Method and device for building personalized singing style spectrum synthesis model

Abstract

The invention discloses a method and a device for building a personalized singing style spectrum synthesis model. The method comprises the following steps: training a reading style spectrum synthesis model and a singing style spectrum synthesis model of a source speaker; acquiring reading style data of a target speaker; determining the tone shift relation between the source speaker and the target speaker according to the reading style spectrum synthesis model of the source speaker and the reading style data of the target speaker; and performing model transformation on the singing style spectrum synthesis model of the source speaker according to the tone shift relation to obtain a singing style spectrum synthesis model of the target speaker. As the singing style synthesis model of the target speaker can be obtained through a small amount of reading style data of the target speaker, the difficulty in building a singing synthesis model of a common user is greatly lowered, and the personalized singing synthesis effect can be further effectively improved.

Description

technical field [0001] The invention relates to the technical field of speech synthesis, in particular to a method and device for constructing a personalized singing style spectrum synthesis model. Background technique [0002] In recent years, singing synthesis systems have been extensively researched and applied, among which parametric synthesis systems based on Hidden Markov Model (HMM) and waveform splicing synthesis systems have reached practical standards. Compared with the synthesis system based on waveform splicing, the parametric synthesis method based on HMM requires less training data and a smaller system scale, so it has better practical value. [0003] The singing synthesis system based on parametric synthesis first determines the duration parameter sequence, fundamental frequency parameter sequence, and spectrum parameter sequence of each basic synthesis unit (such as syllable, phoneme, etc.), and then uses the parametric synthesizer to obtain continuous singin...

AI Technical Summary

Problems solved by technology

[0006] The present invention provides a method and device for constructing a personalized singing style spectrum synthesis model to solve the problems of insufficient training singing data and inaccurate labeling in the construction of a personalized singing style spectrum synthesis model, and to meet the application of personalized singing synthesis need

Images