Study Results
Subjects
The 17 subjects who tested various versions of WiiPaint included students, staff, and children of staff from Wellesley College. Results of the background survey showed a variety of academic and gaming backgrounds. All subjects were at least somewhat comfortable with using new technology. Interest in video games and frequency in playing games varied, as did experience with playing other video game systems. Only one of the subjects played the Wiimore than once a month. The most popular favorite game genres were Puzzles, Role-Playing, and Simulation.
I held 10 1-hour long sessions with one to three people in each session. Each of them was given a consent form and background survey, then asked to explore WiiPaint on their own, but they were allowed to ask questions. WiiPaint was projected onto a screen in the classroom where the study was held.
What do subjects try to do first?
- The first thing most subjects try to do is figure out the basic controls - how to draw and change color. With the 4-Wiimote version, users first tried to figure out which dot on the screen corresponded to their Wiimotes.
- After figuring out basic controls, users experiment with what they can draw. Most of them prefer doodles, abstract, geometic drawings.
WiiPaint-specific issues
These issues pertain to bugs found within WiiPaint. Here are the most pertinent issues.
- Shape selecting does not work well because the boundary box is a rectangle that holds the shape, not around its outline. For example, the user could not draw an X because when drawing the second line, it would pick up the first line instead. This makes drawing on top of existing shapes impossible unless it was drawn outside of the rectangular boundary box, then dragged back in.
- The shape recognizer algorithm needs to be more specific. Sometimes it would draw rectangles in place of circles.
- Scaling/rotating should be instantaneous, not gradual. The acceleration based on the distance between the Wiimotes makes it harder to control.
- WiiPaint should give the user more control, such as keeping multi-colored lines and selecting which strokes to manipulate/delete.
- Significant lag in the 4-Wiimote version that makes the 2-Wiimote versions seem really fast.
Interaction-specific issues
These results were most interesting and valuable because what I thought would be an intuitive technique turned out to be too complex.
- When more than one user was present, they would often look to the other person to see how to use the Wiimotes. If the function of their Wiimote is different from the other one, they are not sure what to do next. Because there are many buttons on the Wiimote, it would probably work best if their functions were consistent across Wiimotes (same mapping), or not use them at all.
- In trying to figure out how to change color, users pointed the Wiimote at an XY plane rather than using 3D wrist or arm gestures to change the yaw, pitch, and roll. It was easier to imagine a color wheel or palette located on a 2D plane than try to get the Wiimote at the correct angle. A separate window for the color patch was suggested.
- The point-and-click interface encouraged point-and-click type movements on the user end. It also made WiiPaint feel like MS Paint. There is some discordance between point-and-click interface and 3D interaction.
- The scaling/rotating mechanism was not intuitive. Users pointed the Wiimotes closer or further away from the screen to try to make the shape bigger or smaller, not closer or further from each other like the iPhone pinch.
- Collaboration was sometimes difficult because users ended up drawing their own shapes instead of composing an image together. I think this issue comes from trying to figure out too many features at once. This shows that if the system is not easily learnable, users will focus too much on working through the system rather than working with each other. Likewise, using two Wiimotes simultaneously can become confusing if the user is too focused on the functions on one of the Wiimotes.
- Visual/tactile feedback can improve usability.
Wiimote Limitations
- Infrared motion sensing is hypersensitive. It picks up every small hand shake, and drawing off-screen creates unwanted sticky and jittery effects. IR is not very forgiving, and makes the user feel like they are trying too hard when drawing something specific. However, it works fine for abstract art.
- The buttons are distracting because users feel like they all have a function, but only a few of them do. Either use all of the buttons, or none at all. A museum setting does not allow for a lot of time to learn and memorize all button functions, so the interface should be kept as simple as possible.
Similar interfaces
Many subjects said that WiiPaint reminded them of MS Paint, and would like to see more features such as an eraser brush or background fill.
WiiPaint also reminded users of:
Users also mentioned they wanted to see the following kinds of applications and features: