LEARNABLE NONLINEAR COMPRESSION FOR ROBUST SPEAKER VERIFICATION

In this study, we focus on nonlinear compression methods in spectral features for speaker verification based on deep neural network. We consider different kinds of channel-dependent (CD) nonlinear compression methods optimized in a data-driven manner. We also propose multi-regime (MR) design on the nonlinearities, aiming of improving robustness. Our methods are based on power nonlinearities and dynamic range compression (DRC). Results on VoxCeleb1 and VoxMovies data demonstrate improvements brought by proposed compression methods over both the commonly-used logarithm and their static counterparts, especially for ones based on power function. While CD generalization improves performance on VoxCeleb1, MR provides more robustness on VoxMovies, with a maximum relative equal error rate reduction of 21.6%.