Relative Transfer Function Inverse Regression from Low Dimensional Manifold

TL;DR

This paper explores a data-driven deep learning approach to model the low-dimensional manifold of room acoustic RTFs, enabling high-dimensional response prediction from low-dimensional source-receiver pose data.

Contribution

It introduces a supervised DNN model for RTF inverse regression, demonstrating its ability to predict RTFs more accurately than free field assumptions.

Findings

DNN achieves lower prediction error than free field assumption.

Model outperforms linear interpolation at larger sampling distances.

Fails to outperform linear interpolation at small sampling distances.

Abstract

In room acoustic environments, the Relative Transfer Functions (RTFs) are controlled by few underlying modes of variability. Accordingly, they are confined to a low-dimensional manifold. In this letter, we investigate a RTF inverse regression problem, the task of which is to generate the high-dimensional responses from their low-dimensional representations. The problem is addressed from a pure data-driven perspective and a supervised Deep Neural Network (DNN) model is applied to learn a mapping from the source-receiver poses (positions and orientations) to the frequency domain RTF vectors. The experiments show promising results: the model achieves lower prediction error of the RTF than the free field assumption. However, it fails to compete with the linear interpolation technique in small sampling distances.