Sasha J. Blondeau : Jérôme Nika 3/4
After taking a sabbatical year, computer music designer and musician Jérôme Nika decided to rejoin IRCAM’s research teams at the beginning of 2024. However, despite his ongoing research on human-machine interaction, which initially brought him to develop Dicy2 with the Musical Representations team, he is now joining the Sound Music Movement Interaction team to undertake new adventures. His projects focus on interactive and generative models for musical creation, with regard to ‘digital instrument-making’ and interplay composition. This new scientific approach led him to take part in Sasha Blondeau’s artistic research residency, focused on the elaboration of a model for dramaturgical prototyping of music.
This new approach might surprise some. In fact, Nika switches from improvisation-oriented research to research focusing on the high-level formalization of the musical discourse. Jérôme Nika however defends a perfect continuity in his work: “Although my research to date has led to the development of tools such as Dicy2 – which tend to be applied in the field of improvisation, with the somewhat catch-all term “AI” thrown in for good measure, my primary focus was on formalizing the controls and writing of electronic discourse.”
“What I find most fascinating in the long-term collaborations with artists using these tools – at least as much as human-machine interaction – is the “composition” of its behavior. In other words, the construction of the material it plays with, the way in which you want it to generate its discourse, and the way in which you want it to listen to your partner, etc. Because the subject of my thesis was the modeling and use of scenarios or underlying structures in improvisation, I was already working on the dimension of narration and storytelling, albeit a local one.
Moreover, contrary to the popular view that improvisation is a creation happening in the instant, the philosophy of my earlier work was to model it as a real-time mobilization of off-line computation processes; i.e. of a compositional nature. Hence the need to “learn” to improvise.
“Improvisation requires short or medium-term projection, often, in certain repertoires at least, according to pre-established rhythmic or harmonic structures. In short, improvisation is just one of the ways in which to use the tools I've been developing up to now, while composition is another. Incidentally, these tools have been used by artists who were not improvisers.”
Jérôme Nika Photo by William Beaucardet
From what most composers told me after working with Dyci2, it allows for a great reflexivity: “At the end of every creative project, they told me: ‘I loved it’. Interestingly, even more than the musical result, it is the process that they had liked. In fact, learning new things with the machine forced them to formalize and think about the creative process in itself. It also helped them realize that what they aspired to was to design a specific type of trajectory in a specific musical dimension.”
Because Sasha Blondeau’s artistic residency focuses on musical dramaturgy, it completely fits with Nika’s scientific concerns.
“I would go even further and say that I have the feeling that Sasha’s simplicial complexes and the models I have developed follow the same approach, at least metaphorically”, says Nika. “Sasha elaborates models of space and imagines paths within this matrix, while Dyci2, for example, builds a memory module, the trick being to know how to walk through it. In both cases, the aim is to model a way of thinking about form, a way of thinking about narrative, at the highest level.”
However, Jérôme Nika has not yet reached the point where he will be able to code a ready-to-use ad-hoc tool for other interested composers at the end of this residency. Admittedly, this experience serves two purposes for Nika: on the one hand, to formalize practices enough in order to establish a taxonomy, and on the other hand, to imagine prototypes for digital tools. But he remains realistic and modest regarding this second goal. He explains:
“One research topic I would like to focus on in the future involves going back to computer assisted composition (CAC), but of a kind that would internalize the artistic project in its process. Hence, the necessity to meet composers in order to gather their thinking of the form and try to extract invariants from that data. Only once that is done can we potentially start developing tools, for instance to facilitate going back and forth between a sketch of a formula and its modification by a given process.
Despite what people might think when considering my previous works, I am not a fierce advocate of machine learning or any other artificial intelligence able to generate massive quantities musical data. But conceiving a tool that enables composers to assimilate a given material, and model its structure parameters by using high-level symbols to then be able to work on it and develop it; that is something much more exciting!”
Jérôme Nika does not hide his enthusiasm to work on the description of the music itself, whether it be on the nature of the material itself, on its function (to make a comparison with the work of a screenwriter: a particular scene plays a specific part in the narration in order to introduce a character or steer the plot in a certain direction) or on coupling relationships possibilities. Nika explains: “Let us imagine two musical lines, that I would call line A and line B. In this scenario, in order to create B, I would choose to describe A and its relationship with B, rather than describing B directly.”
But even more than this almost musicological approach, it is the future collaborations with composers that he is anticipating the most. “Are these avenues that I would like to explore really pertinent? Would they help the composers? And, more importantly, what other paths would they open?”
Indeed, when it comes to research, whether it be scientific, artistic or metaphysical, every given answer raises dozens of new questions, each one broader than the last…
Interview conducted by Jérémie Szpirglas
