Pralong D
University Laboratory of Physiology, Oxford, United Kingdom.
J Acoust Soc Am. 1996 Dec;100(6):3785-93. doi: 10.1121/1.417337.
The auditory system localizes a sound source on the basis of binaural cues as well as monaural cues provided by the outer ear. These cues are described by the head-related transfer functions (HRTFs). The individual variation in the frequency and amplitude of perceptually significant features in the human HRTFs raises the issue of whether a given listener can obtain adequate localization cues from stimuli based on another listener's HRTFs. This question can be addressed in "virtual auditory space," using headphone stimuli which are convolved with HRTFs and the inverse of the headphone transfer function (HpTF). Using an in-ear recording system we have measured HpTFs (Sennheiser 250 Linear) for both ears of ten subjects and found that they also display significant individual differences in the 4-10 kHz region. Therefore, the proper reconstruction of both individualized and nonindividualized HRTFs requires that the listener's HpTF be deconvolved from the stimulus. We have verified this acoustically using the in-ear recording system for both individualized and nonindividualized HRTFs. Finally, we show that the use of nonindividualized HpTFs creates considerable distortions in the 4-10 kHz range, which could partly account for the increased incidence of mislocalizations reported in previous virtual auditory space localization experiments.
听觉系统基于双耳线索以及外耳提供的单耳线索来对声源进行定位。这些线索由头部相关传递函数（HRTF）来描述。人类HRTF中感知上显著特征的频率和幅度的个体差异，引发了一个问题：给定的听众能否从基于另一位听众的HRTF的刺激中获得足够的定位线索。这个问题可以在“虚拟听觉空间”中解决，使用与HRTF以及耳机传递函数（HpTF）的逆进行卷积的耳机刺激。使用入耳式录音系统，我们测量了10名受试者双耳的HpTF（森海塞尔250线性），发现它们在4 - 10千赫兹区域也表现出显著的个体差异。因此，对个性化和非个性化HRTF的正确重建都要求从刺激中对听众的HpTF进行去卷积。我们已经使用入耳式录音系统对个性化和非个性化HRTF在声学上验证了这一点。最后，我们表明使用非个性化的HpTF在4 - 10千赫兹范围内会产生相当大的失真，这可能部分解释了在先前的虚拟听觉空间定位实验中报告的定位错误发生率增加的原因。
