Abstract
This paper provides a guide through the FMP notebooks, a comprehensive collection of educational material for teaching and learning fundamentals of music processing (FMP) with a particular focus on the audio domain. Organized in nine parts that consist of more than 100 individual notebooks, this collection discusses well-established topics in music information retrieval (MIR) such as beat tracking, chord recognition, music synchronization, audio fingerprinting, music segmentation, and source separation, to name a few. These MIR tasks provide motivating and tangible examples that students can hold onto when studying technical aspects in signal processing, information retrieval, or pattern analysis. The FMP notebooks comprise detailed textbook-like explanations of central techniques and algorithms combined with Python code examples that illustrate how to implement the methods. All components, including the introductions of MIR scenarios, illustrations, sound examples, technical concepts, mathematical details, and code examples, are integrated into a unified framework based on Jupyter notebooks. Providing a platform with many baseline implementations, the FMP notebooks are suited for conducting experiments and generating educational material for lectures, thus addressing students, teachers, and researchers. While giving a guide through the notebooks, this paper’s objective is to yield concrete examples on how to use the FMP notebooks to create an enriching, interactive, and interdisciplinary supplement for studies in science, technology, engineering, and mathematics. The FMP notebooks (including HTML exports) are publicly accessible under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Highlights
While logarithmic compression and normalization increase the robustness to variations in timbre or sound intensity, we study in the fundamentals of music processing (FMP) Notebook Temporal Smoothing and Downsampling postprocessing techniques that can be used for making a feature sequence more robust to variations in aspects such as local tempo, articulation, and note execution
As an alternative to the spectral flux, we introduce in the FMP Notebook Phase-Based Novelty an approach that is well suited to study the role of the short-time Fourier transform (STFT)’s phase
([1], Section 6.3.2), we introduce in the FMP Notebook Beat Tracking by Dynamic Programming a genuine beat tracking algorithm that aims at extracting a stable pulse track from a novelty function, given an estimate of the expected tempo
Summary
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Processing [1] (FMP) yields an example of how music may provide a rich and challenging application domain for introducing, teaching, and studying fundamental techniques and algorithms relevant for general courses in computer science, multimedia engineering, information science, and digital humanities. Suitably designed software packages and freely accessible web-based frameworks have made education in computer science and signal processing more interactive. Such novel technology allows for designing courses that help students move. The FMP notebooks are built upon the Jupyter notebook framework, which has become a standard in industry as well as in educational settings [2] This open-source web application allows users to create documents that contain live code, text-based information, mathematical formulas, plots, images, sound examples, and videos. The FMP notebooks closely follow the eight chapters of the textbook [1], and as such, provide an explicit link between structured educational environments and current professional practices, in line with current curricular recommendations for computer science [3]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
