The effect of level and relative frequency region on the recovery of overshoot.

Overshoot--in particular, threshold for a signal near masker onset--can be reduced by presenting a stimulus (precursor) just prior to masker onset. The recovery of overshoot can be examined by varying the delay between the offset of the precursor and the onset of the masker, where "recovery" denotes an increase in the threshold for a signal near masker onset. The present study examined the effects of stimulus level and relative frequency region on this recovery. In all experiments, the masker was a broadband of noise and the signal was a 4-kHz sinusoid. The first experiment examined the effects of masker level on overshoot in order to choose two levels (one "low" and one "high") that produced similar amounts of overshoot; these levels were used in the remaining experiments. In the second experiment, the precursor was identical to the masker, and recovery functions were measured for both low and high masker and precursor levels. there was no consistent difference in the recovery functions between the two levels. In the third experiment, the precursor was divided into two bands (one below and one above 4 kHz); one was presented continuously while the other was gated as in experiment 2. The recovery was more complete when the band above 4 kHz was gated, although the recovery was usually less than that observed in experiment 2 when (effectively) both bands were gated. The results suggest that the frequency regions on both sides of the signal are important for the recovery of overshoot, but that the frequency region above the signal may be more important than the region below.

Development of Phonetically Balanced Phrase Lists Typical of Communication in Noisy Work Environments.

The existing speech corpora are of limited value in assessing the combined effects of hearing loss, hearing protection, and noise on peech intelligibility in the work place. One reason for this is the differences between the acoustic characteristics typical of the elevated vocal effort in the work place and those typical of the normal vocal effort used in recording various existing speech corpora. The other reason for the limited value of the existing tests is related to the differences in nonacoustic factors (linguistic structure, predictability of the messages, etc.) between the available speech corpora and the actual messages used in the work place. In this study a speech test designed for use in industrial work places has been developed. The test material consists of eight 20-phrase lists. The lists are phonetically balanced and of approximately equal difficulty. The variability among lists is greatest at very low sensation levels and decreases progressively as the sensation level increases.