I use AVSpeechSynthesizer and AVSpeechUtterance for text-to-speech. It works fine, but the User must enable SpeakSelection feature in his/her device's accessibility settings. Is there a way to do it for them, or at least check the settings and warn them?
Maybe you can try to say something, attach delegate and measure duration of speaking. If duration of speaking out yours text is near zero, user should check settings/download voice or so!
来源：https://stackoverflow.com/questions/26181764/using-avspeechsynthesizer-avspeechutterance-for-text-to-speech-will-not-work-if