Ishigami H, Suzuki T, Isojima G
Department of Otorhinolaryngology, Aichi Medical University, Japan.
Auris Nasus Larynx. 1995;22(1):1-6. doi: 10.1016/s0385-8146(12)80175-1.
In this experiment, the critical band (CB)-widths were measured using two narrow-band noises (NBNs) with steep cut-off slopes as a masker, varying their spectral level and center-frequency. The following results were obtained: (i) The CB-widths at all center-frequencies were narrower than the classical CB-widths estimated by Zwicker et al [J Acoust Soc Am 29:548-557 (1957)], and decreased continuously toward the lower frequency at the center-frequencies below 500 Hz, though Zwicker's data were constant in the same range. These results barely coincided with the frequency dependence of the equivalent rectangular bandwidth (ERB) which was estimated from the shape of the auditory filter measured directly. (ii) Level dependence of masking noise in CB was constant beyond 20 dB of the masked threshold. This depends on the use of NBN as a masker. (iii) The CB was influenced by distortion products and off-frequency listening. Accordingly, the bandwidths might differ with the experimental method.
在本实验中，使用具有陡峭截止斜率的两个窄带噪声（NBN）作为掩蔽声来测量临界带宽（CB）宽度，改变它们的频谱水平和中心频率。得到了以下结果：（i）所有中心频率处的CB宽度均比Zwicker等人[《美国声学学会杂志》29:548 - 557（1957）]估计的经典CB宽度窄，并且在中心频率低于500 Hz时，朝着较低频率持续减小，尽管Zwicker的数据在相同范围内是恒定的。这些结果几乎与根据直接测量的听觉滤波器形状估计的等效矩形带宽（ERB）的频率依赖性不一致。（ii）在掩蔽阈值超过20 dB时，CB中掩蔽噪声的电平依赖性是恒定的。这取决于使用NBN作为掩蔽声。（iii）CB受失真产物和偏离频率聆听的影响。因此，带宽可能因实验方法而异。
