✨│ai-help

[low shard]

did you install and setup vac lite first?

[storm fern]

whats that stand for

is it virtual audio cabels then yes

[flint vapor]

There is the tutorial 😄 Virtual Audio Cable means

Tutorial on guide

[storm fern]

i did it

[flint vapor]

#✨│ai-help message

[low shard]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#virtual-audio-cable this, be sure you aren't using vb audio cable, it's worse

[flint vapor]

Oh nice you set it well right?

[low shard]

You need specifically vac lite

[flint vapor]

If you installed the vac from the guide :

Run setup64, not 64a, after extracting the zip to a new folder

After installing the Virtual Cable, it changes your default audio system. Click Yes when it asks you to open the audio device settings (or press WIN+R, type "mmsys.cpl" if you closed it already), and change your Recording and Playback devices back to your usual devices. Same for communications device aswell (right click -> set as default communication device)

[storm fern]

i did that one

idk what moniter is tho like in the menu you thing what do i set it as

[low shard]

monitor is to hear urself optionally, either set it to none or ur headphones

!give-media-perms 1h @storm fern

u can share a screenshot of ur settings now btw

[low shard]

just press the stamp button on your pc to do a screenshot

or press windows+shift+s at the same time

just saying for the future lol

[storm fern]

Ik it takes to long

In my opinion

But thx

[low shard]

gpu: rtx 4060 ti
f0: rmvpe without onnx
chunk: 90
extra: 2.7
output: line 1

on wokada deiteris fork, you can **optionally **use more advanced settings for benefits:

• Advanced Settings -> Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Advanced Settings -> Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Reduce the delay via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

[low shard]

it takes just 1 sec and provides better quality tho 😭 but anyways luckily your settings were visible enough

[storm fern]

No it don’t you don’t

You gotta take a screen shot open google open your files then click the image and yea but it’s in MY OPINION

[low shard]

ohh you go on your files? you can just paste the screenshot (ctrl+v) in the discord chat right after doing the screenshot without looking for any files btw, anyways this isn't important

anyways

did you use the settings I told you?

[low shard]

those are important

[low shard]

those are optional, but give benefits

[storm fern]

Ye

Only for the first one

[storm fern]

The one you replied to

[low shard]

I would suggest you to do atleast the 2 first optional settings as they help with performance and quality/stability, but your choice

[storm fern]

I did

Should it work now

[low shard]

find any models, upload then, then just set line 1 as the input in any other games/discord, and click start and you're done

[broken urchin]

wait

@low shard will the FP32 option work on RTX 3050?

[low shard]

yes, want me to also check your wokada deiteris fork settings while at it?

I can give you perms to send screenshots

[broken urchin]

yeah sure

[analog obsidian]

it works on every gpu but it might affect the perfomance by a bit

[low shard]

!give-media-perms 1h @broken urchin

send a screenshot of the normal wokada settings, and advanced ones

yw, lmk for any issues

[true owl]

hey where can i put the model pth and index in appolio for tex to speech

[broken urchin]

@low shard that good?

[low shard]

anyways, any updates?

[broken urchin]

holy shit the FP32 mode really increases the gpu usage

[analog obsidian]

yea bc its more precise

[broken urchin]

what does that even do?

the FP32 mode

[analog obsidian]

more precise predictions by the ai

[low shard]

enable " disable jit compilation", also putting crossfade lenght to 0.15 can help with quality but add more delay

[broken urchin]

so you sound more natural?

[broken urchin]

okay

[low shard]

also, seeing that model image, is that an anime character or you're looking to do e girl trolling/catfishing?

[analog obsidian]

no, but the model is going to reconstruct the high frequencies better

[low shard]

https://docs.aihub.gg/rvc/local/applio/#tts

[analog obsidian]

giving it a more clean sound

[broken urchin]

its an anime character

i dont troll

hey

i just did what you told me @low shard

[low shard]

alright, that's good since catfishing is bad

[broken urchin]

my perf went up to 70 - 75ms

[analog obsidian]

another thing fp32 infer fixes are the random glitches that happen when using fp16 mode, so randomly fp16 will break the output for no reason, fp32 fixes that

[broken urchin]

i put crossfade to 0.15 and disable jit compilation and perf went up to 70

[low shard]

yeah, that forcefp32 mode and crossfade lenght help in quality but add more delay

while disable jit compilation helps with reducing delay

[broken urchin]

is it worth it though?

so i just keep the 90ms and all the options you told me and the voice changer will be ultra high quality

since the perf is still green

[low shard]

depends if you care more about quality or low delay lol

[broken urchin]

well the delay stays the same right

i keep chunk at 90 or do i increase it?

[low shard]

if your perf becomes red, just put the chunk higher than the perf value a bit, that's the general rule

also if you're going to use it in games

play on lowest graphics, 1080p, 60fps

[broken urchin]

yes

[broken urchin]

its green so i dont need to change anything right?

it uses a ton of gpu

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

!give-media-perms 1h @bleak flax

[low shard]

as long as its green, yes

[broken urchin]

alright thanks

[low shard]

you're welcome

you used a youtube video tutorial right?

those settings are outdated asf

and some settings are specific to an over year old version of original wokada

and you have vb audio cable too, right?

yeah, that causes issues on windows, such as crackling and randomly stop working

you can basically uninstall everything,

delete the old original wokada folder and zip, delete the folder and zip of vb audio cable, and uninstall it from windows app settings

uninstall everything, this is a messy old software, it can cause issues and has worse quality and performance

you should NEVER trust youtube video tuts for realtime voice changers

yes, outdated software that is running on hopes and dreams 😭

NEVER use video tutorials for realtime voice changers

there will be never an updated one

[sterile raven]

I got a question

[low shard]

sure, spit it out and elaborate please, this is a general ai server, there's thousands of ai programs

[sterile raven]

Im trying to setup the w okada vc and i created the shortcut or wtv and the program doesnt open

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[sterile raven]

nvm it just took a min to open

[true owl]

i keep getting an error when trying to convert audio anyways to fix this

[sterile raven]

sorry

[low shard]

I hope you didn't use a youtube tutorial, because they use an over year old version of original wokada with worse quality and performance

[low shard]

if you have something like start_http.bat, that is old asf and you should delete it, there's better software

[sterile raven]

it doesnt have the .bat part

[low shard]

but it has start_http right?

