Here’s an example of a sentence where the text-to-speech has the inappropriate stress. The most persistent problem, as I tried to indicate earlier, is between words (rather than being within words). For instance, here clozemaster gives us the key emphasis on the second syllable of ‘manca’, which may be correct, but it elides into a less-stressed ‘marul’, so that we hear something like "CAM-a-rul’. I believe that the first syllable of ‘marul’ should in fact also receive stress, with less elision between the last two words. We should have something like, in my inexact phonetics: mun-cah’ MAH-rul.