The number of face to face conversations is consistently decreasing as new technologies increasingly available. We attempt to add back some of the nonverbal information lost with the addition of visual cues. These cues are in the form of visualized auditory data. That is, one sees what one hears. We will look at evolution of the conversation mapping program in this thesis. Using four visualizations we asked users to describe to us how the visualizations aided or disrupted their conversations. With that feedback, we try to discover any support the program gives users. We conclude with our results on how to continue adding back nonverbal cues in conversations.
With online forms of communication evolving rapidly, sometimes, people feel the real essence of communicating is getting lost in the process. After emails, there was instant messaging. After instant messaging, came voice over IP. People have tried to find ways to bring more meaning to these conversations that are not held face to face. In instant messaging, emoticons were used to express facial expressions. That is just one form of nonverbal cues. However, with voice conversations, emoticons do not make as much sense. This is why we are looking at ways to visualize audio. We are creating visualizations to represent what one hears, just like one emoticons were used to represent what one sees.
To test our visualizations impact on voice over IP conversations, we performed numerous user studies. In these studies, two users were first asked to have a normal conversation. No visualizations were used. Each person was in a separate room. There was no way of seeing each other. After this initial conversation, which is used as a control, the users were then introduced to the program and its visualizations. Users had the basics of program explained, but it was left up to the users to find more meaning what was being displayed. After this conversation, users were asked to fill out a short survey, answering questions on how they felt the visualizations helped or hindered the conversation.
Analysis of the studies showed a few things. We expected that the visualizations would encourage the conversation to be more balanced. That happened to some extent. Users who talked more without the visualizations talked less with them. Users who talked less without the visualizations ended up talking more. However, not so much as to balance the conversation, but rather the conversation significantly tipped towards person who was quieter with the visualizations. We also found out that the visualization that most accurately depicted the conversation also turned out to be the users favorite of the visualizations. Finally, when looking to see how the interface was explored, users seemed to spend little time at one visualization at a time, until they found the one they favored the best. From this we gather that they spent less time trying to find out what the interface might tell them, but rather which interface was more visually appealing at the time.