公开(公告)号：WO2015057757A1

公开(公告)日：2015-04-23

申请号：PCT/US2014/060567

申请日：2014-10-15

Applicant: KNOWLES ELECTRONICS, LLC

Inventor： NANDY, Dibyendu ,  LI, Yang ,  THOMSEN, Henrik ,  FURST, Claus

IPC: G10L25/78

CPC classification number: G10L15/20 ,  G06F1/32 ,  G10L15/28 ,  G10L19/002 ,  G10L25/84 ,  G10L2015/088

Abstract: Streaming audio is received. The streaming audio includes a frame having plurality of samples. An energy estimate is obtained for the plurality of samples. The energy estimate is compared to at least one threshold. In addition, a band pass estimate of the signal is determined. An energy estimate is obtained for the band-passed plurality of samples. The two energy estimates are compared to at least one threshold each. Based upon the comparison operation, a determination is made as to whether speech is detected.

Abstract translation: 接收流音频。 流式音频包括具有多个样本的帧。 对于多个样本获得能量估计。 将能量估计与至少一个阈值进行比较。 此外，确定信号的带通估计。 对带通的多个样本获得能量估计。 将两个能量估计值与至少一个阈值进行比较。 基于比较操作，确定是否检测到语音。

公开(公告)号：WO2017011403A1

公开(公告)日：2017-01-19

申请号：PCT/US2016/041787

申请日：2016-07-11

Applicant: KNOWLES ELECTRONICS, LLC

Inventor： NANDY, Dibyendu ,  LI, Yang ,  RAGHUVIR, Ramanujapuram ,  POPPER, Robert A. ,  PAWLOWSKI, Andrzej ,  BERTHELSEN, Kim Spetzler ,  THOMSEN, Henrik ,  WARREN, Niel, D. ,  ROSSUM, David, P.

IPC: G10L21/0224 ,  G10L21/02

CPC classification number: G10L15/22 ,  G06F1/3206 ,  G10L15/30 ,  G10L25/48 ,  G10L25/78 ,  H04R3/00 ,  H04R3/005 ,  H04R19/04 ,  H04R2201/003

Abstract: At a microphone, voice activity is detected in a data stream while simultaneously buffering audio data from the data stream to create buffered data. A signal is sent to a host indicating the positive detection of voice activity in the data stream. When an external clock signal is received from the host, the internal operation of the microphone is synchronized with the external clock signal. Buffered data stream is selectively sent through a first path, the first path including a buffer having a buffer delay time representing the time the first data stream takes to move through the buffer. The data stream is continuously sent through a second path as a real time data stream, the second path not including the buffer, the real time data stream beginning with the extended buffer data at a given instant in time. The buffered data stream and the real time data stream are multiplexed onto a single data line and transmitting the multiplexed data stream to the host.

Abstract translation: 在麦克风上，在数据流中检测到语音活动，同时从数据流缓冲音频数据以创建缓冲数据。 信号被发送到主机，指示数据流中的语音活动的肯定检测。 当从主机接收到外部时钟信号时，麦克风的内部操作与外部时钟信号同步。 缓冲数据流被选择性地通过第一路径发送，第一路径包括缓冲器，缓冲器具有表示第一数据流在缓冲器中移动的时间的缓冲延迟时间。 数据流通过作为实时数据流的第二路径连续发送，第二路径不包括缓冲器，在给定时刻从扩展缓冲器数据开始的实时数据流。 将缓冲的数据流和实时数据流复用到单个数据线上，并将多路复用的数据流发送到主机。

公开(公告)号：WO2015069878A1

公开(公告)日：2015-05-14

申请号：PCT/US2014/064324

申请日：2014-11-06

Applicant: KNOWLES ELECTRONICS, LLC

Inventor： DAI, WeiWen ,  POPPER, Robert A ,  NANDY, Dibyendu ,  LI, Yang ,  QUTUB, Sarmad ,  FURST, Claus Erdmann ,  THOMSEN, Henrik ,  KASSIR, Thibault

IPC: G10L25/78

CPC classification number: G10L25/78 ,  G10L15/28

Abstract: Analog signals are received from a sound transducer. The analog signals are converted into digitized data. A determination is made as to whether voice activity exists within the digitized signal. Upon the detection of voice activity, an indication of voice activity is sent to a processing device. The indication is sent across a standard interface, and the standard interface is configured to be compatible to be coupled with a plurality of devices from potentially different manufacturers.

Abstract translation: 模拟信号从声音传感器接收。 模拟信号转换为数字化数据。 确定语音活动是否存在于数字化信号内。 在检测到语音活动时，将语音活动的指示发送到处理设备。 该指示通过标准接口发送，并且标准接口被配置为兼容以与来自潜在不同制造商的多个设备耦合。