面向助听器的语音信号处理算法开发平台设计

Design of Speech Processing Platform for Hearing Aids

综合考虑助听器体积小,功耗低等特点,设计了一种基于嵌入式系统的实时语音处理算法的开发平台。该平台的硬件核心主要由Cortex A8嵌入式处理芯片和FPGA芯片综合构成,共包含4个关键模块:音频输入模块、内部时钟模块、FPGA控制模块和信号处理模块。为了提高系统的处理效率,系统设计了基于FPGA的多路语音处理转换模块。为验证平台性能,设计并实现了基于维纳滤波的助听器语音增强算法,并进行了主观测试,实验效果良好。

Considering the characteristic of the hearing-aid with small size,and low power consumption,one development platform based on the real-time speech processing algorithm of embedded systems is designed. The hardware core of the platform is mainly composed of Cortex A8 chip and FPGA chip,which contains four key modules： audio input module,internal clock module,FPGA control module and signal processing module. In order to improve the system processing efficiency,the multi-channel speech processing conversion module based on FPGA is designed.The speech enhancement algorithm of the hearing-aid based on Wiener filtering is designed and realized to verify the platform performance and a subjective test is conducted with good experiment results.