A Study on Acoustic Signal Analysis Based on Cyclostationarity

Abstract

For sounds with temporal energy fluctuations, the A-weighted sound pressure level may differ from the subjective impression when listening to the sound. It may be because humans use sounds with temporal energy fluctuations, such as voice and music, as a means of information transmission. In this presentation, we applied the method of analyzing the time-variability of acoustic signals using cyclostationary analysis to the Japanese cicada sounds and the cat's purr. We report the source mechanism estimation results focusing on the first-order and second-order cyclostationarity and the high-precision time-frequency analysis results.