I’m interested in Cantai but I’ve wanted to skip using it while the quality improves and it gets more reliable.
I’ve had a listen to some of the Cantai demos on Richard’s page but there’s no way of telling AFAIK which version they were created with.
The Mozart Confutatis in some ways sounds lovely in others pretty awful IMO. It uses Ottetto so I’m guessing it’s quite recent. Does anybody know - is that rendering pretty much as good as it gets at the moment?
Does anyone have any renderings of classical choir pieces generated with the latest version posted on the web anywhere? If so, I’d love to listen and I’d be grateful for the opportunity.
Do people find it’s now reliable and predictable to use?
I’ve listened to both those, and to be fair to Cantai, it’s an impressive achievement.
I think, though, that for me Cantai could act now as a good proof reading tool but not as a way of creating renders to share. It flattens the emotions in the music too much, IMO.
Cantai remains more of a toy at this point than a professional tool. Dynamics are not consistent with dynamic ranges in Dorico (I suspect its the same in Sibelius and MuseScore). There are still issues with voices acknowledging and singing the specified rhythms. There are still issues with pitch when moving by intervals - i.e., sliding into pitches rather than clean movement to a new pitch. Rendering will be much better with v 3.0.2 where the new GPU rendering significantly improves rendering time (up to 300% in my tests). Hopefully, the key issues for professional use (dynamics, pitch, and rhythm) will be addressed soon.