Spatial Interpolation of Room Impulse Responses based on Deeper Physics-Informed Neural Networks with Residual Connections

2025/12/28 by Ken Kurata, Gen Sato, Kurata, Ken +5
Computer Science · Engineering · Neuroscience · #Aerodynamics and Acoustics in Jet Flows #Audio and Speech Processing (eess.AS) #FOS: Electrical engineering #Hearing Loss and Rehabilitation #Speech and Audio Processing #electronic engineering #information engineering

paper · doi:10.48550/arxiv.2512.22915

openalex publication_date 2025/12/28 · openalex created_date 2025/12/31 · openalex updated_date 2026/07/28

Abstract

The room impulse response (RIR) characterizes sound propagation in a room from a loudspeaker to a microphone under the linear time-invariant assumption. Estimating RIRs from a limited number of measurement points is crucial for sound propagation analysis and visualization. Physics-informed neural networks (PINNs) have recently been introduced for accurate RIR estimation by embedding governing physical laws into deep learning models; however, the role of network depth has not been systematically investigated. In this study, we developed a deeper PINN architecture with residual connections and analyzed how network depth affects estimation performance. We further compared activation functions, including tanh and sinusoidal activations. Our results indicate that the residual PINN with sinusoidal activations achieves the highest accuracy for both interpolation and extrapolation of RIRs. Moreover, the proposed architecture enables stable training as the depth increases and yields notable improvements in estimating reflection components. These results provide practical guidelines for designing deep and stable PINNs for acoustic-inverse problems.

Citations

• Physics-Informed Neural Network for Volumetric Sound field Reconstruction of Speech Signals
• PirateNets: Physics-informed Deep Learning with Residual Adaptive Networks
• Kernel Learning For Sound Field Estimation With L1 and L2 Regularizations
• Generative Adversarial Networks
• MeshRIR: A Dataset of Room Impulse Responses on Meshed Grid Points For Evaluating Sound Field Analysis and Synthesis Methods
• Compressive Neural Representations of Volumetric Scalar Fields
• Implicit Neural Representations with Periodic Activation Functions
• Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Densely Connected Convolutional Networks
• Deep Residual Learning for Image Recognition
• U-Net: Convolutional Networks for Biomedical Image Segmentation
• ImageNet classification with deep convolutional neural networks
• Active noise control: a tutorial review
• The perceptron: A probabilistic model for information storage and organization in the brain.

Related

• Room impulse response reconstruction with physics-informed deep learning
• Room impulse response reconstruction with physics-informed deep learning
• Room Impulse Response Prediction with Neural Networks: From Energy Decay Curves to Perceptual Validation
• Physics-Informed Direction-Aware Neural Acoustic Fields
• Explicit Context-Driven Neural Acoustic Modeling for High-Fidelity RIR Generation