Separation of Electric Guitar Sound based on stacked Recurrent Neural Network
Wen-Hsing Lai, Siou-Lin Wang · 2020
Most audio source separation-related studies have been speech/singing separation, and those on musical instruments separation have been relatively scant. In this research, a stacked recurrent neural network was used to train a model to extract electric guitar sound from mixtures of electric guitar and other musical instruments. Objective assessment scores indicate that the proposed method can extract the electric guitar sounds from the musical mixtures successfully.