Posters 

Text-Setting in African Opera: A Study of Syllabic and Melismatic Practices in Igbo Works

Abstract: This study quantifies text-setting strategies in Nigerian Igbo opera, examining how composers set linguistic text to music across recitatives, arias, and choruses. Analyzing works by Meki Nzewi, Lazarus Ekwueme, Clement Ezegbe, and Alvan-Ikoku Nwamara, the research measures syllabic density (syllables per note), melismatic passages, and word repetition patterns. Findings reveal that Igbo opera recitatives exhibit higher syllabic density (0.9-1.2 syllables/note) than Western operatic recitative, reflecting Igbo oral narrative influences. Arias demonstrate strategic melismatic treatment (1:3-1:5 ratios) on semantically significant words, serving both musical and textual emphasis functions. Choral sections show moderate density with frequent word repetition. Comparative analysis with Western opera and Igbo traditional vocal music suggests Igbo opera represents a hybrid practice: adopting Western formal structures while prioritizing text intelligibility and speech-rhythm fidelity characteristic of Igbo vocal traditions. The study proposes a framework for analyzing text-music relationships in African art music and discusses implications for understanding compositional decision-making in tone-language opera.

Re:chord: A Transformer-Based Harmonic Inference System for Hummed Melodies

Abstract: Re:chord is an AI-assisted music interaction system that generates harmonic chord suggestions from hummed melodies. The system supports beginner and non-professional users in the early stages of music creation by transforming melodic vocal input into editable harmonic ideas. The pipeline consists of monophonic pitch estimation, symbolic note tokenization, and transformer-based harmonic inference. A transformer encoder–decoder model was trained on a self-constructed symbolic music dataset in MusicXML format. To improve accessibility and training stability, harmonic labels were constrained to 10 representative chord qualities, including major, minor, dominant, diminished, and seventh-based harmonies. This design allows harmonic richness beyond basic major and minor chords while maintaining accessibility for beginner users. Rather than fully automating music composition, Re:chord positions AI as a collaborative creative assistant that supports intuitive musical exploration and user-driven creativity.

Acoustic Substantiation of Pedagogical Language in Gujeon-simsu Oral Transmission: A Case Study of the Twelve Japga <Jipjangga>

Abstract: This study aims to convert the abstract pedagogical language passed from master to student in the teaching of the Gyeonggi japga Jipjangga into acoustic data, thereby scientifically clarifying what actual changes in vocal production these verbal instructions bring about. Gyeonggi japga is a specialized vocal genre that demands more difficult vocal techniques and more delicate ornamentation than ordinary Gyeonggi folk songs, and it has been handed down by singers who learned directly from a master over many years. Among these pieces, Jipjangga features densely packed subtle pitch changes and complex shifts in vocal production, making it a representative work in which the master's abstract verbal instructions are used especially frequently and precisely. As a Certified Trainee of National Intangible Cultural Heritage in Gyeonggi Folk Song, the researcher draws on direct training under a master to capture, from the perspective of an insider who has personally undergone the process, the instructional language exchanged in the actual teaching setting and the corresponding changes in sound.

Persistent Homology of Music Network with Three Different Distances

Abstract: Persistent homology has been widely used to discover hidden topological structures in data across various applications, including music data. To apply persistent homology, a distance or metric must be defined between points in a point cloud or between nodes in a graph network. These definitions are not unique and depend on the specific objectives of a given problem. In other words, selecting different metric definitions allows for multiple topological inferences. In this work, we focus on applying persistent homology to music graph with predefined weights. We examine three distinct distance definitions based on edge-wise pathways and demonstrate how these definitions affect persistent barcodes, persistence diagrams, and birth/death edges. We found that there exist inclusion relations in one-dimensional persistent homology reflected on persistence barcode and diagram among these three distance definitions. We verified these findings using real music data.

Dynamic Reconfiguration of Relative Phase Patterns in ADHD: Resting-State and Naturalistic EEG Analysis

Abstract: 

Introduction

Attention-Deficit/Hyperactivity Disorder (ADHD) lacks reliable neural biomarkers capable of capturing moment-to-moment fluctuations in attention, emotion, and arousal. This limitation constrains both precise diagnosis and the development of adaptive, brain-based interventions. As precision medicine and brain–machine interface applications increasingly demand objective and personalized dynamic neural measures, there is a growing need to characterize brain-state dynamics beyond static, group-averaged metrics.

