I'd work with NNSVS more if it wasn't such a setup to get it started every time. DiffSinger is more convenient as i don't need to run a "server" to do it, but these free AI synths still take a long time to render anything. I don't have the most high-end hardware, so that's probably my fault.
I don't get how I'm supposed to tell DiffSinger what language it's supposed to sing in. In the end, classic UTAU are still the easiest, but remembering all the unique flags present with all the different resamplers is a chore. I wish OpenUTAU already knew what all the flags were for each one, but that's probably a lot of work.
I know that I should probably be using SynthV lite, but I don't want to work with watered down versions of the Pro voices. This really is an expensive hobby.
By the way, is it really true that Vocaloid2 voices sound different than intended in Vocaloid3/4? That sounds like an issue to me. Do the old V2 voices come with the editor? I would like to someday get Luka V2, but I don't want to risk getting the voice and not the editor to use it in.
I don't get how I'm supposed to tell DiffSinger what language it's supposed to sing in. In the end, classic UTAU are still the easiest, but remembering all the unique flags present with all the different resamplers is a chore. I wish OpenUTAU already knew what all the flags were for each one, but that's probably a lot of work.
I know that I should probably be using SynthV lite, but I don't want to work with watered down versions of the Pro voices. This really is an expensive hobby.
By the way, is it really true that Vocaloid2 voices sound different than intended in Vocaloid3/4? That sounds like an issue to me. Do the old V2 voices come with the editor? I would like to someday get Luka V2, but I don't want to risk getting the voice and not the editor to use it in.
If the AI version "shouldn't" sound realistic, then what is the point of upgrading the synthesizer? Why even develop a new renderer? What do you expect AI to be? This isn't genAI we're talking about, it's SynthesizerV and it's equivalents. They're still as much of an instrument as Vocaloid is.
You just seem to be very against SynthV for no reason whatsoever. Is this a cause of brand loyalty/nostalgia?
In the end, we still have to make the background track ourselves, so why force the musician to add more lengthy processes to their work? Isn't the fact they made a whole song already artistic enough? This isn't about cover tuners, you know.