✨│ai-help

[limpid trout]

It appears that my wallet is empty,

Well I do want to Train My own voice
Talking seems too hard

Or use Audio
To upload in local voice changer to change the voice to appear different character

For videos

[hallow thistle]

Your queries still confusing.

Paid request and commission aren't allowed in this server, so you can keep money in your wallet/bank for something better.

[limpid trout]

Ok,
Here's my query

I want RVC voice changer for locally

That has option to any upload Voice models

And ways to Train models (by myself)

And you can suggest me 2 or 3 voice changers or rvc's as well

Realtime Voice changer isn't necessary

That has Both TTs and Audio upload to convert

[hallow thistle]

How am I supposed to decipher these? Do you mean something like an RVC (retrieval-based voice conversion) that could train a voice model and also do AI cover (inference) at the time? Because "Applio RVC" literally has both features I mentioned, with "TTS" a feature in Applio RVC. The retrieval-based voice conversion and realtime voice changer, while both use the initials "RVC", are not the same thing.

[limpid trout]

Yes RVC I mean Retrieval Based Voice conversation...

Guess he gave up

[hallow thistle]

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[hallow thistle]

When I took a break from typing for a few seconds, it doesn't always mean I'm done and giving up, and in this case, I simply didn't know what to respond.

[ivory trout]

I fixed I just had to adjust the settings a bit

[hallow thistle]

[rocky dove]

Sup guys, have a good new year! I'm new on this server, thanks for having me here. My goal here is to seek learning, share knowledge in artificial intelligence, meet people, contribute to open source, and help people with problems with artificial intelligence.

I start learning Artificial intelligence. I have some knowledge in Python, databases, cloud and etc. I looking for some good recommendations and advise for study. If you ask me "which area of artificial intelligence you prefer?", I would say: AI applicability and Machine learning.

It's pleasure to be here, and nice to meet you all!

[regal gate]

Hi, what would be the best program to emulate an anime voice?

[slender grotto]

is vonovox the better version of w-okada or and is there really a difference?

[sonic sand]

Hello how to fix the 20 sec delay ?

I have no delay only in the default voices

[viral mason]

When you say default voices what do you mean

You're most likely using an old program if it came with voices pre installed

[sonic sand]

i am using Voice changer Client demo from github

[viral mason]

Yup that's old

[sonic sand]

its hard to use i got it from a youtube video

[viral mason]

What gpu do you have

[sonic sand]

RX 6600

what program do u use ? where can i install it ?

[viral mason]

Should switch to wokada tg fork, since u use AMD use this download

https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip

I'm assuming you have a virtual cable already from the YouTube video

VB cable

[sonic sand]

i dont think so, i have voicemeeter does it work ?

[sonic sand]

ty

[viral mason]

It could yea

In case not tho I'll get you one

https://software.muzychenko.net/freeware/vac470lite.zip

[sonic sand]

i installed the program on my hdd should i install it on my ssd ?

[hallow haven]

Hello, i have a problem with the installation of Okada in the guide, I have a message who say "pkg_ressources is deprecated as an API [...] The pkg ressources package is slated for removal as early..." I need to instal something ? I have a link with this message

[viral mason]

Idk what that means

[odd flower]

Please help guys, its been 2 years and i am not able to make a good RVC for my voice.

[digital thicket]

Can somebody help me to create a rvc for my voice?

[void flume]

In case people wonder: I am currently waiting for staff to reply to a question about the rules I had, due to namari's moderation decision. In the mean time I'll randomly answer recent help forum posts with what I know. (might as well do something useful.) -- I have not ever set up an environment where I can train models in due to lack of having an expensive GPU and I do rather do things locally, but I can answer stuff when it comes to using some of the existing AI tools. I apologize if my attempts to help are a bother and you rather have an official helper attempting to help you.

[silent stratus]

how much of an impact do transients make on your model

like for example if im training an acapella and the artist agressively compresses his vocals would it turn out way worse than lets say i had the same acapella but with no compression

[silent stratus]

i got u

https://docs.aihub.gg/rvc/cloud/applio-colab/

this is the easiest way for beginners

[viral mason]

kaggle is better

[silent stratus]

ye but its harder for new people to learn

[shy ember]

does Vonovox work better than Okada

for real time v oice

also , i cant find setup.bat in the folder for Vonovox

can u hel.p plz

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 4060 8gb vram desktop)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a legal, safe & ethical community, we will NOT provide help for:

• ANY illegal activities.
• NSFW/Porn.
  Requests for these topics may be ignored, not helped and result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.
• Don't Ask To Ask.

[shy ember]

Okay

I guess not then lol

whats a good index rate to have

[hallow thistle]

Tg Develop's W-Okada fork and Vonovox can be "competitive" because while Vonovox can give better audio quality than other W-Okada versions, Vonovox can be less familar for most people while W-Okada still remains easier to get around.

[hallow thistle]

Not always. While Applio RVC can be harder to get around than those online websites that use "RVC" (Google Colab and Kaggle don't count, though they are known for running Applio RVC on cloud), Applio RVC is better at voice model audio quality where some other online websites (like Kits.ai) don't offer the same level as the former.

[tame oracle]

its just different kinds of hard drives, an HDD is a spinning disc drive and a SSD is a much faster solid state drive

[hallow thistle]

HDD stands for hard disk drive, a more bulky and heavier drive that has spinny disks, storing data on each platters. SSD refers to solid state drive, this stores data similar to a hard drive but uses flash chips instead of spinny disk.

[shy ember]

how do i fine tune the model to sound really good

its cutting out and not clear sometimes

when speakin in Vovonox

sometimes grainy

[hallow thistle]

If you install an operating system (like Microsoft Windows) on SSD, it will be faster than a spinny hard drive because SSD has no moving part to pick one data, whereas hard drive has moving parts where it has to pick data from a platter.

[hallow thistle]

See my message.

[shy ember]

ive went through all the docs lmfao

[hallow thistle]

[shy ember]

I have a 4090
Windows 11
Voice sounds grainy / doesnt pronunciate things that good sometimes

[hallow thistle]

When you mention "how do I finetune the model on Vonovox" after saying "does Vonovox work better than W-Okada", it felt like some steps being skipped, so I'm not sure what to answer.

[shy ember]

well ive gotten better luck with Vonovox so im using Vonovox

I went through this guide

https://docs.aihub.gg/realtime-voice-changer/local/vonovox/

[hallow thistle]

You can send your Vonovox screenshot to here now.

[hallow thistle]

Most of the time people couldn't open those .zip.001, .zip.002 files because they either used the built-in Windows archive extractor instead of WinRAR or 7-Zip, confusing ".zip.001" file as simply ending with ".zip" because they had "Hide extensions for known file types" enabled, or didn't read the guide as reference. To much the headache, I had to provide some solutions for those paradoxical problems

[shy ember]

https://discordapp.com/channels/1159260121998827560/1274015763501416490

this is the model im usiong

[shy ember]

are u saying i didnt extract it right? im confused

[hallow thistle]

I'm talking about Tg Develop's W-Okada fork, not Vonovox. If you have attempted to extract the Tg Develop one before, either one from my statement could be the cause.

[shy ember]

oh no im talking about Vonovox

[shy ember]

its sounding grainy on Vonovox im not using W-Okada anymore. Unless you think that would work better?

[hallow thistle]

I'm more familar with W-Okada voice changer. For Vonovox, I never actually used it as the main voice changer, though I open the program as a visual guide/reference where words alone couldn't imagine.

[shy ember]

Okay, ill try the fork.

[shy ember]

given my specs, i would get the AMD64-cuda.zip.001 or .002??

[hallow thistle]

Now that I open both W-Okada fork and Vonovox. On your Vonovox, what about the "Block size"?

[hallow thistle]

"voice-changer-windows-amd64-cuda.zip.001" and "voice-changer-windows-amd64-cuda.zip.001" are split zips of a single zip file, so you "must" download both.

This is b2397 W-Okada.

[shy ember]

why is one of the zips a .002 f ile

or is that normal

[hallow thistle]

Ok, now that I predicted the common potential issue.

[shy ember]

derp

[hallow thistle]

With "Hide extensions for known file types" disabled, these two files should look like this.

[shy ember]

there we go

now i extract in that folderr

[hallow thistle]

Open the .zip.001 one and extract to somewhere like "C:\Users\cybor\Downloads\MMVCServerSIO".

[shy ember]

ok got it running

[shy ember]

my processing unit should be my gpu correct?

[hallow thistle]

Yes.

[shy ember]

ok its all setup, now how would i fine tune?

this is what my performance looks like

[hallow thistle]

The term "finetuning" is commonly used in training an AI voice model (like Stable Diffusion), and it's rare for someone to take "simple settings" as "finetuning" on W-Okada.

[shy ember]

so its hopeless to keep trying this model?

[hallow thistle]

I'm trying to provide more answer, so don't hurry.

On W-Okada:
Chunk: around 40 - 60 ms
Extra: 2.7 s
GPU: NVIDIA GeForce RTX 4090
Pitch detection: rmvpe

Input: microphone
Output: Line 1 (Virtual Audio Cable)
Monitor: your headphones/speakers, this one is optional

[shy ember]

is virtual audio cable needed

ive just been using my mic

oh wait nvm

sorry

heres my comparisons

[south vortex]

hello, does anyone have a guide on how to use RVC on discord mic on linux?

