Unveiling the New AI Voice Cloner

Unveiling the New AI Voice Cloner | OpenVoice

Рет қаралды 20,144

Wingnut Labs

Күн бұрын

Пікірлер: 39

@esra_erimez 9 ай бұрын

1:15 you look positively chad in that frame.

@xbon1 9 ай бұрын

If you want one that's more uncanny you should use RVC, although it's mildly impressive what OpenVoice can do with just voice audio and not needing to craft an entire voice model around it.

@wingnut_labs 9 ай бұрын

Thanks for the tip! I’ll check that out!

@marcuswebb 8 ай бұрын

What is this RVC and where I can find? Thanks

@jaya7760 9 ай бұрын

Thanks for sharing this, I've been trying to figure out how to clone my own voice but I don't have programming experience and other options have been complicated and hard to find support for. The music you use while you intro has some sort of popping effect that makes it sound like you're having microphone issues. Also, it's overwhelming your voice.

@wingnut_labs 9 ай бұрын

Thanks for the feedback! I'll try to get that sorted for my next video!

@WackFPV 8 ай бұрын

I love you bro, that was well said, show how to leverage AI to make your workflows, lives and overall creativity more streamlined and enjoyable! shit gave me goosebumps! Also its the model. There's already models out there that sound really really damn good. You gotta learn how to finetune them, which i'm trying to learn rn.

@wingnut_labs 8 ай бұрын

Heck yeah! Let me know if you find anything worthwhile! I’d love to learn any models that perform well 💪🏻 thanks for positive feedback!

@paristo 8 ай бұрын

I don't really care about cloning a voice, or doing it real time. I care about a good Text-To-Speech capability with voice from chosen sex, age, accent and some emotional status (normal, panic, etc speed, pitch etc) that would make a voice that sounds more realistic. For a content creation where example you need to produce a voice for device to inform user "Door Open, Door Open" or "Warning, Warning, obstacle behind". etc, such TTS is just beneficial when you don't need to hire a voice actor or manipulate your own voice to get it wanted like.

@nick777-oo1wc 9 ай бұрын

For what I see, the first link is just a demo. Cannot use that, as its too short. And the second has 2 ipynb demos in it, bot none of them works. Any way around this?

@ElumStudios 2 ай бұрын

11:18 This part doesn't work for me, can you help me?

@farsi_vibes_edit 7 ай бұрын

I think you should upload the sound with more time, maybe the output will be better

@alexmitsou3452 9 ай бұрын

I just don't get it. I tried to set up the whole project locally, but there's no option to upload a reference audio. Why's that? Thanks

@frosti7 9 ай бұрын

Can it do real-time voice replacement?

@wingnut_labs 9 ай бұрын

Unfortunately it’s an analyze to synthesize type model, so it doesn’t modulate any voice input in real time. I do want to make a video on some of the ai voice changing programs I’ve been messing around with. Those ones do intact change voice in realtime

@focusedeye 9 ай бұрын

I'm more than a little concerned using a "public" system for free. Does this mean if I upload my voice, that my voice is now available for anyone to access? Gee...what could possibly go wrong?

@chancepaladin 5 ай бұрын

dang it. looking for one where I can clone my voice to narrate my books. limit 200 chars, ;( if I'm running it on my own system, should unlock, hopefully I can find something.

@cloudfashionmani 8 ай бұрын

thank you

@theGreatQAwakening 6 ай бұрын

THANQ!!

@Bliepzz 6 ай бұрын

will dutch belgium will work soon ai cover

@Bliepzz 6 ай бұрын

Warning The detected language pt for your input text is not in our Supported Languages: ['zh', 'en']

@videoeditoranimation1714 Ай бұрын

Dude. My clone sucked too. I don't get it. I can't seem to find a really good voice clone program. Even eleven labs was disappointing.

@khajask8113 5 ай бұрын

It supports Hindi language..?

@TheBunnetHypothesis 9 ай бұрын

Good stuff. Would love to know how to use this in a python program.

@mrquicky 4 ай бұрын

The model 1 version of the python program would execute without too much fiddling. Model 1 is pretty horrible for cloning, though the effort is made with a few exposed parameters. There is a hard limit of 200 characters for TTS in model 1. Model 2 requires an OpenAI API key with funded account.

@ourhomeidea 8 ай бұрын

thank you for the great video. Can this be used in Indonesian? Or is there a special way to change the language? thank you for your answer

@jasonfournier 9 ай бұрын

It is terrible at voice cloning

@wingnut_labs 9 ай бұрын

Agreed 👍🏻 especially with my voice it seems to do a poor job. It gets some samples I give it better than others, but overall I’d say I’m not super impressed with the state of the tech.

@pylotlight 8 ай бұрын

@@wingnut_labs Like the LLM leaderboards, we should have a TTS leaderboard for best quality/efficient to run list to compare self-hosted TTS models. Nothing I've seen yet comes close to 11ai yet unfortunately.

@wingnut_labs 8 ай бұрын

Heck. Yes....! I love that idea. That would be epic. @@pylotlight

@KEKW-lc4xi 8 ай бұрын

So we cannot robustly train this model? like I can easily just narrate for an hour or so if need be in order to produce an actually good model. because these samples given are pretty bad.