A method of cloning accent and rhyme based on speech training
A voice training and rhyme technology, applied in voice analysis, voice synthesis, instruments, etc., can solve the problems of cloning and simulating user rhythm, accent voice content, etc.
Active Publication Date: 2021-06-15
SHENZHEN SHUZIXINGHE TECH CO LTD
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0002] Voice technology is now one of the hottest and most cutting-edge technical fields in scientific and technological research. Voice technology is also a subject of vigorous research by many international technology companies such as Google and Xunfei. There is a LYREBIRD (Lyrebird) website platform abroad that provides voice cloning services. , the user only needs to read aloud through the text of about 30 sentences on the website platform, and the user can get the reading voice of the input text content, and the sound quality of the voice obtained is similar to the user's voice. Now it has become a hot topic, and many users are in major The video website platform releases this voice cloning video, but the LYREBIRD (Lyrebird) website platform is the same as other existing voice mode cloning technologies. Although the sound quality is very close to the user's voice, it is a great breakthrough, but the released The human voice is still a mechanical electronic sound, and acoustic engineers all over the world have been working on making the machine "speak human", but most of the results are still far from the real human voice, which is blunt and unnatural. It is a common problem; at its root, the existing voice simulation cloning technology is basically based on the mandatory consistency of sound quality and pitch, and there is no cloning to simulate the voice content of the user's intonation, rhythm, accent, etc. that can represent the user's personalized characteristics
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
[0042] The specific technical solution first deploys the database, system software and hardware environment, configures necessary modules of conventional systems such as user registration module and user module, and also includes the following modules, steps and elements.
[0043] Step 1. The voice information processing module collects signals from voice hardware such as the function of the voice sensor in the microphone or the smart device, forms the voice of the user into the system, and converts the collected voice into a voice digital signal Push or wait for the system to perform calculations.
[0044] Step 2: Build a low-pass filter algorithm; in order to minimize the noise in the voice and shield the voice information that is not related to the user's voice content, it is necessary to filter to allow signals lower than the cut-off frequency to pass, and the cut-off is higher than the cut-off A signal of frequency; an embodiment: the Butterworth filter algorithm is 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
Login to View More PUM
Login to View More Abstract
A method for cloning accents and rhymes based on speech training, which discloses the use of typical classified texts representing different intonations in training texts, and obtains units containing accents and rhymes under the same phonetic symbol of the user in several different intonations after training, from which phonemes and phonemes are extracted The segment unit is the content of the sound library, which fully retains the accent and rhyme characteristics of the user's personalized voice. When cloning the simulated voice, the corresponding phoneme group is obtained according to the intonation represented by the text features and the training experience. Align and correct synthesized speech to achieve fidelity cloning of accents and tones.
Description
technical field [0001] The invention relates to the technical field of voice cloning, in particular to a method for cloning accents and tones based on voice training. Background technique [0002] Voice technology is now one of the hottest and most cutting-edge technical fields in scientific and technological research. Voice technology is also a subject of vigorous research by many international technology companies such as Google and Xunfei. There is a LYREBIRD (Lyrebird) website platform abroad that provides voice cloning services. , the user only needs to read aloud through the text of about 30 sentences on the website platform, and the user can get the reading voice of the input text content, and the sound quality of the voice obtained is similar to the user's voice. Now it has become a hot topic, and many users are in major The video website platform releases this voice cloning video, but the LYREBIRD (Lyrebird) website platform is the same as other existing voice mode ...
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 Patents(China)
IPC IPC(8): G10L13/02G10L13/08G10L13/10G10L21/0208G10L21/0216
CPCG10L13/02G10L13/08G10L13/10G10L21/0208G10L21/0216
Inventor 邝翠珊
Owner SHENZHEN SHUZIXINGHE TECH CO LTD