论文信息 - A COMPARISON OF PROBABILISTIC MODELS FOR ONLINE PITCH TRACKING

A COMPARISON OF PROBABILISTIC MODELS FOR ONLINE PITCH TRACKING

In this study, we propose and compare two probabilistic models for online pitch tracking: Hidden Markov Model and Change Point Model. In our models each note has a certain characteristic spectral shape which we call spectral templates. Hence the system’s goal is to find the note whose template is active given the audio data. The main focus on this work is the trade off between latency and accuracy of the pitch tracking system. We present the probabilistic models and the inference schemes in detail. Encouraging results are obtained from the experiments that are done on low-pitched monophonic audio.

Umut Şimşekli | Umut Simsekli

[1] Nicola Orio,et al. An HMM-based pitch tracker for audio queries , 2003, ISMIR.

[2] Arshia Cont. Realtime Multiple Pitch Observation using Sparse Non-negative Constraints , 2006, ISMIR.

[3] Anssi Klapuri,et al. Multipitch Analysis of Polyphonic Music and Speech Signals Using an Auditory Model , 2008, IEEE Transactions on Audio, Speech, and Language Processing.

[4] Ali Taylan Cemgil,et al. Bayesian Inference for Nonnegative Matrix Factorisation Models , 2009, Comput. Intell. Neurosci..

[6] Ali Taylan Cemgil,et al. Bayesian Music Transcription , 1997 .

[7] Christopher Raphael,et al. Automatic Transcription of Piano Music , 2002, ISMIR.

[8] Roland Badeau,et al. A tempering approach for Itakura-Saito non-negative matrix factorization. With application to music transcription , 2009, 2009 IEEE International Conference on Acoustics, Speech and Signal Processing.

[9] Hideki Kawahara,et al. YIN, a fundamental frequency estimator for speech and music. , 2002, The Journal of the Acoustical Society of America.

[10] Kunio Kashino,et al. Application of the Bayesian probability network to music scene analysis , 1998 .

[11] Emmanuel Vincent,et al. Harmonic and inharmonic Nonnegative Matrix Factorization for Polyphonic Pitch transcription , 2008, 2008 IEEE International Conference on Acoustics, Speech and Signal Processing.