Simulation-Based Plate-Reverb Parameter Estimation from a Single Impulse Response

2026-08-01

AI Summary

This work addresses the challenge of accurately estimating physical parameters from a single plate reverberation impulse response by proposing a non-iterative, end-to-end parameter estimation method. Trained exclusively on synthetic data, the approach extracts amplitude, spectral, and decay characteristics from the impulse response and employs an ensemble tree regression model to simultaneously predict six target parameters in a single forward pass. Evaluated on two independent synthetic validation sets and a shared test set, the proposed method consistently outperforms conventional baselines and the official default particle swarm optimization (PSO) with a single run, while substantially reducing computational overhead during inference. This demonstrates a compelling balance between estimation accuracy and computational efficiency.

Abstract

We present a simulation-trained, non-iterative estimator for Task A of the 1st DAFx Parameter Estimation Challenge. Each unnormalized plate-reverb impulse response is summarized by amplitude, spectral, and decay descriptors, and an ensemble of tree regressors estimates the six target parameters in one pass. Across two independent synthetic validation sets, the normalized models outperform the training-set mean and an earlier raw-regression baseline. On a shared set, the final ensemble also outperforms a single run of the official default PSO at substantially lower inference cost. Since the official labels are hidden, parameter accuracy is measured on simulator-matched data, and the released responses support only audio-side consistency checks. The estimator returns point estimates without uncertainty.

Problem

• plate reverb
• parameter estimation
• impulse response
• simulation-based estimation
• audio signal processing

Innovation

• simulation-based estimation
• non-iterative
• plate reverb
• ensemble tree regressors
• impulse response parameterization