Research On Binaural Bone Conduction Digital Equalization Technology

Bone conduction technology is widely used in audio player,intelligent hardware,virtual reality,hearing rehabilitation and other fields.In military,bone conduction earphone can transmit voice clearly and efficiently at low signal-to-noise ratio,so it has high practical application value.Bone-conduction Headphone Eardrum Transfer Function(BHETF)has peak and valley structure,which has a great impact on the sound reproduction of bone-conduction headphones.Therefore,it is necessary to optimize the transfer function(BHETF)from the osseous headphone to the tympanic membrane to obtain a flat frequency response curve.In order to evaluate the performance of digital equalization algorithm in bone conduction earphone sound reproduction,subjective experiments are needed to evaluate the balance results.In this paper,through collating the literature of bone conduction and digital equalization at home and abroad,we measured the actual human head bone conduction system and obtained the transfer function from the bone conduction earphone to the tympanic membrane.The important work of this artical is as follows:Firstly,the equalization algorithm is seldom used in bone conduction system.Firstly,an inverse filtering algorithm based on least squares(LS_FIR)is used to equalize the transfer function from bone conduction earphone to tympanic membrane.The algorithm solves the filter coefficients iteratively to obtain the inverse system of bone conduction system.The simulation results show that the inverse filtering algorithm can better approximate the inverse system.When the original bone conduction transfer function is cascaded with the inverse system,the output is a spike pulse,which realizes the system's undistorted transmission,completes the digital equalization,and is conducive to the further realization of bone conduction sound playback.Secondly,aiming at the difference of human auditory perception at high and low frequencies and the nonlinearity of frequency resolution,an equalization algorithm based on Kautz filter is further simulated.By adjusting the position of poles,the weights of different frequency bands are further adjusted,and the target response is better approached by increasing and decreasing the number of poles.The simulation results show that by increasing the order of the filter,the Kautz filtering algorithm can more precisely approximate the target response and achieve the digital equalization of bone conduction.Finally,an experimental platform of bone conduction system is built.The effectiveness of the digital equalization algorithm in bone conduction sound reproduction system is further verified by the subjective experimental listening of the experimental subjects.Subjective evaluation was carried out by using the dual comparison method.After three groups of experiments(male,female and musical)were compared and balanced,the preferences of the experimenters for male,female and musical sounds after different balancing treatments were evaluated.The subjective audiometric results showed that after digital balancing treatment of bone conduction system,the reproduced timbre and clarity of the three kinds of audio(male,female and musical sounds)were different.The same level of improvement.