2025/01/01 by Satoshi Nishimura, Atsushi Marui · 1 voice
Computer Science · Neuroscience · #Music Technology and Sound Studies #Music and Audio Processing #Neuroscience and Music Perception
paper · pdf · doi:10.1080/09298215.2025.2540434
openalex publication_date 2025/01/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/06/15
Pytakt is a new Python library for the text-based description, algorithmic generation, and real-time processing of symbolic (event-level) music information. The library provides embedded textual description that makes it possible to represent scores containing chords, polyphony, and performance information, such as velocity or control changes, in a compact form. Scores can be concatenated, merged, or repeated with operators, and various score transformations, such as diatonic transposition and pattern replacement, are available. Pytakt also has real-time MIDI input/output functions with a priority queue, which are useful in interactive music applications, together with music-theoretic classes such as scales and chords. In addition, it incorporates basic analytic features, including the retrieval of active notes and controllers at any given time with a novel algorithm and a multi-track visualiser with a playback function. This paper introduces the design and features of Pytakt and presents its usage examples, including procedural music description and data preparation for machine learning. Furthermore, the results of performance comparison with other libraries are shown to confirm the lightweight property of our library.