公开(公告)号：WO1998019297A1

公开(公告)日：1998-05-07

申请号：PCT/US1997018761

申请日：1997-10-15

Applicant: MOTOROLA INC.

Inventor： MOTOROLA INC. ,  CORRIGAN, Gerald ,  KARAALI, Orhan ,  MASSEY, Noel

IPC: G10L03/02

CPC classification number: G10L13/08

Abstract: The present invention teaches a method (400), device and system (300) utilizing at least one of: mapping a sequence of phones to a sequence of articulatory features and utilizing prominence and boundary information, in addition to a predetermined set of rules for type, phonetic context, syntactic and prosodic context for phones to provide a system that generates segment durations efficiently with a small training set.

Abstract translation: 本发明教导了一种方法（400），装置和系统（300），其利用以下至少一个方法：将一系列电话映射到一系列发音特征并利用突出和边界信息，以及用于类型的预定规则集 ，语音语境，语法和韵律上下文，以提供一种能够通过小型训练集高效生成段持续时间的系统。

公开(公告)号：WO1995030193A1

公开(公告)日：1995-11-09

申请号：PCT/US1995003492

申请日：1995-03-21

Applicant: MOTOROLA INC.

Inventor： MOTOROLA INC. ,  KARAALI, Orhan ,  CORRIGAN, Gerald, Edward ,  GERSON, Ira, Alan

IPC: G06F15/18

CPC classification number: G10L13/08 ,  G10L25/30

Abstract: Text may be converted to audible signals, such as speech, by first training a neural network using recorded audio messages (204). To begin the training, the recorded audio messages are converted into a series of audio frames (205) having a fixed duration (213). Then, each audio frame is assigned a phonetic representation (203) and a target acoustic representation, where the phonetic representation (203) is a binary word that represents the phone and articulation characteristics of the audio frame, while the target acoustic representation is a vector of audio information such as pitch and energy. After training, the neural network is used in conversion of text into speech. First, text that is to be converted is translated to a series of phonetic frames of the same form as the phonetic representations (203) and having the fixed duration (213). Then the neural network produces acoustic representations in response to context descriptions (207) that include some of the phonetic frames. The acoustic representations are then converted into a speech wave form by a synthesizer.

Abstract translation: 通过首先使用记录的音频消息训练神经网络（204），将文本转换为可听信号，例如语音。 为了开始训练，记录的音频消息被转换成具有固定持续时间的一系列音频帧（205）（213）。 然后，每个音频帧被分配语音表示（203）和目标声学表示，其中语音表示（203）是表示音频帧的电话和发音特性的二进制字，而目标声学表示是矢量 的音频信息，如音调和能量。 训练后，神经网络用于将文本转换成语言。 首先，要转换的文本被转换为与语音表示（203）相同形式并具有固定持续时间（213）的一系列语音帧。 然后，神经网络响应于包括一些语音帧的上下文描述（207）产生声学表示。 然后，声学表示由合成器转换成语音波形。

公开(公告)号：WO1998025260A2

公开(公告)日：1998-06-11

申请号：PCT/US1997018815

申请日：1997-10-15

Applicant: MOTOROLA INC.

Inventor： MOTOROLA INC. ,  KARAALI, Orhan ,  MASSEY, Noel ,  CORRIGAN, Gerald

IPC: G10L00/00

CPC classification number: G10L13/04 ,  G10L15/063 ,  G10L25/30

Abstract: A method (500, 600), device (201 and 206) and system (203) provide, in response to text/linguistic information, efficient generation of a parametric representation of speech. A coder parameter generating system provides a principal set and a supplementary set of speech parameters, the principal set of speech parameters being the parametric representation of speech. Then feedback is provided to the coder parameter generating system using the supplementary set of speech parameters to modify the principal set of speech parameters.

Abstract translation: 响应于文本/语言信息，方法（500,600），设备（201和206）和系统（203）提供语音的参数表示的有效生成。 编码器参数产生系统提供一组主语集和一组辅助语音参数，主要语音参数集是语音的参数表示。 然后使用补充的语音参数集来向编码器参数生成系统提供反馈以修改主要语音参数集。

公开(公告)号：EP0876660A1

公开(公告)日：1998-11-11

申请号：EP97946842.0

申请日：1997-10-15

Applicant: MOTOROLA, INC.

Inventor： CORRIGAN, Gerald ,  KARAALI, Orhan ,  MASSEY, Noel

IPC: G10L13

CPC classification number: G10L13/08

Abstract: The present invention teaches a method (400), device and system (300) utilizing at least one of: mapping a sequence of phones to a sequence of articulatory features and utilizing prominence and boundary information, in addition to a predetermined set of rules for type, phonetic context, syntactic and prosodic context for phones to provide a system that generates segment durations efficiently with a small training set.

公开(公告)号：EP0876660B1

公开(公告)日：2004-01-02

申请号：EP97946842.8

申请日：1997-10-15

Applicant: MOTOROLA, INC.

Inventor： CORRIGAN, Gerald ,  KARAALI, Orhan ,  MASSEY, Noel

IPC: G10L13/08

CPC classification number: G10L13/08

Abstract: The present invention teaches a method (400), device and system (300) utilizing at least one of: mapping a sequence of phones to a sequence of articulatory features and utilizing prominence and boundary information, in addition to a predetermined set of rules for type, phonetic context, syntactic and prosodic context for phones to provide a system that generates segment durations efficiently with a small training set.

公开(公告)号：EP0710378A1

公开(公告)日：1996-05-08

申请号：EP95913782.0

申请日：1995-03-21

Applicant: MOTOROLA, INC.

Inventor： KARAALI, Orhan ,  CORRIGAN, Gerald, Edward ,  GERSON, Ira, Alan

IPC: G06F15 ,  G06N3 ,  G10L13

CPC classification number: G10L13/08 ,  G10L25/30

Abstract: Text may be converted to audible signals, such as speech, by first training a neural network using recorded audio messages (204). To begin the training, the recorded audio messages are converted into a series of audio frames (205) having a fixed duration (213). Then, each audio frame is assigned a phonetic representation (203) and a target acoustic representation, where the phonetic representation (203) is a binary word that represents the phone and articulation characteristics of the audio frame, while the target acoustic representation is a vector of audio information such as pitch and energy. After training, the neural network is used in conversion of text into speech. First, text that is to be converted is translated to a series of phonetic frames of the same form as the phonetic representations (203) and having the fixed duration (213). Then the neural network produces acoustic representations in response to context descriptions (207) that include some of the phonetic frames. The acoustic representations are then converted into a speech wave form by a synthesizer.

公开(公告)号：EP0932896A2

公开(公告)日：1999-08-04

申请号：EP97946261.0

申请日：1997-10-15

Applicant: MOTOROLA, INC.

Inventor： KARAALI, Orhan ,  MASSEY, Noel ,  CORRIGAN, Gerald

IPC: G10L13 ,  G10L15

CPC classification number: G10L13/04 ,  G10L15/063 ,  G10L25/30

Abstract: A method (500, 600), device (201 and 206) and system (203) provide, in response to text/linguistic information, efficient generation of a parametric representation of speech. A coder parameter generating system provides a principal set and a supplementary set of speech parameters, the principal set of speech parameters being the parametric representation of speech. Then feedback is provided to the coder parameter generating system using the supplementary set of speech parameters to modify the principal set of speech parameters.