Computational Music Creativity
Instructor: Sergi Jordà, Lonce Wyse
Credits: 5 ECTS
Computational or artificial creativity is a multidisciplinary field of research located at the intersection of artificial intelligence, cognitive psychology, philosophy, and the arts. Although the term was officially coined in the late 90s, the study of the use of computers for simulating or enhancing human creativity dates back from the 50s, initially inspired by the work of cyberneticists such as Weiner or Shannon. That said, for the last 2 years language models and generative AI have taken the world by storm and the realm of music is no exception, with groundbreaking news and advancements showing emerging regularly. Generative AI is indeed bringing many exciting and unimaginable possibilities as well as many potential risks to the dynamic world of music creation.
In this course, we will study the creation and performance of music using computers and algorithms, with a focus on real-time interaction and control, i.e. never forgetting about "the human in the loop". We will cover these topics from a multilayered perspective, combining historical, conceptual, technical and esthetical viewpoints. We will discuss the profound impact AI can have on the music industry, exploring its transformative potential in areas such as composition, performance, audience engagement or copyright laws.
The course last for 10 weeks, with weekly classes on Mondays (from 17:30-19:00) and Wednesdays (17:00-18:00). Monday's classes will be more theoretical, with lectures and discussions, while Wednesdays will be more practical with hands-on labs. For the first part of the course, in the labs we will use the visual programming language Pd. In the second part of the course, we will study the application of Neural Networks for music generation. At the end of the course, students will have to perform/compose/improvise a musical piece of their own, using the tools they will have developed.
The evaluation of the students is based on weekly lab reports, reading assignments, and the completion of a final project focussing on the design, implementation and performance (or demo) of a system for artificial composition and/or performance. This final project will be developed in groups of 2 or 3 students. Additional students’ contributions, under different formats (such as music, links, class participation, etc.) are also items contributing to the final grade.
Pure data tutorials
- Johannes Kreidler's "Programming Electronic Music in Pd". This website contains a quite complete and good introductory book, translated and available in several languages (ENG, GER, ES...)
- Miller Puckette's "The Theory and Technique of Electronic Music", is a very advanced book on signal processing using Pd, by the creator of the language. All the examples in the book correspond to the help files in the Pd documentation (not the easiest way to learn Pd).
- FLOSS Pure Data manual, focuses on the the audio synthesis, guiding the reader through the creation of a subtractive modular synthesiser.
Music and AI bibliography
- Nakajima, R. and Shibata, A. (2022). An Overview of AI Music Generation and its Potential. Qosmo Inc. This freely downloadable whitepaper, gives a very complete overview of the current state of the art.
- Jean-Pierre Briot, Gaetan Hadjeres and François-David Pachet (2019). Deep Learning Techniques for Music Generation – A Survey. Extensive survey and analysis of different ways of using deep artificial neural networks to generate musical content (full book version)
- Alexandre DuBreuil (2020). Hands on music generation with Magenta. (Github of the book) (Magenta site)
- Artificial Intelligence and Music Ecosystem, edited by Martin Clancy (2023), offers interdisciplinary approaches to pressing ethical and technical questions associated with AI and Music.