公开(公告)号：US09053711B1

公开(公告)日：2015-06-09

申请号：US14023424

申请日：2013-09-10

Applicant: Ampersand, Inc.

Inventor： Stephen Smith ,  George Howitt

IPC: H04M1/64 ,  G10L25/78 ,  H04M3/533 ,  H04M3/493

CPC classification number: G10L25/78 ,  G10L15/20 ,  G10L15/265 ,  G10L25/18 ,  G10L25/51 ,  G10L25/93 ,  H04M3/493 ,  H04M3/533 ,  H04M3/53383 ,  H04M2201/40 ,  H04M2203/2027

Abstract: A system and method matches a digitized stream of audio signals to a known audio recording by receiving an audio stream from an application. The method performs signal analysis on sound and silence in the audio stream to compute a stream audio print of the audio stream, to identify a pattern associated with the audio stream. The method compares the stream audio print to one or more known audio prints from a library of known audio prints. The method first applies a silence constraint to match periods of silence. If application of the silence constraint is successful, the method computes a total scoring metric function to identify a match between the stream audio print and the known audio print. The total scoring metric function employs a scoring metric function that is computed at a subset of temporal points in the non-silent portion of the audio stream audio print.

Abstract translation: 系统和方法通过从应用接收音频流将音频信号的数字化流与已知音频记录相匹配。 该方法对音频流中的声音和静音执行信号分析以计算音频流的流音频打印，以识别与音频流相关联的模式。 该方法将流音频打印与已知音频打印的库中的一个或多个已知的音频打印进行比较。 该方法首先应用沉默约束来匹配沉默时期。 如果沉默约束的应用成功，则该方法计算总评分度量函数以识别流音频打印和已知音频打印之间的匹配。 总评分度量函数使用在音频流音频打印的非静音部分中的时间点的子集处计算的评分度量函数。

公开(公告)号：US09679584B1

公开(公告)日：2017-06-13

申请号：US14677970

申请日：2015-04-02

Applicant: Ampersand, Inc.

Inventor： Stephen Smith ,  George Howitt

IPC: G10L25/78 ,  G10L25/93 ,  G10L15/20 ,  G10L15/26 ,  H04M3/533 ,  H04M3/493

CPC classification number: G10L25/78 ,  G10L15/20 ,  G10L15/265 ,  G10L25/18 ,  G10L25/51 ,  G10L25/93 ,  H04M3/493 ,  H04M3/533 ,  H04M3/53383 ,  H04M2201/40 ,  H04M2203/2027

Abstract: A system and method matches a digitized stream of audio signals to a known audio recording by receiving an audio stream from an application. The method performs signal analysis on sound and silence in the audio stream to compute a stream audio print of the audio stream, to identify a pattern associated with the audio stream. The method compares the stream audio print to one or more known audio prints from a library of known audio prints. The method first applies a silence constraint to match periods of silence. If application of the silence constraint is successful, the method computes a total scoring metric function to identify a match between the stream audio print and the known audio print. The total scoring metric function employs a scoring metric function that is computed at a subset of temporal points in the non-silent portion of the audio stream audio print.

公开(公告)号：US10014006B1

公开(公告)日：2018-07-03

申请号：US15589622

申请日：2017-05-08

Applicant: Ampersand, Inc.

Inventor： Stephen Smith ,  George Howitt ,  Douglas Campbell

IPC: G10L25/78 ,  G10L25/51 ,  G10L25/90 ,  G10L25/18 ,  H04M3/523 ,  H04M3/493

CPC classification number: G10L25/78 ,  G10L17/02 ,  G10L17/08 ,  G10L25/51 ,  H04M3/493 ,  H04M3/53383 ,  H04M2203/2027 ,  H04M2203/306

Abstract: A system receives an audio stream from an application, and performs signal analysis on the audio stream to compute (on sound and silence) a sequence of events. The sequence of events identifies a pattern comprised of at least two of sound, silence, frequency, magnitude, duration, tone, and Dual Tone Multi Frequency (DTMF). The pattern matches a similar pattern of at least two of sound, silence, frequency, magnitude, duration, tone, and DTMF, and the signal analysis is performed on sound and silence within the audio stream. The similar pattern is associated with at least one of an automated device or human answering the call. The system determines whether the sequence of events indicates the phone call is answered by the automated device or the human, based on rules in a heuristic engine and/or a known audio print from a library of known audio prints.