In our project, the machines can both take the role of a performer for us to play with and they can be extensions of our own instruments. They can become both our duet partners and they can be additions, expansions or augmentations of our sound. Playing is core in our investigation, and it is based on this experience we try to articulate thoughts and answers to the following questions:
- How do we improvise with "creative" machines, how do we listen, how do we play?
- How will improvising within an interactive human-machine domain challenge our roles as improvisers?
- What music emerges from the human-machine improvisatory dialogue?
The project's artists are Andrea Neumann (GER), Morten Qvenild (NOR) and Ivar Grydeland (NOR). The artist Sidsel Endresen (NOR) is our observer, commentator, critic and discussion partner. Norwegian Center for Technology in Music and the Arts (NOTAM) is our technological collaborator. Additionally, musician, composer and researcher Henrik Frisk (SWE), writer, musician, composer David Toop (UK) and director and writer Annie Dorsen (US) contributes to the project.
We improvise with algorithms that listen to our music, records it, analyse our artistic choices and respond to us in real-time. When designing algorithms, we decide their attention span, what they should listen to in our music, how they should analyse it and how this information could be turned into responses. George Lewis coins these processes as the system "personality" when describing his algorithmic improvising software Voyager (1993). Even though we as humans are designing this system personality, the vast number of options decided by the algorithms still create unexpected musical output. The algorithms can improvise music we as humans never would have thought of.
In A Piece of Work (2013), a digital rework of Hamlet, theatre director and writer Annie Dorsen and her team developed a computer system that can "autonomously create and perform an adaptation of Shakespeare's piece according to algorithmic principles." She describes the carelessness of the algorithms: "They don't know what they say, or what they have said before. They don't know what grief is, or revenge, or an entrance, or an exit. They make decision after decision, over and over, generating a nonstop flow of effects without causes, and causes without effects."
In a human-human improvisatory dialogue, we expect that most human improvisers would strive at creating a dialogue and music that "works". Obviously, there are as many definitions of "when music works" as there are practitioners and listeners. But still, one major distinction between humans and machines is that machines do not want anything. They just produce numbers. Or put differently, they do not care if the music "works". How do we as human improvisers operate in this domain?
Individual research, regular workshops and LAB concerts are interrelated activities and key elements in the project's work methods. In the LAB concerts, we present different stages of works-in-progress by artists. Every LAB has test performances, followed by comments by invited guests, reflections by the artists and discussions between the artists and the audience. The LAB concerts allow the audience to take part in considerations and reflections, both leading up to, during and after the test performance. The LABs are essential arenas between the informal workspace and the official concert hall, an amalgam of the ease of the workspace and the tension of the podium.
Creating a piece of music, and the reflections behind this process, is in general, not open to the public. Goodbye Intuition seeks to reveal such processes and make them less private. How can unfolding our processes nourish the artistic output and the reflections?
Goodbye Intuition is an artistic research project hosted at The Norwegian Academy of Music in Oslo (NMH). It is financed by NMH, Norwegian Artistic Research Programme, Norwegian Center for Technology in Music and the Arts (NOTAM) and Royal College of Music in Stockholm. The project began in October 2017 and concludes in July 2020.