Embedded-optimization-based loudspeaker compensation using a generic Hammerstein loudspeaker model

This paper presents an embedded-optimization-based algorithm for loudspeaker compensation using a generic Hammerstein loudspeaker model, i.e. a cascade of a memoryless nonlinearity and a linear finite impulse response filter. An optimization procedure is embedded into the algorithm to carry out the loudspeaker compensation on a frame-by-frame basis. In order to minimize the perceptible distortion incurred in the loudspeaker, a perceptually meaningful optimization criterion is constructed by using a psychoacoustic model. The resulting per-frame optimization problems are solved efficiently using a gradient optimization method. Objective evaluation experiments show that the proposed loudspeaker compensation algorithm provides a significant audio quality improvement, and this for all considered amplitude levels.