A Neural Network Approach to the Adaptive Correction of Loudspeaker Nonlinearities
An adaptive loudspeaker correction scheme for subwoofers is presented that is based upon a four-layer neural network algorithm. Compensation for both dominant time-invariant nonlinearities and time-varying changes in cone mass and voice-coil resistance is achieved. Control data for the correction algorithm is derived using a novel method based upon the knowledge of the current and voltage monitored at the drive unit's terminals.
This paper costs $20 for non-members, $5 for AES members and is free for E-Library subscribers.