Equalization of Sound Reproduction System Based on the Human Perception Characteristics

In this paper, we propose an equalization approach for sound reproduction system (e.g., loudspeaker and headphone) based on the human perception mechanism. Traditional FIR-and IIR-based approaches have difficulties in equalizing the low frequencies, the warped filters provide the feasible design for equalization in a non-uniform resolution way and invovle the high complex frequency. To deal with these issues, a novel equalization approach is presented which takes the human perception properties into consideration. Specifically, a weighted-cost function is designed where the weight is calculated based on the perceptual resolution of human beings, which is followed by a modified quais-Newton gradient search algorithm to determine the optimal coefficients of the equalization filter. The performance of the proposed equalization approach is finally compared to those of the state-of-the-art equalizers.