gms | German Medical Science

In Germany the statutory health insurance requires for reimbursement that the benefit of hearing aids needs to be confirmed with speech tests. To this end, the Freiburg monosyllabic speech test can be used to compare the aided and unaided condition in quiet and noise. For the test in noise, speech and noise levels but no loudspeaker configurations are defined. Besides the relative evaluation between aided and unaided condition, also a comparison of speech recognition in both conditions with the performance of normal-hearing listeners is audiologically relevant. Recently, a reference speech recognition curve has been published by [1] for a frontal presentation of speech and noise (S0N0) so that a comparison to normal-hearing listeners is possible for this configuration. However, in daily practice, various loudspeaker configurations are used and the reference curve for S0N0 is not applicable. Therefore, reference speech recognition curves for the Freiburg monosyllabic speech test in noise were determined for various loudspeaker configurations. A study with 100 normal-hearing listeners was conducted for speech coming from 0° and the continuous CCITT noise coming from 0° (S0N0), ±45° (S0N±45), 45° (S0N45), 90° (S0N90), and 180° (S0N180). Speech recognition was measured at four signal-to-noise ratios (SNR) for each loudspeaker configuration. The median speech recognition scores for all SNRs were used to fit a psychometric function to each configuration according to ISO 8353-3:2012. The tests were conducted in an anechoic test room in Oldenburg and in an audiological test room in Lübeck. A comparison of the results of both locations reveals significant differences especially for the configuration S0N90. An investigation of different impacts suggests that the room acoustics has a stronger influence as expected.