Have you ever heard of the Wizard of Oz? Yes, you read it right. We were task to perform behavioral prototype or Wizard of Oz to test gestural interface for apple TV.
Build and test a behavior prototype for the following scenario:
Gesture recognition platform: a gestural user interface for an Apple TV system that allows basic video function controls (play, pause, stop, fast forward, rewind, etc.). The gestural UI can be via a 2D (tablet touch) or a 3D (camera sensor, like Kinect) system.
Your prototype should be designed to explore the following design research and usability questions:
• How can the user effectively control video playback using hand gestures?
• What are the most intuitive gestures for this application?
• What level of accuracy is required in this gesture recognition technology?
In considering the design prompt provided for this assignment, we decided on the following parameters for our behavior prototype:
1. Our prototype will demonstrate a 3D gestural UI for an Apple TV system.
2. Our behavior prototype and user testing will focus on the following 7 functionality of the Apple TV:
3. We will establish a unique gesture for each of the above 7 functionality and an initiate gesture ahead of the user testing.
4. Our user testing will examine and validate the following four areas:
a. Users’ ability to command the Apple TV through gestures without instructions or training on the actual available gestures.
b. Users’ ability to command the Apple TV through gesture after some explicit in structions/training on the actual available gestures.
c. Different range of motions (to determine the ideal range for the users) for the gesture UI.
d. The need for a trigger/initiating motion/gesture to activate the gesture UI.
5. The user scenario is one where the user is standing in front of his large screen TV and casually browsing and sampling through different channels and shows as he is not sure what he would like to watch at the moment.
1. We used a 15-inch Macbook Pro situated on a moving cart (to elevate the screen level) to micmic a large screen TV.
2. We considered the embedded camera on the Macbook Pro to be the gesture sensor that takes in user input.
3. Since we don’t possess an actual Apple TV, we used the iTune player to micmic the Apple TV interface.
4. We placed the user 5-10 feet in front of the Macbook Pro.
5. One of our team members was situated a distance behind the user/tester with a Apple remote to control the UI behavior in response to the user/tester’s gestures.
6. We set up two cameras (one on a tripod and another one on a high desk) to capture the videos of the user testing from two different angles/perspectives.
7. We also recorded the actual screen on the Macbook Pro (we actually did not have enough room to splice the screen recording into our prototype video. Instead, we used visual indicators in our final edited video to show whether a gesture from the tester resulted in a desired action on the Apple TV).
8. We asked the user/tester to think out loud and speak the actions/commands that he was gesturing (even though this is not a voice activated system) so that we can follow the tester’s intention.
Discovery & Learning from the Evaluation Session:
1. Without any instructions or training, majority of the gestures that the user/tester attempt ed during our first evaluation session actually matched closely to the gestures we designed/decided on with the exceptions of play, pause, and stop. The user had a hard time guessing what the gestures should be for those three features.
2. The gestures we designed for the play, pause, and stop were not distinguished enough from each other. This caused some confusion on the user’s part.
3. We found that having a trigger/initiation gesture to activate the gesture UI was important to avoid segmentation ambiguity (where the system confuses intended gestures from unintended gestures.)
4. Our user preferred a range of gesture motion to be approximately within the width of his body. Larger gestures motion not only causes strain on the users but also (as the user insightfully pointed out) can create confusion on the system’s part (e.g. it would be harder to tell where a gesture begins or ends when motion for fast forwarding or rewinding.)
5. With a very low cost setup, we were able to observe the user behavior and validate (or in some cases, invalidate) our assumptions going into the the evaluation session.