The production of speech by a computer, by putting together sounds that are created either algorithmically or from recordings of human speech.