ChoraleBricks: A Modular Multitrack Dataset for Wind Music Research

Abstract

Concert band and wind music are deeply embedded in society and play a significant role in the cultural landscape of many countries, including Germany and Austria, particularly within the amateur music scene. However, this type of music, as well as research on wind and brass instruments in general, remains largely overlooked in the field of music information retrieval (MIR). In this paper, we address this underexplored area by introducing ChoraleBricks, a framework featuring multitrack recordings of ten different chorales, each comprising four musical parts: soprano, alto, tenor, and bass. At its core, ChoraleBricks provides isolated recordings of individual parts performed by a diverse selection of wind instruments, including flute, oboe, clarinet, trumpet, saxophone, baritone horn, trombone, and tuba. These isolated recordings act as building blocks or "bricks" that can be modularly superimposed to create full mixes with varying instrumentation. In addition, ChoraleBricks provides sheet music, time‑aligned symbolic music representations, conducting videos, and reference annotations such as fundamental frequencies and note events. The framework is further enhanced by Python software tools that support parsing, mixing, annotation, and modular combination of the recorded audio material. With all multimedia and software components available as open‑source, ChoraleBricks provides a versatile framework for generating and augmenting datasets for polyphonic wind music. It supports systematic experimentation and facilitates evaluation across various research topics, including multi‑pitch estimation, note transcription, audio alignment, and music education applications.

Authors

• Stefan Balke
• Meinard Müeller