FPGA implementation of multi-band spectral subtraction method for speech enhancement
Mohammed Bahoura · 2017
This paper proposes a hardware architecture of the multi-band spectral subtraction method for real-time speech enhancement. The proposed hardware architecture has been implemented on field programmable gate array (FPGA) device using Xilinx system generator (XSG) and Nexys-4 development board. Multi-band approach is based on the fact the whole speech spectrum does not be affected uniformly by the colored noise. Speech enhancement performances obtained by the hardware architecture are compared to those obtained by MATLAB simulation. The resource utilization and the maximum operating frequency are reported for an Artix-7 FPGA chip.