English speech synthesis method and device

Abstract

The invention relates to an English speech synthesis method and device. The method includes: obtaining an English sound library of pronunciation standards; constructing a preset English synthesis model according to the English sound library of the pronunciation standard; obtaining a target English sound library recorded by non-professionals; Adjust the preset English synthesis model to obtain the target English synthesis model; use the target English synthesis model to synthesize the English text to be synthesized into English speech. Through the technical scheme of the present invention, a target English synthesis model with high synthetic naturalness can be obtained, so that English speech with high naturalness (that is, normal speech speed, normal intonation, and consistent timbre) can be obtained after being convenient, and even mixed in Chinese and English. It can also accurately synthesize English speech in different scenarios to ensure the consistency of Chinese and English synthesized timbres.

Description

technical field [0001] The present invention relates to the technical field of speech synthesis, in particular to a method and device for English speech synthesis. Background technique [0002] At present, when building an English synthesis model, it is generally necessary to invite professional English pronunciation personnel to record the sound library, so as to use the pronunciation standard sound library to build an English synthesis model, but the cost of hiring a professional English pronunciation personnel to record the sound library is very high, and , the English synthesis model constructed from this is relatively rigid and cannot accurately synthesize speech. For example, in some scenarios that require mixed synthesis of Chinese and English, it is difficult to ensure the synthesis of Chinese and English by using the English synthesis model based on the sound library based on pronunciation standards. Consistent sound. SUMMARY OF THE INVENTION [0003] Embodiments...

AI Technical Summary

Problems solved by technology

[0002] At present, when building an English synthesis model, it is generally necessary to invite professional English pronunciation personnel to record the sound library, so as to use the pronunciation standard sound library to build an English synthesis model, but the cost of professional English pronunciation personnel to record the sound library is very high, and , the English synthesis model constructed from this is relatively rigid and cannot be accurately synthesized. For example, in some scenarios that require mixed synthesis of Chinese and English, it is difficult to guarantee the synthesis of Chinese and English by using the English synthesis model constructed based on the sound library based on pronunciation standards. consistency of tone

Images