i was originally using sfspeechrecognizer then switched to whisperkit. neither seem to be great. maybe im doing something wrong? feels like it should be an easy fix. please help