CNNM Based on Speech Tone Synthesis and its Application in Translators

Aiming at the lack of good support for tone in current speech synthesis, a tone speech synthesis model based on Convolution Neural Network Model (CNNM) is proposed, and it is applied to the development and application of simultaneous translators. A tone speech synthesis model was designed using the Statistical Parameter Speech Synthesis (SSPS) method, which includes acoustic models, coders, and other modules to generate speech with tone. Implement a Bluetooth simultaneous voice translator using Java language and Android platform. Tests have shown that the algorithm can perform speech recognition, language translation, and tone speech synthesis.