MonTTS:完全非自回归的实时、高保真蒙古语语音合成模型

MonTTS:A Real-time and High-fidelity Mongolian TTS Model with Pure Non-autoregressive Mechanism

针对现有基于Tacotron模型的蒙古语语音合成系统存在的两个问题:①合成效率较低;②合成语音保真度较低,该文基于FastSpeech2模型提出了完全非自回归的实时、高保真蒙古语语音合成模型MonTTS。为了提高MonTTS模型合成蒙古语语音的韵律自然度/保真度,根据蒙古语声学特点提出以下三点创新改进:①使用蒙古文音素序列来表征蒙古文发音信息;②提出音素级的声学调节器以学习长时韵律变化;③提出基于蒙古语语音识别和自回归语音合成两种时长对齐方法。同时,该文构建了一个当前最大规模的蒙古语语音合成数据库:MonSpeech。实验结果表明,MonTTS在韵律自然度方面的主观平均意见分数(Mean Opinion Score,MOS)达到4.53,显著优于当前最优的基于Tacotron的蒙古语语音合成基线系统和基线FastSpeech2模型;MonTTS合成实时率达3.63×10^(-3),满足实时高保真合成要求。最后,文中涉及的训练脚本和预训练模型全部开源(https://github.com/ttslr/MonTTS)。

Aiming at real-time and high-fidelity Mongolian Text-to-Speech(TTS)generation,a FastSpeech2 based non-autoregressive Mongolian TTS system(short forMonTTS)is proposed.To improve the overall performance in terms of prosody naturalness and fidelity,MonTTS adopts three novel mechanisms:1)Mongolian phoneme sequence is used to represent the Mongolian pronunciation;2)phoneme-level variance adaptor is employed to learn the long-term prosody information;and 3)two duration aligners,i.e.Mongolian speech recognition and Mongolian autoregressive TTS based models,are used to provide the duration supervise signal.Besides,we build a large-scale Mongolian TTS corpus,named MonSpeech.The experimental results show that the MonTTS outperforms the state-of-the-art Tacotron-based Mongolian TTS and standard FastSpeech2 baseline systems significantly,with real-time rate(RTF)of 3.63×10^(-3) and Mean Opinion Score(MOS)of 4.53(see https://github.com/ttslr/MonTTS).