[sterile raven]

yea

[low shard]

.bat is just the file extension, they are the same things

[low shard]

you got outdated software from 2023 😭

[sterile raven]

so where should i download from?

[low shard]

delete the wokada folder and zip, delete the vb audio cable folder, zip and uninstall it from windows app settings

then,

Elaborate:

• your PC GPU
• your operative system
• what you want to do

I will help you get the best and new one, but it depends on your pc gpu and operative system, so i need to know that first

and I also need to know what you want to do, like:

• ai covers
• TTS
• e girl trolling/catfishing in games/vc
• roleplay

because there's different AI programs

@sterile raven could you please elaborate what I asked? I asked those things because they are crucial for me to help you the best

[low shard]

did you follow https://docs.aihub.gg/rvc/local/applio/#tts ? what is the error in the cmd?

[broken urchin]

also i was thinking of getting the RTX 5060 Ti 16GB, i was wondering will i see a significant increase in performance and less delay on the voice changer than i currently have with no games open? I currently have the RTX 3050 6GB

[simple ore]

1. why are you using english text, but non-english tts voice? 2) what is the actual error in the console 3) if you're trying a very large text, it is better to use 'file to speech'

[viral mason]

@zenith mountain training the Combine rn from HL2

[zenith mountain]

Sounds exciting 🙏 lmk when they are out.

[low shard]

well ofc it has better performance generally speaking, but I can't know the exact number

just know theres no such thing as 0 delay tho

[broken urchin]

well, will it be much faster than my current?

[north dirge]

my shitty 3050 laptop 4gb gets like 40-50 ms with no games open. Your gpu should be fine

[broken urchin]

bet

[fluid palm]

Hi

1000 epochs is overtraining for 28 minute dataset?

[simple ore]

it is way too much

[fluid palm]

500 is fine?

[potent bone]

is the e girl tag removed for u guys too?

[analog obsidian]

yep, e-girl tag got removed 1 week ago

[low shard]

yeah it got removed 1 week ago as it was useless

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

that tutorial has been reported by others many times, it's outdated asf

it uses an over year old version of original wokada

and vb audio cable creates issues on windows such as crackling and stopping to work

Also, you need to elaborate on what you want to do to know the best settings and software for you, as many confuse rvc and wokada:

• ai covers
• e girl catfishing/trolling like duckus/novision
• roleplay
• use models on pre-recorded audios

e girl catfishing/trolling like duckus?

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

you need to elaborate first

AI is intensive and complex

Please tell your pc gpu and everything else I asked above

also, you should have a perf value if you got the updated one, but I have the feeling you instead got an old yt tut lmao

the OS, like Windows 10, Windows 11, Linux, Arch linux, MacOS 26, etc

that tutorial is extremely outdated, you're not the first sending it
It uses an over year old version of original wokada with vb audio cable that causes issues on windows

uninstall everything

delete the wokada folder zip, the vb audio cable folder zip and uninstall it from windows app settings

[sonic night]

Hi, does anyone know if there is any update for an RVC?
i have these: vcclient_win_cuda_2.0.76-beta and MMVCServerSIO v.1.5.3.18a

My specs are:
RTX 3050
Ryzen 7 5700G
32GB RAM
Windows 11 - 64bit

[low shard]

that's not rvc, that's 2 different version of original wokada, and both of those are outdated lmao

[sonic night]

oh well im sorry

is there any program thats better? or any update?

[low shard]

RVC = Retrieval-based-Voice-Conversion, the best Few Shots Speech To Speech AI Models (on v2), Inferences (use models) pre-recorded audio (ai covers) and train (make) models. Technically, Mainline RVC does have a go-realtime.bat (aka RVC-GUI), but it's pretty messy and outdated so it's extremely not suggested for realtime.

Wokada = uses RVC for realtime inference. There's 2 main versions, Original made by Wok, and the most suggested one is Deiteris Fork (modified version)

What do you want to do?

• ai covers
• inference on pre-recorded audios
• e girl catfishing/trolling
• tts
• roleplay
• etc

I'm asking since this is a general ai server, so i need to know which program you need

[sonic night]

mmmmmmmm, i mainly use for roleplaying or test and troll friends

but it would be nice the ai covers one, or a voice to voice conversion?

audio to voice or audio to audio

[low shard]

so like peter griffin with friends ?

you'd need 2 different programs to do both AI Covers and Realtime Voice Changing to roleplay as peter griffin for example

[sonic night]

test for looking models and play with it hearing myself
roleplay with friends
and troll, maybe with some srpelo voice lol

[low shard]

great then, as e girl catfishing is bad

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

1st link wokada deiteris fork

[sonic night]

well, ai cover sonds good, its there any local program to do that? maybe with the models i already have in my ai

?

[low shard]

this is the voice changer

[sonic night]

nice

what about the covers one?

like mp3 to mp3 or smth

[low shard]

https://docs.aihub.gg/rvc/local/applio/ applio is an rvc fork, easier with enhancements and uses the same rvc models

[sonic night]

ohh

so it does both things?

or only the ai cover?

[low shard]

applio only ai covers and inference (use models) on pre-recorded audios

wokada deiteris fork only realtime voice changing

[sonic night]

oh nice

ill try it, thanks<3

[knotty moth]

if you care more about quality, it's recommended to do inference on pre-recorded audios

rather than the realtime one

[sonic night]

i think i will do both

[low shard]

he wants to do both ai covers and realtime voice changing

[low shard]

you're welcome

[lament mason]

how do i make it not delayed

[sonic night]

its ok, btw, now i have applio but im trying to find where to put the voice models

nvm im looking the docs

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[lament mason]

Can internet be the cause of the delay?

