*Made an account to talk about this hah.
In regards to it being an untrained singer, it's true, but also untrue. The singer had to be able to hold a pitch for the most part, and actually be able to sing to even be allowed to record for emvoice. Rodolphe also had some singers the reclist multiple times until they were within 1/16th - 1/8th of the pitch of the note they were recording at, and for any errors, they were seemingly fixed with Melodyne. So.. it may be an untrained singer, but it leaves out the fact that they were made into a 'trained vocal' (odd wording, but I hope it gets the point across).
I do also agree that Rodolphe was very... cocky for lack of any other word in regards to how he described his own product which currently sounds like it's being tuned to correct pitches manually. It sounds like a program that is being rendered on one pitch at the same tempo/time for each note/word, and then altered in something like Melodyne to be on pitch/time.
Also, it seems that at least the female speech vocal was recorded at the singer's lowest pitch, and only one pitch, which... isn't good, especially when you're using that singular, lowest pitch to show off the product. They also sound like they may have been sick when recording... but that could just be because they were recording/the voice was rendered with their lowest pitch.
And in regards to the pitch thing, has it been confirmed by anyone that their vocals are multi-pitch capable yet? I know Rodolphe has said they would be (I think) but as this is supposed to come out early 2018, I would hope that the multi-pitch is already a component that is ready and usable...
Finally... the fact that no one has talked about the fact that it's just a VST/AU... I know CHIPSPEECH and A/E did alright/well in sales, but... I don't really see how he thinks a VST/AU plugin would sell as much as a standalone, especially when sold without a DAW. - Also the fact that the VST isn't even made yet? He said it was originally JUST going to be a MAC AU, but now he's adding VST capability, which means they'd have to rebuild it from the ground up for a VST (because of a few distinct differences between VST/AU), so it seems like it may be delayed because of that, or they'll pull a VoxWave Leora and release the product without it being fully done (*without the VST version).