Iracema: a Python library for audio content analysis

Abstract

Iracema is a Python library that aims to provide models for the extraction of meaningful informationfrom recordings of monophonic pieces of music, for purposes of research in music performance. With this objective in mind, we propose an architecture that will provide to users an abstraction level that simplifies the manipulation of different kinds of time series, as well as the extraction of segments from them. In this paper we: (1) introduce some key concepts at the core of the proposed  architecture; (2) describe the current functionalities of the package; (3) give some examples of the application programming interface; and (4) give some brief examples of audio analysis using the system.