[tame oracle]

that's a better explanation

i like da spinny disks even tho SSDs are better

[digital thicket]

Well idk how to sound my voice good without being sound ugly yk what I mean

[autumn cedar]

Does anyone have any experience with German-language voice models for inferencing? Do they get German pronunciation, vowels, consonants, etc. correct?

I suspect the answer is "yes" but I just want to discuss this a little

[outer valley]

Hello chat.
I just wanted to ask nicley after 10 months if there is any new update or delevopment for live voice changer. I used "Modified Okada (b2332)" but i got told by AI there is something new and better named "Applio" ? That even needs newer models and can better work with emotional speaking like laughing and similar?

[digital thicket]

Searching this 2 lol

[outer valley]

I tried models from like 1-2 years ago and they were really bad from quality.
But the model i mixed together like over 1 year ago did sounded pretty good with english and german talking.
But i think with current new builds and development there can be even better stuff? I just got back into it and want to try some new stuff out myself

[hallow thistle]

Applio RVC (RVC stands for retrieval-based voice conversion) is not a realtime voice changer, though there's a realtime mode in version 3.6.0. Applio RVC is typically used for simple AI cover (inference) and even train an RVC voice model.

There's a newer, continuous W-Okada fork version developed by Tg Develop, the latest version is b2397. There's another voice changer named Vonovox, a complete alternative to other W-Okada versions. Vonovox can be competitive at audio quality but its interface and settings can be less familar.

[hallow thistle]

What are you looking for? W-Okada voice changer or Applio RVC?

[outer valley]

so Gemini again told me half baked truths?
It was like saying Applio is New and improved successor to Okada in terms of Realtime conversations....
And that even my old models would probably sound a bit better in it or i should make a new Version 3 model ?

[outer valley]

Are this forks you named focused on real time conversation and better in quality then applio?
I not mind if they look bad from the menu or need some setup.
Important for me is quality and anything that helps with the old problem from 1 year ago with emotional talking breaking the models and similar.
I have a 4090 so i can train myself (even if i never did that before).
So what would be the best direction for me for my goal ?

[hallow thistle]

What? You can't always rely on Google Gemini for anything, what Gemini says to you might be of outdated information surrounding Google search results. By the way, the latest version of RVC voice model is still RVC v2; the RVC v3 is a hypothetical one that hasn't yet been a thing.

[outer valley]

Yes i know that many stuff is outdated or not always total correct from gemini and similar chat ai's.
Reason i got back too this chat and asking the experienced humans how much of what gemini told me is true or not.
Because the talk about applio and how gemini made it sound hyped me a lot

[hallow thistle]

This is b2397 W-Okada. https://cdn.discordapp.com/attachments/1159290139609137264/1457950706500763809/image.png?ex=695ddde7&is=695c8c67&hm=2ee979361639eb78c99688b54bd4c04ff0f5cf2c67d325af567636049db15296&

This is Applio RVC v.3.6.0. You can see one tab labeled "realtime" there. https://cdn.discordapp.com/attachments/1159290139609137264/1455545707204182016/image.png?ex=695dafd3&is=695c5e53&hm=533bce1d646a77e23915599c99c404eb81538a5f5d6e7ddc47093a40ce45a33f&

For actual realtime voice changer, it's best to go for W-Okada voice changer fork or Vonovox, where Applio RVC's realtime mode might not always be reliable for realtime task.

For training a voice model, Applio RVC is preferred. Other voice changers I mentioned can only do inferencing (convert your voice in realtime using RVC voice model).

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[hallow thistle]

You can see https://docs.aihub.gg for more information and guide docs about related programs.

[outer valley]

When i asked gemini about it with what you told me it really like told me the same you just did.
Interesting
Okay Applio is the model maker and newer Okada fork and Vonovox are the software to use the model.
Can i expect far better quality when i make a new model with Applio and then use one of the newer okada forks?
My model is like 1-2 years old and mixed from 2 different voices so i have something of my own that is not just someone else voice. My goal is to create something new that is my own not just simple copy pasta.

Is Applio good in mixing 2 data sets of 2 different voices? Gemini said it can do that.

[hallow thistle]

Sure, Applio RVC can do "voice models mixing into one. Aside from original pretrain voice models, Applio RVC also lets you to use any custom pretrain model from third-party sources for training a model, although the voice model result is often questionable by the nature of those custom pretrains, where a few models trained with custom pretrains might work as intended while few others fail or ended up lower quality.

While I use Gemini for some parts, I wouldn't go for something like "what does Namari mean about their statement?" and then show them chatbot with some screenshots. While I'm more focus on some basic settings and inference parts that generally accessible for most people, for more information about training a voice model, you can ask some fellow members who have pink "model maker" role here, where they might know how to achieve a better working model intended.

[outer valley]

that sound really great! And i think with my 4090 training wouldnt not need that long?
I remember 1 years ago training myself sounded hyper complex... Is it with applio now much easier and user friendly?

[outer valley]

ohh i did some.... commissions back then when it was a "thing" on this discord too some of this Master model maker and... some did what they promised and... some not really...

But i assume since then and with how training now works i maybe better with going the "train myself " route now

[hallow thistle]

Yes, NVIDIA GeForce RTX 4090 is so far above than the minimum/recommended GPU stated on official Applio RVC website, so any RVC task (even training) will be super fast than some entry-level to mid-tier ones with 8 GB VRAM. While paid request and commission were a thing here, today they are not allowed, only free ones acceptable. Some time around when paid requests weren't allowed, the model maker role channel went missing, until December 2025, the similar channel has finally returned as #📤│model-maker-role.

[outer valley]

Well i can approve some... individuals tried to "scam" other had a lot of passion and it was fun working with them

[digital thicket]

Real time voice changer I would say so W-Okada right?

[outer valley]

That is really great to hear. I think i will use this week and work on my nano banana projects and this applio + Voxvox software. I want too see how much i can improve my old stuff with the newer ways

[hallow thistle]

W-Okada is realtime voice changer. What is your PC GPU? For better working audio quality, it has to happen with settings, the program version you use and a supposed GPU.

[outer valley]

i think it was a good decision to remove the paid commission system. It has to much risk for this interesting community and what it works on.
I personally a still shocked how crazy fast image and video AI gets pushed by all the big companies.
But when it comes to voice.... maybe only "text to speech" is a thing but realtime voice changing with AI ???
I mean it would be a big useful thing for movie companies and similar. But the development it that area.... is idk... it feels very very slow compared to chat,video,image ai development.

[digital thicket]

Nvidia GeForce 360ti

[hallow thistle]

Do you mean "NVIDIA GeForce RTX 3060 Ti"? The "NVIDIA GeForce 360 Ti" doesn't seem to exist as actual GPU. For better working W-Okada, try Tg Develop's fork. https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001 https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002 There's the guide. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/

[marble minnow]

Bro how to upload download rvc in applio . I'm confusing plz help

[hallow thistle]

Do you mean "to download a voice model" and put them into Applio RVC? Simply, copy a Hugging Face link from a thread from #1175430844685484042, go to Applio RVC, paste it in "model link", click "download model", go to "inference" tab, refresh model list and there you go.

"Drop files" option there is possible, but you'd have to extract files (.pth, .index) from the zip before uploading them "one by one".

[digital thicket]

Yes sorry my autocorrect did some BS

[digital thicket]

But idk if it work yk just searching for a real time voice changer without problems yk

[tame oracle]

yes i do, thats why i said “i like da spinny disks even tho SSDs are better”

[shy ember]

hey weight

for some reason when i open teh fork it opens terminal then closes immediately

[shy ember]

ah its because i moved it out of the server file

is there anyway i can make a shortcut on my desktop so i dont have to go into the oflder

[viral mason]

Their name is Namari lol

[crimson depot]

@viscid moss hey bro. Can U DM me?

[nova folio]

Hi! I have a question:
I’ve had Realtime Voice Changer for the past year or two, and I haven’t updated it yet. Could someone tell me where I can go to download the latest version? the version I'm on issss- v.1.5.3.18a [onnxgpu-cuda]

[viral mason]

this is so old, what gpu does your pc have?

[nova folio]

Currently NVIDIA GeForce GTX 1080
am needing to upgrade it

[viral mason]

hm

[viral mason]

I used to have a similar gpu so u may be able to run this but idk
https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001

https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002

u need both

download them then extract the 001 zip, then drag the 002 zip into the folder for 001 and run mmvcserversio

[rocky spruce]

Can someone help me and voice a small text in a ai voice of dottore from genshin impact? I've spent the whole day trying to figure this out... if it's not difficult for someone, please help me

[jovial urchin]

thevoicemodels.com when i try to download a model it says
you must gave the @AI Caylagi role to download

how can i get the role

[hardy tide]

how do i hear myself to hear what it sounds like im new to this

[hallow thistle]

For RVC voice models, see #1175430844685484042 and Weights.com.

