We extract high-level features from a set of 16 commercial audio recordings (and corresponding music scores) of jazz guitarist Grant Green in order to characterize the expression in the pieces. Here, we present a data driven computational approach to induce expressive performance rule models for note duration, onset, energy, and ornamentation transformations in jazz guitar music. By using mobile device programming, the project aims at developing an application using the predictions of the models,providing beginner-level students with a reference for playing expressively.Įxpert musicians introduce expression in their performances by manipulating sound properties such as timing, energy, pitch, and timbre. Shifting from traditional synthesis goals, the project is developed through a didactic perspective. By means of machine learning techniques, the project pro-poses a study on vibrato detection and vibrato parameters prediction. The project aims at studying whichinformation carried in the scoreis relevant for a musician whens/he decides to perform a vibrato. Here vibrato performed on violin is considered as an expressive resource. Studies on expressive performances try to identify what information encoded in the scores is relevant for a musician when playing expressively. This lack is left by a gap between the com-poser’s initial musical intent and the available tools of music representation. The control over expressiveness during a music performance is an element used by the performer to fill an existing lack of information in the score.
Please refer to the conference proceeding: 10.5281/zenodo.1422607
0 kommentar(er)