Also I use nvidia idk how to spell it

[low shard]

elaborate the things I asked you

We don't even know what program you're talking about

[low shard]

You need to check your PC GPU, operative system, what you want to do, the tutorial link you followed and send a screenshot

You can check your pc gpu on Windows via:

ctrl+shift+esc (task manager) -> Performance tab -> GPU

[low shard]

Nvidia Is Just a GPU brand, you need to give the full name

There are Nvidia gpus that are good, some old ones that arent

[lament mason]

Okay so, it says the gpu is Nvidia Geforce RTX 4060 Ti

https://www.youtube.com/watch?v=LX5en3pZJwM

this is the tut i was using

recommended by my friend

to download the program i assume

[knotty moth]

never trust any video tutorials about voice changer

I'd recommend getting the better version here

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

[knotty moth]

and that video is for using beatrice model

it is faster and can run on average cpu

but the quality is too mediocre compared to RVC models

[lament mason]

Thanks ill try it out

[brittle wing]

Is there a software that scans the media files i have like images or videos, and looks them up on imagebooru sites or social media, and redownloads them in higher quality and with the imagebooru tags?

[slender python]

thanks!

[oak edge]

can I change pretrain model in applio's google colab public url

(I'm trying to voice swap songs from south asia so need some advices thanks in advance)

[clear citrus]

can someone give me the link for deitries

[oak edge]

some one tell what do these mean in applio rvc

[rose leaf]

how i remove the ai staff from my pc i wanna ask ?

[flint vapor]

Copilot?

[flint vapor]

-wokada

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[flint vapor]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/ 😄

[gentle forum]

why cant i find run-install on applio?

[flint vapor]

You have downloaded from the huggingface repo right?

Or from github zip?

[flint vapor]

https://huggingface.co/IAHispano/Applio/resolve/main/Compiled/Windows/ApplioV3.2.9.zip

Download from here

[simple ore]

you did not follow the doc

what exactly did you download?

why

what's you gpu?

oh.. so you did not read ai hub tutorial after all

[flint vapor]

No no, the one have you downloaded is outdated now and i don't reccomend you anymore, i suggest you to download Deiteris

[simple ore]

-rt

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[flint vapor]

And to follow the guide

[flint vapor]

That @simple ore sent to you now

Yes follow the deiteris one 😄

Deiteris can have better performance and is more updated than the original, i suggest you to use deiteris

You welcome 😄

[hollow kraken]

Does anyone have a Goku voice that sounds good in Latin?

[flint vapor]

There is download link on the guide that Noobies sent you

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

Which gpu do you have?

You are on Mac?

Wait i give you download link directly

Wait you have mac silicon or mac intel?

I'm not so practical with these things about mac

You should search on info of mac. i think is it on apple logo on high left

You can see from cpu, if your cpu is intel, or is mac silicon

Ok is silicon

https://github.com/deiteris/voice-changer/releases/download/b2332/voice-changer-macos-arm64-cpu.tar.gz

Download from here 😄

[low shard]

you and your friend use an over year old version of original wokada, with vb audio cable that creates issues on windows, never use video tuts for voice changers

[low shard]

both of yall delete the old program yall got first

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

what program? there's thousands, and you need to elaborate which amd gpu

elaborate first

[pastel token]

can i have cell perfect voice

[low shard]

Firstly we need to check that your pc is good enough to run it, and to be sure you're using the right program for what you want to do, since many confuse wokada, deiteris and rvc all together

Elaborate:

• your PC GPU
• your operative system
• what you want to do

[low shard]

what ai stuff? this is a general ai server

NEVER trust video tuts for voice changers

thats outdated, its original wokada

tell your OS, what you want to do, and what you did

what's your OS? and what do you want to do? RVC doesn't mean Realtime Voice Changer, and RVC & Wokada are 2 different things:

RVC = Retrieval-based-Voice-Conversion, the best Few Shots Speech To Speech AI Models (on v2), Inferences (use models) pre-recorded audio (ai covers) and train (make) models. Technically, Mainline RVC does have a go-realtime.bat (aka RVC-GUI), but it's pretty messy and outdated so it's extremely not suggested for realtime.

Wokada = uses RVC for realtime inference. There's 2 main versions, Original made by Wok, and the most suggested one is Deiteris Fork (modified version)

which operative system?

yeah there are different versions
you need atleast macOS 12 Monterey or later, check what version do you have

also, im guessing you want realtime voice changer in games/vc right? since wokada isn't for pre-recorded audios

did you do also the apple quarantine stuff?

alr great

you also got https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#virtual-audio-cable ?

also be aware that mac isn't great for AI, so don't expect to be using it in lowest delay on games

the program runs locally on ur hardware, but uses a web user interface like most ai programs

I just explained you, you can read https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#why-does-it-run-in-a-browser-and-not-its-own-window for more info

it is an app, it just uses a web user interface, it's completely normal for AI

you can't open it in its own window

what browser do you use btw?

great, show a screenshot of the settings

it's not a website... it runs on your hardware, it's a normal program like everything else, it just uses a web user interface that is hosted on your own pc only for yourself

please read the link i sent for more info

it's not connected to anything, it would work even completely offline (after download ofc)

can u show a screenshot of ur settings so i can help up?

!give-media-perms 1h @quick furnace

f0: fcpe
extra: 1.0
chunk: depends on the perf value at the top left after you click start, while running

Advanced Settings -> Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)

You can search rvc ai voice models at:

• https://discord.com/channels/1159260121998827560/1175430844685484042
• In https://discord.com/channels/1159260121998827560/1163592055830880266 , Do /find with @earnest musk
• https://weights.com/ (login required)
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• Suggested Models for Realtime Voice Changing (Wokada)
• https://voice-models.com/
• https://thevoicemodels.com/ (for Turkish Models, login required with discord and level 2 on their server)

if there isnt one, you can:

• make it yourself with our docs guides
• Ask a free request in https://discord.com/channels/1159260121998827560/1159289738314919936
• Be aware that we don't allow any paid comms

[earnest musk]

:wave: @low shard, How can I help?

Available Commands:
• @weights find <query> or /find <query> - Search for RVC Voice Models

[low shard]

Do you need help with finding a specific RVC voice model?

Alr lmk

[flint vapor]

Don't use mainline wokada, i recommend you to use deiteris, first all which GPU and OS you have?

[proven hill]

i need help

[low shard]

no, chunk controls delay

share a screenshot of ur settings

[spice sluice]

How to change voice on music with ai?