[hallow thistle]

On W-Okada voice changer, this might give you some ideas. https://cdn.discordapp.com/attachments/1159290139609137264/1456249974525526185/image.png?ex=695eee39&is=695d9cb9&hm=855eb487bc4b36c0b98020ee4b2f524a0500e4d3e0764487e0a5f04e825471fb&

[hardy tide]

how do i fix rvc voice models not working

[hallow thistle]

What is your PC GPU? And did you follow any tutorial or guide before?

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 4060 8gb vram desktop)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a legal, safe & ethical community, we will NOT provide help for:

• ANY illegal activities.
• NSFW/Porn.
  Requests for these topics may be ignored, not helped and result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.
• Don't Ask To Ask.

[hardy tide]

5050

rtx

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 4060 8gb vram desktop)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a legal, safe & ethical community, we will NOT provide help for:

• ANY illegal activities.
• NSFW/Porn.
  Requests for these topics may be ignored, not helped and result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.
• Don't Ask To Ask.

[hallow thistle]

Older versions like b2332 and even v.1.5.3.18a weren't compiled for NVIDIA GeForce RTX 50 series GPU, so better try Tg Develop's W-Okada fork (b2397) or Vonovox (1.6.9) where they are compiled to work with RTX 50.

-realtime

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE

⚔️ Wokada Tg-Develop Fork vs Vonovox

For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[hardy tide]

how do i find the donwload

[hallow thistle]

This is Tg Develop voice changer. https://github.com/tg-develop/voice-changer/releases/tag/b2397

[hardy tide]

ok

[hallow thistle]

On GitHub, click download both "voice-changer-windows-amd64-cuda.zip.001" and "voice-changer-windows-amd64-cuda.zip.002".

[hardy tide]

i download both

[hallow thistle]

These two files have to be in the same folder. Use 7-Zip or WinRAR to open and extract .zip.001 one.

[hardy tide]

ok

there donwloading rn

[hallow thistle]

Go inside MMVCServerSIO folder, and try double click the program of the same name.

[hardy tide]

its loading

is this it

?

so i cant use it

[viral mason]

don't use og wokada, use what u have in the picture u just sent

[hardy tide]

how do i make it sound better

[viral mason]

use your gpu instead of cpu, switch to rmvpe (not onnx) extra at 2.7 and chunk around 122.7

[hallow thistle]

The original W-Okada (v.1.5.3.18a) once supported Beatrice voice model alongside RVC, while fork W-Okada (like b2397) has Beatrice support "removed", there's only RVC voice model.

[hardy tide]

i put it on gpu when i speak says this

[viral mason]

did u set up your microphone settings and have a voice installed

[viral mason]

you're missing the virtual cable

[hardy tide]

where do i put the cable

on mic or output

[hallow thistle]

Did you not install Virtual Audio Cable from start?

[hardy tide]

i have it

[viral mason]

it goes in ouput

line 1

or if u have vb cable cable input

[hallow thistle]

This might give you an idea, for second time. https://cdn.discordapp.com/attachments/1159290139609137264/1456249974525526185/image.png?ex=695eee39&is=695d9cb9&hm=855eb487bc4b36c0b98020ee4b2f524a0500e4d3e0764487e0a5f04e825471fb&

[hardy tide]

i have vb not line 1

how do i get that one

[hallow thistle]

For Virtual Audio Cable lite, download from this link. https://software.muzychenko.net/freeware/vac470lite.zip

[hardy tide]

ima restart and comeback

aight

[hallow thistle]

Click stop server, try set Extra number to 2.7 s and sample rate to 48000.

[hardy tide]

did it

and is there a way to controll the dealy

[hardy tide]

whats the best dealy

[viral mason]

the smallest amount

[hardy tide]

so like

is 80 good

[hallow thistle]

Around 110 ms might work.

[random juniper]

different voice changer? i can only seem to find w-okadas beta from 2 years ago

[hardy tide]

aight and also the voice is stuck for me i delete the last one i had and it stays the same aft i put a new one

[hallow thistle]

That one is b2397 W-Okada fork.

[viral mason]

wokada tg fork is more simple

[random juniper]

how would someone download it

i mean where*

[viral mason]

through the link on the github

what gpu do u have tho, there is an amd version and an nvidia version

[random juniper]

nvidia

[viral mason]

then u would need these two
https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001

https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002

[hallow thistle]

Which NVIDIA GeForce RTX?

[viral mason]

exract the first one then place the second one in the folder made by it

[random juniper]

4070

[hallow thistle]

[random juniper]

that says amd64 though

[hardy tide]

thx for the help

[viral mason]

they're labled weird

they all say amd when they aren't for it, I had the same thought

[hardy tide]

also how do i turn this on

[hallow thistle]

AMD64 is another name for x86-64 or x64, a CPU architecture used in most Intel and AMD CPUs.

[viral mason]

it won't work unless you use client mode which just doesn't work at all anymore

[hardy tide]

dang

well thx

[viral mason]

yea client mode quit working since november 2025

[hallow thistle]

If you set to "server" audio mode on the voice changer, these three options will gray out; if you switch to "client" audio mode, these three options will available. It's a quirk to all W-Okada versions, not just this specific version.

[hardy tide]

thx

[random juniper]

thank you

[random juniper]

why does ts sound so ass when i tab out 💔

[hallow thistle]

On your W-Okada voice changer, here's your settings:
Chunk: around 60 ms
Extra: 2.7 s
GPU: NVIDIA GeForce RTX 4070
Pitch detecting: rmvpe

[viral mason]

what settings do u have, and what games/software are u trying to use it with

[random juniper]

thank you

[random juniper]

trying out games rn, it gets a bit laggy

[viral mason]

lower your game graphics to the lowest settings to get better performance while using it

[hallow thistle]

Try increase chunk number up, until the perf number in "Performance Stats" section is green and stable.

[autumn cedar]

German language progress report: I tried a voice model, Heinz Erhardt, from voice-models.com and it sounded pretty good to my German-learner ears... it seems to get the German fricatives and sibilants correct (ich, auch, acht, Schlange, etc.)

[autumn cedar]

Is there any particular reason why VAC Virtual Audio Cable is recommended? There are other softwares which do the same thing. Is VAC the best? Is it the easiest to use?

[brave torrent]

just need help with voice model stuff my things are always called out for voic changer]

[autumn cedar]

Is there a good help / setup guide for VAC Virtual Audio Cable?

[compact panther]

— now, i'm not an expert, but —

you literally just install it

and you can later set it to act as your input gateway for VC inside other applications

[heady spade]

Is there anyone who can make a voice model?

Ru

[compact panther]

#1159289738314919936

[royal talon]

im so lost.

:I

[visual nimbus]

yo

[royal talon]

what

.

[visual nimbus]

wdym what

[royal talon]

idk

i-

[visual nimbus]

alg gng

where i find the downlaod

[royal talon]

idek\

[visual nimbus]

u tryna find it too?

[royal talon]

yep

[visual nimbus]

fah

@hallow thistle help us pls

[royal talon]

😛

[hardy tide]

whats a good chunk to use nvidia geforce rtx 5050

[hallow thistle]

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 4060 8gb vram desktop)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a legal, safe & ethical community, we will NOT provide help for:

• ANY illegal activities.
• NSFW/Porn.
  Requests for these topics may be ignored, not helped and result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.
• Don't Ask To Ask.

[visual nimbus]

where is the download link twin

[hallow thistle]

@visual nimbus @royal talon you, both of you, make sure to read help guidelines before start asking. What is y'all PC GPU?

[visual nimbus]

uh how do i check

wait nvm

it says gpu 0 and gpu 1

gpu 0 is nvidia and gpu 1 is amd

[royal talon]

nvidia geforce rtx 3060 ti-

[hallow thistle]

Click on GPU 0 to reveal the full name of NVIDIA GeForce RTX/GTX on the right side.

[visual nimbus]

nvidia geforce gtx 1650

[hallow thistle]

Try Tg Develop's W-Okada fork. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/
Download links: https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001 and https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002

[hallow thistle]

See #✨│ai-help message.

[hardy tide]

whats a good chunk to use nvidia geforce rtx 5050 cuz the chunks i been useing sound super ass

[royal talon]

erm oki

[royal talon]

is this only the audio cable?

[visual nimbus]

i do this one?

do i click both links?

[hallow thistle]

For downloads, download both voice-changer-windows-amd64-cuda.zip.001 and voice-changer-windows-amd64-cuda.zip.002. The docs.aihub.gg one is a guide doc about W-Okada.

[hallow thistle]

This link https://software.muzychenko.net/freeware/vac470lite.zip is the download for Virtual Audio Cable.

[royal talon]

im confused because when i download voice-changer-windows-amd64-cuda.zip.001 it doesn't download in a zip

[hallow thistle]

.zip.001 and .zip.002 are split parts of a single zip file. https://cdn.discordapp.com/attachments/1159290139609137264/1457951470505955562/image.png?ex=695f301d&is=695dde9d&hm=0785a4a4002ce97446aaf0a4f148530595ac06f52649c6b2fb2e7abc7cd4e9e7& If you have "Hide extensions for known file types" enabled, the .zip.001 one would simply show just .zip, making things confusing.

