F0 synthesis based on a quantitative model of German intonation