Relative loudness of speech feedback during speech production.

Abstract

Relative loudness of speech feedback during speech production was investigated using an adaptive, two-track, two interval forced choice procedure. The participants indicated which of the two signals, the speech feedback (listening while speaking interval) or its audio recording (listening only interval), sounded louder. Based on the subject’s choice, the gain [indeabel sound pressure level (SPL) on the recording was raised or lowered on the subsequent trial. Relative loudness of the speech feedback was defined as the difference in the sound pressure level of each signal (feedback and replay), when the two were experienced as equally loud. The subject’s speech was recorded with a microphone placed above his/her right ear, and both speech signals were presented via insert earphones. In the canal speech recordings were collected for later analysis. Fifteen normally hearing adults were tested. At the point of subjective equality, the SPL of the recorded and the live signals were equal for some of the subjects, while for others, the SPL of the live speech exceeded the SPL of the recorded speech by as much as 4 dB. The subjects’ preferred speaking level and their acoustic reflex thresholds did not account for the observed individual differences. [Work supported by RERC-HE to Gallaudet University.]