Blind Acoustic Parameter Estimation Through Task-Agnostic Embeddings Using Latent Approximations

TL;DR

This paper introduces a three-stage method using latent representations and auto-encoders for blind acoustic parameter estimation from reverberant speech, outperforming baseline models.

Contribution

It proposes a novel approach combining variational auto-encoders and task-agnostic speech embeddings for acoustic parameter estimation.

Findings

Outperforms end-to-end baseline models.

Effective in estimating acoustic parameters from reverberant speech.

Uses latent representations to improve estimation accuracy.

Abstract

We present a method for blind acoustic parameter estimation from single-channel reverberant speech. The method is structured into three stages. In the first stage, a variational auto-encoder is trained to extract latent representations of acoustic impulse responses represented as mel-spectrograms. In the second stage, a separate speech encoder is trained to estimate low-dimensional representations from short segments of reverberant speech. Finally, the pre-trained speech encoder is combined with a small regression model and evaluated on two parameter regression tasks. Experimentally, the proposed method is shown to outperform a fully end-to-end trained baseline model.