Figure 4

<p> <b>Top</b>: Examples of periodic and noisy sine wave vibrations used in Experiment 4. Noisy vibrations were created by adding a positive or negative random interval (from a Gaussian distribution with mean = 0) to the length of each cycle of the regular sine wave (from 0° to 360°). For example, for a 40 Hz noisy vibration, the length of each cycle varied randomly around a mean of 25 msec. Note, the end of the last cycle was fixed, so that the vibration always had a total duration of 1000 msec. In the examples shown, two vibrations differing in frequency by 8 Hz also differed in amplitude such that they had matched velocity. <b>Bottom</b>: Results of Experiment 4 in which subjects compared two periodic or two noisy vibrations that either had the same amplitude or matched velocity (vertical bars represent within-subject SEM). Frequency discrimination was worse (thresholds increased) if the vibrations were noisy or had matched velocity, and these two effects combined super-additively, in that noise increased thresholds more when subjects were comparing vibrations with matched velocity than when comparing vibrations with equal amplitude.