The Influence of Audiovisual Ceiling Performance on the Relationship Between Reverberation and Directional Benefit

Abstract

Acoustically, directional microphone hearing aids are less able to improve signal-to-noise ratio in more-reverberant environments. However, because auditory-visual (AV) omnidirectional performance is more likely to approach the ceiling level in less-reverberant environments, listeners may perceive a smaller directional benefit in these environments. The first objective of this study was to investigate the influence of AV ceiling performance on the relationship between reverberation and directional benefit. The second objective was to determine whether a modified speech intelligibility index (mSII) model could predict directional benefit across AV listening conditions with different levels of reverberation. Nineteen adults with sensorineural hearing loss were recruited. Directional benefit was measured using the auditory-only (AO) Hearing in Noise Test (HINT) and the AV version of the Connected Speech Test (CST) in environments containing either low or moderate levels of reverberation (reverberation time = 0.2 sec and 0.7 sec, respectively). The AO mSII was calculated for both speech materials. To predict AV directional benefit, the AO mSII was converted to the AV mSII using the correction equations specified by American National Standards Institute S3.5-1997 (R2007) and the new equations developed based on the data collected in a low-reverberation sound booth in a previous study. As the test environment became more reverberant, a decrease in directional benefit was measured using the HINT. In contrast, the results of the AV CST revealed that directional benefit was higher in the more-reverberant condition. Although the mSII model accurately predicted the HINT-measured directional benefit, the model using the American National Standards Institute correction equations underestimated the directional benefit measured using the AV CST in both reverberation conditions. Applying the newer equations, the model resulted in a more accurate prediction in the low-reverberation condition. However, AV directional benefit in the moderately reverberant condition was still underestimated. Hearing aid users would achieve the greatest directional benefit in situations in which they do not reach ceiling performance. The present study suggests that in the real world these situations may include face-to-face communication occurring in environments with higher reverberation. The opposite trends demonstrated by the HINT and AV CST highlight the importance of using ecologically valid laboratory testing to estimate the real-world effectiveness of a hearing aid and its technology. As a likely result of the difference in the effect on speech cues between noise and reverberation, the new equations developed from the low-reverberation booth failed to predict AV directional benefit in the more-reverberant environment. The mSII model with visual cue correction equations would be able to provide accurate predictions for AV directional benefit when the acoustic environments are similar to the ones used to derive the equations.