I just listened to a podcaster complaining about a person who usually says “what?” the moment you stop talking. In this case it's a young person unlikely to have hearing problems. But maybe that “what?” person has a semantic issue that plagued me for a few years, a long time ago.
I think it's natural to assume that as we hear speech, we understand it. It's the speech-to-text computer programs, who may need five to ten seconds of mad analysis to understand the previous sentence, who are unnatural. But in fact I think it takes most of us some lag time to decode audible phrases, and the meaning of a sentence you hear might pop into your head a full second after the sentence is over.
The moment a person stops talking, you may (I certainly did!) find yourself saying “what?” because you know you don't understand it. You've scarcely finished saying “what” when you DO understand it, and now you have to be polite and listen to the whole sentence again, even though you know exactly what the speaker is going to say. The point? You may have to suppress your “what” reaction, give your brain another second to translate before you ask the speaker to repeat.