公开(公告)号：US11134341B1

公开(公告)日：2021-09-28

申请号：US16865900

申请日：2020-05-04

Applicant: MOTOROLA SOLUTIONS, INC.

Inventor： Kar Meng Tang ,  Kurt S. Fienberg ,  Geng Xiang Lee ,  Lian Kooi Ng ,  Thean Hai Ooi

IPC: H04R3/00

Abstract: A method and apparatus for processing audio signals. One system includes a communication device including a transceiver configured to send and receive audio data, and a microphone configured to convert sound waves to a first audio signal. A speaker is configured to convert received electrical signals to an acoustic output and is configured to convert sound waves to a second audio signal. An electronic processor connected to the microphone and the speaker is configured to receive the first audio signal from the microphone, receive the second audio signal from the speaker, determine a correlation value between the first audio signal and the second audio signal, and compare the correlation value to a correlation threshold. In response to the correlation value being below the correlation threshold, the electronic processor generates an output signal based on the first audio signal and the second audio signal, and transmits the output signal.

公开(公告)号：US10381024B2

公开(公告)日：2019-08-13

申请号：US15498560

申请日：2017-04-27

Applicant: MOTOROLA SOLUTIONS, INC

Inventor： Cheah Heng Tan ,  Thean Hai Ooi ,  Wei Qing Ong ,  Alan Wee Chiat Tan

IPC: G10L25/84 ,  G10L25/18 ,  G10L25/78 ,  G10L25/03

Abstract: A voice activity detection system (100) filters audio input frames (102), on a frame=by-frame basis through a gammatone filterbank (104) to generate filtered gammatone output signals (106). A signal energy calculator (108) takes the filtered gammatone output signals and generates a plurality of energy envelopes. Weighting factors are constructed (112) are applied to each of the energy envelopes thereby producing normalized weighted signal (116), in which voice regions are emphasized and noise regions are minimized. An entropy measurement (118) is taken to extract information from the normalized weighted signals (116) and generate an entropy signal (120). The entropy signal (120) is averaged and compared to an adaptive entropy threshold (122), indicative of a noise floor. Decision logic (124) is used to identifying speech and noise from the comparison of the averaged entropy signal to the adaptive entropy threshold.