Visualizing vocal expression

Sound, especially speech, is ephemeral. It is a highspeed, ordered, multichannel stream that plays for a time, and leaves shadows of its presence in our memories. When we communicate, we exchange semantic, expressive, and relational messages. Most of our communicative power lies outside semantics, yet these expressive and relational exchanges are underexplored. We and others have experimented with visual representations of speech, yet little is known about the interpretability, usability, and efficacy of the visualizations; here we focus on interpretability. We provide a system for expressive vocal analysis, new voice visualizations which map vocal parameters to different designs, and a study focusing on the interpretability of the resulting voice visualizations.