That's part of tts (text to speech). It;s not audio files, it's instruction files for the tts software to turn the strings of characters into sounds. Maps sends the characters "turn" to tts, and it pronounces the word according to its rules (which often fail for names - Fayetteville NC has a lot of street names mispronounced - and mispronounced differently by different tts engines. Sometimes it's almost comical. (If there are 2 legitimate ways to pronounce a name, the phone has no way of knowing which way it's pronounced where you are. The same street name may be pronounced the other way somewhere else. How do you pronounce "Murchison"? Here, the 'ch; is pronounced like a 'k'. Maps, using the Samsung engine, pronounces it as 'ch'. If there were audio files for the maps, it would be pronounced the local way all the time - k here and ch where it's pronounced that way.)