Using simultaneous nasometry and standard audio recordings to detect the acoustic onsets and offsets of speech

Simultaneously collected nasometric and standard audio recording data were compared to determine the extent to which the acoustic onsets and offsets of speech could be determined from Nasometric signals. Three male and 2 female subjects produced six repetitions of 12 utterances that were initiated and terminated by vowels and consonants of differing phonetic features. The onset of Nasometer output typically followed output from the standard audio recording system by an average of 29 ms (SD = 36.1 ms). A greater temporal discrepancy was observed for signal offsets (mean Nasometer lag = 100 ms; SD = 83.3 ms). Large differences between the two instruments seemed due to differences in their overall sensitivities. The results suggest that the Nasometer may be of limited value in determining the acoustic onsets and offsets of speech. The Nasometer was not designed to sense the acoustic events studied here, and the current results have no bearing whatsoever on the validity of this instrument in providing information concerning nasal resonance.