Speech translation method and device

A speech translation and speech technology, applied in the computer field, can solve the problems of inappropriate English text, unable to express the style and characteristics of the source speaker, unable to express the style and characteristics of the source speaker, etc., so as to facilitate the understanding of semantics and context. , expressive

Active Publication Date: 2018-06-29
IFLYTEK CO LTD
View PDF10 Cites 14 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, in the current text translation technology, most of the translation technologies only realize the literal translation of the text, that is to say, when the speech data of the source speaker is translated into the text, the translated text often cannot express the source Speaker's style and characteristics
For example, when translating a Chinese speech into English text, since the Chinese text of the Chinese speech may correspond to different English texts, but the language style and emotional characteristics expressed by different English texts may be different, and the actual translated English text Often inappropriate, i.e. the translated text fails to convey the style and character of the source speaker

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Speech translation method and device
  • Speech translation method and device
  • Speech translation method and device

Examples

Experimental program
Comparison scheme
Effect test

no. 1 example

[0076] see figure 1 , a schematic flowchart of a speech translation method provided in an embodiment of the present application, the method includes the following steps:

[0077] S101: Acquire first voice data.

[0078] In this embodiment, the voice data that needs to be translated into text is defined as the first voice data.

[0079] This embodiment does not limit the source of the first voice data. For example, the first voice data may be the real voice or the recorded voice of the source speaker, or may be the real voice or the recorded voice. The processed special effect voice.

[0080] This embodiment also does not limit the length of the first voice data. For example, the first voice data may be a word, a sentence, or a paragraph.

[0081] S102: Generate speech recognition text by performing speech recognition on the first speech data.

[0082] After the first voice data is acquired, the first voice data is converted into voice recognition text by using a voice reco...

no. 2 example

[0089] This embodiment will focus on the specific implementation of S103 in the above-mentioned first embodiment. For other relevant parts, please refer to the introduction of the first embodiment.

[0090] see figure 2 , a schematic flowchart of a speech translation method provided in an embodiment of the present application, the method includes the following steps:

[0091] S201: Acquire first voice data.

[0092] S202: Generate speech recognition text by performing speech recognition on the first speech data.

[0093] It should be noted that S201 and S202 in this embodiment are the same as S101 and S102 in the first embodiment. For related descriptions, please refer to the first embodiment, which will not be repeated here.

[0094] S203: Use the speech recognition text as a unit recognition text, or use each text segment forming the speech recognition text as a unit recognition text respectively.

[0095] In this embodiment, the speech recognition text as a whole may be...

no. 3 example

[0131] In the current speech translation technology, the audio synthesized by the machine after translation is completely the speaking style of the speaker trained in the synthesis model. The effect of the synthesized audio has a very low correlation with the speaking style of the source speaker before translation. Sometimes, simply translating It is difficult for the incoming audio to express the style and characteristics of the source speaker.

[0132] In order to solve this defect, this embodiment provides a voice translation method, which can translate the voice data of the source speaker (that is, the first voice data in the first embodiment and the second embodiment), obtain the translated text, and Audio synthesis is performed by combining the acoustic features in the speech data, so that the synthesized audio is adapted to the speech style of the source speaker, so as to achieve a more natural and expressive speech translation. This method is suitable for real-time spo...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The application discloses a speech translation method and device. The method comprises the following steps: with respect to speech data on which text translation requires to be carried out, generatinga speech recognition text through carrying out speech recognition on the speech data; extracting acoustic characteristics from the speech data, translating the speech recognition text according to extracted acoustic characteristics to obtain a translation text which has a speech style of the speech data. Therefore, as the acoustic characteristics of the speech data per se are considered during carrying out text translation on the speech data, the translation text can conform to the style and characteristics of the speech data, so that the translation text is more natural and has stronger expressive power, and then the meaning and context are convenient for a text reader to understand.

Description

technical field [0001] The present application relates to the field of computer technology, and in particular, to a method and device for speech translation. Background technique [0002] With the increasing maturity of artificial intelligence technology, people are increasingly pursuing the use of intelligent technology to solve some problems. For example, it used to take a lot of time to learn a new language in order to be compatible with native speakers of the language. People communicate with each other, and now, people can directly use the translation machine to realize oral input, text translation, and pronunciation to say the translated meaning around speech recognition, intelligent translation and speech synthesis technology. [0003] However, in the current text translation technologies, most translation technologies only realize the literal translation of the text, that is to say, when text translation is performed on the source speaker's voice data, the translated...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G10L15/00G06F17/28
CPCG10L15/005G06F40/58
Inventor 王雨蒙周良江源胡国平
Owner IFLYTEK CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap