...
首页> 外文期刊>電子情報通信学会技術研究報告. 音声. Speech >Automatic estimation of pronunciation habits using a single word utterance with HMM
【24h】

Automatic estimation of pronunciation habits using a single word utterance with HMM

机译:使用单词话语与HMM的单词话语自动估算发音习惯

获取原文
获取原文并翻译 | 示例
   

获取外文期刊封面封底 >>

       

摘要

Because the word accent of the Japanese language is based on controlling pitch, it is reported that Japanese learners tend to generate the English word stress mainly by controlling F{sub}0. So in our previous study, we have been focusing on this habits. And we developed a stressed syllable detector, a pronunciation habit estimator and visualizer of the habit. However, the methods in our previous study need several dozens of pronunciations about word or large amount of computations. In this paper, we investigated methods which required only a single word utterance and a little computation cost. Results showed that we can observe the similar results from native utterances, to that obtained in our previous study. But results from Japanese utterances are not good, and we require further examination.
机译:由于日语的重点基于控制音调,因此报告日本学习者倾向于主要通过控制F {sub} 0来产生英语词应力。 所以在我们以前的一项研究中,我们一直专注于这种习惯。 我们开发了一个压力的音节探测器,习惯的发音习惯估计和可视化器。 但是,我们以前的研究中的方法需要几十个关于单词或大量计算的发音。 在本文中,我们调查了只需要单词话语的方法和一点计算成本。 结果表明,我们可以观察到原生话语的类似结果,以获得之前的研究。 但是日本话语的结果不好,我们需要进一步检查。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号