arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.09914cs.CLeess.AS
收藏

交替结尾:预测未来文本输入的增量神经TTS韵律改进

Alternate Endings: Improving Prosody for Incremental Neural TTS with Predicted Future Text Input

Université Grenoble Alpes(格勒诺布尔阿尔卑斯大学);CNRS(法国国家科学研究中心);Grenoble INP(格勒诺布尔国立理工学院);GIPSA-lab(GIPSA实验室);LIG(格勒诺布尔信息学实验室);UGA(格勒诺布尔阿尔卑斯大学);G-INP(格勒诺布尔国立理工学院);INRIA(法国国家信息与自动化研究所);NAVER LABS Europe(NAVER欧洲实验室)

Brooke Stephenson, Thomas Hueber, Laurent Girin, Laurent Besacier

摘要:

一个口语单词的韵律是由它周围的语境决定的。在增量文本到语音合成中,其中合成器在其访问完整输入之前产生输出,完整上下文通常是未知的,这可能导致合成语音中的自然度损失。在本文中,我们调查是否使用预测未来的文本可以减轻这种损失。我们比较了下一个未来词的几种测试条件:(a)未知(零词),(b)语言模型预测,(c)随机预测和(d)地面实况。我们测量的韵律功能(音高,能量和持续时间),并发现预测文本提供了显着的改善零字前瞻,但只有轻微的收益随机字前瞻。我们证实了这些结果与感知测试。

英文摘要:

The prosody of a spoken word is determined by its surrounding context. In incremental text-to-speech synthesis, where the synthesizer produces an output before it has access to the complete input, the full context is often unknown which can result in a loss of naturalness in the synthesized speech. In this paper, we investigate whether the use of predicted future text can attenuate this loss. We compare several test conditions of next future word: (a) unknown (zero-word), (b) language model predicted, (c) randomly predicted and (d) ground-truth. We measure the prosodic features (pitch, energy and duration) and find that predicted text provides significant improvements over a zero-word lookahead, but only slight gains over random-word lookahead. We confirm these results with a perceptive test.

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。

↑