Audio Denoising Using Machine Learning
摘要
Audio denoising is the process of removing the unwanted background noise from audio signals. There are various applications of audio denoising like speech recognition and music production. In the case of U-net, the model takes a downsample as input and produces denoised signals as output. For the process of Audio Denoising, we will use the datasets LibriSpeech and ESC-50. U-net is a powerful neural network architecture that has been used for various signal-processing tasks, which is an ASR corpus consisting of public-domain audiobooks.