A beat tracking system for acoustic signals of music

Abstract

This paper presents a beat tracking system that processes acoustic signals of music and recognizes temporal positions of beats in real time. Musical beat tracking is needed by various multimedia applications such as video editing, audio editing, and stage lighting control. Previous systems were not able to deal with acoustic signals that contained sounds of various instruments, especially drums. They dealt with either MIDI signals or acoustic signals played on a few instruments, and in the latter case, did not work in real time. Our system deals with popular music in which drums maintain the beat. Because our system examines multiple hypotheses in parallel, it can follow beats without losing track of them, even if some hypotheses become wrong. Our system has been implemented on a parallel computer, the Fujitsu AP1000. In our experiment, the system correctly tracked beats in 27 out of 30 commercially distributed popular songs.

Real-time beat tracking for drumless audio signals: Chord change detection for musical decisions

“…Our previous system (Goto and Muraoka, 1995b) found beat times by ®rst using frequency analysis to detect onset times and then using autocorrelation and crosscorrelation of the onset times. The cues for tracking the higher-level beat structure of drumless audio signals, however, were not dealt with.…”

“…The multiple-hypothesis issue was addressed in our previous system (Goto and Muraoka, 1994, 1995a, 1996 by managing multiple agents that, according to di erent strategies, examined parallel hypotheses about beat positions. This multiple-agent architecture enables the system to cope with di cult beat-tracking situations: even if some agents lose track of beats, the system will track beats correctly as long as other agents maintain the correct hypothesis.…”

“…Multiple agents interpret the sequences of onset-time vectors according to di erent strategies and maintain their own hypotheses (Goto and Muraoka, 1994, 1995a, 1996. Each hypothesis consists of a predicted next-beat time, its beat types (half-note-level type and measure-level type), and the current interbeat interval (Fig.…”

“…We then compared the performance of this system with that of our previous system (Goto and Muraoka, 1995b) for music with drum-sounds. The previous system was also tested on the same 40 songs, and the results of this comparison are listed in Table 5, which shows that the beat detection rates were remarkably improved by our system extension.…”

Partial Identification Analysis for MP3 Music

“…Even though people cannot completely identify every audio component, we can track musical beats and keep time to music by foot-tapping. Therefore, we can build a computational model of beats and track the beats [7], [8].…”