arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2011.02252eess.AScs.CLcs.SD
收藏

神经文本转语音的韵律表征学习和上下文采样

Prosodic Representation Learning and Contextual Sampling for Neural Text-to-Speech

Amazon Research(亚马逊研究院)

Sri Karlapati, Ammar Abbas, Zack Hodari, Alexis Moinet, Arnaud Joly, Penny Karanasou, Thomas Drugman

摘要:

在本文中,我们介绍了Kathaka,一个新的两阶段训练过程的神经语音合成与上下文适当的韵律训练模型。在第一阶段,我们从训练过程中可用的梅尔频谱图中学习句子级别的韵律分布。在第二阶段,我们提出了一种新的方法,从这个学习的韵律分布使用文本中的上下文信息进行采样。为此,我们在文本上使用BERT,在从文本中提取的解析树上使用图形注意力网络。与录音相比,我们显示出$13.2\%$在强基线上自然度的统计学显著相对改善。我们还对我们的采样技术的变化进行了消融研究,并显示在每种情况下,在基线上有统计学意义的改善。

英文摘要:

In this paper, we introduce Kathaka, a model trained with a novel two-stage training process for neural speech synthesis with contextually appropriate prosody. In Stage I, we learn a prosodic distribution at the sentence level from mel-spectrograms available during training. In Stage II, we propose a novel method to sample from this learnt prosodic distribution using the contextual information available in text. To do this, we use BERT on text, and graph-attention networks on parse trees extracted from text. We show a statistically significant relative improvement of $13.2\%$ in naturalness over a strong baseline when compared to recordings. We also conduct an ablation study on variations of our sampling technique, and show a statistically significant improvement over the baseline in each case.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