Moore B C, Sek A
Department of Experimental Psychology, University of Cambridge, England.
J Acoust Soc Am. 1994 Aug;96(2 Pt 1):741-51. doi: 10.1121/1.410312.
This article is concerned with the mechanisms underlying the detection of amplitude modulation (AM), frequency modulation (FM), and mixed modulation (MM), i.e., simultaneously occurring AM and FM. In a previous study [B. C. J. Moore and A. Sek, J. Acoust. Soc. Am. 92, 3119-3131 (1992)], psychometric functions were measured for the detection of AM alone and FM alone, using a 10-Hz modulation rate and a 1-kHz carrier frequency. Detectability was then measured for combined AM and FM, with modulation depths selected so that each type of modulation would be equally detectable if presented alone. The detectability of the combined AM and FM was better than would be predicted if the two types of modulation were coded completely independently. Significant effects of relative modulator phase were found when detectability was relatively high, but these effects were not correctly predicted by either of two excitation-pattern models considered. The first experiment reported here was similar to the earlier experiment, but performance was compared for carrier frequencies of 1 and 6 kHz; at the latter frequency, neural synchrony to the stimulus fine structure (phase locking) does not occur. The results at both carrier frequencies were similar to those of our earlier experiment, suggesting that the presence or absence of phase-locking information plays little role in the detection of MM. The second experiment was again similar, but bands of noise were used to mask selectively either the upper or lower side of the excitation pattern of the modulated carrier. The phase effects in this case were in the direction predicted by excitation pattern models. The overall pattern of the results could be predicted reasonably well using a multichannel excitation pattern model based on the assumption that listeners use an unweighted sum of decision variables across all suprathreshold channels with a positive signal-to-noise ratio.
本文关注幅度调制（AM）、频率调制（FM）以及混合调制（MM，即同时出现的AM和FM）检测背后的机制。在之前的一项研究[B. C. J. 摩尔和A. 塞克，《美国声学学会杂志》92, 3119 - 3131 (1992)]中，使用10赫兹的调制率和1千赫兹的载波频率，测量了单独检测AM和单独检测FM的心理测量函数。然后测量了AM和FM组合的可检测性，调制深度的选择使得如果单独呈现每种调制类型，其可检测性相同。AM和FM组合的可检测性比两种调制类型完全独立编码时的预测要好。当可检测性相对较高时，发现了相对调制器相位的显著影响，但所考虑的两种激发模式模型均未正确预测这些影响。此处报道的第一个实验与早期实验类似，但比较了1千赫兹和6千赫兹载波频率下的性能；在后者频率下，不会出现与刺激精细结构的神经同步（锁相）。两个载波频率下的结果与我们早期实验的结果相似，表明锁相信息的有无在MM检测中作用不大。第二个实验同样类似，但使用噪声带选择性地掩蔽调制载波激发模式的上侧或下侧。在这种情况下的相位效应与激发模式模型预测的方向一致。基于听众对所有具有正信噪比的阈上通道的决策变量进行未加权求和这一假设，使用多通道激发模式模型可以相当好地预测结果的总体模式。
