Paper Session 2b: AI & Music
Three papers will be presented and discussed:
Hiroshi Yamato, OrbitScore: “A Domain-Specific Language for Polymetric Live Coding Based on Multilayered Temporal Structures”
This paper presents OrbitScore, a domain-specific language (DSL) for live coding polymetric rhythm patterns based on the theory of Multilayered Temporal Structures (MLTS). While existing live coding languages such as TidalCycles and Sonic Pi provide rich pattern manipulation capabilities including polyrhythmic support, OrbitScore offers an intuitive syntax where the beat(n by m) notation directly
maps to the theoretical 4:(n/4) framework, enabling each sequence to maintain its own meter and allowing performers to create intricate polyrhythmic textures in real-time. The system integrates with SuperCollider for low-latency audio synthesis and provides a declarative, method-chaining syntax designed for live performance. We describe the theoretical foundation, DSL design, implementation architecture, and demonstrate the system’s capabilities through a live coding performance. Our contribution lies in bridging the gap between the theoretical framework of Multilayered Temporal Structures and practical live coding tools, making polymetric expressions accessible to performers.
Yuan Zhang and Xinran Zhang, “Hexagram-Based Semantic Composition: Discretizing Embedding Spaces into Symbolic Compositional States for Improvised Performance”
Diffusion-based text-to-audio (TTA) systems such as Udio have introduced a mode of musical making in which linguistic prompts activate high-dimensional latent manifolds to yield contingent, non-repeatable sonic artefacts. This generative architecture—operating through interse-
miotic translation between linguistic signs and high-dimensional latent space—produces distinctive aesthetic conditions that have yet to be adequately theorized. This paper introduces latent music as an emergent aesthetic form produced through generative text-to-audio systems such as Udio. Latent music arises from processes of interpolation, recombination, and associative drift within
high-dimensional latent spaces—existing in states of perpetual becoming characterized by gradient identities, interreferential drift, asignifying ruptures, and ontological indeterminacy. These emergent sonic forms occupy interstitial spaces between recognizable musical signs, resisting categorical stability while revealing distinctive possibilities for sonic expression. The result is a field of sonic objects marked by spectrality, liminality, and cross-material entanglement—sounds that hover between genres, gestures, and perceptual thresholds. Drawing on Deleuzian aesthetics, philosophy, and an extensive corpus of prompt-generated sonic artifacts, the paper situates these emergent forms as products of asignifying rupture and aesthetic drift, where sonic identities dissolve and recombine in unstable assemblage determined by intersemiotic translation between linguistic prompts and audio materiality. This research offers a theoretical framework and critical vocabulary for engaging with these uncanny sonic entities, proposing that latent music invites listening practices attuned to indeterminacy, associative resonance, and the productive tensions of the not-yet-formed.
Colton Arnold, Zhaohan Cheng and Ajay Kapur: “AI Framework for Dynamic Robotic Instrument Calibration”
This paper presents a data-driven calibration framework for robotic musical instruments based on a hybrid ensemble model that combines K-nearest neighbors (KNN) and a multi-layer perceptron (MLP). KNN anchors predictions to recorded acoustic measurements, while the MLP enables nonlinear generalization and smooth interpolation across the instrument’s playable range. A distance-dependent blending strategy integrates the two models, improving consistency across sparse and dense data. The proposed approach produces stable and repeatable calibration estimates for both pitched and non-pitched instruments, outperforming standalone models across a range of sampling conditions. This work establishes a scalable foundation for automated calibration in robotic musical systems.