do an updated version on how to get the cleanest version in vocals / in depth / all the recent stuff

[novel vine]

I’ve seen a lot of people requesting for guides and stuff. Especially in making ai covers / getting the vocals and what’s the best way. Any way we can get some running? Not sure if this would be in other changes or in sever changes but lmk

[final sky]

Vote for this suggestion!

[fathom cradle]

-audio

[final sky]

Suggestions for @novel vine

📚 Audio Guides

• Perfecting Audio Isolation on Low-End Rigs: A Practical Guide, by Litsa The Dancer and Faze Masta Google Docs
• Gathering and Isolating Audio, by SCRFilms Google Docs
• Creating Datasets for RVC using iZotope RX, by LUSBERT Rentry
• Vocal Mixing Tutorial, by Roomie YouTube

🛠️ Tools

Audio Separation/Isolation

• https://mvsep.com/
• https://multimedia.easeus.com/vocal-remover/
• MDX-NET Colab

More commands: -colab, -uvr, -karafan, -overtrain, -help

[novel vine]

Outdated I beileve 🗿

Yes outdated

Need them to be updated

[fathom cradle]

What is outdated about them?

What has changed since then?

[novel vine]

U don’t stack ai isolated models any more

You don’t use voc_ft

or any of those

And then the isotope outdated too

there’s now an 11 version out I beileve

Oh wait the first one kinda updated

[small iris]

everything

[rough loom]

actually I could write some of it up. I'll be giving the pass out so you guys can contribute

[fathom cradle]

If you know that everything is outdated why don’t you correct it

[novel vine]

Dawg he ain’t the one suggesting to update it 😭😭🙏🙏

Juss stating everything outdated mane

Somebody do it or some shit

[rough loom]

man, this came out looking like a high barrier of entry now that I can see what I wrote At least there's a consistent theme of telling people to stop damaging their audio https://rentry.co/RVC-dataset-RX11

info straight out of the server logs

[novel vine]

Well thing is I used this method and my models still came out ass

I literally followed every step

And was left with a very bad ai model

[rough loom]

The main point on the guide is that you don't have to use every method available and use the bare minimum on Rx. There's an extent to how much you can blame Hifigan ig. Like codename said, it gets way more advanced in the workflow for singer models. That means modifying the Rx settings to see what works best for his model and using additional modules (dehum-debuzz or deconstruct)

[rough loom]

It's still heaps better than the previous guide because it encouraged compression. It just assumes you lurked the server long enough without further context. Many of those models are out there now

[potent narwhal]

great job

[rough loom]

Yes I learned a lot from all of you guys at model maker chat. I feel like I graduated with my knowledge lmao

[potent narwhal]

very nice

albeit some stuff needs changing

[rough loom]

welp I don't know what happened to the modify key. If you want to help fill things out I can give you the edit keys in dm

[small iris]

hey I'm a big fan

[potent narwhal]

This you?