Advances in Music Performance Synthesis and Analysis

The field of music performance synthesis and analysis is rapidly evolving, with a focus on developing innovative models and techniques to generate high-quality music performances and analyze musical structures. Recent research has explored the use of neural networks and machine learning algorithms to improve the expressiveness and generalization of music performance synthesis systems. Additionally, there has been a growing interest in developing digital tools and interfaces for music creation and performance, such as gamified instruments and platforms for harnessing microtonal and justly intonated sounds. Furthermore, researchers have been working on creating datasets and frameworks for computational music analysis, including the development of corpora for non-metric Iranian classical music and algorithms for parsing musical structure. Noteworthy papers in this area include MIDI-VALLE, which proposes a neural codec language model for improving expressive piano performance synthesis, and RUMAA, which introduces a transformer-based framework for unified music audio analysis. LIMITER is also notable for its gamified interface for harnessing just intonation systems. These developments have the potential to significantly advance the field of music performance synthesis and analysis, enabling the creation of more realistic and expressive music performances, and providing new tools and insights for music creators and analysts.

Sources

MIDI-VALLE: Improving Expressive Piano Performance Synthesis Through Neural Codec Language Modelling

LIMITER: A Gamified Interface for Harnessing Just Intonation Systems

ASTAR-NTU solution to AudioMOS Challenge 2025 Track1

Radif corpus: a symbolic dataset for non-metric iranian classical music

Grammatical Structure and Grammatical Variations in Non-Metric Iranian Classical Music

Parsing Musical Structure to Enable Meaningful Variations

MambaRate: Speech Quality Assessment Across Different Sampling Rates

RUMAA: Repeat-Aware Unified Music Audio Analysis for Score-Performance Alignment, Transcription, and Mistake Detection

Large Language Models' Internal Perception of Symbolic Music

Built with on top of