Time Domain Noise Filtering to Enhance Speech in Noise

时域噪声过滤可增强噪声中的语音

• 批准号: 6523436
• 负责人: DAVID L FRANKLIN
• 金额: $ 37.39万
• 依托单位: AUDIOLOGICAL ENGINEERING CORPORATION
• 依托单位国家: 美国
• 财政年份: 1997
• 资助国家: 美国
• 起止时间: 1997-09-01 至 2004-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/6523436
• 关键词: behavioral /social science research tag; biological signal transduction; biomedical equipment development; clinical biomedical equipment; clinical research; cochlear implants; computer simulation; deafness; electrotactile communication; hearing aids; human subject; microprocessor /microchip; noise; questionnaires; speech recognition

DESCRIPTION (provided by the applicant): The Phase I study demonstrated that combining a directional microphone with a new noise processing method resulted in SIN improvements on the order of 10 dB as compared to an omni-directional microphone and no processing used in the same setting. Consequences for hearing impaired subjects with losses ranging from moderate to severe was to restore their performances to be statistically equivalent to that measured in quiet. Even more powerful was that for three out of four of the subjects evaluated using the best realization of the method, performances were numerically equivalent in both noise and quiet where the original SIN was 0 dB. The processing method is called Time Domain Noise Filtering (TDNF) and has been granted a US Patent. During this Phase II work we will implement the method in digital technology using a Digital Signal Processor (DSP) to make it compatible with digital hearing aid technology. At the same time, the greater flexibility of DSP technology, as compared to the analog methods used in the Phase I work, will be exploited to make the method adaptable to changing noise levels and to explore the dynamics of the method that might improve its performance even more. We will study a range of hearing impaired subjects including cochlear implant users. We have contacted both a hearing aid company and a cochlear implant company regarding use of the TDNF method in their devices. Letters from the companies are included in this application. PROPOSED COMMERCIAL APPLICATION: Not Available

描述（由申请人提供）：第一阶段研究表明 将定向麦克风与新的噪声处理方法相结合 与Omni方向相比 麦克风和在同一环境中没有使用的处理。听力的后果 受试者受损，损失的损失范围从中度到重度恢复 他们的表现在统计学上等同于以安静的方式测量的表现。 更强大的是，在评估的四分之三的主题中，有三分之一 使用该方法的最佳实现，性能是数值的 原始罪是0 dB的噪音和安静的等效。这 处理方法称为时域噪声滤波（TDNF），已经是 授予美国专利。在这二阶工作中，我们将在 使用数字信号处理器（DSP）使其兼容的数字技术 使用数字助听器技术。同时，更大的灵活性 与I阶段工作中使用的模拟方法相比 将被利用以使该方法适应不断变化的噪声水平和 探索该方法的动态，甚至可以提高其性能 更多的。我们将研究一系列听力受损的受试者 植入用户。我们已经联系了助听器公司和人工耳蜗 植入公司在其设备中使用TDNF方法。来自 这些公司包括在此应用程序中。 拟议的商业应用： 无法使用