Dept. of Speech Commun., Penn State Univ., University Park, PA 16802
Acoustic coupling between the oral and nasal cavities results in the addition of pole-zero pairs to the vocal-tract transfer function, which interacts with variations in vocal-tract shaping across different vowels and speakers. Presentation of the myraid of acoustic features associated with nasalization based on vocal-tract models can confuse students with little mathematical background. One successful approach allows students to discover a few established cues via waveform and spectrographic analysis. Broadband spectrograms of students' productions of ``bead--bean'' clearly illustrate a nasal murmur (200--300 Hz) and the attenuation of formant frequencies is the nasalized vowel compared to the oral vowel. Another broadband spectrogram is made of the tape-recorded speech of a female speaker of Texan English, exhibiting extreme attenuation in the F1 region. The decreased intensity of the nasalized vowel in ``bean'' (compared to the oral vowel in ``bead'') is seen as rounded waveform cycles with lower amplitude using waveform-editing software. Changes in vowel quality due to nasalization can be heard as the students proceed screen-by-screen through the waveform.