gms | German Medical Science

Airflow or wind reaching the membrane of a microphone generates low-frequency noise with potentially high levels at the electrical output. Due to space limitations with hearing aids, the use of passive wind shields, e.g., made of foam or furry materials, is difficult. Therefore, many hearing aids provide digital wind noise reduction (WNR). Technical measurements showed that this feature strongly reduces wind noise in the range of 20–40 dB at the lower frequencies (< 1 kHz). Moreover, in combination with a noise reduction for speech enhancement (NRSE) also moderate improvements in SNR in the range of 1–5 dB can be observed [1]. This clearly demonstrates the effect of WNR but does not allow for a direct interpretation in terms of benefit for the user. Therefore, audiological experiments with test subjects, e.g., to measure speech intelligibility, listening effort or the perceived annoyance of wind noise are very helpful [2]. Usually, those experiments last more than 30 min up to several hours and we believe that exposing test subjects to a windy environment for this long is not ethical. Therefore, we established a setup where the hearing aids are attached to a head and torso simulator (HAT) in one room, and the output signals measured at the internal ear simulators are streamed to the test subject via headphones sitting in another room in real time. In the HATs room, sounds can be presented via loudspeaker from 0° and wind can be generated with a fan from 180°. In the other room, an audiologist can perform measurements via free field or headphones in test subjects room but also via free field in the HATs room. In this talk, we explain the possibilities of the setup and show results of a first study with 18 hearing impaired test subjects wearing 6 different commercial hearing aids. Within this study, speech intelligibility with the Oldenburger Sentences Test (OLSA), listening effort using Adaptive CAtegorical Listening Effort Scaling (ACALES), and perceived annoyance of wind noise using a 9-point scale were measured. These results show surprisingly high differences in speech intelligibility which is partly also seen in the ratings on listening effort whereas the differences in perceived annoyance are less prominent.