Beyond the beat: modeling metric structure in music and performance

Area of Science:

• Music Information Retrieval
• Computational Auditory Processing
• Rhythm and Meter Analysis

Background:

• Existing music analysis models primarily focus on tempo and beat estimation.
• The potential for extracting broader metric and rhythmic information remains largely untapped.

Purpose of the Study:

• To present a novel model for describing the general metric structure of audio signals and behavioral data.
• To investigate the suitability of reson filters for extracting multiple metric levels from acoustic signals.

Main Methods:

• Developed a new model employing reson filters, differing from previous comb filter approaches.
• Investigated the oscillatory properties of reson filters for multi-level metric extraction.
• Tested the model with Dirac impulse sequences and natural stimuli (music excerpts, tapping data).

Main Results:

• The model demonstrates sensitivity to timing variations and accent structures in input data.
• Illustrates responses to natural stimuli, including musical recordings and bimanual tapping tasks.
• Provides a framework for analyzing complex metric structures in acoustic and behavioral patterns.

Conclusions:

• Reson filters offer a promising approach for advanced metric and rhythmic analysis in audio.
• The proposed model extends beyond basic beat tracking to capture deeper structural elements.
• Suggests potential applications and future extensions for comprehensive rhythm analysis.