I have searched and tried two suggestions, which are changing the voice model, and changing the transcriber provider. Both of those did not change anything.