Subjective and Objective Speech Intelligibility During the Tracking of Noise Tolerance Test.

Abstract

Background: The Tracking of Noise Tolerance (TNT) test instructs listeners to set the background noise to a level that they can tolerate while still understanding >90 percent of speech. It is unclear whether listeners follow the 90 percent intelligibility criterion, however.Purpose: To measure objective and subjective intelligibility of TNT speech materials presented with background noise at signal-to-noise ratios (SNRs) experienced by listeners during the TNT test.Research Design: Single-blind within-subjects design.Study Sample: Twenty-four older listeners with a four-frequency pure-tone average (4PTA) ≤20 dB HL (reference hearing [RH]) and 17 listeners with hearing impairment (HI) of sensorineural origin comprised the sample.Data Collection and Analysis: All participants were tested unaided, and listeners with HI were also tested while aided. Listeners performed the TNT test at 75 and 82 dB sound pressure level (SPL) speech levels. Then, listeners completed the objective–subjective intelligibility difference (OSID) test at the same speech levels across five SNRs to generate objective and subjective speech intelligibility performance intensity (P-I) functions. Next, speech intelligibility was estimated to determine the average TNT noise level and the peaks and valleys of TNT tracings using the individualized P-I functions.Results: Subjective intelligibility ranged between 81 and 90 percent at noise valleys (TNT_Valley,SUB) and between 30 and 45 percent at noise peaks (TNT_Peak,SUB), with the average intelligibility (TNT_Ave,SUB) between 58 and 69 percent. Objective intelligibility was similar to subjective intelligibility in listeners with RH. For unaided listeners with HI, objective intelligibility across TNT indices was 12 to 18 percent poorer than subjective intelligibility. Objective and subjective intelligibility were within 1 to 8 percent for aided listeners with HI. Except for aided listeners with HI at 75 dB SPL, the TNT_Valley,SUB was not significantly different from 90 percent. Speech intelligibility (both subjective and objective) measured at the TNT_Ave and TNT_Peak were statistically different from 90 percent.Conclusions: Listeners generally followed the TNT instructions and achieved perceived (i.e., subjective) intelligibility of approximately 90 percent at the TNT valleys. They likely judged the TNT peaks based on other criteria (e.g., loudness) as reflected by the <50 percent subjective intelligibility. These results suggest that TNT informs the clinicians of the SNR that a listener needs to support various degrees of subjective intelligibility.