Authorship and Inventorship

If the sampling clearance system had a consequential effect on sampling culture, how should we understand the moral, ethical, and legal mediations of contemporary and historical uses of synthesis? WP5 examines the evolving legal and corporate determinants of music synthesis methods and their impact on creative practice, attending to the mechanisms, including patents, trademarks, copyrights, and corporate secrets, through which novelty in synthesis design is policed.

Music synthesis operates at the thresholds between categories of thing, idea, and expression. Whether algorithms defining instruments or corporate presets specifying sounds, synthesis's ontology is frequently determined by the ownership regimes in play in a given domain. The question of whether innovation is protected by authorship (as with sampling), inventorship (as with signal processing), or corporate contracts (as with creation on contemporary music platforms) is pressing. Yet musicology's recent material turn, for all its productive engagement with thing theories and new materialisms, has displayed little interest in how a musical entity becomes a thing as opposed to an idea or expression.

Approach

WP5a reviews key contemporary and historical synthesis patents, including the Triadex Muse (Fredkin, 1971), FM synthesis (Chowning, 1977), concatenative synthesis (Jehan, 2010), and text-to-speech systems. It additionally examines the commercial and legal status of corporate presets and digital clones of corporate signal processing routines (such as guitar effects pedals). Across these cases, WP5a seeks to understand how lines are drawn between inventions, business models, and physical processes, with particular attention to controversies where this distinction has become a point of contention.

WP5b examines musical cases demonstrating a folk ethics of fairness with respect to authorship, whether through collaboration with engineers responsible for creating key sounds and instruments, or the explicit citation of the technical context surrounding a synthesis procedure. The PI and PD2 additionally study how these divisions of labour are complicated by machine-learning-based timbre generation, which raises the question of whom the title of a work should be granted to: musician, engineer, algorithm, or dataset.

Methods

Archival research, interviews, case law analysis, music analysis. Historical work situates generative AI not as a rupture but as a case where an already-precarious balance between engineering practice and artistic creation was destabilised.