A frequency-domain nonlinear echo processing algorithm for high quality hands-free voice communication devices

Wang, Qingyun; Chen, Xin; Liang, Ruiyu; Liu, Haicheng

doi:10.1007/s11042-020-10230-y

Cited by 2 publications

(1 citation statement)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The central concepts of these algorithms include the minimum mean square error criterion [ 6 ], affine projection [ 16 ], and various popular objective functions [ 35 ], e.g., least squares, weighted least squares, and mean absolute error. A real-time AEC algorithm based on variable step-size partitioned block frequency-domain adaptive filtering and frequency-domain nonlinear echo processing algorithm is proposed for hands-free voice communication devices [ 33 ].…”

Section: Audio Processing Issuesmentioning

confidence: 99%

CoMAC: A cooperation-based multiparty audio conferencing system for mobile users

Wang

Chen

Sheu

2022

Multimed Tools Appl

View full text Add to dashboard Cite

The existing architectures used in the multiparty audio conferencing systems are typically categorized as either centralized or decentralized. These architectures expose a trade-off between processing latency and system capacity, namely the number of participants. This paper proposes a multiparty audio conferencing system for mobile users to improve the processing latency and system capacity. Instead of using the pure centralized or decentralized architecture, the proposed system adopts a novel cooperation-based architecture, in which only some participants are selected as the central controllers to deal with the tasks such as acoustic echo cancellation, encoding, decoding, mixing and de-mixing. The proposed system also uses a buffer reordering scheme to solve the problems of network jitters and out-of-order packets. This study analyzed the processing latency of the multiparty audio conferencing systems using the existing and proposed architectures. We also implemented these systems on diverse mobile platforms to compare the processing latency and number of participants. Performance evaluation results confirmed that the proposed cooperation-based architecture can not only reduce the processing latency but also support more participants, compared to the existing architectures.

show abstract

Section: Audio Processing Issuesmentioning

confidence: 99%