Auditory events in language and music
Yoshitaka Nakajima, Kazuo Ueda · The Journal of the Acoustical Society of America · 2006
Perceptual organization of both language and music is based on temporal units that can be called auditory events. Speech syllables and musical notes correspond typically to auditory events. The listener categorizes these events in various ways in order to extract information. We are at a preliminary stage of searching for acoustic cues that the auditory system can utilize to determine auditory events. Averaged intensity of a speech signal within a moving Gaussian temporal window of 15 ms gives cues of syllable boundaries, especially when this averaged intensity is compared with the intensity change obtained with a Gaussian temporal window of 200 ms. The change of intensity obtained through a filter of about 600–2200 Hz compared with the intensity change obtained through an almost flat filter also gives cues of syllable boundaries. Thus, intensity cues and spectral cues both work to mark syllables. This holds both in Japanese and in English. Such intensity cues may work when applied to monophonic or homophonic music, but spectral changes do not correspond clearly to musical notes. Thus, the perceptual mechanisms that determine speech syllables and musical notes can be different. [Work supported by JSPS and 21st Century COE, Kyushu University.]