[low shard]

that's original wokada, that's outdated, never trust video tutorials for voice changers

you didn't use the settings i told you #✨│ai-help message

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

can you elaborate with what you need help with?

alright, are you going to use it in a game? if so, put lowest graphics, 1080/720p, 60/30 fps cap

[spice sluice]

I want to make Michael jackson smooth criminal cover with greek rapper name arab voice but i cant find any program-site to do it

[low shard]

what games are you using it with?

close everything else in the background, show a screenshot of u using it so i can check ur settings again and perf value

that's just the input sense, you don't have to touch it, it's related only if you have background issues

you can check every settings in the writtten guide, or just by hovering your mouse over it lol

the chunk has to ALWAYS be higher than the perf value

yeah that's bc you put ur chunk as 2.7 ms

set it to like 230

perf always have to be green

[proven hill]

i followed a youtube tutorial

no

[low shard]

close everything in the background, make always sure your chunk is higher than perf

play with the pitch, higher is more feminine, lower is more masculine

try other models too

there's thousands of rvc models, not all of them are good

[proven hill]

catfish

[low shard]

also there aren't models who can do realistic laughing or screams, rvc limitation

[proven hill]

bye bye

[flint vapor]

He put me the last emoji, goodbye chillax

Oh the emoji gone haha

[proven hill]

everything gone

[proven hill]

https://tenor.com/view/drug-anime-girl-gif-26426197

[low shard]

smh

[proven hill]

catfish hub ong ong

[low shard]

do you actually need help or just joking

[low shard]

nuh uh, catfishers banned, this is a general ai hub

[proven hill]

im not stupid

[proven hill]

yea but most of users are them

[low shard]

you can't say that while users doing those things get banned, they aren't our users lol

[proven hill]

ye but they come here

[low shard]

people who come here for catfishing get banned as you saw rn

[low shard]

just asked if u actually needed help like that time with flux comfyui 😭

[proven hill]

comfyui my beloved

[abstract ivy]

how to use voice changer?

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using (if any)
• a screenshot of the program (if any)

[astral snow]

I want to know if it is possyble to remake an instrumental with AI

[solemn burrow]

hello all, i downloaded a nick eh 30 voice model and when i speak its so crunchy and lagging. im on a macbook m1 pro by the way.

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[odd isle]

why did realism get removed from docs?

[solemn burrow]

i wanna use this new nick eh 30 voice changer but it sounds so bad when i downloaded

i’m using the official guide

[low shard]

you need to elaborate what i asked, there could be over 10 programs and 'official guides' you could be talking about lol

[low shard]

It wasn't needed, a good model doesn't really need those post processing things, plus we had no staff wanting to maintain it, and barely any users ever asked help for it

[solemn burrow]

i meant the download guide from the bot in this server

[warped quail]

how to lower the weights models to put on the mangio rvc?

[low shard]

please elaborate what I asked

[low shard]

mangio rvc is abandoned since 2023, dont use it and dont trust yt tuts

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[warped quail]

ouh, ok

[solemn burrow]

i wanna use the nick eh 30 one

[low shard]

ai is intensive and complex

you have to elaborate to let me understand how to help you

[low shard]

please elaborate what I asked to let me help you, uninstall everything you got off youtube

[sterile needle]

how can i change my accent to different language or smth like that

[warped quail]

My GPU: RTX 4060

My operating system: Windows 11 Pro

What I want to do: Use voices from anime and famous characters

Which tutorial link am I using: None

A screenshot of which program?

[low shard]

please elaborate more

[low shard]

you got a nice gpu

but please elaborate more on what you want to do, like:

• TTS
• Realtime Voice Changing For Roleplay
• inference (use models for pre-recorded audios)
• ai covers
• etc

since we do many programs, and users may confuse them, since there isn't a program that does everything

[sterile needle]

i have rtx 3060ti i want to get polish accent

[warped quail]

I want Realtime Voice Changing For Roleplay

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

1st link, wokada deiteris fork

[warped quail]

ready, what now?

[low shard]

did you read the guide?

[low shard]

ehh there isn't really a program that does that, you can try a polish model in the realtime voice changer, wokada deiteris fork, if that's what you're looking for, but it will sound weird speaking anythng that isn't polish

[warped quail]

no, but I will read it now

[low shard]

please read it carefully, missing a step can fuck up the whole audio system

[warped quail]

ok

[low shard]

let me know for any issues

[low shard]

I'm guessing you're on windows 11,
What tutorial link did you use? or may you show a screenshot of the program settings?

I have a feeling you used a yt video tuto, they are all outdated and use an over year old version of original wokada with vb audio cable that creates issues on windows

!give-media-perms 1h @vague cave

yep, that's the one I was talking about

It's outdated asf

it uses an over year old version of original wokada (worse performance and quality), uses vb audio cable that creates crackling and random issues on windows

NEVER trust video tuts for realtime voice changers

delete the wokada folder zip, delete the vb audio cable folder zip and uninstall it from windows app settings

[flat zealot]

im lowkey like really outdated with this stuff

how do i run Music Source Separation Training on my pc

instead of like having to use a colab

i would use uvr5 ui which i use rn but like it doesnt have all the models i want

also idk how to use a colab and i dont feel like learning rn lol

[viral mason]

tall mario

[viral mason]

true it's not catfish hub

[proven hill]

catfish hub

[viral mason]

https://cdn.discordapp.com/attachments/1344423628954927126/1393351946877796473/ezgif.com-animated-gif-maker_2.gif

[viral mason]

@runic stump I saw your deleted msg, rvc disconnected is not being maintained anymore and hasn't for a while

[runic stump]

I've been using rvc v2 disconnected to train and mangio-rvc-v2 locally, but i took a long break and rvc disconnected was discontinued. What's the current best colab/method I could use for training quality-wise that's still compatible with a local gui/if i should switch that as well.
(I use windows 10 and don't have much knowledge setting complicated code/environment things up myself) Thanks so much for anyone who can help get me up to speed.

[runic stump]

Just seeing this, yeah I'm aware. Sorry had to make sure I was posting in the right place.

