语音信号响度补偿是数字助听器的关键功能之一.由于患者的听力损失常随信号频率改变,故响度补偿增益应随语音信号频率变化而调整.该文提出针对人耳听觉特征进行子带划分响度补偿的方法,设计了实现响度补偿的完美重构子带滤波器组,并针对老年性听损患者给出子带滤波器组参数.实验与仿真结果显示,该方法很好地补偿了听损患者在不同频段的语音能量损失,使患者对语句的辨识率得到显著提升.
Loudness compensation is a key function of hearing aids. Because the degree of hearing loss varies with frequency, the amount of loudness compensation should be adjusted to adapt the variation. In this paper a new method for sub-band loudness compensation is proposed based on human auditory. Design method for near perfect reconstruction oversampled filter bands is presented. Parameters of the filter bands for senile hearing loss are given. Results of simulation and experiments show that energy loss in different bands of hearing loss patients is well compensated and the recognition rate of sentences is significantly improved.