#Questions about getting clean AI vocals, Using youtube dry vocals to train

1 messages · Page 1 of 1 (latest)

formal anvil
#

Hi I have some questions and I hope someone answers me
What is RVC Titan ? What's the difference between it and the regular RVC?

Is there any way to make the generated AI vocals sound less robotic in DAW ? I mean is there a plugin can do it and make it sound more natural?

Is it okay if I used vocals from youtube to train a model? I found Dolby Atmos stems there and there are lots of clean and dry vocals in these stems. But I can't get them anywhere except youtube and isolating lead vocals using ai models like those in mvsep.com doesn't always work

radiant garnet
#

TITAN is a state-of-the-art pretrained model designed for RVC training. It offers a robust solution for transforming voice characteristics from one speaker to another, providing high-quality results with minimal training effort. It was fintuned with RVC v2 original pretrains with pitch guidance, Rmvpe, leveraging an 11.15-hour dataset sourced from Expresso at 530 (48k rate ver), 467 (40k rate ver) and 463 (32k rate ver) epochs each variant.

48k is the variant with the highest sample rate but it may produce artifacts and errors, 40k has a medium sample rate but its meh, and 32k is the most stable variant for almost all models.

formal anvil
#

it's not too robotic it's good ..It's just thai I hear like 1% robotic voice and 99% normal voice and I wanted to get rid of this 1% 🙂

#

thx tho

#

oh ok

#

so there are lots of pre trained

#

models

#

I want to know .. for a female singer voice

#

what should I use ?

#

Oh I get it

#

then which one has a better quality ? 😅

#

I read them but I kinda didn't understand🤣

#

ok .. thank you so much