MULTIPLE FUNDAMENTAL FREQUENCY ESTIMATION & TRACKING IN POLYPHONIC MUSIC FOR MIREX 2010

The goal of the MIREX (Music Information Retrieval Evaluation eXchange) contest is to compare state-of-the-art algorithms and systems relevant for Music Information Retrieval. This paper briefly describes our algorithm for the MIREX 2010 Multiple Fundamental Frequency (multipleF0) Estimation & Tracking task. Specifically, it is submitted for the first two subtasks: i) estimation of several fundamental frequencies (F0s) at frame-level and ii) Tracking note contours on a continuous time basis. Our algorithm, designed as a trade-off between accuracy rate and computational cost, is based on the minimization of a spectral distance which is calculated using the convolution between candidates and harmonic patterns belonging to the same combination at frame-level and maximization of temporal continuity at temporal interval-level.

[1]  Mark R. Every,et al.  Separation of synchronous pitched notes by spectral filtering of harmonics , 2006, IEEE Transactions on Audio, Speech, and Language Processing.

[2]  D. Wang,et al.  Computational Auditory Scene Analysis: Principles, Algorithms, and Applications , 2008, IEEE Trans. Neural Networks.

[3]  Guy J. Brown,et al.  Computational auditory scene analysis , 1994, Comput. Speech Lang..

[4]  F. J. Cañadas Quesada,et al.  A Multiple-F0 Estimation Approach Based on Gaussian Spectral Modelling for Polyphonic Music Transcription , 2010 .

[5]  José Manuel Iñesta Quereda,et al.  Multiple fundamental frequency estimation using Gaussian smoothness , 2008, 2008 IEEE International Conference on Acoustics, Speech and Signal Processing.

[6]  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.

[7]  Anssi Klapuri,et al.  Signal Processing Methods for Music Transcription , 2006 .