Method and device for optimizing end-to-end speech synthesis model, electronic device

Abstract

The present invention provides an optimization method and device for an end-to-end speech synthesis model, an electronic device and a storage medium, wherein the method includes: according to a first preset rule, performing an optimization on the text input in the end-to-end speech synthesis model. The included phonemes are subjected to the first soft occlusion to generate the second text; the phoneme encoder is used to encode the second text, and the variable information predictor is used to perform prediction processing on the encoded second text to obtain the first text. an output; according to the second preset rule, perform second soft occlusion on the first output; input the first output processed by the second soft occlusion into a preset decoder to decode to obtain a mel spectrum . The optimization method of the end-to-end speech synthesis model provided by the present invention adds soft occlusion to the input of the end-to-end speech synthesis model and the decoder input respectively, thereby increasing data disturbance and improving the robustness of the end-to-end speech synthesis model. .

Description

technical field [0001] The present invention relates to the technical field of speech synthesis, in particular to a method and device for optimizing an end-to-end speech synthesis model, and electronic equipment. Background technique [0002] Generally speaking, as figure 1 As shown, TTS (text-to-speech, speech synthesis) is divided into several parts such as text analysis (eg, text regularization, polyphonic word disambiguation, etc.) module, prosody prediction module, duration model, acoustic model and vocoder . The processed text goes through the prosody prediction module to output the text with prosodic symbols, and then performs word-to-sound conversion and other links. The current mainstream end-to-end model combines the duration model and the acoustic model into one model. The text generates phoneme information through the front end, while the end-to-end model uses the phoneme information as input to generate a Mel spectrum, and then an external vocoder. Convert ac...

AI Technical Summary

Problems solved by technology

However, there are also certain problems in end-to-end synthesis, which are related to the pure black-box structure of the end-to-end model

For example, for a model with an explicit duration module, its decoder is prone to overfitting to wrong information, thus affecting the quality of the final synthesized speech

It can be seen that the current end-to-end model has poor robustness, and it is urgent for those skilled in the art to provide a solution to improve the robustness of the end-to-end model

Images