FREE AI Voice Tool - Best Open Source AI Text-to-Speech is out!

Bark is a transformer-based text-to-audio model created by Suno. Bark can generate highly realistic, multilingual speech as well as other audio - including music, background noise and simple sound effects. The model can also produce nonverbal communications like laughing, sighing and crying.

❤️ If you want to support the channel ❤️
Support here:

[nekoeko]

I'm at the point of not having time to test every new AI model that comes out. This one is amazing and it had slipped under my radar, thank you very much

[giustex]

This is amazing 🤯. And as always, thanks for the quick video

[jersainpasaran]

really impressed! even all the realistic voice options are restrictive, this changes everything!! Thank you for detecting these advances.

[MrDevidu]

This is the first open source text to speech I've heard of. Nice.

[AIEverLife]

Abdul... That moment when you dubbed the smile/laugh on your cloned voice😂😂... I enjoyed the video as always and had a blast of laughter as well...

Eid Mubarak... Get going

[VRDynamite]

ohh it's getting exciting! they sound good but the echo on each one takes away from just how impressive this is! cant wait till someone makes a fancy UI for this!!

[anonymousmuskox]

Can you make a video on how to install this locally?

[joannot]

That's 🤯🤯🤯
Especially the laugh!

[alicem]

Omg, these are like something out of a fever dream. You get wildly different results every time you generate it. I'm sure that's terrible if you were trying to use this seriously, but the silliness of it is very entertaining. Most TTS fail because they sound too robotic, but this one sounds absolutely nuts.

[Mirsab]

Perfect timing! I wanted an free/open source alternative to elevenlabs, was literally intending on getting the paid version of elevenlabs.

But you mentioned it's not available for commercial use, is that just for the voice cloning or everything in it?

[anamomo]

I'm in search, text to AI finally .. Big Thanx

[tradingwithwill]

Got torch version install errors in the colab

[michaelberger]

Thanks for what you do! I always look forward to your videos.

[jayhu]

I am very glad for find your channel. I am interest in Text to Audio. So as Whisper and Eleven labs. But this one is great because it is Open Source.
Hopefully you make a streaming to make this run with code examples.

[VincentVonDudler]

Can you speak to what hardware you're running it on? ...and how long the process took? How long would it take to generate an audio version of a news article or an audiobook of Sherlock Holmes?

[stavsap]

this is amazing for indie game developers, they can interduce lots of voices this way into their game, is there some guide how to fine tune the voice? some of them got lots of noise in them

[saito]

Idk how to feel about the fact that they decided to limit the voice cloning API.
Would be cool to see a comparison between this vs Elevenlabs and Tortoise

[serta]

Cool stuff. Also Never take a call from an unknown number anymore! 😅

[Tron]

Bro, it's only allowing for a max of 14 seconds audio output, is the the commercial limitation you're referring about?

[scottmiller]

The prosody is a lot better than the other open source TTS codes out there, but the sound itself has an odd synthetic quality to it - subjectively, it feels like there is something spiky in the voices. It's a little like low resolution quantization, a little like a comb filter. It's definitely easy to tell it from human speech, and it's not clear if I would listen to it for an extended period. Also, that text rendered as music hurt my soul. On the whole, however, it sounds like a "two more papers down the line" problem - this will get fixed up soon.