Cross-language speech synthesis method based on end-to-end tone and emotion migration

The invention discloses a cross-language speech synthesis method based on end-to-end tone and emotion migration, and the method comprises the following steps: S1, collecting and processing Chinese and English speech training data, and extracting required speech features; s2, training a learning network architecture for Chinese and English speech synthesis, wherein the learning network architecture comprises a speaker encoder, a synthesizer and a vocoder; and S3, performing cross-language speech synthesis on the real-time speech input by the speaker by using the trained learning network architecture, so that the synthesized speech can effectively retain the tone and emotion of the speaker. According to the cross-language speech synthesis method provided by the invention, the cross-language speech can be synthesized under the condition that a small amount of speech is given to the speaker, and the tone and emotion of the speaker can be kept in the synthesized speech. 本发明公开了一种基于端到端的音色及情感迁移的跨语言语音合成方法，步骤如下：S1、采集并处