MIT Jerk-o-Meter etc.

Groupmedia research: Extract four basic signals from speech dialogues, possible inside a mobile phone. Use them to provide feedback on the quality of a conversation (Am I a jerk?) or to predict whether people feel attracted or are going to close a contract.

Audio and Mood

Basic idea of a number of works: play back or create the right sound or music for the user's current mood or to change the user's current mood in a determined way. This is a problem of affective computing. These are the basic issues:

What is emotion? What is mood? Can we define them by numbers, for instance by reducing them to a blend of basic emotions?
How can we measure emotion? Typical inputs: heart rate, breath rate, skin resistance, skin temperature. Need signal processing and pattern classification to go from these inputs to emotions.
How can we determine the mood of a piece of music? On may ask listeners, experts, try to compute the mood from the musical score or from the wave data.