Method

In this study, we applied Relative Phase Analysis (RPA: Park et al., 2025, https://www.biorxiv.org/content/10.1101/2025.03.12.642768) to high-density EEG data to identify dynamic information-processing modes in children with ADHD compared to typically developing (TD) controls. We analyzed 128-channel EEG recordings from the Child Mind Institute Healthy Brain Network dataset, including children aged 11 years and older with inattentive-type ADHD (n = 37) and typically developing controls (TD: n = 48). EEG data were collected during resting-state conditions (eyes closed and eyes open) and during a naturalistic movie-watching task (Despicable Me). Relative phase relationships between each EEG signal and the global mean phase were extracted, and k-means clustering was used to identify four dominant processing modes corresponding to top-down, bottom-up, and two other transient intermediate dynamics. Mode occurrence probability, dwell time, and transition patterns were quantified.

Result

Results revealed that TD children exhibited dominant top-down processing modes, particularly during eyes-closed resting state. In contrast, children with ADHD showed reduced top-down mode occupancy and increased dominance and persistence of bottom-up processing modes, consistent with theories of impaired top-down control. Notably, during movie watching, the exaggerated bottom-up dominance observed in ADHD diminished, and the balance between top-down and bottom-up modes became more similar to that of TD children. Additionally, increased occurrence of transient modes during movie viewing suggests stimulus-triggered transitions leading to reconfiguration of neural dynamics.

Conclusion & Discussion

These findings indicate that naturalistic stimuli may non-invasively stabilize attentional dynamics in ADHD by modulating underlying brain-state organization. Relative Phase Analysis offers a sensitive framework for capturing fine-grained neural dynamics and holds promise as a real-time biomarker for neurofeedback and digital therapeutics targeting attentional regulation, motivating future translational investigations.

Behavioral Signatures of Musical Complexity and Expertise in Music Perception

Abstract: Musical preferences vary widely across individuals, yet quantifying the perceptual dimensions underlying these differences remains challenging. Prior work has shown that music appreciation is systematically shaped by stimulus complexity, often formalized as Information Content (IC) within the Information Dynamics of Music (IDyOM) framework (Gold, Pearce, et al., 2019). In non-expert listeners, pleasure and liking responses to music have been reported to follow an inverted U-shaped function of IC, with stimuli of intermediate predictability eliciting the highest ratings. In parallel, the Goldsmiths Musical Sophistication Index (Gold-MSI) provides a validated measure of individual differences in musical sophistication (Müllensiefen, Gingras, et al., 2014). Building on these frameworks, the present study investigated how musical complexity and expertise are reflected in behavioral responses to music, and whether these responses provide a principled basis for subsequent analyses of neural activity.

Participants (N = 43) listened to 17 classical music excerpts during simultaneous EEG–fMRI recording and provided six behavioral ratings for each excerpt. Ratings were analyzed in relation to IC using linear mixed-effects models, with musical expertise (classical music majors versus non-majors) and individual Gold-MSI scores entered as separate predictors. Experts showed a stronger correspondence between perceived complexity and IC than non-experts, indicating tighter alignment between model-derived musical information and subjective complexity judgments. Higher Gold-MSI scores were associated with greater overall familiarity and increased differentiation of perceived complexity across stimuli, particularly for high-IC excerpts, suggesting that individual musical sophistication modulates sensitivity to musical structure. Receiver operating characteristic (ROC) analysis further showed that Gold-MSI scores reliably distinguished experts from non-experts (AUC = 0.82). Finally, principal component analysis (PCA) of behavioral ratings identified a primary familiarity-related component and an orthogonal component dominated by perceived complexity, supporting the relative independence of perceived complexity from other ratings.

Together, these findings indicate that perceived complexity is a robust perceptual dimension that tracks IC across listeners while remaining distinguishable from other subjective evaluations, such as familiarity and liking. These results support the use of perceived complexity as a relatively specific behavioral variable for analyzing neural responses to music. More broadly, this behavioral foundation enables future work to examine how neural activity relates to model-derived and subjectively perceived musical complexity, particularly as a function of musical expertise and individual musical sophistication.

