Adaptive correction of loudspeaker using recurrent neural network

摘要

An audio system is described that corrects for linear and nonlinear distortions. The system can include a physical loudspeaker system responsive to an audio input signal, an adaptive circuit, e.g., with a recurrent neural network, to correct for non-linear distortions from the loudspeaker.