RHYNE, Casey: Long Term Pulse Synchronization: Average Pulse and the Effects of Random Variance on Pulse Salience

Wikis > Final Projects > RHYNE, Casey: Long Term Pulse Synchronization: Average Pulse and the Effects of Random Variance on Pulse Salience

Proposal
By Casey Rhyne

It is widely accepted that performed music rhythmically varies through human imperfection as well as through expression and that these minor fluctuations in tempo present no challenge in following for the human ear. Before the 20th century, pulse salience was rarely tested to its limits. However, with the rise of modern music, pieces now may seriously challenge the brain’s ability to follow a pulse. While research has shown that the following of a pulse is largely based on correlation of fairly recent beats (the brain quickly adapts to new tempi within a few beats), it may be possible to test for a more long term pulse recognition system through the introduction of controlled random variance. This variance will shift each pulse a random amount within a time window v. As v is increased, the brain’s synchronization ability will be able to rely less on short-term intervals and would instead have to rely more on average interval length. If a longer-term system does not exist, the tests would likely show synchronization difficulties beyond average intervals of approximately one half to one third the length of the echoic memory storage. If, however, the brain also averages over a period of time beyond the echoic memory, results would show greater ability to cope with fluctuations on a larger scale.

Pieces of interest would most almost certainly be partially computer generated (for elements of randomness) and would thus be a fairly recent composition. Beyond the generated beat samples, such a piece could be also tested for pulse synchronization to verify whether the principles of average pulse synchronization apply to more complex examples. For simpler musical samples (those without secondary pulse layers that would be distracting to test subjects ), one could analyze the average number of pulses per unit of time as well as the average deviation from this tempo. Comparisons to results from the general experiment described below could indicate whether the brain truly perceives highly contemporary pieces as random or not.

Before analyzing specific pieces, however, information would need to be gathered regarding the perception of randomness and average pulse with controlled test samples. The experiment would be carried out with three groups. The first group would be a control group, and would be asked to tap to a steady pulse at various tempi. The second would be tested with a series of pulses each with random variance within a max variance v. Both tempo and variance would be altered to gather data, which would be analyzed in terms of adherence to the average pulse and adherence to individual local pulses (governed by short-term synchronization). Emphasis would be placed on determining whether a significant drop occurs when the number of pulses within the time-span of the echoic memory drops below three (beyond which a reliance on long term averaging would be necessary). The third group would be a second control group assigned to attempt to tap a pulse (presumably as regularly as possible) against a truly random distribution of beats within a set amount of time. The results for this group would function as a baseline to which the second group could be compared to, especially beyond the echoic memory storage. (Posted by Casey Rhyne on 3/31/2010 and edited 4/4/10)

Annotated Bibliography
Demany, Laurent, and Catherine Semal. “Limits of rhythm perception.” Quarterly Journal of Experimental Psychology A: Human Experimental Psychology. 55A.2 (2002): 643-657. Print.

This article will prove important if I am to examine the cognition of rhythms beyond those between consecutive beats. Demany and Semal’s experiments tested individuals for the ability to be aware of second order beats. While the first experiment did not reveal any significant ability to sense second order intervals, the second, which had dealt with more regular intervals, did. Thus, while a series with every other interval being of a random duration might not produce an audible pulse, my proposed experiment may still induce a sense of pulse in the listener due to the absence of a truly random interval separating the beats in the pulse (For clarity, the interval is not entirely random because it is limited to an extent to temporal points around where the beat should land).

Madison, Guy, and Björn Merker. “On the limits of anisochrony in pulse attribution.” Psychological Research/Psychologische Forschung 66.3 (2002): 201-207. Web. 10 Apr 2010.

This study will provide a baseline interval variance for the random variance experiment. Testing showed that musicians were able to detect constant perturbations in a constant tempo beyond 3.5% of the IOI, while non-musicians began at 8.6%. Because unnecessary testing is undesirable and because participants tested below these values would not be able to detect any tempo variance, the higher value, 8.6%, could serve as the initial variance.

Repp, Bruno. “Comments on ‘Rapid motor adaptations to subliminal frequency shifts during syncopated rhythmic sensorimotor synchronization’ by Michael H. Thaut and Gary P. Kenyon (Human Movement Science 22 2003? 321-338)..” Human Movement Science. 23.1 (2004): 61-77. Print.

In contrast to Thaut and Kenyon’s article, Repp’s article is supportive of my suggestion for a longer term tempo synchronization mechanism. Repp claims that period adjustment and not phase adjustment functions on a longer time-span. Although the IOI percentage shift is less than what would likely be covered in my experiment (2%), my hope is that this phase adjustment mechanism also applies for more audible changes in tempo.

Thaut, Michael, and Gary Kenyon. “Rapid motor adaptations to subliminal frequency shifts during syncopated rhythmic sensorimotor synchronization .” Human Movement Science. 22.3 (2003): 321-338. Print.

This study will be important in my study because it provides the conclusion that I aim to dispute, namely, that period synchronization functions on a beat to beat basis, and that the brain does not take longer time intervals into account when adjusting to tempo changes. While my study would be on synchronization, the important point is the time-span throughout which the mind is attuned to tempo.