公开(公告)号：JP2002328695A

公开(公告)日：2002-11-15

申请号：JP2002085138

申请日：2002-03-26

Applicant: IBM

Inventor： TANG DONALD T ,  SHEN LIGIN ,  SHI QIN ,  ZHANG WEI

IPC: G10L13/06 ,  G10L13/02 ,  G10L13/08 ,  G10L21/00

Abstract: PROBLEM TO BE SOLVED: To provide a method for generating a personalized voice from a text. SOLUTION: The method for generating the personalized voice from the text includes a step for analyzing the input text and obtaining standard parameters of a voice to be synthesized from a standard text voice database, a step for mapping the standard parameters to personalized voice parameters with a personalized model obtained in a training process, and a step for synthesizing a voice corresponding to the input text according to the personalized voice parameters. This method is used to simulate the voice of an object person and turn the voice generated by a TTS system into a more attractive and personalized voice.

公开(公告)号：DE112019004076T5

公开(公告)日：2021-05-06

申请号：DE112019004076

申请日：2019-11-05

Applicant: IBM

Inventor： ZHANG WEI ,  ZHANG LI ,  FINKLER ULRICH ,  CHO MINSIK ,  KUNG DAVID

IPC: G06N3/08

Abstract: Verschiedene Ausführungsformen werden für dezentralisiertes verteiltes Deep Learning durch einen oder mehrere Prozessoren in einem Datenverarbeitungssystem bereitgestellt. Asynchrones verteiltes Schulen von einem oder mehreren Maschinenlernmodellen kann durch Generieren einer Liste von Nachbarknoten für jeden Knoten in einer Mehrzahl von Knoten und Erstellen eines ersten Threads für kontinuierliche Datenübertragung gemäß einer Operation zum Gewichtungsmanagement und eines zweiten Threads für kontinuierliche Berechnung eines Gradienten für jeden Knoten ausgeführt werden. Eine oder mehrere Variablen werden zwischen dem ersten Thread und dem zweiten Thread gemeinsam genutzt.

公开(公告)号：DE60216069T2

公开(公告)日：2007-05-31

申请号：DE60216069

申请日：2002-03-15

Applicant: IBM

Inventor： TANG DONALD ,  SHEN LIQIN ,  SHI QIN ,  ZHANG WEI

IPC: G06F17/28 ,  G10L13/04 ,  G10L11/00 ,  G10L13/00 ,  G10L13/08 ,  G10L15/00

Abstract: An expressive speech-to-speech generation system which can generate expressive speech output by using expressive parameters extracted from the original speech signal to drive the standard TTS system. The system comprises: speech recognition means, machine translation means, text-to-speech generation means, expressive parameter detection means for extracting expressive parameters from the speech of language A, and expressive parameter mapping means for mapping the expressive parameters extracted by the expressive parameter detection means from language A to language B, and driving the text-to-speech generation means by the mapping results to synthesize expressive speech.

公开(公告)号：AT345561T

公开(公告)日：2006-12-15

申请号：AT02708485

申请日：2002-03-15

Applicant: IBM

Inventor： TANG DONALD ,  SHEN LIQIN ,  SHI QIN ,  ZHANG WEI

IPC: G06F17/28 ,  G10L11/00 ,  G10L13/00 ,  G10L13/08 ,  G10L15/00 ,  G10L13/04

Abstract: An expressive speech-to-speech generation system which can generate expressive speech output by using expressive parameters extracted from the original speech signal to drive the standard TTS system. The system comprises: speech recognition means, machine translation means, text-to-speech generation means, expressive parameter detection means for extracting expressive parameters from the speech of language A, and expressive parameter mapping means for mapping the expressive parameters extracted by the expressive parameter detection means from language A to language B, and driving the text-to-speech generation means by the mapping results to synthesize expressive speech.

公开(公告)号：DE60216069D1

公开(公告)日：2006-12-28

申请号：DE60216069

申请日：2002-03-15

Applicant: IBM

Inventor： TANG DONALD ,  SHEN LIQIN ,  SHI QIN ,  ZHANG WEI

IPC: G06F17/28 ,  G10L13/04 ,  G10L11/00 ,  G10L13/00 ,  G10L13/08 ,  G10L15/00

Abstract: An expressive speech-to-speech generation system which can generate expressive speech output by using expressive parameters extracted from the original speech signal to drive the standard TTS system. The system comprises: speech recognition means, machine translation means, text-to-speech generation means, expressive parameter detection means for extracting expressive parameters from the speech of language A, and expressive parameter mapping means for mapping the expressive parameters extracted by the expressive parameter detection means from language A to language B, and driving the text-to-speech generation means by the mapping results to synthesize expressive speech.