Visual comparison of music composing AI

Abstract

In this thesis, we describe a tool that allows a user to visualize and abstract data from melodies generated by multiple AIs. Related work in the field of music composing with AI is investigated and background information for the tool is given. We introduce different use cases and problems that the tool is supposed to solve and consecutively we present different visualization approaches and glyphs used to display the data computed from the AI melodies. We evaluate our approaches and demonstrate usability and readability with exemplary case studies. An outlook and a perspective on future work are given, exploring different additional features and refinements that can be done.