Phase Reconstruction as Trajectory: Comparing Griffin-Lim and Diffusion-Based STFT Phase Recovery

Abstract: Phase is often treated as a technical byproduct in audio analysis, yet it helps determine whether sound components reinforce, cancel, or alter perceived timbre. It encodes fine timing relationships among frequency components that are largely invisible in magnitude-only representations. This can be heard in familiar audio phenomena: a phaser effect changes instrumental color by mixing a dry signal with frequency-dependent phase-shifted copies, while polarity inversion provides an extreme case in which a signal summed with its inverted copy cancels to silence despite having the same magnitude spectrum. In many music and speech pipelines, however, phase is discarded, reconstructed after the fact, or delegated to vocoders and waveform generators, while magnitude-based representations remain the main object of analysis. This poster examines phase reconstruction not only as a problem of final audio quality, but as a trajectory through a circular phase space. Two iterative approaches to recovering phase from magnitude are compared: the Griffin-Lim Algorithm, a classical method based on repeated STFT consistency projections, and DiffPhase, a diffusion-based method that reconstructs phase through a learned reverse process. Rather than evaluating solely which method sounds better at convergence, the investigation focuses on how each method moves through phase space during reconstruction. Using examples from speech and musical audio, circular phase displacement, wrapping behavior, discontinuity propagation, and STFT consistency are tracked across iterative steps. Through this exploratory comparison, the objective is to clarify how classical signal-processing methods and contemporary generative models differ in their reconstruction dynamics, and why phase remains a structurally important but often hidden component of audio representation.

Beyond the Notes: Exploring Tool-bounded Musical Creativity through a Cognitive Science Lens

Abstract: Creativity has long been considered a uniquely human capacity. Yet, the emergence of AI-generated music has challenged assumptions regarding the exclusively human nature of creativity. Such debates often reflect a product-oriented understanding of creativity, in which creativity is evaluated primarily through acoustic outcomes produced by gifted or talented individuals.

Recently, however, music scholars have argued that creativity cannot be reduced to the musical product alone (van der Schyff et al., 2018). Instead, musical creativity is increasingly conceptualized as a dynamic human process shaped by a constellation of biological, psychological, and cultural forces—what John Blacking described as “music-making.”

In this presentation, I explore musical creativity as a tool-bounded cognitive process, with particular attention to how the emergence of musical writing system or musical instrument transforms creative thought and practice. I argue that musical creativity should be understood not simply as a sonic artifact, but as an embodied process arising through interactions among the musical mind, body, and tools. Such an approach broadens predominant conceptions of creativity beyond the final musical product and and offers new perspectives on musical creativity in the age of AI.

* This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (2026S1A5B5A16004352).

Zipf-Mandelbrot Scaling in Korean Court Music: Universal Patterns in Music

Abstract: Zipf's law, originally discovered in natural language and later generalized to the Zipf-Mandelbrot law, describes a power-law relationship between the frequency of a Zipfian element and its rank. Due to the semantic characteristics of this law, it has also been observed in musical data. However, most such studies have focused on Western music, and its applicability to non-Western music remains not well investigated. We analyzed 43 Korean court music pieces called Jeong-ak, spanning several centuries and written in the traditional Korean musical notation Jeongganbo. These pieces were transcribed into Western staff notation, and musical data such as pitch and duration were extracted. Using pitch, duration, and their paired combinations as Zipfian units, we found that Korean music also fits the Zipf-Mandelbrot law to a high degree, particularly for the paired pitch-duration unit. Korean music has evolved collectively over long periods, smoothing idiosyncratic variations and producing forms that are widely understandable among people. This collective evolution appears to have played a significant role in shaping the characteristics that lead to the satisfaction of Zipf-Mandelbrot law. Our findings provide additional evidence that Zipf-Mandelbrot scaling in musical data is universal across cultures. We further show that the joint distribution of two independent Zipfian data sets follows the Zipf-Mandelbrot law; in this sense, our result does not merely extend Zipf's law but deepens our understanding of how scaling laws behave under composition and interaction, offering a more unified perspective on rank-based statistical regularities.

A Diffusion Model for Generating Unknown Pinax Data of Athanasius Kircher’s Arca Musarithmica

Abstract: Automatic music generation has a long history. One significant example from the

1600s is Athanasius Kircher’s concept of a music-generating machine. Today, automatic music generation has advanced significantly thanks to the rapid growth of

machine learning and AI techniques. While Kircher’s machine was quite primitive,

relying on direct combinatorial methods, the core idea shares similarities with mod-

ern AI algorithms for music composition. In this study, we focus on Kircher’s music

theory for automatic generation, specifically his set of musical arrays, Syntagma,

and the creation of Kircher-style music not found in his original tables. The number

of Pinax he provided is limited and insufficient for further training to generate more

phrase blocks. To enhance training, we first transform the Pinax data into image

form and apply a diffusion model. For this transformation, we use nearest-neighbor

interpolation, paying special attention to boundary values in the images. By employ-

ing the diffusion model, we demonstrate that it is possible to generate new phrase

blocks beyond Kircher’s original set. This approach not only enables the creation

of a wider variety of Kircher-style music but also potentially provides new insights

through the generated compositions.