Neural Kalman Filters for Acoustic Echo Cancellation: Comparison of deep neural network-based extensions [Special Issue On Model-Based and Data-Driven Audio Signal Processing]

被引：1

作者：

Seidel, Ernst ^{[1
]}

Enzner, Gerald ^{[2
]}

Mowlaee, Pejman ^{[3
]}

Fingscheidt, Tim ^{[4
]}

机构：

[1] Tech Univ Carolo Wilhelmina Braunschweig, Inst Commun Technol, D-38106 Braunschweig, Germany

[2] Carl von Ossietzky Univ Oldenburg, Dept Med Phys & Acoust, Div Speech Technol & Hearing Aids, D-26111 Oldenburg, Germany

[3] GN Adv Sci, DK-2750 Ballerup, Denmark

[4] Tech Univ Carolo Wilhelmina Braunschweig, D-38106 Braunschweig, Germany

来源：

IEEE SIGNAL PROCESSING MAGAZINE | 2024年 / 41卷 / 06期

关键词：

Training; Measurement; Echo cancellers; Filtering; Special issues and sections; Noise; Training data; Acoustics; Kalman filters; Speech processing;

D O I：

10.1109/MSP.2024.3449557

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

Kalman filtering is a powerful approach to adaptive filtering for various problems in signal processing. The frequency-domain adaptive Kalman filter (FDKF), based on the concept of the acoustic state space, provides a unifying solution to the adaptive filter update and the related stepsize control. It was conceived for the problem of acoustic echo cancellation and, as such, is frequently applied in hands-free systems. This article motivates and briefly recapitulates the linear FDKF and investigates how it can be further supported by deep neural networks (DNNs) in various ways, specifically to overcome the challenges and limitations related to the usually required estimation of process and observation noise covariances for the Kalman filter. While the mere FDKF comes with very low computational complexity, its neural Kalman filter variants may deliver faster (re)convergence, better echo cancellation, and even exceed the FDKF in its excellent double-talk near-end speech preservation both under linear and nonlinear loudspeaker conditions. To provide a synopsis of the state of the art, this article contributes a comparison of a range of DNN-based extensions of FDKF in the same training framework and using the same data. © 1991-2012 IEEE.

引用

页码：24 / 38

页数：15

共 54 条

[1] Hansler E., Schmidt G., Acoustic Echo and Noise Control: A Practical Approach, (2004)
[2] Kellermann W., Analysis and design of multirate systems for cancellation of acoustical echoes, Proc. Int. Conf. Acoust., Speech, Signal Process. (ICASSP), pp. 2570-2573, (1988)
[3] Breining C., Et al., Acoustic echo control: An application of very-high-order adaptive filters, IEEE Signal Process. Mag., 16, 4, pp. 42-69, (1999)
[4] Benesty J., Gansler T., Morgan D., Sondhi M., Gay S., Advances in Network and Acoustic Echo Cancellation, (2001)
[5] Enzner G., Vary P., Frequency-domain adaptive Kalman filter for acoustic echo control in hands-free telephones, Signal Process., 86, 6, pp. 1140-1156, (2006)
[6] Richard G., Et al., Audio signal processing in the 21st century: The important outcomes of the past 25 years, IEEE Signal Process. Mag., 40, 5, pp. 12-26, (2023)
[7] Franzen J., Seidel E., Fingscheidt T., AEC in a Netshell: On target and topology choices for FCRN acoustic echo cancellation, Proc. Int. Conf. Acoust., Speech, Signal Process. (ICASSP), pp. 156-160, (2021)
[8] Seidel E., Franzen J., Strake M., Fingscheidt T., Y2-Net FCRN for acoustic echo and noise suppression, Proc. Interspeech, Brno, Czech Republic, pp. 4763-4767, (2021)
[9] Braun S., Valero M.L., Task splitting for DNN-based acoustic echo and noise removal, Proc. Int. Workshop Acoust. Signal Enhancement (IWAENC), pp. 386-390, (2022)
[10] Seidel E., Mowlaee P., Fingscheidt T., Efficient deep acoustic echo suppression with condition-aware training, Proc. IEEE Workshop Appl. Signal Process. Audio Acoust. (WASPAA), pp. 1-5, (2023)

← 1 2 3 4 5 6 →