• Home

Conference_programme: 5: Audiology

Lecture: Modern pitch detection methods in singing voices analyzes

Author(s): Gawlik Mateusz, Wszołek Wiesław

Voice analyzing problems, presented in the literature, concern mainly recognition, identification, verification or diagnostics of speakers. There has been a plethora of work focused only on speech surveys. Unfortunately, a range of studies of singing was not the subject of great interest by scientists. Results from presented research show that the most elementary parameter used in voice analyzing is fundamental frequency (F0). In this paper we presented pitch detection methods of singing signal in hope of choosing the best one. We decided to compare four detection methods, which according to the literature present promising results: Zero Frequency Band Filtering (ZBF), SEDREAMS algorithm, Modified Zero Frequency Resonator (MZFR) and DYPSA algorithm. The first conclusion from series of experiments is necessity of choosing probes to analyze in the proper way, because singing voice can change its basic frequency approximately by two octaves. This paper presents a precise analysis of the pitch detection by selected methods in singing voices analysis. As the most precisely method for singing voices analyzing after the study we classified Zero Frequency Band Filtering method.

Download the full paper

Corresponding author

Name: Mr Mateusz Gawlik

e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Country: Poland