Greene Nathaniel T, Paige Gary D
Biomedical Engineering Department, University of Rochester, Rochester, NY 14642, USA.
Annu Int Conf IEEE Eng Med Biol Soc. 2012;2012:6455-8. doi: 10.1109/EMBC.2012.6347472.
Free-field sound localization experiments generally assume that a loudspeaker can be approximated by a point-source; however, a large loudspeaker may extend beyond the width that two sources can be discriminated. Humans can accurately discriminate sound source locations within a few degrees, thus one might expect localization precision to decrease as a function of sound source diameter, much as precision is lower for localizing the center of a wide, blurry light source. In order to test the degree to which humans differentially localize small and large sound sources, auditory targets were presented using a single 25.4 cm by 10.2 cm elliptical loudspeaker with the primary axis oriented both horizontally and vertically in different sessions. Subjects were seated with their heads fixed by a bite bar in a darkened, echo-attenuating room facing a cylindrical, acoustically transparent screen at a distance of 2 meters. Auditory targets consisted of repeating bursts (5 Hz) of low frequency band-pass noise (0.2 - 1 kHz, 75 dB SPL). Subjects were instructed to quickly and accurately guide a laser pointer mounted on a cylindrical joystick towards targets, presented randomly within a field ± 40° in azimuth by ± 10° in elevation, with oversampled points located every ten degrees along the primary meridians. Localization accuracy and precision (mean and standard deviation of localization error at oversampled locations) were not significantly different between speaker orientations, and were comparable to baseline measurements recorded using a 7.6 cm circular speaker. We conclude that low frequency sound localization performance is not dependent upon the size of the sound source as predicted theoretically, and is well approximated by a point source.
自由场声音定位实验通常假定扬声器可近似为点声源；然而，大型扬声器的宽度可能超出能区分两个声源的范围。人类能够在几度范围内精确区分声源位置，因此可以预期定位精度会随着声源直径的增大而降低，这与定位宽而模糊的光源中心时精度较低的情况类似。为了测试人类对大小不同声源进行差异定位的程度，在不同实验环节中，使用一个25.4厘米×10.2厘米的椭圆形扬声器呈现听觉目标，其主轴分别水平和垂直放置。受试者坐在黑暗、消声的房间里，头部用咬杆固定，面对一个2米远的圆柱形、声学透明的屏幕。听觉目标由重复的低频带通噪声脉冲串（5赫兹）组成（0.2 - 1千赫，75分贝声压级）。受试者被要求迅速准确地将安装在圆柱形操纵杆上的激光指向目标，目标在方位角±40°、仰角±10°的范围内随机呈现，沿主要子午线每隔十度有过采样点。扬声器不同方向之间的定位精度和精确性（过采样位置处定位误差的均值和标准差）没有显著差异，并且与使用7.6厘米圆形扬声器记录的基线测量结果相当。我们得出结论，低频声音定位性能并不像理论预测的那样依赖于声源大小，并且用点声源可以很好地近似。
