[HarryT]

One thing I find interesting with Voice Dream is that the "premium" Acapella voice I use with it (the female British English voice "Rachel") reads so much better than the built-in Siri voices. Not in the voice quality - there's nothing to choose in that - but in the actual reading. The Siri voice is a flat reading voice, whereas Rachel changes speed, volume, and pitch to match the punctuation of the text. You can clearly hear in the reading when a sentence ends in an exclamation mark or a question mark, for example, just as you would with a human reader. It seems that there's a lot more "information" stored in the Acapella voice than the Siri voices to allow them to do this, and presumably the same would also be the case with the Ivona voices.

I wouldn't claim that it's on a par with a human reader, but it's certainly the best TTS I've found so far.