Ubiquitous Musical Signal Processing, Machine Learning, and Large Language Models

Despite recent advances in musical signal processing, little attention has been given to the demands of nontechnical stakeholders. The reduction of technical requirements in audio processing and synthesis frequently involves a musical-genre bias. Timbre has not yet been thoroughly explored as a meaningful handle for musical interaction processes. The recent developments in machine learning as applied to digital signal processing only compound these issues. Ubiquitous music (ubimus) signal processing frameworks may encompass the incorporation of emergent adaptive techniques in human-computer interaction. Design qualities such as aesthetic pliability fostered by second-wave ubimus frameworks may provide targets for research in ubimus signal processing.Ubiquitous music (ubimus) is an expanding field of research that has a track-record of almost two decades of development. A first wave of proposals (between 2007 and 2014) featured exploratory initiatives targeting expanded notions of music-making, including the incorporation of nonmusicians in creative activities and the deployment of musical resources in everyday settings. Second-wave ubimus features the emergence of conceptual and methodological frameworks based on empirical evidence with the explicit goal of reducing demands for creative participation, increasing the circulation of knowledge and expanding our understanding of musical experiences in ubiquitous contexts.Our goal is to explore the synergies between musical signal processing, ubiquitous music, machine learning and new techniques of sound transformation. In particular, we are interested in approaches that reduce the technical requirements for the stakeholders and increase the range of aesthetically relevant results.The following topics are within the range of this topic:- New applications of classic musical signal processing techniques within the context of ubimus- New methodologies for adopting ML methods in digital signal processing- Approaches to explainable AI within the context of ubiquitous music and musical signal processing- New applications of established ML/AI approaches (such as DDSP, ANNs etc)- Music programming environments for signal processing- Issues of ubimus creativity within the scope of sound design and musical signal processing. Topic editors: Victor Lazzarini, Joseph Timoney, Damián Keller, Marcelo Caetano Source: https://www.frontiersin.org/research-topics/80293/ubiquitous-musical-signal-processing-machine-learning-and-large-language-models