Joint Estimation of Piano Dynamics and Metrical Structure with a Multi-task Multi-Scale Network

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: He, Zhanhong, Meng, Hanyu, Huang, David, Togneri, Roberto
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910010354696192
author He, Zhanhong
Meng, Hanyu
Huang, David
Togneri, Roberto
author_facet He, Zhanhong
Meng, Hanyu
Huang, David
Togneri, Roberto
contents Estimating piano dynamic from audio recordings is a fundamental challenge in computational music analysis. In this paper, we propose an efficient multi-task network that jointly predicts dynamic levels, change points, beats, and downbeats from a shared latent representation. These four targets form the metrical structure of dynamics in the music score. Inspired by recent vocal dynamic research, we use a multi-scale network as the backbone, which takes Bark-scale specific loudness as the input feature. Compared to log-Mel as input, this reduces model size from 14.7 M to 0.5 M, enabling long sequential input. We use a 60-second audio length in audio segmentation, which doubled the length of beat tracking commonly used. Evaluated on the public MazurkaBL dataset, our model achieves state-of-the-art results across all tasks. This work sets a new benchmark for piano dynamic estimation and delivers a powerful and compact tool, paving the way for large-scale, resource-efficient analysis of musical expression.
format Preprint
id arxiv_https___arxiv_org_abs_2510_18190
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Joint Estimation of Piano Dynamics and Metrical Structure with a Multi-task Multi-Scale Network
He, Zhanhong
Meng, Hanyu
Huang, David
Togneri, Roberto
Audio and Speech Processing
Machine Learning
Sound
H.5.5; I.2.6; I.5.4
Estimating piano dynamic from audio recordings is a fundamental challenge in computational music analysis. In this paper, we propose an efficient multi-task network that jointly predicts dynamic levels, change points, beats, and downbeats from a shared latent representation. These four targets form the metrical structure of dynamics in the music score. Inspired by recent vocal dynamic research, we use a multi-scale network as the backbone, which takes Bark-scale specific loudness as the input feature. Compared to log-Mel as input, this reduces model size from 14.7 M to 0.5 M, enabling long sequential input. We use a 60-second audio length in audio segmentation, which doubled the length of beat tracking commonly used. Evaluated on the public MazurkaBL dataset, our model achieves state-of-the-art results across all tasks. This work sets a new benchmark for piano dynamic estimation and delivers a powerful and compact tool, paving the way for large-scale, resource-efficient analysis of musical expression.
title Joint Estimation of Piano Dynamics and Metrical Structure with a Multi-task Multi-Scale Network
topic Audio and Speech Processing
Machine Learning
Sound
H.5.5; I.2.6; I.5.4
url https://arxiv.org/abs/2510.18190