- This event has passed.
Paper Session 3a: Music Notation & Representation I
Session Chair: Axel Berndt
Paper abstracts
Tianze Zhang, Shingyui He and Lei Xuan, “CNN-BiLSTM Hybrid Model with Physical Constraints for Automatic Piano Fingering Generation”
Piano fingering is a pivotal technique that piano learners must master. To address the difficulties in the application and arrangement of fingering during practice and early stages of learning, this paper proposes an artificial intelligence hybrid model based on Bidirectional Long Short-Term Memory (BiLSTM) network, Convolutional Neural Network(CNN) and Attention, with the aim of automatically generating piano fingerings. The model extracts physical features, including spatial, temporal, hand motion, and fingering information, and integrates biomechanics constraints during neural network training for the first time. Based on the aforementioned algorithms, the model achieved a good result. This innovative methodology enhances predictive performance by accurately capturing the complex
physical interactions inherent in piano fingering. This paper also compares the model with fingerings generated by other algorithms to verify the reasonableness and effectiveness of the hybrid model in piano fingering prediction. In conclusion, the model can efficiently and conveniently provide fingering support for piano learners, and has strong application prospects and practical value.
Juan Carlos Vasquez and Zhonghao Chen: “Recursive Radiance: Multimedia Interpretations of Traditional Chinese Aesthetics”
inspired by the Jianzi notation system. The sonic component transforms improvisations based on the traditional guqin piece ”Cai Zhen You” (Wandering in True Essence) through fab synthesis, spatial diffusion, and electronic processing. The composition employs chaotic attractors for spatial movement, creating an immersive soundscape that embodies Daoist principles of fluidity and transformation. The visual component features a series of hanging scrolls and fragments functioning as both notation and artistic extension. These graphic scores emerge from a hybrid methodology combining traditional ink boxes with tensioned strings, cyanotype printing, 3D environmental scans, and AI-generated imagery created through LoRA models trained on interpretative readings of the Guqin’s notation system. Recursive Radiance functions as both a hybrid physical-digital installation and a framework for cultural preservation.
This paper documents the completed musical composition, graphic scores, and conceptual approach to public engagement through immersive multimedia. Our research demonstrates how computational tools and engineering techniques can support artistic expression while preserving cultural heritage, offering new pathways for audience interaction with traditional art forms through contemporary multimedia experiences.
Rob Canning: “Scores That Run: Graphic Notation with Embedded Performance Semantics”
This paper presents an approach to digital graphic notation in which performance semantics are embedded directly into the visual surface of the score. Working in standard SVG and authored entirely in Inkscape, the score is composed as a graphic–semantic document: visual elements carry lightweight cue structures encoded in their identifiers, and these cues are executed in real time by a browser-based runtime. The score therefore functions simultaneously as image, temporal structure,
and performative interface, without reliance on symbolic engraving or external playback systems.
The framework supports hybrid formal topologies, including continuous scrolling trajectories, page-based local environments, and patterned navigation between sectional states. Animated motion fields provide shared gestural resources for ensemble coordination and may
optionally drive live electronic processes, enabling a unified grammar of acoustic and electronic gesture. All cue semantics—structural, temporal, gestural, textual, and media-based—are authored within the same executable layer as the notation, so behaviour and interpretation
arise from a single surface. Because the system is based entirely on open web standards, it enables a direct draw-and-perform workflow accessible to composers and performers without specialised technical infrastructure.