[analog obsidian]

applio

[flat zealot]

yo so like

im trying to use the revive 2 model right

i put it in uvr

it genuinely just crashes my pc

idk like what to do about that but like

[tropic plank]

sa

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[tropic plank]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[flat zealot]

rtx 5080
windows 11 24h2
I want to use unwa's revive 2 model with gpu conversion YET my pc black screens (obviously this is a really heavy model BUT i didn't expect it to make my pc crash)
no tutorial link

i just like imported it manually

[simple ore]

if you're using version 5.6 it may be way outdated

[flat zealot]

how do i update

[simple ore]

you can try this instead - https://github.com/Eddycrack864/UVR5-UI/releases

[flat zealot]

got it

[simple ore]

it has 5000 series support

[flat zealot]

i actually have been using this BUT im unsure on how to import my own models

so i just went back to uvr5

[simple ore]

download the compiled version

[tropic plank]

anyone can help me rvc

@timid olivefav helperım

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[tropic plank]

i downloaded this

i want realtime voice changer

How should I adjust the settings section?

[low shard]

what's your pc gpu? operative system? what you want to do?

[tropic plank]

windows

nvidia

[low shard]

which nvidia gpu?

[low shard]

which? windows 11?

[tropic plank]

11

[low shard]

you need to elaborate since AI is complex and intensive

alr good

[low shard]

yes that is wokada deiteris fork

[low shard]

did u do https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#virtual-audio-cable

[wooden wharf]

[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[tropic plank]

no

[low shard]

you have to, read it carefully because missing a step can fuck up your whole audio system

[low shard]

@wooden wharf I hope you didn't use a youtube video tut, elaborate what I asked so I can help you please

[tropic plank]

done

now ill run mmvcserver

[low shard]

did you do setup64 and the recording and playback thing?

[low shard]

that's wrong

[tropic plank]

right?

how

[low shard]

you have to set your usual input/output device as default in both recording/playback

Line 1 shouldn't be the default one

[tropic plank]

i appplied both

[low shard]

you didn't in the screenshot, Line 1 is still default

[tropic plank]

hmm

?

right?

[low shard]

show a screenshot of the entire playback and recording tab

[wooden wharf]

понимаешь ли ты на русском?

[low shard]

This server is english only

[tropic plank]

@low shard

[low shard]

great

[low shard]

gpu: rtx 4060
extra: 2.7
f0: rmvpe without onnx
chunk: 90
input: microphone
output: line 1
monitor: optional to ur headphones

on wokada deiteris fork, you can **optionally **use more advanced settings for benefits:

• Advanced Settings -> Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Advanced Settings -> Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Reduce the delay on Windows via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

[low shard]

@wooden wharf I'm not russian, and this server is english only, It's best you either speak english or use a translator

[wooden wharf]

To understand the app, I used a Russian video on YouTube
All preset voices don't work except for Beatrice-JVS speaker:1
I'm using a different audio cable, but the sound is outputting to Discord (again, using Beatrice-JVS as an example)
When I installed other AI voices, I encountered the same error ([Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting to generate pipeline...)
RTX 3060ti. Ryzen 5 3600. 32 gb RAM

[low shard]

All youtube tutorials are outdated, you're using an over year old version of original Wokada, and "vb audio cable" creates issues on Windows
Beatrice models are lightweight but lower quality, there's almost none becuase most are RVC

It's best that you delete everything and uninstall it from the Windows App Settings

You got a good PC,
what are you looking to do? AI Covers? Realtime Voice Changer? TTS? Use Voice Models on Pre-recorded Audios?
I'm asking since there are many ai programs

[wooden wharf]

I only need it for the rp server.

[low shard]

So realtime voice changer for roleplay right?

[wooden wharf]

yes

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

Please read carefully the 1st link, Wokada Deiteris fork

@wooden wharf let me know for any issues!

[wooden wharf]

Okay, thanks, I'll try to understand.

[tropic plank]

@low shardthanks

[low shard]

you're welcome, need any other help?

[tropic plank]

no thanks

[pastel oak]

Almost 0 delay is never possible, closest to lowest delay is with an RTX xx90

• using deiteris wokada version

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

that gpu is old asf, share also the tutorial link you used a nd a screenshot of the program

[sterile yacht]

Is there any tutorial on how to make our own models?

[low shard]

im guessing RVC models, in the https://docs.aihub.gg

but what's your pc gpu?

[sterile yacht]

AMD Radeon TM graphics

Not the best for the job I guess

[low shard]

That's integrated graphics, won't ever get recognized for any local ai task

use cloud options

[sterile yacht]

Is it expensive?

[low shard]

cloud is remote good pc with limited free time

[low shard]

Ai is intensive and complex, it ofcourse needs good hardware

Train (make) RVC Models on cloud:

1. Prepare the Dataset
2. Setup RVC:
   Choose a cloud way to use RVC,

• Google Colabs (max 4 hours of daily T4 16gb gpu not granted for free, not much hours for training, but easy to use, there's a paid tier):
  • Applio (UI)
  • Mainline (U, broken right now)
• Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus, either T4x2 16gb each or P100 16gb, only free):
  • Applio (UI)
  • Mainline (UI, broken right now)
• Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly, Free Studios run 24/7 but require restart every 4 hours. There's a paid tier):
  • Applio (UI)
  • Mainline (UI, No guide as of right now)

3. Be sure to know about the tensorboard

Google Colab = Easier but risk of getting disconnected
Kaggle = Harder but way more gpu time
If you are looking for the easiest way and for free, is using https://weights.com/ which ofc uses RVC

RVC Inference (use models) on pre-recorded audio on Cloud

You can use either:

• Weights.com: Easiest Possible Ever Automatic
• Ilaria RVC Zero: Fastest free on cloud
• Applio UI Colab: RVC Fork with some extra features like TTS
• RVC AI Cover Maker UI: Automatically Separates the Vocals and Instrumentals, converts the voice and mixes them back

[idle osprey]

I think its because that, you have to delete Model_dir folder in the voice changer folder, which is located next to where is MMVC is.

