hi, i just download and installed it all .. but idk how to make a stems... theres no kind of button to separate stems... theres no manual or something that may i missing?

EDIT1: so first, i figured out the models are the converters/separators? so, i upload one song that i want to take the stems separated. but i choose model of kanye west for example, and the final result was two files , wav and mp3 of converted vocals, the vocal preview mp3 file its not working. and the final files wav and mp3 works fine but i though wrong cause now i have a kanye west singing instead of drake for example.. so i dont know how to use this correctly, but this its working nice. my cpu its a ryzen 5 3600, rarely touch the 70% of usage, at some momento took some ram , about 4gb but not always... but how can i make stems separation of any song that i want? thanks in advance, i hope i dont sound bad at all, im not a native english speaker.

 

Last edited: Nov 10, 2023

hi there curtified! thanks for all you have done. is there anyway to put the download on another drive?

 

whats is free, then gets paid

also,

 

Works well without problems

v1.81 here

Good Job @curtified

 

Very Cool! I'd like to use this on an offline computer.The program won't go past the models download,even if i have manually installed a model

 

I love it :D

 

Huh?? Confused.

 

Hmmm..

let me try to fix this

 

I haven’t done this myself. But you should be able to if you copy all the resources. Or possibly use a symlink.

if you figure it out please let us know.

 

There should be a stem separation switch in the settings.


If you click on the song you rendered in the left hand side it will show all of the stems and vocal conversions.

 

I waiting




Thanks for your job

 

this is the best!!! ..

 

Sorry if these are dumb questions, but as a total newbie to this sorta thing its all a bit confusing tbh.

Is there a step-by-step "how to" installation and basic operation manual or reference guide somewhere?

My Windows 10 PC is offline, will it work without needing a internet connection?

I only use the stock GPU for my i7 12th gen CPU, is it OK?

As this is standalone program, how to sync everything with the project BPM?



*I downloaded the program and also 5 models. Now what?

-install program (custom folder?)
-add models (where?)
-any other steps before it all works?

Now how do I actually use the program - type words and/or record my vocal to audio in my DAW (with BPM matched to the song it will be used in), then drop the vocal audio file into the Replay standalone app and clone/process/replace with the model voice, render WAV file, drop into DAW etc., or something like that?

And is it the same process for all audio cloning ie instrument sounds?

 

Last edited by a moderator: Nov 11, 2023

I used a symlink and it works!

 

It works pretty well as a vocal replacement, but sounds absolutely nothing like the people modeled. sounds more like it just alters the pitch of the vocal.

 

I think I understand what you're trying to say. In all cloning methods, you will encounter the same "issue." You have to consider that each singer has a unique style, character, expression, etc., when singing. So, if you try to sing like your favorite singer and then clone it, it won't sound exactly like them. Why? Because you add your own nuance, your expression; it will sound like a mix of both. For example, if someone has a voice very similar to singer "x" and then clones the voice, it will sound very similar. However, you can do interesting things that are actually available on the web, like imagining how "x" singer would sound singing another singer's song. It's pretty "crazy" to hear the different interpretations. Try to mimic as closely as possible the voice you want to emulate/clone, including their expressions, melodic turns, whether they sing loudly, softly, etc.

 

does anyone have help on how to get the trainer installed locally?
git clone https://github.com/ddPn08/rvc-webui.git
then run webui.bat
but webui.bat gives errors and cant finish
python is found
torch is installed as the bat excutes
mc builder is installed

may i ask why it does not fininsh? does anyone have a manual or video to follow?
btw Chad Kroeger from nickleback sounds great replacing Michael Poulsen in volbeat songs

ps. is Harvest the best FO ?

 

try running it as a virtual environment:


cd rvc-webui
rm -rf venv
python3 -m venv venv
source venv/bin/activate

sudo ./webui.sh -f --share

 

Thank you for this!

I want to note the same goes for instruments.

Also try different pitch algos. If you are a male trying to sound like a female model use the pitch slider etc etc.

 

glad to see it did!

for all my mac users trying to make symlinks with a GUI:
https://www.macupdate.com/app/mac/41493/symlinker