A machine learning approach to ornamentation modeling and synthesis in jazz guitar

<p>We present a machine learning approach to automatically generate expressive (ornamented) jazz performances from un-expressive music scores. Features extracted from the scores and the corresponding audio recordings performed by a professional guitarist were used to train computational models for predicting melody ornamentation. As a first step, several machine learning techniques were explored to induce regression models for timing, onset, and dynamics (i.e. note duration and energy) transformations, and an ornamentation model for classifying notes as ornamented or non-ornamented. In a second step, the most suitable ornament for predicted ornamented notes was selected based on note context similarity. Finally, <i>concatenative synthesis</i> was used to automatically synthesize expressive performances of new pieces using the induced models. Supplemental online material for this article containing musical examples of the automatically generated ornamented pieces can be accessed at doi: <a href="http://dx.doi.org/10.1080/17459737.2016.1207814" target="_blank">10.1080/17459737.2016.1207814</a> and <a href="http://https://soundcloud.com/machine-learning-and-jazz" target="_blank">https://soundcloud.com/machine-learning-and-jazz</a>. In the Online Supplement we present an example of the musical piece <i>Yesterdays</i> by Jerome Kern, which was modeled using our methodology for expressive music performance in jazz guitar.