[fleet cedar]

@low shard

[low shard]

#🧬│ai-chat message from what you said it might be cutting off issues

[fleet cedar]

by the way can u help me ban this creep randomly dming me this who dmed me when i talked here

[low shard]

Try to set extra to 2.0, and be sure to not speak quieter at the end of your words (which you might have been doing since you have monitor to headphones to hear urself lol)

[fleet cedar]

no its working fine, i meant like when i talk in a voice message its like at the start its normally loud then while continue speaking its quiet

[low shard]

Oh wtf 💀 thanks for reporting, please next time dm @hushed perch for creeps or other stuff

[fleet cedar]

wont that make low quality

[low shard]

it is a slight 0.7 difference, and also high extra can cause cutoff issues, the quality depends mostly by the model tbh

please try it first

[fleet cedar]

i tried it and its so high pitched

wait wtf

why is line 1 making my voice models sound like a trash mic quality

that makes it high pitched

[fleet cedar]

dude

line 1 is making my voice model sound like a child

😭

at the start it was working fine

idk why

i only went afk for a min then came back

and ts shi happens

[low shard]

this is happening after setting extra to 2.0? maybe you accidentally changed the pitch

[fleet cedar]

no

[fleet cedar]

great now my line 1 doesnt pick up audio

after restarting my pc

wtf is happening

[low shard]

could you please restart and show a screenshot again of ur settings?

[fleet cedar]

restart what

[low shard]

the pc and voice changer

[fleet cedar]

ok

doesnt work wtf

[low shard]

set extra to 2.0
mon to 100

[fleet cedar]

still doesnt work

[low shard]

out to 150

[low shard]

what doesnt work

[fleet cedar]

line 1 literally doesnt pick up audio

[simple ore]

did you download 5000-series version?

[fleet cedar]

yes

i did

[low shard]

do you have any other voice changer related programs such as voice.ai, vb audio cable, voicemod, dubbingAI, etc

[simple ore]

i'd try server mode first

[fleet cedar]

no i dont

[simple ore]

in case mic access disabled by browser

[low shard]

what browser do u use?

[fleet cedar]

firefox

[low shard]

maybe try chrome and be sure microphone perms are allowed
else it's better to use server with wasapi/asio, which is a bit harder to setup but even makes you have less delay

this doesn't seeem related to the VAC

[simple ore]

click that thingie

and see what you have

[fleet cedar]

its the same thing with chrome

[low shard]

temporarily
that explains lol, seems to reset everytime

be sure its allowed like in @simple ore 's screenshot

[simple ore]

well, it is allowed

[fleet cedar]

weird it works in chrome

but i prefer firefox tho

wtf

i dont wanna use chrome

[low shard]

might be related to some browsers to some specific versions and settings, the best is that you just use chrome for the voice changer only

[fleet cedar]

it worked before

but not now

wtf

wym

[low shard]

might be a firefox update, not sure I don't use firefox

[fleet cedar]

firefox is not same as google chrome

cause it makes my voice models sound better

for some reason

[simple ore]

microphone privacy settings?

[low shard]

some users reported specific errors to some specific browsers version and settings tbf, like OperaGX which is also common to not work, while usually either firefox or chrome works

[fleet cedar]

where do i find microphone privacy settings

[simple ore]

looks good

[fleet cedar]

lol i fixed it by clicking clear settings in wokada

@low shard but why is line 1 making it sound like a weird mic

like a little high pitched

[remote mist]

Hello I dont have the exe when I want to download okada

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

maybe try playing with the pitch, this is weird

[fading lodge]

This isnt help but a question. Where did the model master shop go??

[solemn burrow]

@low shard can you give me media perms for like an hour i wanna share so I can elaborate on yesterdays question

[knotty moth]

paid commissions are no longer allowed in this server

[sand bison]

A question, Google Collab isn't available yet to create AI RVC voice models, right? Sorry if I've asked this before.

[oak edge]

anyone know how to mount drive in applio colab

[simple ore]

you click mount drive cell, you get popup to pick an account and give permissions

[oak edge]

😭 i'm not getting anything

it's just spinning and nothing else

[knotty moth]

@oak edge make sure you confirm those two popups instead of dismissing it

if not sure, try disconnecting the session and start it again

[west rampart]

do you guys use vcclient for ai voice changer?

[cerulean berry]

What concepts and what do I need to get into ai if I use c++?

[knotty moth]

cuda GPGPU programming

or even contributing into ZLUDA development

[wooden wharf]

an error occurs after renaming model

Error
unhandledrejection
no error stack

TypeError: can't access property "modelSlots", e is null

[craggy bough]

they do tho

then dont use them

they dont seem particularly trustworthy

i fail to see how ai generations relate to health insurance

i dont care, im not signing in

i dont trust them simply by looking at their faq

if you dont trust them, dont use them

its not owned by this server

nothing anyone here can do about it

great, people here use the big name ones

i think thats all the info you need to know about this server's relation to chatllm

the only mention of it is from you

okay then, go ahead and believe what you want

not worth my time explaining

personally, i wouldnt trust them

no support?

even less trustworthy

even the shithole that is grok has support

1. no support
2. their faq makes claims but fails to back them up ("we give you about 10x more tokens than any other LLM service", which other service(s)?
3. also lists privacy things that are completely unrelated to ai prompts (what does generations have to do with health insurance)

[quiet mulch]

my pc is stuck with saying activating like it's been 15 minutes

i mean the ai voice changer

[flint vapor]

Which voice changer are you using? are you using deiteris fork or mainline okada?

[quiet mulch]

okada

i fixed it

but now it keeps saying client closed

and it doesn't open at all

[flint vapor]

I recommend you to use deiteris fork, is always okada

Which GPU do you have?

[quiet mulch]

amd

[flint vapor]

Which model ?

[quiet mulch]

but it's not really one of the best ones

don't judge me

AMD ryzen 4000 series 7

[flint vapor]

I not judge no one don't worry i need for help you

[quiet mulch]

aight thank you!

[flint vapor]

This is a CPU Not gpu, as gpu you have integrated like vega? or you have another GPU?

[quiet mulch]

i don't think i have a cpu

it's a hp laptop

[flint vapor]

No the one you mentioned me now is a cpu

[quiet mulch]

I MEA GPU

sorry my bad

[flint vapor]

If you not have cpu your pc can’t work

[flint vapor]

You have gpu too

But you have to go on task manager

[woeful basin]

does anyone know why my thingy just sounds weird compared to what i hear on videos, i cant send a recording (.wav) cuz idk pls help

[flint vapor]

Then select performance and see the model 😄

[woeful basin]

i got a recording though

[flint vapor]

Voice changer ?

