Transcribing Frequency Modulated Musical Expressions from Polyphonic Music Using HMM Constrained Shift Invariant PLCA

Abstract

In recent years, there has been a lot of work in transcribing polyphonic music using non-negative spectrogram factorization. However, most of them focus on transcribing audio signal into the occurrence of notes, onset and pitch of notes. In this paper, a concept for automatic transcription of frequncy modulated muscial expressions such as vibrato, glissando is proposed. To transcribe those musical expressions from polyphonic music signal, hidden Markov model constrained shift-invariant probablistic latent component analysis is used. From a impulse distribution which reveals the frequency variation of each note, each expression can be modelled in accordance with designed rules. Experiments showed that the impusle distribution can be used to transcribe expressions from polyphonic music signals.

Audio classification for music information retrieval of Hindustani vocal music

“…Hidden Markov models (HMM) have also been used extensively for music transcription and musical expression detection tasks. For example, a remarkable method of transcribing frequency modulated musical expressions using HMM and spectrogram factorization approach is suggested in [8]. The limitation, however, of the feature-based computational methods is the need of hand engineering of the features and their accurate computation.…”

Understanding the Probabilistic Latent Component Analysis Framework

“…Within this framework, Probabilistic Latent Component Analysis (PLCA) has then been developed as a general method for feature extraction from non-negative data, with pioneer applications to audio [5] and image [6]. Following studies in audio research have in particular dealt with the tasks of multipitch estimation [7,8,9,10,11,12], sound source separation [13,14], instrument identi cation [15], melody extraction [16,17], temporal music structure [18] and speech processing [19,20].…”