MMario --

The factor of accenting syllables is certainly part of it, but I have a theory that it also has to do with just simply the sounds of the words -- the vowels and the consonants -- that we hear when the words are spoken or sung (and what we use to distinguish one word from another), that are completely missing in the digital version of a tune.

Also, I don't know of any human who can sing (or play, for that matter) with the mathematical precision of a computer. We subconsciously shorten or lengthen the syllable of a word based on its context, meaning, or emotional impact. The time value of a dot on a score is only aproximate to our ears, but unvaringly precise to a computer. And those minute differences seemed magnified when they're overlapped in the brain.

So -- have any of you folks found a song in the DT that you just loved in theory (beautiful melody, powerful lyrics) and wanted to sing, but you couldn't get to work when the two halves are put together? And if so, how do you get around that?