Evaluation of Neural Surrogates for Physical Modelling Synthesis of Nonlinear Elastic Plates

TL;DR

This paper compares neural network models for simulating nonlinear elastic plates in physical modeling synthesis, highlighting their limitations and discussing future improvements for real-time audio applications.

Contribution

It provides a comprehensive evaluation of neural surrogates for nonlinear elastic plate vibration modeling, emphasizing their current limitations and proposing directions for enhancement.

Findings

Neural models can predict short sequences but struggle with long-term accuracy.

Time domain error metrics are insufficient to evaluate model performance.

Limitations impact real-time synthesis of nonlinear vibrating structures.

Abstract

Physical modelling synthesis aims to generate audio from physical simulations of vibrating structures. Thin elastic plates are a common model for drum membranes. Traditional numerical methods like finite differences and finite elements offer high accuracy but are computationally demanding, limiting their use in real-time audio applications. This paper presents a comparative analysis of neural network-based approaches for solving the vibration of nonlinear elastic plates. We evaluate several state-of-the-art models, trained on short sequences, for prediction of long sequences in an autoregressive fashion. We show some of the limitations of these models, and why is not enough to look at the prediction error in the time domain. We discuss the implications for real-time audio synthesis and propose future directions for improving neural approaches to model nonlinear vibration.