2009/02/11

Povel, D.-J. & Okkerman, H. (1981). Accents in equitone sequences. Perception & Psychophysics, 30 (6), 565–572.

Presentation: Shelly

Summary: Sally

Equitones are defined as sequences generated from tones identical in all aspects. Accents perceived when hearing equitones show the perceived structure of the sequence. In a pilot study, the authors found that when presented with an isochronic sequence of 50-ms tones and intervals of the same length, subjects heard groups of two tones as the duration of the second intervals exceeded the first by 5-10%. If the second intervals were further lengthened, participants would remain the same grouping, but their perceived accents would move from the first tone to the second one. Explanations for this interval-induced accent phenomenon include “integration” (the residual sensation from the first auditory elements is added to the sensation of the second ones), “loudness enhancement” (under certain conditions the loudness of a tone is enhanced if it is preceded by another one), “partial masking” (the loudness of an auditory element is decreased by the presence of another element) and echoic memory/preperceptual auditory store (the information in the stimulus is kept in a store for a short amount of time after termination of the stimulus, waiting for being processed). 

In this study, more details about the interval-produced accent in equitone sequences were investigated. Experiment 1 was designed to determine the conditions under which this kind of accent occurs. Two-tone groups (tone + 1^st interval + tone + 2^nd interval) were chosen due to its simplicity in structure and feasibility for forced choice responses. The 2^nd intervals (i₂) were always longer than the 1^st (i₁) under different levels of temporal manipulation. Results showed that if there was only small difference between i₁ and i₂, subjects perceived the accent on the first tone; as the difference increased, more accents were heard on the second; once the difference exceeded 180 ms, the accent on the second tone became apparent. In Experiment 2, each sequence was preceded by a 500-ms tone in order to test whether perceiving accents on the first tones was a result of an orienting response (the tendency to hear an accent on the first tone). Results confirmed the hypothesis in that with this added tone, the percentage in perceiving accents on the first tone lowered, as compared to the results of Experiment 1. In Experiment 3, tones were lengthened to 150 ms for a test on the integration hypothesis. Exactly the same accent perception as Experiment 1 was found, showing that energy integration was not the only cause for the interval-produced accent. Experiment 4 was then conducted to directly determine the strength of the accents found in equitone sequences, especially of the interval-produced accents. The strength of the accent on the second tone was measured by the amount of intensity of the first tone to be raised to balance out the accent. By turning a knob and adjusting the intensity to make the two tones sound equally accented, subjects showed a robust interval-produced accent effect: an increase of 4 dB was needed to balance it out. Accents perceived on the first tone in less differentiated sequences, on the other hand, were easier to compensate for.