ABSTRACT
Hearing impaired listeners heavily rely on facial expressions and unfiltered speech signals for a sufficient speech understanding. Due to the current Cov-id-19 pandemic face masks are worn by almost everyone every day. They therefore interfere especially with the communication of hearing impaired listeners. We used a modified audiovisual German matrix sentence test (AV-OLSA) 1 to further evaluate the effect of face masks on speech intelligibility in a cohort of cochlear implant users (CI, n = 15) and a control group of normal hearing listeners (NH, n = 5). Besides “audio-only”, “visual-only” and “audiovisual” conditions, we added modified conditions. These consisted of an audiovisual condition with a simulated mask and an audio signal, which was edited according to the acoustic filter properties of different face masks (surgical and FFP-2). Our preliminary data show a gain of 5.3 dB (CI) and 2.6 dB (NH) in speech reception thresholds at 80 % word recognition (SRT80%) respectively, if a speaker's video is added to an unfiltered audio signal. Different types of face masks (surgical and FFP-2) led to a deterioration in SRT80% in both groups of up to 7.6 dB (CI) and 4.2 dB (NH) when compared to an unfiltered audiovisual condition without a face mask. The acoustic filter properties of the face masks accounted for up to 2.2 dB (CI) and 1.5 dB (NH) of the deterioration in SRT80%. This effect was therefore less distinct compared to the effect of a missing video signal. Face masks complicate the daily communication. These effects are already detectable in a cohort of normal hearing listeners 2 and even more pronounced in hearing impaired listeners, e.g. cochlear implant users.