End-to-end accent conversion method

A conversion method and accent technology, applied in speech analysis, speech synthesis, speech recognition, etc., can solve problems such as authentic accent troubles

Pending Publication Date: 2020-07-28
深圳市达旦数生科技有限公司
View PDF6 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to provide an end-to-end accent conversion method to solve the troublesome probl...

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
  • End-to-end accent conversion method
  • End-to-end accent conversion method
  • End-to-end accent conversion method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0015] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings. It should be noted here that the descriptions of these embodiments are used to help understand the present invention, but are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not constitute a conflict with each other.

[0016] This embodiment proposes an end-to-end accent conversion method, including an accent conversion system that implements the accent conversion method. The accent conversion system includes a speech recognition module, a speaker encoder, a speech synthesis module, a neural network vocoder, and a speech recognition system. The module is used to adjust the acoustic features of the input non-native accent to the signal parameters of the non-native accent, and the signal...

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 invention discloses an end-to-end accent conversion method, and belongs to the technical field of voice processing. The method converts a non-standard accent into a standard accent, and can be used for converting the voice of a patient with pronunciation disorder into standard voice. An accent conversion system for realizing the accent conversion method comprises a voice recognition module, aspeaker encoder, a voice synthesis module and a neural network vocoder, wherein the voice recognition module is used for adjusting the acoustic characteristics of input non-standard accent into the signal parameters of a standard accent, wherein the signal parameters are only related to the speaking content of the non-standard accent; and inputting the signal parameters of the non-standard accentand the speaker vector into the voice synthesis module, and synthesizing the standard accent of the specific speaker through the neural network vocoder after the voice is processed by the voice synthesis module. The method has the advantages that in the conversion process, the non-standard accent can be converted into the standard accent without any guidance of standard accent reference audios, and the original tone of a speaker is kept.

Description

technical field [0001] The invention relates to the technical field of speech processing, in particular to an end-to-end accent conversion method. Background technique [0002] Speech recognition technology is more and more widely used. The current speech recognition library is basically based on the standard Mandarin speech. The standard Mandarin of the speaker is converted into text, and the accuracy rate is relatively high. However, in real life, the pronunciation of most people is not standard, and they carry some local accents more or less. In order to communicate with them better, it is necessary to change the non-authentic accent into an authentic accent. In addition, there are currently many patients with dysphonia (such as Chinese patients) who cannot communicate with other people normally. It is particularly important for patients' daily communication to convert their non-standard pronunciation into standard pronunciation. The traditional speech conversion method ...

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
IPC IPC(8): G10L21/013G10L19/00G10L19/16G10L25/24G10L25/30G10L15/02G10L15/06G10L13/02G10L13/04
CPCG10L21/013G10L19/0018G10L19/16G10L25/24G10L25/30G10L15/02G10L15/063G10L13/02G10L2021/0135Y02T10/40
Inventor 刘颂湘王迪松曹悦雯孙立发吴锡欣康世胤吴志勇刘循英蒙美玲
Owner 深圳市达旦数生科技有限公司
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