Frequency and fluctuation structure of various environmental sounds

The conventional method of measuring environmental sounds according to sound pressure level (SPL) alone is not adequate for a qualitative evaluation of environmental sounds. To improve such evaluation, we measured the power spectra of various environmental sounds and their fluctuation in a microtemporal domain of several milliseconds using Fast Fourier Transform (FFT) analysis and the maximum entropy spectral array method (MESAM). We recorded environmental sounds in an urban area, a country village, and a tropical rain forest with a B&K 4939 microphone and a high-speed one-bit coding signal processor with a good frequency response over 100kHz, developed by Y. Yamasaki. The environmental sounds of the urban area fell below 15kHz and showed a monotonous appearance with poor fluctuation. In contrast, the power spectrum of the country village sounds showed a rich high-frequency component above 50kHz with complex fluctuation. The power spectrum of the tropical rain forest showed a rich high-frequency component above 100kHz with complex, non-periodical fluctuation. These results indicate that the inaudible-tohumans high-frequency component having a complex fluctuation structure in a micro-temporal domain is lost as the environment becomes more artificial.