Get rid of the noise! : A speech enhancement technology using microphone arrays

• Main Author: Cai, Zihui
• Other Authors: Andy Khong W H
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/149292
• Institution: Nanyang Technological University

This report is to experiment and compare the different performances of two different deep learning networks on denoising speech signals using MATLAB. Through this experiment, Short-time Fourier Transform is used to process the speech audio signals to analyze the clean and noise signals, then generate the targets to make predictions for noise signals, and finally remove the noise from speech signals. The two networks used are Fully Connected Network (FCN) and Convolutional Neural Network (CNN). The parameters compared are the training time duration, the total number of weights, the quality of the denoised signal compared to the clean signal.