Examining the Perceptual Effect of Alternative Objective Functions for Deep Learning Based Music Source Separation

摘要

In this study, we examine the effect of various objective functions used to optimize the recently proposed deep learning architecture for singing voice separation MaD - Masker and Denoiser. The parameters of the MaD architecture are optimized using an objective function that contains a reconstruction criterion between predicted and true magnitude spectra of the singing voice, and a regularization term. We examine various reconstruction criteria such as the generalized Kullback-Leibler, mean squared error, and noise to mask ratio. We also explore recently proposed, for optimizing MaD, regularization terms such as sparsity and TwinNetwork regularization. Results from both objective assessment and listening tests suggest that the TwinNetwork regularization results in improved singing voice separation quality.