Piano Legato-Pedal Onset Detection Based on a Sympathetic Resonance Measure

Abstract

In this paper, the problem of legato pedalling technique detection in polyphonic piano music is addressed. We propose a novel detection method exploiting the effect of sympathetic resonance which can be enhanced by a legato-pedal onset. To measure the effect, a specific piano transcription was performed using the templates of pre-recorded isolated notes, from which partial frequencies were estimated. This promotes the acquisition of residual components associated to the weak coexcitation of damped notes due to the legato pedalling technique. Features that represent the sympathetic resonance measure were extracted from residuals. We finally used a logistic regression classifier to determine the existence of legato-pedal onsets.

Piano Sustain-pedal Detection Using Convolutional Neural Networks

“…According to piano acoustics and the observations in our prior works [4] and [5], musical features can be different in the frames around a pedal onset versus the ones within a pedalled segment. Here, a pedalled segment is referred to as samples between a pedal onset time and its corresponding offset time.…”

“…Since these features were extracted from isolated notes, dedicated new features were developed to detect pedalling in polyphonic music. A method for detecting onset times of legato pedalling (pressing the sustain pedal immediately after the note onset) was first proposed in [5] based on a measure of sympathetic resonance. Since a piano transcription technique was used as an intermediate step, the robustness of this method may be reduced by note transcription errors.…”