Development of an Open Speech Signal Processing Platform

开放式语音信号处理平台的开发

• 批准号: 9548311
• 负责人: Ross Kenneth Snider
• 金额: $ 64.85万
• 依托单位: FLAT EARTH, INC.
• 依托单位国家: 美国
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-07-01 至 2021-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/9548311
• 关键词: Acoustics; Algorithms; Auditory; Back; Code; Communities; Custom; Development; Devices; Digital Signal Processing; Family; Feedback; Future; Hearing; Hearing Aids; Human; Libraries; Neurosciences; Noise; Performance; Phase; Physiologic pulse; Play; Process; Research; Research Personnel; Research Support; Resources; Running; Science; Signal Transduction; Software Tools; Speech; Standardization; Stream; System; Testing; Textiles; Universities; Work; analog; computer framework; cost; design; next generation; programs; signal processing; software development; sound; speech processing; web site

This proposal will create an open speech signal processing platform. The open computational framework will be used for developing algorithms that will help accelerate acoustic and hearing science research. It will permit researchers to more rapidly and easily test new algorithms with greater control, performance, and ease. Several example systems will be created to show that the platform can be adapted to a wide range of applications. Example designs will include a hearing aid processor, a cochlear processor, and a general purpose sound processor that can connect to a wide range of audio I/O interfaces. The open framework will leverage a new class of computational devices that represent the future for embedded systems needing significant digital signal processing (DSP) capability. These devices, known as system-on-chip field programmable gate arrays (SoC FPGAs), enable algorithms to run with deterministic and extremely low-latency abilities that are not possible with conventional CPUs. The open platform will enable the research community to create the next generation hearing aids and sound processors.

该提案将创建一个开放的语音信号处理平台。开放的 计算框架将用于开发有助于加速的算法 声学和听力科学研究。它将允许研究人员更快、更 轻松测试新算法，具有更强的控制力、性能和易用性。一些 将创建示例系统以表明该平台可以适应广泛的环境 应用范围。示例设计包括助听器处理器、人工耳蜗 处理器和通用声音处理器，可以连接到各种 音频输入/输出接口。开放框架将利用一类新的计算 代表需要大量数字化的嵌入式系统的未来的设备 信号处理（DSP）能力。这些器件被称为片上系统领域 可编程门阵列 (SoC FPGA)，使算法能够确定性地运行 以及传统 CPU 无法实现的极低延迟能力。这 开放平台将使研究界能够创建下一代听力 辅助设备和声音处理器。