Analysis of music: controlled random music and probability distribution function of recurrence time of amplitude peaks

TL;DR

This paper investigates the correlations in monophonic waveforms of random music by analyzing the probability distribution of recurrence times of amplitude peaks, revealing a power law behavior in specific time ranges.

Contribution

It introduces a method to analyze waveform correlations in random music and explores their origins by isolating factors like frequency, intensity, and duration.

Findings

Power law distribution observed in recurrence times from 0.1 ms to 20 ms

Correlations are influenced by specific factors such as frequencies and note durations

Randomization of factors affects the distribution, indicating their role in waveform correlations

Abstract

Correlations in music that exist within its waveform are studied. Monophonic wave files of random music are generated and the probability distribution function of time interval between large signal values is analyzed. A power law behavior for the distribution function in the range from 0.1 millisecond to 20 milliseconds is observed. An attempt is made to investigate the origin of these correlations by randomizing each of the factors (frequencies, intensities and durations of notes of the random music files) separately.