公开(公告)号：EP3158775A4

公开(公告)日：2018-02-21

申请号：EP15809800

申请日：2015-06-12

申请人： CYPHER LLC

发明人： GEIGER FREDRICK ,  BUNDERSON BRYANT ,  GRUNDSTROM CARL

IPC分类号： G10L21/02 ,  G10L21/0216 ,  G10L25/27 ,  H04R9/08 ,  H04R9/10 ,  H04R19/04

CPC分类号： H04R3/00 ,  G10L21/02 ,  G10L25/27 ,  G10L2021/02165 ,  H04R2410/05 ,  H04R2499/11

摘要： Techniques for processing audio signals include removing noise from the audio signals or otherwise clarifying the audio signals prior to outputting the audio signals. The disclosed techniques may employ minimum mean squared error (MMSE) analyses on audio signals received from a primary microphone and at least one reference microphone, and to techniques in which the MMSE analyses are used to reduce or eliminate noise from audio signals received by the primary microphone. Optionally, confidence intervals may be assigned to different frequency bands of an audio signal, with each confidence interval corresponding to a likelihood that its respective frequency band includes targeted audio, and each confidence interval representing a contribution of its respective frequency band in a reconstructed audio signal from which noise has been removed.

公开(公告)号：EP3198592A1

公开(公告)日：2017-08-02

申请号：EP15844365.5

申请日：2015-09-26

申请人： Cypher, LLC

发明人： VICKERS, Earl

IPC分类号： G10L15/16 ,  G10L25/27 ,  G10L25/78

CPC分类号： G10L21/0264 ,  G10L21/0224 ,  G10L25/30 ,  G10L25/60 ,  G10L25/78 ,  G10L25/84 ,  G10L2015/0636

摘要： A “running range normalization” method includes computing running estimates of the range of values of features useful for voice activity detection (VAD) and normalizing the features by mapping them to a desired range. Running range normalization includes computation of running estimates of the minimum and maximum values of VAD features and normalizing the feature values by mapping the original range to a desired range. Smoothing coefficients are optionally selected to directionally bias a rate of change of at least one of the running estimates of the minimum and maximum values. The normalized VAD feature parameters are used to train a machine learning algorithm to detect voice activity and to use the trained machine learning algorithm to isolate or enhance the speech component of the audio data.

摘要翻译： “运行范围归一化”方法包括计算对语音活动检测（VAD）有用的特征值的范围的运行估计，并通过将特征映射到期望范围来对特征进行归一化。 运行范围标准化包括计算VAD特征的最小值和最大值的运行估计值，并通过将原始范围映射到期望范围来标准化特征值。 平滑系数可选地被选择为定向地偏置最小值和最大值的运行估计中的至少一个的变化率。 归一化的VAD特征参数用于训练机器学习算法以检测语音活动并使用训练的机器学习算法来隔离或增强音频数据的语音分量。

公开(公告)号：EP3198592A4

公开(公告)日：2018-05-16

申请号：EP15844365

申请日：2015-09-26

申请人： CYPHER LLC

发明人： VICKERS EARL

IPC分类号： G10L25/78 ,  G10L25/30

CPC分类号： G10L21/0264 ,  G10L21/0224 ,  G10L25/30 ,  G10L25/60 ,  G10L25/78 ,  G10L25/84 ,  G10L2015/0636

摘要： A “running range normalization” method includes computing running estimates of the range of values of features useful for voice activity detection (VAD) and normalizing the features by mapping them to a desired range. Running range normalization includes computation of running estimates of the minimum and maximum values of VAD features and normalizing the feature values by mapping the original range to a desired range. Smoothing coefficients are optionally selected to directionally bias a rate of change of at least one of the running estimates of the minimum and maximum values. The normalized VAD feature parameters are used to train a machine learning algorithm to detect voice activity and to use the trained machine learning algorithm to isolate or enhance the speech component of the audio data.

公开(公告)号：EP2681691A2

公开(公告)日：2014-01-08

申请号：EP12751851.2

申请日：2012-03-03

申请人： Cypher, LLC

发明人： EDWARDS, Tyson

IPC分类号： G06K9/00

CPC分类号： G06K9/00744 ,  G10L25/51

摘要： A data interpretation and separation system for identifying data elements within a data set that have common features, and separating those data elements from other data elements not sharing such common features. Commonalities relative to methods and/or rates of change within a data set may be used to determine which elements share common features. Determining the commonalities may be performed autonomously by referencing data elements within the data set, and need not be matched against algorithmic or predetermined definitions. Interpreted and separated data may be used to reconstruct an output that includes only separated data. Such reconstruction may be non-destructive. Interpreted and separated data may also be used to retroactively build on existing element sets associated with a particular source.

公开(公告)号：EP3158775A1

公开(公告)日：2017-04-26

申请号：EP15809800.4

申请日：2015-06-12

申请人： Cypher, LLC

发明人： GEIGER, Fredrick ,  BUNDERSON, Bryant ,  GRUNDSTROM, Carl

IPC分类号： H04R9/08 ,  H04R9/10 ,  H04R19/04

CPC分类号： H04R3/00 ,  G10L21/02 ,  G10L25/27 ,  G10L2021/02165 ,  H04R2410/05 ,  H04R2499/11

摘要： Techniques for processing audio signals include removing noise from the audio signals or otherwise clarifying the audio signals prior to outputting the audio signals. The disclosed techniques may employ minimum mean squared error (MMSE) analyses on audio signals received from a primary microphone and at least one reference microphone, and to techniques in which the MMSE analyses are used to reduce or eliminate noise from audio signals received by the primary microphone. Optionally, confidence intervals may be assigned to different frequency bands of an audio signal, with each confidence interval corresponding to a likelihood that its respective frequency band includes targeted audio, and each confidence interval representing a contribution of its respective frequency band in a reconstructed audio signal from which noise has been removed.

摘要翻译： 用于处理音频信号的技术包括从音频信号中去除噪声或者在输出音频信号之前澄清音频信号。 所公开的技术可以对从主要麦克风和至少一个参考麦克风接收到的音频信号采用最小均方误差（MMSE）分析，并且涉及其中使用MMSE分析来减少或消除来自主要麦克风 麦克风。 可选地，可以将置信区间分配给音频信号的不同频带，每个置信区间对应于其各自频带包括目标音频的可能性，并且每个置信区间表示其重建音频信号中其各自频带的贡献 从哪个噪音已被删除。

公开(公告)号：EP2681691A4

公开(公告)日：2015-06-03

申请号：EP12751851

申请日：2012-03-03

申请人： CYPHER LLC

发明人： EDWARDS TYSON

IPC分类号： G06K9/00 ,  G10L25/51

CPC分类号： G06K9/00744 ,  G10L25/51