Exploring Jazz Standards with Web Visualisation for Improvisation Training

The availability of music data is increasing, along with the increase in computer capabilities, internet bandwidths and enhancement of web-related technologies. This increase creates an overload of information. While methods that allow exploration and retrieval of music evolve to compensate for this overload, such methods focus on commercial services and mainly involve community-driven data processing. This paper presents a methodology that focuses on the direction of content-based exploration of music material in symbolic form, for facilitating the retrieval of Jazz standards for music improvisation training. The objective is to allow users to visualise a large dataset of Jazz standards with colourful 3D representations, while setting checkpoints on specific pieces for making the exploration more e cient. The underlying methods employ harmonic similarity metrics, which is an important factor of Jazz improvisation training, and the presented implementation incorporates custom web technologies that improve the educational perspective of the music playback experience. Evaluation of the UI and UX provides pointers for future improvements.