基于WaveNet的端到端语音合成方法  

End-to-end speech synthesis based on WaveNet

作　　者：邱泽宇 屈丹[1] 张连海[1] QIU Zeyu;QU Dan;ZHANG Lianhai(College of Information Systems Engineering, PLA Strategic Force Information Engineering University, Zhengzhou Henan 450000, China)

机构地区：[1]战略支援部队信息工程大学信息系统工程学院,郑州450000

出　　处：《计算机应用》2019年第5期1325-1329,共5页journal of Computer Applications

基　　金：国家自然科学基金资助项目(61673395)~~

摘　　要：针对端到端语音合成系统中Griffin-Lim算法恢复相位信息合成语音保真度较低、人工处理痕迹明显的问题,提出了一种基于WaveNet网络架构的端到端语音合成方法。以序列映射Seq2Seq结构为基础,首先将输入文本转化为one-hot向量,然后引入注意力机制获取梅尔声谱图,最后利用WaveNet后端处理网络重构语音信号的相位信息,从而将梅尔频谱特征逆变换为时域波形样本。实验的测试语料为LJSpeech-1.0和THchs-30,针对英语、汉语两个语种进行了实验,实验结果表明平均意见得分(MOS)分别为3.31、3.02,在合成自然度方面优于采用Griffin-Lim算法的端到端语音合成系统以及参数式语音合成系统。Griffin-Lim algorithm is widely used in end-to-end speech synthesis with phase estimation, which always produces obviously artificial speech with low fidelity. Aiming at this problem, a system for end-to-end speech synthesis based on WaveNet network architecture was proposed. Based on Seq2 Seq(Sequence-to-Sequence) structure, firstly the input text was converted into a one-hot vector, then, the attention mechanism was introduced to obtain a Mel spectrogram, finally WaveNet network was used to reconstruct phase information to generate time-domain waveform samples from the Mel spectrogram features. Aiming at English and Chinese, the proposed method achieves a Mean Opinion Score(MOS) of 3.31 on LJSpeech-1.0 corpus and 3.02 on THchs-30 corpus, which outperforms the end-to-end systems based on Griffin-Lim algorithm and parametric systems in terms of naturalness.

关 键 词：语音合成 端到端 Seq2Seq Griffin-Lim算法 WaveNet 

分 类 号：TN912.33[电子电信—通信与信息系统]