Low-Complexity Acoustic Echo Cancellation with Neural Kalman Filtering

TL;DR

This paper introduces neural Kalman filtering (NKF), a low-resource, neural network-based approach to acoustic echo cancellation that improves convergence and re-convergence performance over traditional methods, especially during abrupt echo path changes.

Contribution

The paper presents a novel neural Kalman filtering method that models noise covariances implicitly, achieving superior performance and low computational complexity for acoustic echo cancellation.

Findings

NKF outperforms state-of-the-art methods in convergence and re-convergence.

Model size is only 5.3 K, suitable for low-resource platforms.

RTF as low as 0.09 indicates high efficiency.

Abstract

The Kalman filter has been adopted in acoustic echo cancellation due to its robustness to double-talk, fast convergence, and good steady-state performance. The performance of Kalman filter is closely related to the estimation accuracy of the state noise covariance and the observation noise covariance. The estimation error may lead to unacceptable results, especially when the echo path suffers abrupt changes, the tracking performance of the Kalman filter could be degraded significantly. In this paper, we propose the neural Kalman filtering (NKF), which uses neural networks to implicitly model the covariance of the state noise and observation noise and to output the Kalman gain in real-time. Experimental results on both synthetic test sets and real-recorded test sets show that, the proposed NKF has superior convergence and re-convergence performance while ensuring low near-end speech degradation comparing with the state-of-the-art model-based methods. Moreover, the model size of the proposed NKF is merely 5.3 K and the RTF is as low as 0.09, which indicates that it can be deployed in low-resource platforms.