[quiet mulch]

where can i check it?

[flint vapor]

Search on your windows bar, Task Manager

[quiet mulch]

aight

[flint vapor]

Then go on 3 lines on up after you open it

[woeful basin]

yeah with this program

[flint vapor]

And select performance

[flint vapor]

You have downloaded mainline okada or deiteris fork ?

[woeful basin]

?

[quiet mulch]

oh i see

[woeful basin]

i got it and i can hear the voice but it just sounds quiet and odd

[flint vapor]

Oh which voice changer did you download ?

[quiet mulch]

AMD Radeon(TM) Graphics

[woeful basin]

i tried coolkid from forsaken roblox awile ago but rn im trying hatsune miku

[quiet mulch]

that's my gpu

[flint vapor]

Oh not the model, but the voice changer are you using

[flint vapor]

There isn’t a number ?

[woeful basin]

wdym

[quiet mulch]

no.

it's a normal laptop

[quiet mulch]

he means are u using okaida or deiteris fork

[woeful basin]

uhhh

[flint vapor]

You can’t use it on local if you have integrated but only in cloud I will send you now a link 😄

[quiet mulch]

okay!

[flint vapor]

https://docs.aihub.gg/rvc-voice-changer/cloud/w-okada-kaggle/

[woeful basin]

its the voice changer from this video https://youtu.be/_JXbvSTGPoo?si=NpZwLpeNtMUQfxqG

[flint vapor]

Is very old I will send you now a new updated one

[woeful basin]

i forgot what its called but this is where i got the program

[flint vapor]

Which gpu do you have ?

[woeful basin]

whats a gpu i forgot

[quiet mulch]

does it cost money?

or is it free?

[flint vapor]

Graphic cards, you can see on task manager

[flint vapor]

Is free it have some limits

[quiet mulch]

oh okay so i don't have to pay for it right?

[flint vapor]

No don’t worry only if you use it more than 30 hours in week 😄

[quiet mulch]

OOOHH

[woeful basin]

amd i think

[quiet mulch]

that's really good

[flint vapor]

Can I see the model? You can see it going on task manager, then go on 3 line above left

Select performance

[woeful basin]

where do i show u i dont have image perms here

[flint vapor]

Oh don’t worry you have only to say to me the model

[quiet mulch]

should i just open the setup.exe instead with the 64 if i have system32

[flint vapor]

Setup64

Not the setup64a

[quiet mulch]

oh

i opened the setup 🙊

the one without the 64

and 64a

[flint vapor]

System32 is In every version of windows, you have 64 bit don’t worry

You can open the 64 one

[quiet mulch]

oh okay

[flint vapor]

Also if you have wrong don’t worry

[quiet mulch]

oh okay!

[alpine onyx]

Is it normal it says Index search FAILED or disabled?

[quiet mulch]

what should i do now?

[flint vapor]

Are you using rvc or voice changer ?

[flint vapor]

Oh there is all in the guide 😄

[alpine onyx]

RVC

[woeful basin]

in gpu tab it says AMD Radeon RX 5700

[quiet mulch]

oh ok!

[alpine onyx]

Couldn't get the one from w-okada to work

[flint vapor]

After installing the Virtual Cable, it changes your default audio system. Click Yes when it asks you to open the audio device settings (or press WIN+R, type "mmsys.cpl" if you closed it already), and change your Recording and Playback devices back to your usual devices. Same for communications device aswell (right click -> set as default communication device)

[flint vapor]

Wait now I will check

[woeful basin]

?

wdym

[flint vapor]

You can use it in local

[flint vapor]

If is compatible 😄

I will send you now link as guide and download link

[quiet mulch]

so i have to change my input and output to like normal?

[alpine onyx]

I followed the guide step by step too

[flint vapor]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/ Guide

https://github.com/deiteris/voice-changer/releases/download/b2332/voice-changer-windows-amd64-dml.zip Download for AMD ONLY

[woeful basin]

whats this, i already have the voice changer installed

[flint vapor]

Yes 😄

[flint vapor]

Which gpu you have ?

[alpine onyx]

3060

[flint vapor]

Uninstall the old one

You have old version

[flint vapor]

You want use voice changer right ?

[alpine onyx]

yep

[quiet mulch]

ok just did it

what now?

[woeful basin]

can i dm you what it currently sounds like since i cant send it here

[flint vapor]

Follow this guide https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

[alpine onyx]

But no matter if I use w-okada or RVC there is always an issue with index

[flint vapor]

I send you download link for your gpu

https://huggingface.co/Shadicti/deiteris-Fork/resolve/main/voice-changer-windows-nvidia-b2332.zip?download=true

Try maybe this version will work

[flint vapor]

Sure I’m using phone now so I will listen in a bit

[alpine onyx]

So just not use RVC and use the voice changer from W-okada ?

[flint vapor]

Now only follow guide step by step😄

[quiet mulch]

ok

[flint vapor]

Rvc is another thing, is only for train model and voice conversion, voice changer use rvc model for speak in real time

[alpine onyx]

hmmm ok

I will try w-okada ones again

but I had index issues with it too

[flint vapor]

Try the okada from
The guide I send you

Not the mainline one

[alpine onyx]

which one ?

cuda001 or 002

[flint vapor]

No one of this, follow the guide I send you

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

[alpine onyx]

That is the fork from the guide you sent me

[flint vapor]

https://huggingface.co/Shadicti/deiteris-Fork/resolve/main/voice-changer-windows-nvidia-b2332.zip?download=true

[flint vapor]

This is the download 😄

[woeful basin]

ok i sentit

[flint vapor]

Ok in a bit I will listen to it

[alpine onyx]

thanks!

[quiet mulch]

okay i'm done on the things in the guide what should i do now?

[knotty moth]

depend on your hardware

• cuda: Nvidia gpu
• rocm: AMD gpu for linux
• dml: AMD/intel gpu for windows

[flint vapor]

Oh he have 3060 😄

[alpine onyx]

Yeah but there is cuda 001 and cuda002

[flint vapor]

Your voice changer interface is opened now ?

