FPGA Design of Speech Compression by Using Discrete Wavelet Transform

This paper presents the Discrete Wavelet Transform (DWT) for real-world speech compression design by using the Field Programmable Gate Array (FPGA) device. Compared with many reports which only focused on the DWT architecture design, this paper gives the speech compression system design detail on how to interface DWT block with SDRAM and audio codec chip. Speech compression was achieved by keeping only the approximation part of the DWT result. The compressed speech signal was read back after upsampling was performed. The resulting compressed speech can be heard clearly with some introduced background noise Future work is to reduce noise.