Abstract
We built a skull vibration detection system and attempted to detect direct bone-conducted sounds with skin penetration via a bone-anchored pick-up which consisted of a bone-anchored titanium implant and a piezoelectric acceleration meter. This study clarified the following. 1) In the oscillation experiment on the skull, the output level almost linearly increased more than 20 dBHL above the input level in the direct bone-conducted sounds. In comparison with non-direct bone-conducted sounds, there was no severe attenuation of output level in the high frequency band, from 1000 to 4000 Hz, in the direct bone-conducted sounds. 2) In long-term average spectrum (LTAS) analysis (sentence readings of 28 sec), decreasing of output power was shown at more than 1000 Hz in the non-direct bone-conducted sounds. However, few differences in output power (approximately 0-10 dB) were seen between direct bone-conducted sounds and air-conducted sounds. 3) In the monosyllable speech discrimination test, there was a significant difference (p<.01) in the percentages of correct answers for vowel sounds between non-direct and direct bone-conducted sounds. In comparison with non-direct bone-conducted sounds, there was no tendency to confuse vowel sounds in the case of direct bone-conducted sounds.
