It sounds like they worked out some formula for what list of sounds and syllables to have the real singers sing in order to get all the right raw materials to work with. The control over time alignment, softness/assertiveness, mouth opening, etc. is pretty staggering. And it seems like where ever pronunciation was strange they could fix some of it. What I didn’t see was how effectively the mouth open parameter can shape various vowels and phonemes.

$100 for the app plus $70 per voice actually seems pretty reasonable to me. BTW, I just checked and there are 8 different English voices, 8 Japanese, and 8 Chinese voices.

I get the impression that this software could generate a convincingly realistic vocal track, but it would take effort– on the editing side. What goes unsaid is that you could also create some nightmarishly fun avant garde sounds.

the owner of the studio where i worked used to compose music for this kids educational company, and he was always busy with some project because they were always releasing stuff

he wrote and composed all parts (including keys, woodwinds, brass, drums, etc. etc) using a MIDI guitar… then he would export the tablature from logic and have actual session people come in to record their parts

id imagine this software is for people like him

Ok, really only half-serious with that comment. Guaranteed there are folks who will do amazing stuff with this tool, and hey, maybe fewer vocalists will be injured during any dangerously monotonous vocal recording sessions.