Spectral properties of Japanese whispered vowels referred to pitch
Hideaki Konno, Hideo Kanemitsu, Jun Toyama, Masaru Shimbo · The Journal of the Acoustical Society of America · 2006
Whispered speech can communicate the same linguistic information as ordinary speech in spite of the great difference of their respective acoustic characteristics. In this respect, whispering is an interesting object for studying speech perception and recognition. In this study, we investigate the spectral properties of five whispered Japanese vowels uttered in isolation and having different pitch. Pitch of ordinary whispered vowels was measured in terms of the manner in which the talkers were listening to pure tones while uttering and adjusting its frequency so that the pitch matched the utterance. For other samples, talkers changed the pitch of utterances to match the given pure tones. Acoustic analyses were carried out on formant frequencies, a spectral tilt, and a peak frequency of wide-ranging spectral shape using second-order LPC method called a global peak. Preliminary results show a tendency of upward shift of F1, F2, and global peaks, and flattened spectral tilts on overall vowels with increasing pitch. The intended pitch by a talker is not correspondent to a specific formant frequency. Results of pitch comparison tests among samples and pitch matching test with pure tones will also be discussed. [Work supported by MEXT Japan.]