The next batch of Videos are up!!!

[s3yang]

We're not there. It sounds unnatural in that you never really hear her take a breath. Also, I think they should take a good hard look at human sinus systems and larynxes (really, the entire makeup of everything from the neck on up) and learn to properly understand how sound interacts with everything to create a proper human voice.

Pretty sure it wouldn't take too much effort to get CAT scans (or the like) from famous singers... say Megumi Hayashibara for example? Pretty sure that they can get a large enough number of pop stars, voice actresses, and divas for something like that.

Once they do that we will be much closer to Macross's Sharon Apple.

And we will also be that much closer to a technological singularity. And I, for one, will welcome our new machine masters who have the voices of angels. ;D

I also suspect that we as English speakers are probably missing what might sound out of place to someone who actually speaks/understands Japanese. Because I know the English volcaloids don't sound even close to that good to me. It's also possible Japanese itself is easier to simulate, because they've got a very limited and distinct set of syllables compared to english.