Use WinRAR or 7-Zip to open the .zip.001.

[royal talon]

o okay

[hot hinge]

audio keeps cutting out

like it sounds fake

cause of that

[hallow thistle]

There's no exact best working "chunk (or chunk size)" value in W-Okada voice changer. The chunk value can be of any number, as long your GPU can handle, see #1458165876263223457 message for a bit more information.

[hallow thistle]

What is your PC GPU? And did you follow an tutorial or guide before?

[royal talon]

now it wont even let me download it....... um

[hallow thistle]

Check if your internet is fast enough and any antivirus program that might still scanning your PC.

[hot hinge]

yes i followed a guide but my gpu is 4060 ti and cpu is 12400F i5

[hallow thistle]

Follow this #✨│ai-help message.

[royal talon]

ok

nope its just not working it ways couldn't download - download error

:I

ummm

oh now its working

.

nvmmm

[hallow thistle]

You can send a screenshot to here now, otherwise I won't be able to identify the issue from words alone.

[royal talon]

sorry nvm its now working

[hallow thistle]

[royal talon]

is this what it is supposed to look like?

[hot hinge]

ok thanks i’ll try

[hallow thistle]

The usual behavior would look something like in my screenshot, one folder of full size and two batch files all intact. If you see one folder of halfed size and one batch file there, it can indicate that the program might load only .zip.001 while the other part (.zip.002) missing.

[royal talon]

so i get .zip.001 and .zip.002?

[hallow thistle]

Another potential cause is that your WinRAR is under "evaluation copy (unlicensed)" mode. If .zip.001 and .zip.002 are in different folders you should put them in the same download folder. If you didn't download .zip.002 you must download it, then follow the earlier step.

[royal talon]

oki

[royal talon]

so like this?

[hallow thistle]

This is good now.

[royal talon]

then what do i do?

[hallow thistle]

[royal talon]

.

are there steps?

[hallow thistle]

"Extracting a zip file" is one of the most basics anyone can do. In WinRAR, click "Extract To", set "destination path" to somewhere like "C:\Users\your username\Downloads\MMVCServerSIO" or "D:\MMVCServerSIO" and then click "ok".

[royal talon]

it said this

[hallow thistle]

Click on the first button.

[royal talon]

ok

okay so i did it but it made a windows sound but nothing popped up @hallow thistle

how do ik if it extracted

?

[hallow thistle]

Double click on "MMVCServerSIO" program again and try screenshot it.

[hallow thistle]

That's inside "MMVCServerSIO" folder within the zip file.

If you extracted the files to a folder outside the zip, it should look like this.

[royal talon]

i dont see anything out of it

[quartz moon]

hi so im on linux but i wanted a voice changer but also my hardware sucks so maybe idk an online voice changer(free since im broke)

[hallow thistle]

[hallow thistle]

What is your PC GPU? If there any.

[quartz moon]

rx 570

[royal talon]

is it this ?

[hallow thistle]

Yes.

[royal talon]

oh

.

okay then i just double click that?

[hallow thistle]

While "AMD Radeon" driver is known to work well with Linux distros, some have observed, this "AMD Radeon RX 570" might likely struggle with W-Okada voice changer even on Linux. For a cloud service, see https://docs.aihub.gg/realtime-voice-changer/cloud/tg-develops-w-okada-fork-kaggle/.

[hallow thistle]

Double click the program.

[quartz moon]

soo it will js be the same process as windows?

[royal talon]

is this what i am supposed to see

[hallow thistle]

This is Tg Develop's W-Okada (b2397). It's completely normal for this specific version to look like this.

[royal talon]

oki

[floral escarp]

if I wanted to get rid of my current installation of RVC, and get a newer more updated one, how would I do that?

pretty sure my current app is very outdated, and quite broken.

[royal talon]

now how do i put a model in?

[hallow thistle]

Ehh, kinda, but some additional steps are required for Linux, like installing portaudio in your terminal for a "virtual cable" to work on Linux similar to "Virtual Audio Cable or VB-Cable" in Microsoft Windows.

[quartz moon]

ye i saw that
ty

[hallow thistle]

You're welcome.

[hallow thistle]

W-Okada voice changer or Applio RVC? Generally, completely delete the older folder while extract the newer one (from zip) into the same folder is the only acceptable workflow steps. When you extract the newer one overlapping the older one, there's a less chance the program still works although mostly fail to even run.

[royal talon]

i dont have any yet where do i download them? because i look in the voice models but i never see the download links...

[floral escarp]

it was W-Okada, does deleting the folder it was initially extracted into delete it fully?

[hallow thistle]

For how to download a voice model from #1175430844685484042, go to #1175430844685484042, go to a post/thread there, spot a link (usually being Hugging Face) and then click it to download a file. https://cdn.discordapp.com/attachments/1159290139609137264/1456902989422923959/image.png?ex=695f5424&is=695e02a4&hm=76558fc8b12998acbc1077d022ef01918c2eaf5625dbcfd46b22be643506f541&

[hot hinge]

can some go in dms with me and tell me if my voice sounds real (ai) ill send audio clip thanks

[hallow thistle]

If you completely delete (without going into Recycle Bin) the whole MMVCServerSIO folder, all files should gone with no trace.

[hallow thistle]

No.

[hot hinge]

i need one opinion plsss

[hallow thistle]

If you go into one's direct message, you're potentially limiting those knowledge for just me and you. In #✨│ai-help is better for such query.

[hot hinge]

can i post a audio clip of me there tho

[royal talon]

im so confused none of them have a download option......

[royal talon]

oop i meant to reply to this

[hallow thistle]

Read again. If there's no "download" button alongside a URL link, it doesn't mean you can't download a file.

If you see this message, click "Continue to download".

[royal talon]

yea i dont see any of that all i see is a link to weights.com?

[hallow thistle]

You can send your file here until your username turns blue.

[quartz moon]

hi
i ran the third cell in kaggle voicechanger and its stuck on /kaggle/working/MMVServerSIO for like 10 minutes is that normal or do i procced to the next step or what

i cant send pics so i cant send the exact thing lol

[quartz moon]

or nvm im js dumb it doesnt show more than that i needed to procced to the next step

hi
my voice changer isnt working for some reason like its not changing

i dont have image perms so i cant send images

[hot hinge]

wait wdym

[quartz moon]

she meant u cant

or he idk

[hot hinge]

"If you go into one's direct message, you're potentially limiting those knowledge for just me and you" then this doesnt matter... bruh

[quartz moon]

lol js test it in discord

or the web page of ur voice changer

or gui idk

[short wagon]

guys help why does it looks like that

well i cant attach photos

[quartz moon]

lmao

wait for a staff member to give u perms(i hope they can give perms)

[short wagon]

https://imgur.com/a/6hKlJrp

[quartz moon]

oo ur smart

[short wagon]

there are no models in list, but they downloaded as i check in the path

[quartz moon]

wait for the admins im too dumb for this

[quartz moon]

.

im pretty sure i did everything correctly

god why is linux so complicated it worked perfectly fine on windows

https://imgur.com/a/Rh052U6

its with kaggle btw not local

now i cant even start i get this when i start
Error

An error occurred during voice conversion. Check command line window for more details.

idk where im even supposed to check cuz im on cloud

changing from tesla p100 to cpu worked

but now i cant hear my voice

[hallow thistle]

On Kaggle, use "T4 x 2" instead of P100.

[hallow thistle]

What is your PC GPU? You seem to be using the older W-Okada voice changer version.

[quartz moon]

ill try it

[hallow thistle]

Talk a bit more, so when your username turns blue here you'll be able to attach a file (like an image).

[hot hinge]

ohh ok thanks

can u pls give me the permission tho

[hallow thistle]

[hallow thistle]

I'm a Helper, not a moderator, so I can't give you a permission.

[hot hinge]

ok allg thx

wait no way

finally

[hallow thistle]

Your username now turned blue.

[hot hinge]

lol

[short wagon]

4060

[hallow thistle]

Try follow this #✨│ai-help message message.

[void flume]

@hallow thistle
Referring to: #1458165876263223457 message
Because his windows username corrupted
and because he claimed that more problems were happening on his system.
I mean his appdata local folder has changed location on the disk outside of his control
His disk is broken
both you ans the moderator have been unhelpful; you made the voice client work for one day, maybe two or three if lucky, but his pc is dying in all likelihood. First you need to get the foundation in proper order (his pc) after that you can work in getting the voice changer to work. Ignoring the foundation will screw things up.

Like what is this comment anyway? #1458165876263223457 message
the OS doesn't matter, it's his username and other issues.

[visual nimbus]

my thing isnt working

it js shows up w this and then closes after like 10 seconds

[hallow thistle]

Still not done trying to get me demoted yet? The original poster got helped as intended, so it's just over for you now.

[visual nimbus]

can i get help pls

[hallow thistle]

The v.1.5.3.18a and b2397 are two different versions. Check again if you have downloaded Tg Develop's W-Okada fork (b2397) instead of the v.1.5.3.18a one.

[void flume]

Where the heck did you get that idea? I only question the the way you seem to react and your actions, and the rules.

