Voice synthesis system and method for simulating singing of user

Abstract

The invention discloses a voice synthesis system and method for simulating the singing of a user and belongs to the voice simulation technological field. According to the principles of the voice synthesis system and method, the normal speaking voices of the user which are inputted from the outside are acquired and converted into a pronunciation text, and a phone sequence is formed according to thepronunciation text; processing is performed according to the phone sequence, so that the original time section information of basic phonons is obtained; the original time section information of the basic phonons is adjusted to beat time section information corresponding to music score information; the original time section information is compared with the beat time section information according to the basic phonons, the voice synthesis parameters of the basic phonons are adjusted correspondingly according to a judgment result; and synthesis voices for simulating the singing of the user are generated according to the adjusted voice synthesis parameters of the basic phonons and the pronunciation text, and the synthesis voices are outputted. With the voice synthesis system and method provided by the technical schemes of the invention adopted, the singing of the user can be simulated at the absence of modeling; the efficiency of voice simulation is improved; an effect near real-time feedback is achieved; the timbre information of the user is retained; details are rich; an effect is vivid; and user experience can be improved.

Description

technical field [0001] The invention relates to the technical field of speech simulation, in particular to a speech synthesis system and method for simulating a user's singing voice. Background technique [0002] With the continuous development of speech synthesis technology, more and more application software began to use speech synthesis technology to simulate the content of human speech, such as using speech synthesis technology to simulate the content of human speech for the purpose of "learning the tongue", or using speech synthesis Technology to simulate human singing and other voices that are different from ordinary speaking scenes. [0003] Specifically, in the prior art, in the scene of simulating a user singing, the usual practice is to use the inherent timbre in the speech synthesis database to generate a song, and it is necessary to model the user’s timbre information, and use the timbre transformation technology in the inherent timbre of the song. Based on the ...

AI Technical Summary

Problems solved by technology

[0004] 1. It is necessary to model the user's timbre information in advance, which makes the process of speech synthesis more complicated;

[0005] 2. It is necessary to realize the transformation of the user's voice according to the constructed model, so as to obtain the synthesized singing voice. The processing speed is slow, the processing efficiency is low, and the effect of real-time processing and output singing voice cannot be realized;

[0006] 3. Using the inherent timbre information in the synthesizer database to realize speech synthesis and simulation cannot retain the user's own timbre characteristics, making the result of voice simulation relatively dull, and the simulation effect does not match the actual timbre

Images