Abstract
Beat scaling is an important characteristic of classical music and gives it the sense of meter. The sense of meter gives the music its feeling and emotional qualities.

The beat scaling factor is the difference in amplitude between the accented and unaccented notes in a group. In this work, we will find beat scaling factors typically used in classical music by analyzing the performances of professional musicians. We will also interpret these results and identity context-dependent results from more salient features. Computer music performance can benefit from the use of these beat scaling factors to simulate human performance.