Ratings of speed in real music as a function of both original and manipulated beat tempo.

There is an apparent contradiction between the narrow range of tempi optimal for perceptual judgment and motor synchronization and the wide range of beat tempi found in real music. The relation between listeners' perception of speed and beat tempo was therefore investigated, both for real music excerpts (ME) and metronome sequences. Tempi ranged from 42 to 200 beats per minute (BPM), and some excerpts were further tempo manipulated in four levels from ±5 to ±20%. Regression analyses showed that speed was a shallower function of original tempo for fast (>150 BPM) and slow (<95 BPM) MEs than for MEs with intermediate tempi, describing a non-linear, sigmoid function. Manipulated tempo had twice as large an effect on speed as had original tempo. In contrast, speed was an almost linear function of tempo for metronome sequences. Taken together, these results show that the non-linearity stems from properties of the musical signal, rather than being a subjective perceptual effect. They indicate an inverse relation between tempo and relative event density in real music, and demonstrate that the perception of periodic signals is affected not only by the beat level, but also by faster and slower levels.