I asked the staff to glance at the situation and was waiting for a reply using the ticket system, that's it

[visual nimbus]

i downloaded the right one the v.1.5.3.18a one

[hallow thistle]

I didn't tell you to download and use v.1.5.3.18a. I told you to download https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001 and https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002.

[visual nimbus]

o shit my bad gng

[void flume]

I asked for clarity specifically. I wondered what I was and wasn't allowed to do or post; If you have seen my ticket as a staff member, then you should know.
You're just doing your job, I assume, so it really isn't about you.

[quartz moon]

yall i keep getting an error but its pretty long so idk if i should sent it here(yes very dumb lol)

ye i cant even send it even if i want lmao

[hallow thistle]

When you cry about what I did one more time, for real though, you would never be a good "Helper" as me, and the moderation might be considered for your actions, since past few days.

[quartz moon]

chatgpt fixed it lmao

[hallow thistle]

OpenAI ChatGPT can be useful for few parts, in this case it's a bit better.

[quartz moon]

uh so the sound is still not working even with t4

maybe im testing it wrong?

i prolly am tbf

also why am i still pink i have almost 200 messages bro

[hallow thistle]

Try talk a bit until your usename turns blue.

[quartz moon]

uh ive been talking for a while

well

do u know the solution to my problem

also i can send imgurs

or whatever they r called

https://imgur.com/a/uVIuoj3

im pretty sure i set voice cable wrong cuz its not in my microphone list

[hallow thistle]

On your voice changer, these settings might work:
Chunk: around 64.0 ms
Extra: 2.7 s
GPU: T4
Pitch extraction: rmvpe

Input: microphone
Output: a virtual audio line in Linux
Monitor: your speakers/headphones

[quartz moon]

i dont have a virtual audio line even tho i downloaded one in the tut

from the guide i meant

i js have 2 microphones(a monitor and a normal one idk why i had them since i installed linux) and my speaker

ohh monitor of is the virtual one

im so dumb

how do i even test it cuz its not working when i record using serverio analyzer(idk what that is so maybe im wrong)

[hallow thistle]

This thread https://discord.com/channels/1159260121998827560/1442151712390512701 might be your answer on how to achieve that virtual audio line. Because while some people here use Linux distros for specific tasks, it's rare for some to actually get W-Okada voice changer to work on Linux like in Windows, where it would be a new knowledge for everyone.

[quartz moon]

ty ill try it out

this did smthn cuz now firefox keeps crashing lmao

i love when my voice changer ruins my browser and audio

i deleted the conf file and now i have the same issue but working browser yay

[visual nimbus]

what do i do here

[hallow thistle]

If you have extracted the MMVCServerSIO onto your desktop folder, better don't do that. Instead, try extract to somewhere like "D:\MMVCServerSIO".

Also make sure your NVIDIA GeForce driver stays up to date.

[brittle wing]

Can anyone help with perfecting for w okada voice changer

Gain in and out value

I use E girl rvc model. think thats best english voice model

[carmine thorn]

what model is used for this ? https://www.tiktok.com/@peterpat15/video/7580899860405095710?is_from_webapp=1&sender_device=pc

[sly zephyr]

can some one help me get the voice working in discord because its not

[proper flame]

i extracted the zip001 with the zip002 in the same folder, that worked, i tried opening the application inside but it just closes immediately, what do i do?

[hallow thistle]

What is your PC GPU? Are you following Tg Develop W-Okada fork guide?

[hallow thistle]

What is your PC GPU? And did you follow any tutorial or guide before? Make sure to read help guidelines before start asking.

[hallow thistle]

[proper flame]

nvidia rtx 3050, tried a few guides and tried some stuff in this channel when i ran into problems. im welcome to start fresh though because i dont really know what im doing lol

[hallow thistle]

In my screenshot, this Tg Develop's W-Okada fork (b2397). https://cdn.discordapp.com/attachments/1159290139609137264/1457950706500763809/image.png?ex=695f2f67&is=695ddde7&hm=62f9be1ebec429bbe87fecf29cc942ea136b815f2b569be2a934f53cdfe5accd& If you're referring to this one, you're looking for right one. Review your steps: both voice-changer-windows-amd64-cuda.zip.001 and voice-changer-windows-amd64-cuda.zip.002 have to be in the same folder, use WinRAR or 7-Zip to open the .zip.001 one, try extract files to somewhere like "D:\MMVCServerSIO", and if you extracted the program files on your desktop, better don't do that.

[proper flame]

i extracted it to the same folder as the 001 + 002 was in, is that not ok then

[proper flame]

what about it?

[hallow thistle]

Go inside MMVCServerSIO folder, try double click on MMVCServerSIO.exe.

[proper flame]

yes thats what i said originally, when i open it, it just closes

[hallow thistle]

The program won't open has to happen before when you extract the files. An incomplete zip file, where inside the zip file, there's one folder of halfed size and a single batch file present, generally indicates a bad sign where .zip.002 might be missing. Meanwhile, the complete one (where .zip.002 is present alongside .zip.001) has a folder of full size and two batch files present there.

[proper flame]

yes i have the version with both

[hallow thistle]

What do you mean by "both"?

[proper flame]

i have 2 batch files present

[hallow thistle]

Try open MMVCServerSIO.exe within CMD, and copy the error message from terminal if there any.

For much better identifying the issue, talk a bit in here until your username turns blue so you would be able to send your screenshot and any attachment in here.

[proper flame]

no theres no error

[dry dawn]

using vonovox , my delay starts off at like 700ms then builds up to 1200ms over time, anyone know how i can keep it at 700ms constantly?

[viral mason]

Ew

[visual nimbus]

whenever i try it it does this

[lapis lintel]

Guys, if it's written in the guide epoch 300, what does that mean? and other stuff
(pitch extraction: rmvpe
steps: 21k
batch size: 6
pretrain: legacy core v1.5 / 40k)

[plush trellis]

#✨│ai-help message

[jaunty shale]

any idea why it doesn't work? I used old okada before and these input and output worked fine for me.

(I use online method)

[viral mason]

This is not a guide, this is what it was trained on

The information is for viewing the models settings used when it was trained

[proper flame]

i got vonovox, extracted it, ran the setup.bat but now when i do the start.bat it just says: C:\Users\fart\Downloads\Vonovox169>runtime\python.exe launcher.py
Initializing CPU limits: 4 cores, 4 threads [ID:8259]
PyTorch thread limits set to 4
CPU affinity set to 4 cores (0-3) out of 16 available (total/4 = 4)

C:\Users\fart\Downloads\Vonovox169>pause
Press any key to continue . . .

[graceful patio]

"\fart" 😭 💔

[somber bough]

what is the best free tool that i can use to faceswap a video

[tame oracle]

@light swift

-rt

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE

⚔️ Wokada Tg-Develop Fork vs Vonovox

For Windows Nvidia, Both Wokada Tg-Develop fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Tg-Develop Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[tame oracle]

the second guide would be perfect for you then

[light swift]

Wokada Tg-Develop Fork?

[tame oracle]

yes

[random juniper]

what server settings would be best for a 4070?

[swift glacier]

My voice doesnt go through even tho i Selected the right input the Vol indicator is at 0 all the time

[viral mason]

is it on client mode or server mode?

[swift glacier]

Client mode

[viral mason]

client hasn't worked since last november

try using server mode

[swift glacier]

Ill let you know if it worked

[swift glacier]

[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
[Voice Changer] ex: could not broadcast input array from shape (1024,8) into shape (32768,8)
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
[Voice Changer] ex: could not broadcast input array from shape (1024,8) into shape (32768,8)

Im getting this error code

[viral mason]

after switching to server u set your mic settings back up right?

[swift glacier]

I guess so, gonna be honest i dont really know what im doing but it should be right i followed a tut

[viral mason]

oh

so you probably have something outdated then

what tgpu do u have

[swift glacier]

Rx 570

[viral mason]

damn

[swift glacier]

hahahaha its good tho i can play everything i want to

[viral mason]

yea that small gpu is not really good for ai tbh

gaming is ok ig then

[swift glacier]

So u think its bc of the gpu?

[viral mason]

most likely

[swift glacier]

Is there any good alternative Voice changer, im lowkey desperate for that free content

[merry eagle]

<@&1159293204038955078> can you guys give me line in software download link

[viral mason]

do u just need the virtual audio cable or

[merry eagle]

yea that thing

[viral mason]

1 sec

[viral mason]

https://software.muzychenko.net/freeware/vac470lite.zip

[merry eagle]

thank you soo much my lovely friend

[viral mason]

not any good free ones, I can give u the one I use but I doubt it'll work with ur gpu

the one I use is free tho, it's wokada tg fork

[swift glacier]

Imma try it out thanks tho!

[viral mason]

https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip

there's the download

[plush trellis]

yo im using vonovox i can hear my model fine when im using headphones but when i put it on cable audio my voice also overlaps

fixed after a restart, i had passthrough on

[swift thunder]

FP16+TF32

What are the training benefits of this?

[zealous pendant]

@tame oracle

hi

can u help me with something

when i open the realtime voice changer

it says to wait for web server perma

it worked just fine some time ago but now it wont open

[viral mason]

I'm not the person you pinged but I am somewhat good at helping with the voice changer, what gpu do you have?

[zealous pendant]

thank you

[viral mason]

And did you get the voice changer off a YouTube tutorial

[zealous pendant]

i have an rx5700xt

amd

[zealous pendant]

ye

[viral mason]

It's outdated then

[zealous pendant]

:(

[viral mason]

Pretty much anything they show off on YouTube isn't the good stuff that's up to date code wise

[zealous pendant]

ooh

so where should i get the new one

[viral mason]

I'll get the link for the one I use

[zealous pendant]

thank you so much :D

can i use the same models tho?

nice

[viral mason]

Also do you have a virtual audio cable like VB cable?

[zealous pendant]

yeye

[viral mason]

Ok cool

https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip

[zealous pendant]

thank you very much friend

[viral mason]

You're welcome! To run it just start mmvcserversio.exe

Btw make sure to switch to server mode because client mode hasn't worked since November 2025

[zealous pendant]

okey

[tame oracle]

my apologies

was not at my pc

glad you got help tho

[zealous pendant]

no problems cutie

[teal otter]

Upload a Voice Model can no longer be used in Weights.com?

[viral mason]

wdym?

[viral mason]

Did you get it working or need anymore help? :3

[teal otter]

Why can't you upload trained models to the page anymore on Weights.com?

[viral mason]

I'm still kinda confused, do you mean models trained on appio?

[zealous pendant]

im trying to figure it out

how to be as smooth as possible

[distant hamlet]

Weights removed the ability to upload your own voice model, as of yesterday

[viral mason]

fr?

greedy fuckin company

[graceful tinsel]

guys now what to do

@hallow thistle

BROOO MY TWIN MY GLORIOUS KING HELP ME

TWIN

i have to go to school by 8:30 am say fast!!!

its 7:44am now!!

[graceful tinsel]

HELP MEE

SEE

[hallow thistle]

That's the "file explorer" instance coming from Applio RVC notebook on Kaggle, it's not the actual Applio RVC interface in this case.

[graceful tinsel]

im dumb

how to open rvc interface

[hallow thistle]

This is Applio RVC. https://cdn.discordapp.com/attachments/1159290139609137264/1455545707204182016/image.png?ex=696052d3&is=695f0153&hm=cbba5f99b5e802b5ee1603dfbdc3f4b40af324333b87cbe91d49660a4dcb4209&

[graceful tinsel]

mine looks like this

[viral mason]

?

[graceful tinsel]

read the chats

[hallow thistle]

Check on your Kaggle terminal, there's a URL link if you run the last code cell; if you use ngrok, there's ngrok; if you use "Gradio + Localtunned" there's Gradio.

[hallow thistle]

Yes, it's in the "Applio Public URL" part, not a local IP address consisting of numbers.

[hallow thistle]

That's Tensorboard, still not actual Applio RVC.

[graceful tinsel]

wah😭💔✌

[graceful tinsel]

how to get applio RVC now

twinn

[hallow thistle]

[graceful tinsel]

now clicking applio public url got me here

[viral mason]

just

replce ngrok

[graceful tinsel]

Twin are you einstein

are you tesla?

you solved it

💔✌😭

@hallow thistle gng ngl u got a L here✌ still thanks for ur help

[hallow thistle]

This is getting too confusing, because all links in your terminal are the same one. If you replace if Tunnel = "Ngrok" to Tunnel = "Gradio + LocalTunnel", mine would look like this on Kaggle.

[graceful tinsel]

nvm its solved

[hallow thistle]

Sorry, I'm too slow to even type one message.

[graceful tinsel]

its okkk its okk twin we are not perfect

[viral mason]

https://cdn.discordapp.com/attachments/1261530277591777322/1290670122003857469/ToKzrs.gif

[graceful tinsel]

the modern day einstein

💔✌🥀

how to add voice models

help me twins

Legend is that @hallow thistle is still typing

[viral mason]

u doing covers?

[hallow thistle]

Recently, there's a user here who trying to help people of his own will and getting me demoted from as a staff member because I clarified to him once to not enter an old thread that has long solved, if you see me no longer have Helper and staff role, it can mean the judgement has been decided I was wrong on that one.

[graceful tinsel]

idk just 4 fun

[graceful tinsel]

bruh wthhh

its not that deep bruhh

now help me add voice models twin

[hallow thistle]

To this day, that one person has been spamming in Modmail account talking about my actions as if did something too far terrorism to this server, as well as trying to drag another mod as if neither staff member available at the time.

[hallow thistle]

See #✨│ai-help message.

[viral mason]

u adding them to applio?

[graceful tinsel]

ya ya

[graceful tinsel]

yup thankss

[white sierra]

Could someone please do a vocal swap for me? I'm having a little trouble as I haven't worked with RVC since 2024.

I need help transforming the vocals from the song "Fate of Ophelia" by Taylor Swift into the vocals of Tobias Forge (Papa Emeritus) for a musical project. Any kind soul to help me?

Sorry for my bad English.

[hallow thistle]

What is your PC GPU? There's Applio RVC.

[graceful tinsel]

it doesnt take my gpu power right it uses cloud right of kaggle?

damn its fast afff😭🙏🙏 when I used to use rvc ui locally mine used to take 2 minutes for 40sec

[hallow thistle]

Anything on Kaggle always use resources from Kaggle and related cloud servers.

[white sierra]

RX 5600 XT (6GB)

[graceful tinsel]

its solved thanks

[graceful tinsel]

DAMMNNNN ts took 10 sec to convert so fast!!! i sang ts btw

[hallow thistle]

Applio RVC can be quirky when you put an output path to a non-existent directory (like /results/Me Too male part Yurizono Seia.wav rather than /asset/audios/Me Too male part Yurizono Seia.wav), where the Applio RVC might show successful message on frontend (GUI part), it doesn't actually create anything aside from throwing up an error in backend (terminal part).

[hallow thistle]

[graceful tinsel]

is ts good or not twin😭✌💔

[graceful tinsel]

applio giving me trust issues now

[hallow thistle]

I'm at work, so I don't have any headphones to listen.

[graceful tinsel]

[viral mason]

do u have other good song samples to test ur model

[graceful tinsel]

ts not my model

just doing inference

[viral mason]

o

[graceful tinsel]

did i sing good😭✌

[viral mason]

that was u?

[graceful tinsel]

ya but converted to frank ocean

bye gng have to go to school😭💔🙏

[viral mason]

It didn't artifact much so I'd say yea

[viral mason]

I graduated already

[visual nimbus]

whenever i try it it does this

[pearl anvil]

If you don't mind answering, what should I get then? If I have a Nvidia Rtx graphics card?

[inner island]

Resberry pie

[hallow thistle]

Which NVIDIA GeForce RTX?

[pearl anvil]

Rtx 4060

[hallow thistle]

https://upload.wikimedia.org/wikipedia/commons/thumb/1/10/Raspberry_Pi_4_Model_B_-_Top.jpg/480px-Raspberry_Pi_4_Model_B_-_Top.jpg

[hallow thistle]

Tg Develop's W-Okada fork. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/ https://github.com/tg-develop/voice-changer/releases/tag/b2397

[viral mason]

@pearl anvil you'll need these two zips for wokada tg fork
https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001

https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002

[stark heart]

When i speak nothign in game or call comes out

its like on mute

[inner island]

Crazy

[hallow thistle]

Did you follow any tutorial or guide before? And what is your PC GPU?

[graceful tinsel]

heyyy help me gng

since u are not a model maker

tell someone to help me make a model

[graceful tinsel]

help me

[brittle wing]

i did everything correctly,and when i try in discord it doesnt get my voice,nothing comes out

[hallow thistle]

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 4060 8gb vram desktop)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a legal, safe & ethical community, we will NOT provide help for:

• ANY illegal activities.
• NSFW/Porn.
  Requests for these topics may be ignored, not helped and result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.
• Don't Ask To Ask.

[night ibex]

guys im training a model

From what I understand from the guide, I should prioritize using the pth closest to the top and with the most steps, right?

I'm looking at it through TensorBoard; I can't send image

[candid urchin]

“Guys, how do I train it? I need help

[oak edge]

-kaggle

[patent trellis]

📘 Kaggle Notebooks

Kaggle is a Cloud (Remote Good PC) Service that offers 30 hours of GPU weekly, but needs a phone number verification

• **Applio Notebook**

by IAHispano
Kaggle

• **Hina Mod Original Wokada**

by Hina
Kaggle

• **Wokada Deiteris Fork**

by Hina & Deiteris
Kaggle

• **UVR5 UI**

by Eddy, ArisDev & Nick088
Kaggle

• **UVR5 NO UI**

by Eddy
Kaggle

• **RVC AI Cover Maker UI**

by Shirou & ArisDev
Kaggle

• **Music Source Separation**

by Shirou
Kaggle

[burnt matrix]

My laptop gpu is NVIDIA RTX 1050 windows 11. I am trying to use the ai voice changer wokada but whenever I use it when in game or record it lags or glitches, sometimes voice is not picked up by the changer. Some people have told me about upgrading my system if that’s the case, can someone suggest me a budget cheap build that can run the ai voice changer better than my laptop or maybe can I send the options that I was looking at and let me know if that would work. I don’t want to upgrade and then find out that it’s not better than my old laptop.

[candid urchin]

I don’t know much English. If possible, could you please give me a specific video tutorial on training RVCI don’t know much English. If possible, could you please give me a specific video tutorial on training RVC

[warped wave]

What is currently best in terms of quality realtime voice changer?

[tender kite]

Any recommendations for what I should use for RVC interference if I have an AMD GPU (RX 7900 GRE) and I'm on Linux or an Apple Silicon Mac?

[tender kite]

Anything local with balanced buffer works best

[warped wave]

so what would it be?

[viral mason]

Nope

Just test each one until it sounds good to you

With a sample, tensorboard is pretty much useless to look at for training voice models

[viral mason]

Vonovox, but that's Nvidia only

And for both Nvidia and amd there's wokada tg fork

[warped wave]

do you have a link for it?

[viral mason]

I will only be sending a link for one of them if I know what gpu u have

Since I don't wanna send something that won't work

Soo what gpu does your PC have

[warped wave]

Sry it's rtx 3090

[viral mason]

ur good, what are you looking to do with the voice changer btw :p

[warped wave]

Probably gonna just use it from time to time when hanging out with my friends for shits and giggles

[viral mason]

https://tenor.com/view/gif-emoji-emoji-meme-emoji-memes-spin-gif-3355519892839036421

so vonovox is the current best for nvidia, would you like that one?

if ye then u need this and this
https://github.com/dr87/Vonovox/archive/refs/tags/v1.6.9.zip

https://software.muzychenko.net/freeware/vac470lite.zip

@warped wave

[night ibex]

Is there any way to record audio model output in Vonovox?

like okada

[viral mason]

not currently no, if you'd want to just record it using obs or smth tho that could work

[warped wave]

What is the second link for?

If it's for virtual canble then I have it installed

[viral mason]

It's a virtual cable, vac lite

[rigid fiber]

can sm1 help me, when i run "start_http.bat" it closes

[burnt matrix]

Any kind soul ?

[rigid fiber]

hey david could u help rq?

[viral mason]

what you're using is old what gpu do u have

[rigid fiber]

uhh

[viral mason]

an rtx 3060 may be able to do what u want properly

[rigid fiber]

NVIDA GeForce RTX 3050 Ti Laptop

[rigid fiber]

do u have a newer one i can use??

[viral mason]

U can try wokada tg fork, idk how well laptops will run this stuff

[rigid fiber]

oh ok do u have the link?

[viral mason]

u need these twohttps://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001

https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002

[rigid fiber]

ok

do i download them at the same time?

[viral mason]

extract the first one then put the second one in the folder for the first one

[rigid fiber]

ok ok

[viral mason]

bc the second one can't be extracted but u still need it

[viral mason]

if u want

[rigid fiber]

alright

ones its done can i dm u and show u? like a video

its my first time using rvc so i dont really know what to do

[viral mason]

I guess? I can't call tho

[rigid fiber]

okay ill send u a video dw

[rigid fiber]

ok i got them all downloaded

[rigid fiber]

check dms

[burnt matrix]

So is that the only thing that matters ? The graphics card ? What about the intel core ? Like i5 or i7? Is there any other component i should consider for the upgrade for my purposes ?

[viral mason]

intel cannot run ai like at all

it's bad

only amd and nvidia can run realtime stuff like that

anything that's like equivilent to an rtx 30 series and up for amd is recommended

so for nvidia 30 series and up

10 series CAN work but not well

[stark heart]

yes and i got a gtx 1650

[lusty merlin]

pls help my ai is not opening

[viral mason]

very helpful

[rigid fiber]

can anyone help me? anyone

[graceful tinsel]

hey broo

help me

should we use UVR compulsory??

@hallow thistle

you know twin?

[viral mason]

what?

[graceful tinsel]

UVR

[graceful tinsel]

this thing

for vocal isloating

[viral mason]

I use the google colab of uvr

[graceful tinsel]

YOOO SEND IT TO ME

🙏✌

[graceful tinsel]

whatever but since i got the song now how to mix it and make a dataset

[graceful tinsel]

bro you here🙏😭✌

[graceful tinsel]

bro someone help me mix vocals

please

[tame dawn]

Is this a virus

[graceful tinsel]

what

[tame dawn]

Why do i have to open my windows command prompt

like what

[graceful tinsel]

in what

[hallow thistle]

Intel Arc Axxx/Bxxx (dedicated GPU) can, while other integrated Intel GPUs don't. Intel Core/Core Ultra are the CPU.

Newer Intel Core Ultra CPUs, while offer integrated Arc GPU for most models, they generally can run smaller AI tasks (like some camera background removal) and won't be the same level as dedicated Arc. There's an Intel NPU as well but it's primary for offloading some tasks from CPU and GPU (like noise suppression from microphone audio) and still not optimized for high-computing AI task like a voice changer either way.

[viral mason]

hmm ok, ty for the info

[wintry spire]

hi, how can i get this to run on my gpu instead of cpu ? i have an amd rx 6700 xt

[hallow thistle]

W-Okada voice changer (MMVCServerSIO) or Applio RVC (retrieval-based voice conversion)? Generally, both programs are free and open-source, use RVC voice models, though they are of different purposes.

[hallow thistle]

Try Tg Develop's W-Okada fork, especially its DirectML variant. https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/

[wintry spire]

i can use it in game ?

[hallow thistle]

Yes, given input/output audio settings look like this in your game.

[hallow thistle]

Try Tg Develop's W-Okada fork. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/ https://github.com/tg-develop/voice-changer/releases/tag/b2397

[rigid fiber]

can u help me download and help setting up vrc??

[hallow thistle]

Are you following Tg Develop's W-Okada fork? On this GitHub repository https://github.com/tg-develop/voice-changer/releases/tag/b2397, there are two files named voice-changer-windows-amd64-cuda.zip.001 and voice-changer-windows-amd64-cuda.zip.002 there, click both links to download them.

[rigid fiber]

can we go to dms and u ca help me there??

[hallow thistle]

Check the voice changer guide for more information. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/
These two files have to be in the same download folder Use WinRAR or 7-Zip to open .zip.001 one.

No, stay here. Going into one's direct message generally limits those knowledge only just for you and me, and if it's nothing personal I wouldn't do that.

[rigid fiber]

okayy

ill let you know when my 2 files are downloaded

[knotty grove]

w-okada has been having a TON of errors and getting worse lately. i have NO IDEA HOW, considering it's all LOCAL, i do not understand how anything could have changed, but it observably has. does anyone know if there's a reason why?

[hallow thistle]

You have been asking for W-Okada voice changer before. Is b2397 W-Okada still not working or something?

[rigid fiber]

@hallow thistle their downloaded what do i do now??

[hallow thistle]

Use WinRAR or 7-Zip to open .zip.001 one.

[rigid fiber]

openwith option right?

[knotty grove]

well client mode stopped working out of the blue last time i asked, (which still BAFFLES me because nothing in my files changed), but having downloaded a new version of both forks, the server version is now stuttering immensely, and seems to be acting like there's a harsh noise cancellation function cutting off my sentences. last time you helpfully advised me to make sure the sample rate was good, which helped at the time, but now it's getting worse regardless, the performance is worse as well, and it feels to me like the quality is just degrading due to some factor i cannot comprehend.

[hallow thistle]

Either double click or right click menu will work.

[rigid fiber]

ok

[rigid fiber]

ok im on the thing

what do i do now

[hallow thistle]

Is your PC GPU still AMD Radeon RX 7800 XT? On your W-Okada, try these settings:
Chunk: around 100 ms
Extra: 2.7 s
GPU: your GPU
Pitch extraction: rmvpe_onnx

[rigid fiber]

@hallow thistle

[hallow thistle]

Which step you're at?

[rigid fiber]

im in the file

where it says

mmvcserversio

[knotty grove]

these are and always have been my settings. like, i've been using this software with ease for months without changing anything. no new model, no new hardware, no new software, things just started not working. i'd love to find a solution but i still have to wonder WHAT happened, why, how.

[rigid fiber]

forece_gpu

reset_gpu

[hallow thistle]

Calm down, and don't be hurry. You can send your screenshot to here to make sure anything there is alright.

[knotty grove]

sorry, didn't mean to come off as upset at you or anything, i'm calm, it's just frustrating and confusing you know? just wanted to explain that i am doing everything pretty by the books here.

[hallow thistle]

Extract files to somewhere like "D:\MMVCServerSIO". https://cdn.discordapp.com/attachments/1159290139609137264/1458384995956818122/image.png?ex=69616c9e&is=69601b1e&hm=fdb4027ce1e885a726458bbf157fb69d2de4f5b38380a3d06d8e057020b9381d&

[knotty grove]

i'll let you help the newbie get set up, i can troubleshoot this on my own fairly well, i just was wondering if you knew how a software could be altered to function differently without me downloading anything new. like, how did client mode break? i don't comprehend it.

