Vibrato detection in Byzantine Chant Music

Vibrato´s existence has been recently presented in the context of Byzantine Chant Music (BMC); but further studies on the phenomenon require a robust method for the detection of vibrato segments inside large acoustic corpora. In this work we present a digital signal processing approach for vibrato detection in (BCM) recordings and report the accuracy of the results. The two main tasks of the process are: onset-offset detection and vibrato identification and extraction. For the first task sound level and frequency level of the acoustical signal are used to determine all possible points for such transitions. The second task was based on fundamental frequency analysis using the YIN algorithm. Extra smoothing was applied to resolve issues from high frequency noise. Evaluation was performed by conducting a psychoacoustic experiment. Results show that the detection process proved to be accurate enough for processing large BMC audio recording corpora.