Rhythm Transcription of Polyphonic Piano Music Based on Merged-Output HMM for Multiple Voices

Abstract

Abstract-In a recent conference paper, we have reported a rhythm transcription method based on a merged-output hidden Markov model (HMM) that explicitly describes the multiple-voice structure of polyphonic music. This model solves a major problem of conventional methods that could not properly describe the nature of multiple voices as in polyrhythmic scores or in the phenomenon of loose synchrony between voices. In this paper we present a complete description of the proposed model and develop an inference technique, which is valid for any merged-output HMMs for which output probabilities depend on past events. We also examine the influence of the architecture and parameters of the method in terms of accuracies of rhythm transcription and voice separation and perform comparative evaluations with six other algorithms. Using MIDI recordings of classical piano pieces, we found that the proposed model outperformed other methods by more than 12 points in the accuracy for polyrhythmic performances and performed almost as good as the best one for non-polyrhythmic performances. This reveals the state-ofthe-art methods of rhythm transcription for the first time in the literature. Publicly available source codes are also provided for future comparisons.Index Terms-Rhythm transcription, statistical music language model, model for polyphonic music scores, hidden Markov models, music performance model.

Proceedings of the DMRN+11 Digital Music Research Network Workshop

“…The average (first, third quantiles) was 53.7% (43.8%, 67.3%) for the polyrhythmic data and 38.9% (28.2%, 47.3%) for the standard polyphony data. From the results we can find that methods using statistical learning tends to have better accuracies and an HMM-based model with multiple-voice structure [9] had the best accuracy for the polyrhythmic data. Example results and discussions will be given in the poster.…”

“…There have been many studies on converting a music audio signal into a piano-roll representation based on acoustic modelling of musical sound (see [1,2] for reviews). To obtain a music score, we must recognise quantised note lengths (or note values) of the musical notes in piano rolls, which is called rhythm transcription [3][4][5][6][7][8][9].…”

“…Extending his previous study [4], Temperley developed a model unifying metrical analysis, harmony analysis and stream segregation [7]. To capture the multi-stream structure of polyphonic music, models based on probabilistic context-free grammar (PCFG) [8] and an extension of hidden Markov model (HMM) [9] have been proposed recently.…”

Towards Complete Polyphonic Music Transcription: Integrating Multi-Pitch Detection and Rhythm Quantization

“…Rhythm quantization methods receive note-track data or performed MIDI data (human performance recorded by a MIDI device) and output quantized MIDI data in which notes are associated with quantized onset and offset score times (in beats). Onset score times are usually estimated by removing temporal deviations in the input data, and approaches based on hand-crafted rules [10,11], statistical models [12][13][14][15][16][17][18], and a connectionist approach [19] have been studied. A recent study [18] has shown that methods based on hidden Markov models (HMMs) are currently state of the art.…”

“…Onset score times are usually estimated by removing temporal deviations in the input data, and approaches based on hand-crafted rules [10,11], statistical models [12][13][14][15][16][17][18], and a connectionist approach [19] have been studied. A recent study [18] has shown that methods based on hidden Markov models (HMMs) are currently state of the art. Especially, the metrical HMM [13,14] has the advantage of being able to estimate the metre and bar lines and avoid grammatically incorrect score representations (e.g.…”