This page is currently under construction! Check back soon for more.
Nikhil Singh*, Manuel Cherep*, and Jessica Shand, “Creative Text-to-Audio Generation via Synthesizer Programming,” in NeurIPS Workshop on Machine Learning for Audio (2023).
Abstract: Sound designers have long harnessed the power of abstraction to distill and highlight the semantic essence of real-world auditory phenomena, akin to how simple sketches can vividly convey visual concepts. However, current neural audio synthesis methods lean heavily towards capturing acoustic realism. We introduce an open-source novel method centered on meaningful abstraction. Our approach takes a text prompt and iteratively refines the parameters of a virtual modular synthesizer to produce sounds with high semantic alignment, as predicted by a pretrained audio-language model. Our results underscore the distinctiveness of our method compared with both real recordings and state-of-the-art generative models.
Previous Work
Contributor: Systematic Suppression: Hungary’s Arts and Culture in Crisis (March 2022)
[Overview] Artistic Freedom Initiative (AFI) has published this report in the context of the existential crises facing artists and arts institutions in Hungary. Our aim is to draw attention to the nature of Hungary’s restrictions on artistic freedom, and influence stakeholders—both in Hungary and in the region—to take measures to thwart and reverse these dangerous and anti-democratic trends.
[Abstract] In this paper, we ask: what can art and artists experimenting with adversarial designs contribute to anti-surveillance movements, and in particular, to notions of privacy in public? First, we establish a basic technical understanding of facial recognition systems to provide context for the designs presented in this paper, which are tailored to exploit specific features of these systems. We then examine both how aestheticized forms of resistance fail to challenge the disproportionate impacts of automated facial recognition technologies by law enforcement agencies, and the strategic fallacies of implementing adversarial designs in the first place. In light of this, we argue that aestheticized forms of resistance ought not to be examined uncritically. We consider a performative theory of privacy to suggest that these kinds of artistic interventions may further anti-subordination goals insofar as they highlight the expressive dimension of functional, quotidian demands for privacy in public. To conclude, we offer recommendations for future research in expressive resistance to state surveillance—reaching beyond physical space, and toward cyberspace.
Author: Artificial Creativity: Musically Improvising Computers and the Listening Subject (December 2019)
[Abstract] In this paper, I do not attempt to define an absolute idea of creativity; rather, I turn toward musically improvising computers to unpack notions of creativity within a very specific context (namely, 21st-century contemporary art music) using American composer Pauline Oliveros’s Expanded Instrument System (EIS) as a case study. I allow for a creativity that is co-determined by a non-human, machinic agency, temporarily setting aside Lovelace’s famous assumption that machines cannot be creative. In order to do this, I rely on the method of actor-network theory (ANT), which examines both human and non-human actors in terms of the natural and social networks of relations in which they exist and thereby assigns agency to all such actors. Through a close reading of the role of computer-improvisers in the music of Pauline Oliveros as performed by her close friend and collaborator, the flutist Claire Chase, I ask: how is creativity constructed in Oliveros’s context? How has software and hardware shaped and been shaped by that construction? As in this case study the notion of creativity is strictly tied to that of improvisation, I also question the relationship between the two. To conclude, I directly respond to recent developments in the theorization of the listening subject as put forth by musicologist Nina Eidsheim – namely, her proposal for a vibrational practice of listening in which “music is predominantly understood…as material and intermaterial vibration.” I argue that Eidsheim’s model breaks down in the face of computer temporalities that operate outside of, even beyond, the material here-and-now: indeed, any contemporary practice of listening must be pliable enough to capture the agency, however immaterial, of the listening machine.
Author: Game Music: Game-Theoretic Principles in Music Composition (May 2020)
[Abstract] The history of the application of mathematical concepts in musical composition and theory dates back to at least Greek antiquity. However, it was only in the twentieth century that a handful of composers of contemporary classical music, largely led by Iannis Xenakis, began to formalize sophisticated stochastic methods within their work. In this paper, we survey two pieces of game music within the contemporary classical repertoire: Xenakis’s own Achorripsis – a group improvisation based on a matrix generated by a Poisson distribution – and a more recent Xenakis-inspired installation by Davide Morelli and Marco Liuni, a zero-sum game between opposing players. We then proceed to design and implement our own set of musical games related directly to the tension between determinism and stochasticity in Xenakis, Morelli, and Liuni’s work, concluding that the delicate balance between the two is the foundation of a score’s success.
Author: The Sound of Occupation: Music and The Spectacle of Collective Action (March 2020)
[Abstract] In an age of global socio-political campaigns led largely by young people and made possible only by digital media and smartphone technology, the 2019 Harvard-Yale sit-in stands out for the energy it galvanized both on and off the field. In this paper, however, I grapple with the site of protest itself. What unusual circumstances, if any, enabled students to overcome the collective action problem that stalls the vast majority of spontaneous political demonstrations? Taking into consideration the blatant incongruency between the sound technologies employed by sound engineers in response to the protest and those available to student protestors, I focus on what I believe to have been the role of sound and music in solidifying student morale. I contend that in combination, extreme volume and the cultural salience of certain popular music – the former corresponding to a materiality of sound, the latter bringing us to the level of the intellect – played a vital part in the turn of events on November 23rd, 2019, subverting the intentions of stadium personnel and exemplifying the ability of music to compel action.