Annotation for Miller, Benjamin O.
Time Perception in Musical Meter Perception
Annotation (by Bill Tilghman):
- Miller presents a simple model that attempts to integrate time
perception into meter perception. The model proposes (1) that
listeners are better able to perceive an event's temporal
position if the event coincides with a beat at a higher metrical
level; (2) that these abilities decrease as the time interval
between the event and the preceding event increases, since our
perception of duration grows less accurate (i.e., our "temporal
resolution" decreases) as the duration increases; and (3) that
this "temporal resolution" varies among different listeners
according to each listener's personal "Weber fraction" (i.e.,
what proportion of difference between two stimuli is required in
order for the subject to perceive the stimuli as different).
Three experiments are offered as empirical support for the model.
Each experiment was constructed using the popular "probe tone"
technique, in which each stimulus begins with the presentation of
a context or "stem" (in this case, a short rhythm that is
presumed to establish a meter) followed immediately by some
variable "probe" (in this case, a final note, which may occur at
any of seven temporal positions with various metrical strengths).
Subjects were asked to judge the metrical strength of the probe
(by rating its "goodness of fit"), to identify its temporal
position (by identifying the duration of the preceding note), or
to reproduce its temporal position (by tapping out the just-heard
rhythm). In general, the subjects performed as the model
predicts: those performing the rating task gave higher ratings to
stronger metrical positions, and those performing the
identification and reproduction tasks did so more accurately when
the probe was at a stronger metrical positions, but performance
on all three tasks declined as the probe latency increased.