I got UTAU running on my ancient Mac:
This is essentially Animenz’s arrangement of “Sis Puella Magica!,” with “Camila Melodia” doing the the quasi-vocals. She just sings “na na na” because I didn’t have the lyrics at hand, and I didn’t manipulate the sound beyond changing the syllables from the default. The harp and synth were added in Logic.
So UTAU works, sorta. Making it sound like music is another matter. Not the least of the challenges is that the GUI is in Japanese, which I don’t read. Fortunately, the manual has been translated into English with numerous screen shots; unfortunately, the interface for the Mac version was significantly revised, so many of the instructions are no longer quite correct. ((There does exist a hack to convert the UTAU GUI to English, but it looks like it only works in Windows.)) The range of topics in the manual suggest that the skilled user has a great degree of control over the sound. I’ve heard some UTAUloid recordings that are much more lifelike than the above, though none are as convincing as the better Vocaloid efforts. It should be possible to make Camila sound more alive and intelligible, though that probably means tweaking multiple parameters for each and every individual syllable.
It might be less trouble to just take voice lessons and pitch-shift recordings as necessary.