In some embodiments, the set of audio characteristics comprises an audio energy value and the first motion state is randomly selected from a subset of the plurality of motion states based on the audio energy value, wherein the subset of the plurality of motion states comprise an audio energy match characteristic associated with the audio energy value. In some such embodiments, the set of audio characteristics further comprises a beat value, a tempo value, a melody value, and a danceability value. Similarly, in some embodiments, the audio energy value and the danceability value are based on a beat consistency over time value and a fundamental frequency value associated with a melody of the audio data. Other embodiments can use other such combinations or other data characteristics.