No one trained voice model from this actor yet. And to train a voice, you need to find good samples. For the game ones, you can find samples in any game files, without background noise, music, etc. But for the cartoon one, you havd to filter the voice, and is not easy ( in some cases almost imposible to get good audio samples) ps: unless you talking about new sonic actor for tge games LOL. If so, is simply that nobody trained it (yet)
Well, I found a model for jaleen a few minutes ago... but the method used to train was an old one (svc) so is incompatible with the new rvc, that is in current use :(
This...needs work. I could not understand more than maybe three words in the song. Each one had the same problem. Also the ai voices in this one are sounding way to strained.
Is the actual technology ( and by that i mean the ai training avaliable). Things are going to sound better, with every update ( for example, last update bring a variation of crepe called "mangio", that make ai voices ( both training and make them talk or sing) sound way better fhan before. The othe point nobody talks about is: the majority of AI covers you see on youtube is from people who take other creators trained AI voice, and the voice was not well trained. My videos are made with voices i trained myself, and i still still learning. But yeah, things are going to feel better in the future.
@@hikarushepard Thanks for the info! I was pointing out that stuff so you could try and train the ai out of doing such things. I am not sure how training ai's work. As most ai i have seen have to have a source to mimic. Example an ai robot arm cut fruit in the same way a swordsmen had. by watching via a camera. It took 3 to 4 try's for the ai to get it right. You are essentially teaching it three things at once. 1) the voice 2) how to recrate a song in said voice 3) sing with voice with a new song plus new genre. Keep up the good work! 👍🦾
Oh these are awesome!!! They all sound so cute!!! Especially the ryan one//^//^///
Jason and Ryan Versions both sound similar to each other, AND each sounds at least 50% Jaleel White.
Way… past…COOL 😎🤩🤟
Awesome, what about some Sonic underground songs with the game sonic's 👀 i think that would a nice twist
omg I never thought of that! That's a great idea! :D
YES
🎉🎉🎉🎉
I love this stuff, but how come there is no Jaleel White version?
No one trained voice model from this actor yet. And to train a voice, you need to find good samples. For the game ones, you can find samples in any game files, without background noise, music, etc. But for the cartoon one, you havd to filter the voice, and is not easy ( in some cases almost imposible to get good audio samples) ps: unless you talking about new sonic actor for tge games LOL. If so, is simply that nobody trained it (yet)
@@hikarushepard Oh.
Well,
I found a model for jaleen a few minutes ago... but the method used to train was an old one (svc) so is incompatible with the new rvc, that is in current use :(
This...needs work. I could not understand more than maybe three words in the song.
Each one had the same problem.
Also the ai voices in this one are sounding way to strained.
Is the actual technology ( and by that i mean the ai training avaliable). Things are going to sound better, with every update ( for example, last update bring a variation of crepe called "mangio", that make ai voices ( both training and make them talk or sing) sound way better fhan before. The othe point nobody talks about is: the majority of AI covers you see on youtube is from people who take other creators trained AI voice, and the voice was not well trained. My videos are made with voices i trained myself, and i still still learning. But yeah, things are going to feel better in the future.
@@hikarushepard Thanks for the info!
I was pointing out that stuff so you could try and train the ai out of doing such things.
I am not sure how training ai's work.
As most ai i have seen have to have a source to mimic.
Example an ai robot arm cut fruit in the same way a swordsmen had.
by watching via a camera. It took 3 to 4 try's for the ai to get it right.
You are essentially teaching it three things at once.
1) the voice 2) how to recrate a song in said voice 3) sing with voice with a new song plus new genre.
Keep up the good work! 👍🦾