Computer synthesis of music performance
Johan Sundberg · Oxford University Press eBooks · 2001
This chapter describes some ingenious studies where a computer ‘learns’ to play expressively by being supplied with a set of rules for modulating timing and loudness in certain musical contexts. These rules are modified by playing the results to a panel of human judges and asking for judgements of appropriateness. In summary, it seems that analysis of music performance is beginning to reveal what exactly it is that musicians add to written music, what the code they use is, and from where this code is taken. Therefore, this chapter proposes that analysis of music performance can shed light on basic aspects of music communication.