Harmonic tracking using sequential monte carlo
Corentin Dubois, Manuel Davy · IEEE/SP 13th Workshop on Statistical Signal Processing, 2005 · 2005
This paper extends our previous work about individual frequencies detection and tracking to the extraction of harmonic components from multicomponent signals. Once harmonic components are detected, we study their time evolution applying statistical filtering to track them in the time-frequency plane. A nonlinear observation model based on spectrogram is defined and a state-of-the-art particle filtering algorithm is developed. Applications are mainly audio signals