Evolution Models and Partial Tracking of Music Signals

H. Satar-Boroujeni and B. Shafai (USA)


Modelling and simulation strategies, statistical and probabilistic modelling, music analysis, tracking.


In this paper we propose a state-space model for evolution of frequency and power partials in music signals. This model is employed for tracking of partials using Kalman filter. In fact, our state-space model comprises a set of instrument-specific and frequency-dependant models. For different classes of melodic instruments different models are introduced. Parameters of each model can be estimated through an extensive statistical analysis of a large number of musical sounds. These parameters are frequency-dependant and for each frequency bin a different set of parameters are estimated. Our Kalman tracker estimates frequency and power tracks using these evolution models.

Important Links:

Go Back