基于FPGA的语音预处理系统  

Audio signal pre⁃processing system based on FPGA

作　　者：张皓然 张涛 万书芹 蒋颖丹 苏小波 ZHANG Haoran;ZHANG Tao;WAN Shuqin;JIANG Yingdan;SU Xiaobo(China Key System and Integrated Circuit Corporation,Wuxi 214000,China)

机构地区：[1]中科芯集成电路有限公司,江苏无锡214000

出　　处：《电子设计工程》2023年第14期182-186,共5页Electronic Design Engineering

摘　　要：文中介绍了一种语音预处理系统的设计思路以及在FPGA平台下的实现方法。系统由模拟前端电路(Analog Front End,AFE)、高速Flash存储电路和PC端的数据处理模块构成。该系统主要功能是对语音信号实行预处理,通过AFE中的模拟回环功能获得预处理过的语音模拟信号,同时可向PC端发送未经处理的原始数据,并结合LMS自适应滤波算法可滤除杂波,得到较为纯净的语音数字信号。相比较其他语音处理方式,该系统既可将原始语音信号处理成低噪的数字信号,也可通过内部回环模式得到底噪且增益可放大的模拟信号。A kind of design about audio signal pre⁃processing system and the design thought are introduced,along with it’s realization based on FPGA.The system is consisted of several modules including Analog Front End(AFE),high⁃speed Flash circuit and data processing module on PC.The system’s main function is pre⁃processing audio signal.It can acquire the pre⁃processed analog audio data using AFE’s loop mode,and in the same time,send the raw data unprocessed to PC.Combined with the LMS self⁃adapted filter algorithm,these raw data could be filtered into purer digital audio data.Compared with other audio processing methods,the system could process the raw signal into digital data with lower noise and purer gain⁃adjustable analog data by AFE’s loop mode.

关 键 词：模拟前端电路 LMS自适应滤波算法 语音预处理系统 FPGA 

分 类 号：TN98[电子电信—信息与通信工程]