Prosody enhanced TTS for Dysarthric Speakers

Prosody 增强构音障碍人士的 TTS

• 批准号: 7670599
• 负责人: THOMAS JAKOBS
• 金额: $ 12.95万
• 依托单位: INVOTEK, INC.
• 依托单位国家: 美国
• 财政年份: 2009
• 资助国家: 美国
• 起止时间: 2009-05-12 至 2011-04-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7670599
• 关键词: Acoustics; Communication; Communication Methods; Complex; Computer software; Cueing for speech; Cues; Custom; Dysarthria; Effectiveness; Ensure; Feasibility Studies; Generations; Loudness; Measures; Orthography; Output; Partner Communications; Persons; Process; Research; Savings; Speech; Speech Synthesizers; System; Techniques; Technology; Text; Vendor; Writing; alternative communication; base; computer generated; design; improved; programs; public health relevance; software development; speech recognition; tool

DESCRIPTION (provided by applicant): This feasibility project investigates whether people with complex communication needs can control the prosody of a text-to-speech synthesizer within a unique speech-aware word prediction program. The research team will design a second-generation speech recognition system for dysarthric speakers that includes the ability to measure whether the speaker emphasizes a spoken word. Words that are emphasized by the speaker will then be emphasized by a text-to-speech synthesizer. We will incorporate this speech recognition technology into a "prosody aware" word prediction program so that people with dysarthria can use their speech, in combination with other techniques, to control the prosody of computer generated speech. People with complex communication needs, their partners, and third party listeners will evaluate this system to determine if prosody control improves the communication process. PUBLIC HEALTH RELEVANCE: The aim of this research is to increase the communication effectiveness of persons who have complex communication needs. We propose to do this by developing technology that enables them to control the prosody of a commercial text-to-speech synthesizer using their dysarthric speech and written cues.

描述（由申请人提供）：这个可行性项目调查了具有复杂沟通需求的人是否可以控制独特的语音感知单词预测程序中文本到语音合成器的韵律。研究团队将针对违反言语的人设计第二代语音识别系统，其中包括衡量说话者是否强调口语的能力。然后，讲话者强调的单词将由文本到语音合成器强调。我们将将这种语音识别技术纳入“韵律意识到”单词预测计划中，以便具有构想障碍的人可以将其语音与其他技术结合使用，以控制计算机生成的语音的韵律。有复杂沟通需求的人，他们的合作伙伴和第三方听众将评估该系统，以确定Prosody Control是否可以改善沟通过程。 公共卫生相关性：这项研究的目的是提高具有复杂沟通需求的人的沟通效率。我们建议通过开发技术来实现这一目标，以使他们能够使用质心语音和书面提示来控制商业文本到语音合成器的韵律。