in real time as well. AR Karaoke provides four kinds of information to help a participant: the text (dialog to speak, and sometimes stage direction), a progress bar which indicates how long the user has to speak the current line, a map view with a ghost for blocking information (and sometimes dots or other info that is visible on the ground), and a direction display (arrows) for helping players with prop movement.

The paper has 3 main examples: a scene with the cowardly lion from the “Wizard of Oz”, a scene from the “Princess Bride” with a swordfight between Inago Montoya and the Man in Black, and a scene from “Four Angry Men.” Some conclusions / future work: can we set up multiplayer experiences? How can we help participants better predict upcoming lines and actions? How can we involve the audience more (i.e., what can they watch of the experience?)? How would one go about getting this into a deployable state?