A method and apparatus for generating speech through neural text-to-speech (TTS) synthesis. A text input may be obtained (