Judgments of interval size by musically trained and untrained listeners

The current experiment sought to clarify the appropriateness of the musical scale for representing phenomenological judgments of interval size. Musically trained and untrained listeners were asked to judge the size of ascending and descending intervals that varied in log frequency distance from 50 cents (1/2 semitone) to 2400 cents (2 octaves). All intervals were presented melodically (two-tone sequence) in both a lower pitch register and a higher pitch register. Although interval size judgments were well predicted by log frequency distance, two unexpected findings were revealed. First, the rate of expansion in size judgments for increasing interval sizes presented in the upper register was greater than the rate of expansion in interval size judgments for increasing
interval sizes presented in the lower register. Second, for trained listeners, the rate of expansion in size judgments for increasing interval sizes that were smaller than an octave (e.g., 100 to 400 cents) was greater than the rate of expansion in size judgments for increasing interval sizes that were larger than an octave (e.g., 1300 to 1600 cents). For untrained listeners, the slope of the best fitting line did not vary as a function of whether intervals were smaller than or larger than an octave.