摘要
针对麦克风阵列语音增强系统对高精度和强实时性数据采集的需要,设计并实现了一种基于FPGA的麦克风阵列数据采集系统。其主要包括数模转换模块、数据接收处理模块和以太网控制模块3部分,实现了对16路语音信号的高质量采样和传输。系统中,将UDP数据报协议用硬件编程语言verilog在FPGA上实现,与基于操作系统的TCP/IP协议族实现UDP协议相比,大大提高了资源利用率。测试结果表明,系统能完成16路语音信号的高精度、高可靠性实时采集和传输,以太网传输速率达2.3MByte/s,满足了麦克风阵列语音增强系统的研究需要。
Due to the requirement of data acquisition with high precision and strong real-time performance for microphone array-based speech enhancement system,this paper designed and implemented a microphone array data acquisition system,which is based on FPGA.This system included three parts:digital-analog conversion module,data receiving and processing module and Ethemet control module,which implemented the functions of sampling and transmission with high quality for 16-channel-voice signals.In this system,UDP datagram protocol was implemented on FPGA by using Verilog,which was a hardware programming language.Compared with the method that implemented UDP datagram protocol by TCP/IP protocol family,which is based on operating system,the method mentioned above greatly improved the resource utilization efficiency.The test results demonstrate that the system could implemented the functions of real-time data sampling and transmission with high precision and high reliability for 16-channel-voice signals,in which case,the Ethernet transmission rate could be 2.3MByte/s,which meets the research requirement of microphone array-based speech enhancement system.
出处
《信号处理》
CSCD
北大核心
2013年第10期1362-1367,共6页
Journal of Signal Processing
基金
国家自然科学基金(61071159)