Question ahout neuro-sama’s singing voice

[coral tree]

Hi! I’m a pretty new fan, and I’m really interested in the tech behind Neuro. While listening to the covers in her “starting soon” screens, it made me wonder what program Vedal & team uses for her voice because they sound really good. I know that the general opinion of the Neuro-sama community is against generative AI that takes from human artists, which I agree with, so that got me questioning how you create her voice. Is it an UTAU/vocaloid software thing? Sorry if this has been asked before. I checked the FAQ and couldn’t find an answer but I was pretty curious so I decided to make a post.

damn neuro… whyd you have to essay react me like that…

[white patrol]

I've been working a lot with copying Neuro's voice for everyone to use, and from that I have gained two main guesses; a likely one, and an unlikely one:
more likely Neuro's singing voice is some sort of vocal synthesizer using an open-source vocal synthesizer architecture, but not UTAU or Vocaloid, as Vocaloid is proprietary and not AI-based, and UTAU is also not AI-based, it's also not SynthV due to the cost of getting access to the SynthV SDK being in the thousands and only for companies
less likely Neuro's voice is some sort of vocal synthesizer combined with a SynthV voice bank, similar to how we use an RVC model and the SynthV Solaria voice bank to effectively get a near-1:1 Neuro voice

[coral tree]

this is exactly the answer i was looking for thank you so much!!!

[white patrol]

Working on cloning Neuro's voice kinda gets a person deep into voice AI tech

[quick spindle]

That wasn’t Neuro. That was the Neurobot, which is a primitive discord bot unrelated to her. Neuro’s in here but she’s not active. She only typed in chat a year ago during the Ironmouse Canvas War event.

Evil’s in here but she’s also not active.

To answer your question, yeah she uses similar technology to a UTAU/Vocaloid, and the singing is tuned by QueenPB, who’s a friend of Vedal and is a known Vocaloid producer.

[oblique mason]

Can you expand on your point "cost of getting access to the SynthV SDK"? Why does Vedal need access to the SynthV SDK? All he, but in this case he would delegate that task to QueenPB, so all QueenPB needs to do is to have a copy of "Synthesizer V Studio 2 Pro", which is ~$100 right now.
QueenPB would tune the song with SynthV with her preferred voice bank. Then they would use RVC to convert the SynthV vocal to either Neuro or Evil, then QueenPB would take that converted vocal for final song mixing.
Why would the SynthV SDK ever be needed? Neuro and Evil never break out into singing random sentences mid stream.

[white patrol]

The SynthV SDK would be for making a custom vocal synthesizer
I'm about 95% sure Neuro and Evil are not RVC

More likely Neuro and Evil are some open source SVS arch

[oblique mason]

That's new to me. Thanks for sharing.