State-of-the-art in fundamental frequency tracking
Stéphane Rossignol, Peter Desain, Henkjan Honing · UvA-DARE (University of Amsterdam) · 2001
Obtaining accurate fundamental frequency information from audio data is a hard problem, especially when, for example, sympathetic resonance of open strings in string instruments interfere with some harmonics of the main sound, when transitions are so fast that tracks of different harmonics are connected, or when the signal is polyphonic. Two kinds of knowledge are used. Firstly, the score of the music is used as a guide. This score information constrains signal processing systems. Secondly, these systems can take advantage also of the fact that other information, like the timing, the instrument and sometimes even the fingering is known. The usefulness of musical-knowledge-based signal processing systems is discussed. Our primary motivation of using knowledge in f0 tracking is to obtain reliable and precise fundamental frequency trajectories that can be used for the analysis and modeling of the vibrato shape. This idea can be useful for other computer music systems as well that need to track, for instance, the fundamental frequency in a live situation where score and timing information is available.