Hybrid accelerated synthesis method of TTS system

A synthesis method, a technology of speech synthesis, applied in speech synthesis, special data processing applications, speech analysis, etc., can solve the problems of inability to eliminate the objective influence of huge amount of calculation, the response speed of the first packet, the slow synthesis speed, and the impact of TTS landing effect, etc. , to achieve the effect of optimizing speech synthesis efficiency, reducing computing pressure, and solving high cost

Pending Publication Date: 2020-11-24
普强时代(珠海横琴)信息技术有限公司
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Due to focusing on the clarity, naturalness, fluency and other indicators of sound, the existing technology cannot eliminate the objective impact brought by the huge amount of calculation. The response speed of the first packet and the synthesis speed are getting slower and slower. The requirements for computer performance indicators are getting higher and higher
Problems such as high cost, slow response speed and easy lag directly affect the landing effect of TTS when it is applied to human-computer interaction scenarios
How to make the synthesized voice not be limited by computer performance on the basis of maintaining a high degree of anthropomorphism, improve the efficiency of TTS synthesis, and avoid the phenomenon of voice broadcast stuttering in the process of human-computer interaction has always been an important aspect of the implementation of TTS technology. problem

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
  • Hybrid accelerated synthesis method of TTS system
  • Hybrid accelerated synthesis method of TTS system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0021] The present invention will be further described below in conjunction with specific embodiments, a hybrid accelerated synthesis method of a TTS system, which adopts the mode of recording+TTS real-time synthesis to perform speech synthesis on the entire text requested for synthesis, such as figure 1 As shown, the method specifically includes the following steps:

[0022] Step 1: Check the compliance of the entire requested synthesized text; after the TTS system receives the speech synthesis request, it first checks whether the entire requested synthesized text contains voice-marked text, and if there is no voice-marked text, perform speech synthesis , if it contains speech markup text, then check the compliance of the text requested to be synthesized to check whether there are incomplete tags, if there are incomplete tags, the system will report an error and stop speech synthesis, if the speech markup of the text is complete, the compliance is correct Format requirements,...

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 provides a hybrid accelerated synthesis method of a TTS system. The hybrid accelerated synthesis method comprises the following steps: step 1, carrying out compliance check on a whole text requested to be synthesized; 2, judging and selecting a synthesis mode; 3, calling an existing record, and broadcasting the record; step 4, handling abnormal recording; 5, normally synthesizing a text; 6, ending a voice synthesis action; according to the method, the voice synthesis efficiency is optimized, mixed broadcasting is carried out in different synthesis modes, the calculation pressureof real-time synthesis is greatly reduced, the response speed of first sentence character broadcasting is increased, the voice broadcasting lag phenomenon is avoided, and the hardware cost of a computer is reduced.

Description

technical field [0001] The invention relates to the technical field of speech synthesis, in particular to a hybrid accelerated synthesis method of a TTS system. Background technique [0002] In recent years, with the rapid development of computer technology, speech technology has gradually integrated into people's lives. Products such as telephone robots, AI assistants, and children's story machines all rely on speech synthesis technology to broadcast clear and easy-to-understand messages in the process of human-computer interaction. machine-synthesized sounds. With the widespread application of speech technology, people have higher and higher requirements for the clarity, naturalness and fluency of synthesized voice. TTS technology is constantly innovating. The TTS synthesis method has changed from the initial splicing method to the parametric synthesis method in recent years, and then to the recently emerging end-to-end synthesis method. The effect of synthesized sound ha...

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): G10L13/04G06F16/68G06F16/638
CPCG06F16/638G06F16/68
Inventor 沈彬彬丁妍王添宝何国涛蒲瑶李全忠
Owner 普强时代(珠海横琴)信息技术有限公司
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