The FPGA Implementation Of The Speech Signal Processing Process Of The Distributed Microphone Array

At present,with the rapid development of voice signal processing technology,the traditional signal processing devices have been unable to meet the market demand.To improve the shortcomings of DSP and other processing devices,such as not supporting parallel operation,few register resources and poor real-time processing,this paper proposes a speech signal processing scheme based on FPGA devices.This paper mainly studies the FPGA implementation of speech signal processing from the three parts of theoretical research,design implementation and algorithm simulation analysis.Based on the basic theory of microphone array,the software modeling and hardware design of distributed microphone array are carried out.In terms of hardware,It includes the signal conditioning circuit with the functions of signal amplification and prefiltering,the AD chip with the functions of multi-channel signal synchronous acquisition and analog-digital conversion,and the power supply circuit of the acquisition system designs.In terms of software program,two parts including hardware circuit control and voice enhancement algorithm implementation are designed.Hardware control includes clock divider,inter-board data transmission protocol,etc.Speech enhancement algorithm: multi-channel adaptive beamforming algorithm,power spectrum subtraction in frequency domain,FIR filter design.In this paper,MATLAB and ModelSim simulation software are combined with Quartus ? for joint debugging simulation and results analyzed.The experimental test results show that the speech enhancement algorithm based on FPGA can improve the SNR optimization of the collected original speech by about 2.4dB,and the processing and transmission rate reaches 9600bits/s.It shows that the design of the voice processing system is basically successful,and some aspects still need to be optimized.At the same time,it also reflects the advantages of FPGA,such as sufficient resources,parallel processing,stronger floating point computing ability,good real-time performance of the signal processing system,and better system reusability.