[rigid fiber]

okk im extracting it noww

[viral mason]

no explanation for client mode

[rigid fiber]

its done extracting and i extracted it to the example

what do I do noww

[knotty grove]

so everyone is confused? alright, at least i'm not alone lol

[hallow thistle]

Go inside MMVCServerSIO folder, there's a program of the same there, double click the program to launch the voice changer.

[rigid fiber]

okay

[hallow thistle]

https://cdn.discordapp.com/attachments/1159290139609137264/1458387597209501777/image.png?ex=69616f0a&is=69601d8a&hm=594d8eaeff9f21a6fb3dfa844e699edecb9d5d368c68ec48d2da5b988b0038d7&

[rigid fiber]

its loading now

[hallow thistle]

This is not Windows Explorer.

[rigid fiber]

oh

[hallow thistle]

The actual folder should look like this.

[rigid fiber]

ok ok i got it

imma let it download

the thing stopped

does it do that?

[hallow thistle]

Wait for a bit.

[rigid fiber]

ok ok

ok

it took me here

@hallow thistle

[hallow thistle]

Chunk: around 120 ms
Extra: 2.7 s
GPU: NVIDIA GeForce RTX 3050 Ti Mobile
Pitch extraction: rmvpe

Input: microphone
Output: Line 1 (Virtual Audio Cable)
Monitor: your speakers/headphones, this one is optional.

This is where you upload a voice model. https://cdn.discordapp.com/attachments/1159290139609137264/1458389948796375070/image.png?ex=6961713b&is=69601fbb&hm=d5d20516b18f32c2305f92debf554570057ac324ca0b6942c8b2c2c8c31b4d4e&

[frank shuttle]

Yo what’s the current best fully offline, local voice model trainer? I wanna make a very high quality model but i dont want the audio to be leaked since it’s all unreleased stuff.

[hallow thistle]

Applio RVC.

[frank shuttle]

is there a tutorial for it?

[hallow thistle]

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[hallow thistle]

Check out AI Hub docs and Applio official website.

[rigid fiber]

ok im done now how do i put models in

[frank shuttle]

ok thanj you

[hallow thistle]

#✨│ai-help message

[rigid fiber]

oh ok wait

look

[hallow thistle]

You're welcome.

[rigid fiber]

i dont have line 1 VIrtual audio cable

and i use voicemeeter output

how can i

uh

[hallow thistle]

Either Voicemeeter or VB-Cable might work, though I often use "Virtual Audio Cable lite" because it simply offers one single virtual line for input/output at the time. Voicemeeter/VB-Cable seem to offer many virtual lines than necessary.

This is the download link for Virtual Audio Cabe. https://software.muzychenko.net/freeware/vac470lite.zip

[rigid fiber]

wait soo voicemeeter is another choice to put?

like is it good to use it

[frank shuttle]

just to be clear none of the files that i use in the local version of applio have any way of being found online right?

[hallow thistle]

I don't know, there are many instances of Voicemeeter's virtual lines in your screenshot which is confusing, so I'd not using it.

[rigid fiber]

ok ill use virtual

[hallow thistle]

What does this mean if online cloud (like Google Colab) doesn't use repositories from GitHub? Both locally and online ones generally use roughly identical or same code repos from GitHub, but designed to work on different environments.

[rigid fiber]

@hallow thistle

[hallow thistle]

Extract these files into "C:\Users\your username\Downloads\vac470lite".

[rigid fiber]

ok done

[hallow thistle]

If you mean something like this: when you run Applio locally, user data (like voice models and dataset audio files) would still stay in that folder, your files won't go or being shared anywhere outside your PC. Same goes to online cloud services but the data there might be removed all after the session ended.

[rigid fiber]

i extracted it

[hallow thistle]

Go inside "vac470lite" folder, double click on "setup64.exe" to install Virtual Audio Cable lite.

[rigid fiber]

do i have to restart?

[hallow thistle]

Generally, I am not supposed to tell every step to members even if the guide docs exist, although I still do it anyway if the poster asked for it.

[rigid fiber]

i installed it

and took me to my sound

[hallow thistle]

Right click on your speakers/headphones, set it to as "default device". Same goes to "Recording" tab with microphone.

[rigid fiber]

i dont have speakers/headphones but i have them sep

[hallow thistle]

I wonder which output device you're using at this moment.

[rigid fiber]

im using voicemeeter

output

im confused:sob

😭

[hallow thistle]

No, I mean the physical device, not a virtual cable.

[rigid fiber]

do i set this

on default?

@hallow thistle

you there

uh @hallow thistle

@hallow thistle hellooo

[hard cobalt]

how do i fix "restart session?'

[rigid fiber]

wait i figured it out but when i talk it keeps glitching and it doesn't sound like the voice i want

and i cant hear anything when i go to youtube

it keeps cutting out

[candid urchin]

Traceback (most recent call last):
File "F:\2\Applio-3.6.0\app.py", line 6, in <module>
import gradio as gr
ModuleNotFoundError: No module named 'gradio'

help me

[hallow thistle]

[rigid fiber]

wait i got it

i no need help ansymoree

but thank u for helpingg

[hallow thistle]

Try remove your current Applio RVC folder, re-download the zip from this https://huggingface.co/IAHispano/Applio/blob/main/Compiled/Windows/ApplioV3.6.0.zip link, and re-extract again.

[hallow thistle]

You're welcome.

[candid urchin]

https://ibb.co/7t679Bmj

I’ve downloaded it, but I don’t know how to install and configure it yet, because it’s quite different from the videos I’ve watched

[autumn cedar]

is there a guide or tips for noise reduction in .wav files?

I have Audacity and I know how to use its noise reduction but I wonder if there are other recommended methods?

[hallow thistle]

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[hallow thistle]

See AI Hub docs and Applio website for guide docs.

[autumn cedar]

Important question - does RVC have to work harder to do inferencing on FLAC files, instead of .WAV files?

[hard cobalt]

can't we use rvc on roblox?

[graceful tinsel]

heyy bro

[rigid haven]

Do i need headphones for it to work

[candid urchin]

How can I fix my voice when it keeps cutting out

What software are you using for PTH with Index? Can you share your settings? My voice keeps cutting out

[quasi condor]

i mean dude u can just put vac lite on voicemeeter

that's what i did :p

[graceful tinsel]

guys i made a model

how to post it

[hallow thistle]

See #📤│model-maker-role.

[hallow thistle]

Without a headphone or speaker, you won't be hearing anything from a desktop PC, unless it's a laptop with integrated speakers.

[hallow thistle]

What is your PC GPU? W-Okada realtime voice changer or Applio RVC? And did you follow any tutorial or guide before?

[graceful tinsel]

how to upload a model to hugginface

so people can download

[hallow thistle]

You didn't register a Hugging Face account? You should've done that sooner.

[graceful tinsel]

i already have an account but tell me how

to add models so people can download

[hallow thistle]

On Hugging Face.

[graceful tinsel]

@hallow thistle

select model?

or data

which one?

@hallow thistle

bruhh✌😭

don't ghost me bruhh

[graceful tinsel]

"skibidi toilet💔🥀"

and what to do next

[hallow thistle]

Generally, I have no idea what to put in "License" part, so please stop asking me further.

[graceful tinsel]

ok but you know how to upload a model right??

or no

[graceful tinsel]

bro giving help suggestion to helper✌

[graceful tinsel]

thanks gng

u my g

u a goat fr

[knotty moth]

just openrail

[rigid haven]

Ty for the information my laptop is Windows 11 Lenovo

[hallow thistle]

What is your laptop GPU?

[rigid haven]

How do I check for that

[hallow thistle]

Open Task Managaer, go to Performance tab, check if there any GPU 0 or GPU 1.

[radiant vessel]

Hi I keep having a problem with the Voice changer I seem to have fix it then it comes back.
What happens is it change my voice fine but then I get a echo of the voice again the same as if I play sounds on the PC it will hear that changed too.
I have bluetooth headphones so it shouldn't be speaker echo and I have tried turning the monitor sound off.

Has anyone had anything like this before I don’t know what is going wrong. Thanks

[rigid fiber]

can someone help I dont even know how to adjust it and when i try adjusting it t js sounds fake 😭

[viral mason]

Ew

Choose something better

Can't believe I tried to help you

I'm very disappointed

[tame hare]

is there an rvc realtime voice changer app i can use for free?
nvm!

[tame hare]

-# i actually downloaded it just for your cyn voice lol hi

[viral mason]

lol haiii

I have a secret one I haven't posted that has no effects, you'd need to add an autotune effect and pitch it up some with effects tho in a seperate program like fl studio :3

I use it personally as it's better than the current public ones

[tame hare]

cool!

[viral mason]

ye it's fun :3

[feral sierra]

hi, is the guide about the deiteris fork still up to date?
https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/

[viral mason]

tbh wokada tg fork is a better choice now, it is better from my experience switching from deiteris to it

[astral junco]

ValueError: could not broadcast input array from shape (4468,2) into shape (5056,2)
How to fix

[feral sierra]

oh okay thanks, I'll definitely try that one out

[viral mason]

👍

If you use Nvidia you will need to download the zip files labeled 001 and 002

If u use AMD download the zip file right below those two

The other downloads idk what they do