Neff D L
J Acoust Soc Am. 1985 Dec;78(6):1966-76. doi: 10.1121/1.392653.
In forward masking, performance may be affected by confusion, that is, by the difficulty of discriminating a suprathreshold signal from the preceding masker. This study investigated confusion effects for forward maskers composed of repeated bursts of a 100-Hz sinusoid followed by sinusoidal signals; such "pulsing" maskers produce confusion when the properties of the signal are identical to those of an individual masker "pulse." The level, frequency, and duration of the signal relative to an individual masker pulse, as well as offset-onset delay, were varied to determine the minimum change necessary to eliminate confusion. For maskers composed of 20-ms pulses, confusion was eliminated by changes in signal level of 5 dB or changes in signal frequency of 30 to 40 Hz. For maskers composed of 10-, 20-, or 40-ms pulses, confusion was eliminated by signal delays of 8 to 16 ms or by signal durations less than half or greater than twice the masker-pulse duration. Results with adaptive procedures designed to measure confusion-free or confusion-determined thresholds suggest that confusion effects can be minimized or avoided by extensive listener training with a procedure in which the signal and masker are not presented at similar intensities.
在前向掩蔽中,性能可能会受到混淆的影响,也就是说,受到区分阈上信号与先前掩蔽声的难度的影响。本研究调查了由100赫兹正弦波的重复脉冲串后跟正弦信号组成的前向掩蔽声的混淆效应;当信号特性与单个掩蔽声“脉冲”的特性相同时,这种“脉冲式”掩蔽声会产生混淆。相对于单个掩蔽声脉冲改变信号的电平、频率和持续时间,以及起始-偏移延迟,以确定消除混淆所需的最小变化。对于由20毫秒脉冲组成的掩蔽声,信号电平变化5分贝或信号频率变化30至40赫兹可消除混淆。对于由10、20或40毫秒脉冲组成的掩蔽声,信号延迟8至16毫秒或信号持续时间小于掩蔽声脉冲持续时间的一半或大于两倍可消除混淆。旨在测量无混淆或由混淆确定的阈值的自适应程序的结果表明,通过使用信号和掩蔽声不以相似强度呈现的程序对听众进行广泛训练,可以将混淆效应降至最低或避免。