Study on 3-dimensional spatial hearing and its control for highly advanced acoustic communication

在本财年，详细分析了上一财年提出的听觉显示的声音再现精度。实验结果表明，该方法可以达到-20分贝左右的误差水平。去年10月在熊本召开的国际声学会议上也提到了这一点。还研究了房间传递函数(RTF)和头部相关传递函数(HRTF)的有效编码方法。Haneda等人提出的传递函数的共极点零点模型。被应用于我们开发的模型中使用的RTF和HRTF集合。结果表明，与传统的FIR滤波器组相比，该方法能够以更少的参数有效地编码这些传递函数集所包含的信息，由于我们提出的声场再现系统是基于声音传输系统的直接建模，当考虑到听者头部运动引起的系统动态变化时，这种方法可能是冗余的。因此，本课题重点研究了听者头部运动对定向听力的影响。听力实验是利用实时系统根据磁传感器检测到的听者头部运动引起的声源和耳朵的相对位置来改变与声源信号卷积的传递函数。结果表明，听者的头部运动是非常有效的，在高精度的听觉显示中，考虑到动态变化是重要的。

In this fiscal year, sound reproduction accuracy of the auditory display proposed last fiscal year was analyzed in detail. The results showed that the error level of around-20 dB can be achieved. This was represented in an international conference on acoustics held last October in Kumamoto.The Effective coding methods of Room Transfer Functions (RTFs) and Head-Related Transfer Functions (HRTFs) were also investigated. Common-pole zero modeling of transfer functions proposed by Haneda et al. was applied to the set of RTFs and HRTFs used in our developed model. As a result, it was shown that this method has a potential to code the information involved in these set of transfer functions effectively with less parameters than those modeled by using the conventional FIR filter banks.Since our proposed sound field reproduction system is based on the straightforward modeling of the sound transmission system, this may be redundant when the dynamic change of the system due to the listener's head movement is taken account. Hence the influence of the listener's head movement in directional hearing was focused in our project. The hearing experiment was executed with the real time system changing the transfer functions convolved with the sound source signal according to the relative position of the sound source and the listener's ears due to his/her head movement sensed by a magnetic sensor. The results showed that the listener's head movement is significantly effective and taking into account the dynamic change is important in the highly accurate auditory display.