Macro-rhythm was introduced as a means to distinguish typologically different languages according to their overall tonal rhythmicity at the phrase level. The concept of macro-rhythm has been established theoretically, although little work has attempted to quantify it acoustically and led to mixed results. More research is needed to obtain an acoustic approximation of macro-rhythm using more diverse types of speech from different languages. This is done in the current paper by investigating phrases taken from TED talks in Greek, German and Portuguese, languages that are assumed to differ in their degree of macro-rhythm. The current approach takes the length of a stylized and speaker-corrected intonation contour as a core measure of macro-rhythm. Results show that this method indeed ranks the three languages in the expected way, although more work capturing tonal regularity is needed to improve the outcomes.