Voice synthesis database pause information automatic marking method and system
A technology of speech synthesis and speech data, applied in speech synthesis, speech analysis, instruments, etc., can solve the problems of high cost, long cycle, poor economic benefits, etc., achieve short cycle, high labeling accuracy, save cost and time Effect
Active Publication Date: 2016-06-01
UNISOUND SHANGHAI INTELLIGENT TECH CO LTD
View PDF11 Cites 19 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] The purpose of the present invention is to overcome the defects of the prior art, provide a method and system for automatically labeling pause information in a speech synthesis database, and solve the problem of long cycle, high cost, and economic benefits of the method of marking pauses by manual listening in the prior art bad question
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0048] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.
[0049]The present invention provides a system and method for automatically labeling pause information in a speech synthesis database, which is used to solve the problems of long labeling cycle, high cost, low labeling accuracy and poor economic benefit in the existing artificial listening to voice to judge the presence of labeling pause information. question. The present invention uses the smoothed speech feature frame energy sequence combined with the minimum length to judge the pause information in the speech, and then generates a phoneme sequence with label information based on the mapping relationship between the automatically segmented phoneme sequence and the pause information on the time axis. With the post-processing of labeling information, the automatic labeling of pause information is realized, with high accuracy and short labeling t...
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
Login to View More Abstract
The invention relates to a voice synthesis database pause information automatic marking method and system, and the method comprises the steps: obtaining to-be-marked voice data in a voice synthesis database; enabling the voice data to be converted into a voice feature frame sequence, judging the pause information in the voice feature frame sequence, and forming a prediction position of the pause information; converting the voice data into text data; calculating the mapping relation of the voice data and the text data in a time axis through employing an automatic cutting alignment method; enabling the pause information to be inserted into the text data based on the mapping relation, so as to form mark text information; and calculating the mapping relation of the voice data and the mark text information in a time axis through employing the automatic cutting alignment method. The method achieves the automatic marking of the pause information of the voice data in data, saves the cost and time in manually marking the pause, is short in period, is high in accuracy, and is good in economic benefit.
Description
technical field [0001] The invention relates to the field of speech synthesis, in particular to a method and system for automatically marking pause information in a speech synthesis database. Background technique [0002] Speech synthesis refers to a system that converts input text information into sound. The speech synthesis system is divided into two modules, the front-end processing module and the back-end module. The text is analyzed in the front end, and information related to prosodic pauses such as pronunciation, word segmentation, part of speech, etc. is output. The back-end module uses the output information of the front-end module and the features proposed by the original speech to train the cepstrum parameter model, fundamental frequency parameter model and duration parameter model respectively. The advantage of the parametric speech synthesis system is that the model is small, it is convenient for synthesis and customization, and it is convenient for offline imp...
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
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G10L13/08
CPCG10L13/08G10L2013/083
Inventor 刘青松许东星王鸣黄盼
Owner UNISOUND SHANGHAI INTELLIGENT TECH CO LTD