[deleted] Switch uses a speech recognition engine to generate a rough approximation of the words spoken in an audio file. Choose a engine and profile from the list. You can click on the drop down menu for "Output Format", and select the option from there.