Dreamtonics dropped some new demos of their two new AI voices: Yuma and Natalie!
Yuma is a new masculine Japanese database that’s described to have a young, fresh texture and an androgynous voice
Natalie is a new feminine English database that’s described to have a soft, clear voice with rich expressions in the mid range. She’s suitable for easy listening and music with strong backing tracks like EDM.
Both of them will be compatible with cross-lingual singing synthesis, vocal modes, and AI retakes.
We currently don’t have a release date as of this post.
Basically, people were commenting that his name shares similarity to VY2 “Yuuma / Yuma” (which isn’t pointed out in the article, but it’s obvious and there’s still comments dropping the name) and Dreamtonics said that because their new vocal had an androgynous tone, they wanted a name that was easy to distinguish. When investigating other products to make sure they don’t have similarities, they failed to take codenames into account (names like “Mizki” and “Yuma” for VY1 and VY2 for example), so they ended up in this situation. Dreamtonics said they’ll pay more attention to names of other voice synth products in the future.
Also any rude and uncivil comments were removed to maintain a friendlier atmosphere for discussions.
EDIT: See @LinR_PN 's reply; apology was for the behavior.
Correction: there were a few very over-the-top accusations when the video was first posted, those comments were removed, and Dreamtonics is apologizing for the behavior.
We were given such high quality voices with Weina and Solaria, it makes me feel a bit lukewarm for these to be honest, I’d much prefer either a ‘gold’ level male Solaria/Weina, or at least a different vocal style.
Hope other people are more enthusiastic for these voices and that they at least break-even in terms of revenue.
While the quality is slightly lower than that of Solaria and Weina I don’t think it’s enough to “matter” for many producers. Of course there are those who are looking for the absolute most realistic vocals possible, but imo the quality difference isn’t so big as to put off vocal synth fans who are accustomed to competing engines like Vocaloid or CeVIO.
Put differently, I think many users will be more happy to see increasing variety in voice types than they will be disappointed by the quality difference.
I don’t think they’re different or impressive enough to pick them up myself, but I’m just a casual hobbyist that hasn’t actually finished a project in ages
Anyway, as for the voice types I think Yuma works really well as a high masculine vocal, and I can immediately imagine him being used for a number of genres.
I find it a bit more difficult to imagine what sort of songs Natalie would be suited to, but perhaps that means she’s suited to things we don’t tend to see in the vocal synth space, which would still be a positive thing if true.