arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2012.03763cs.CL
收藏

利用先前的声学上下文改进文本到语音合成

Using previous acoustic context to improve Text-to-Speech synthesis

University of Edinburgh(爱丁堡大学)

Pilar Oplustil-Gallegos, Simon King

摘要:

许多语音合成数据集,特别是从有声读物导出的数据集,自然包括话语序列。然而,在训练模型和推理时,这些数据通常被视为个体的、无序的话语。这就抛弃了话语层面之上的重要韵律现象。在本文中,我们利用数据的连续性,使用声学上下文编码器,产生嵌入的前一个话语音频。这是输入到Tacotron 2模型中的解码器。嵌入也用于次要任务,提供额外的监督。我们比较了两个次要任务:预测话语对的排序,以及预测当前话语音频的嵌入。结果表明,连续的话语之间的关系是信息:我们提出的模型显着提高了自然度超过Tacotron 2基线。

英文摘要:

Many speech synthesis datasets, especially those derived from audiobooks, naturally comprise sequences of utterances. Nevertheless, such data are commonly treated as individual, unordered utterances both when training a model and at inference time. This discards important prosodic phenomena above the utterance level. In this paper, we leverage the sequential nature of the data using an acoustic context encoder that produces an embedding of the previous utterance audio. This is input to the decoder in a Tacotron 2 model. The embedding is also used for a secondary task, providing additional supervision. We compare two secondary tasks: predicting the ordering of utterance pairs, and predicting the embedding of the current utterance audio. Results show that the relation between consecutive utterances is informative: our proposed model significantly improves naturalness over a Tacotron 2 baseline.

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。

↑