[alpine onyx]

anyway kakao sent me the download link needed

[quiet mulch]

should i open the okaida?

[flint vapor]

Yes 😄

[knotty moth]

download through this link

https://huggingface.co/Shadicti/deiteris-Fork/blob/main/voice-changer-windows-nvidia-b2332.zip

[quiet mulch]

sadly it's still crashing 😦

[flint vapor]

What it says ?

[flint vapor]

@alpine onyx is the same link don’t worry

[knotty moth]

the difference is it is just a single file to extract

[alpine onyx]

what exactly do I start now? I have these.

Do I just start the MMVCServerSIO.exe ?

[flint vapor]

Yes run it

[alpine onyx]

ok it's open in a browser now instead of locally

[flint vapor]

You followed all correct ? Did you put your ngrok token?

[flint vapor]

Yes don’t worry is always local

[quiet mulch]

uhhh what's that?

[flint vapor]

Now in the guide you will see which chart put for chunks and extra

[flint vapor]

Is in the guide always 😄

[alpine onyx]

nope it doesn't change the voice

it's still my voice

listening through passthru

and on speaker I only have targert[0] instead of target[1]

[knotty moth]

if the passthru is red, click it to change into green

[alpine onyx]

The passthru doesn't turn green at all. If I click on it, it blinks red and when I click again it turns gray

[quiet mulch]

i need a phone number but i don't have one 😦

i can't verify😭😭

[knotty moth]

stop the voice changer first before changing it

[quiet mulch]

on the kaggle.com site

[alpine onyx]

it is stopped. Still, the same thing happens I just mentioned

[woeful basin]

i dont wanna sound like im rushing u but can u check now

[knotty moth]

please show the screenshot, the passthru button should be changeable while stopped

!give-media-perms 10m @alpine onyx

[flint vapor]

Yes 5 min max

[alpine onyx]

@knotty moth

[flint vapor]

I think you didn’t select a model

You have to put a model first

[alpine onyx]

I did

[woeful basin]

can someone help my voice changer sounds quiet and odd compared to what it should be

[alpine onyx]

I loaded in the index and pth of the model

[woeful basin]

i dont know how to explain it

[alpine onyx]

the model name is even on top of those buttons

[woeful basin]

but i have a recording of it too

[flint vapor]

Have you installed VAC lite? And also set chunks and extra ?

And you click start ?

[knotty moth]

try clicking the passthru button and change the setting so it would turn into green

[alpine onyx]

I have VAC lite installed and I couldn't find anything to chunks and extra in the guide

[knotty moth]

then show also the audio settings

[flint vapor]

I give you one but are not permanent just for try, put as chunk 298 and extra 2.7

[flint vapor]

I know chunks looks high but is just temporary

F0 put rmvpe

[alpine onyx]

ok it works a tiny bit now? But the pronouncication is really bad

[flint vapor]

Pronunciations is all due to model

[alpine onyx]

where exactly is this ?

[flint vapor]

Maybe try put index 0.5 but it should not fix so much

[woeful basin]

can someone help my voice changer sounds weird, i dont know how to explain it but i have a recording

[flint vapor]

On left

[knotty moth]

well...

it mostly depends on the model quality itself

[flint vapor]

Yes in a bit I will check

[knotty moth]

so try another better one

[alpine onyx]

any other changes ? Also where exactly do I find the chunk thigns? A higher chunk with this model did not work.

[woeful basin]

sorry im getting impatiant

[flint vapor]

Is all written on interface

Chunk and extra

And F0 detector

[flint vapor]

Look on right and left

[flint vapor]

Don’t worry just 1 min

[alpine onyx]

where 😭

[flint vapor]

I will go to pc

[woeful basin]

dude i wish i had embed perms here 😔

[flint vapor]

Not here

[flint vapor]

Here

[alpine onyx]

one last question. There isn't a thing such as the more you use the model, the more it gets trained right ?

[knotty moth]

there's no such false belief like that

[solemn burrow]

@flint vapor can you give me media perms for like an hour i wanna share so I can elaborate a question from 2 days ago

[flint vapor]

I still can’t give permission I don’t have powers for do it

[knotty moth]

!give-media-perms 10m @solemn burrow

[flint vapor]

The more is gets trained you can risk overtraining, it depends from dataset mostly, correct epochs, there are many good models you can try

[woeful basin]

i just checked why do i have media perms i didnt earlier

[knotty moth]

or you could also consider making a thread in https://discord.com/channels/1159260121998827560/1192011222023950368

[alpine onyx]

oh one more thing. what's the difference in letting the audio run through client and server? I'm seeing that for WASAPI you have to let it run through server?

[flint vapor]

Server can run a bit better but quality is not so good

[alpine onyx]

ah oki

[solemn burrow]

Im trying to use nick eh 30 voice changer but i cant hear my voice clearly im on a macbook

[knotty moth]

show also the console window

[solemn burrow]

gimmie a sec

[woeful basin]

whatever il put this here ig
im trying to use a miku voice changer, why does it sound like this

[solemn burrow]

wdym by console window

nvm

[solemn burrow]

as in the whole terminal thing?

[flint vapor]

Which chunk did you put?

And extra

[knotty moth]

• use Sup2 (in client audio mode) or a 3rd party noise suppression
• turn up chunk depending on your gpu performance
• F0 should be rmvpe/onnx

[woeful basin]

i think in that recordin i had this or 700

[knotty moth]

well that's a bad extra setting and the version itself

[woeful basin]

wdym

[flint vapor]

But i send you a link for download the correct version i said to not use anymore old okada

[solemn burrow]

did u mean the whole terminal thingy ma jiggly thingly?

[woeful basin]

is that really the issue though, also can u send it in dms il do it later

[knotty moth]

it's called terminal in mac

[solemn burrow]

oh ok

[flint vapor]

Please uninstall the current voice changer you have now

https://github.com/deiteris/voice-changer/releases/download/b2332/voice-changer-windows-amd64-dml.zip And install this

[woeful basin]

how do i uninstall it

[flint vapor]

Delete all the folder

Of the voice changer you have now

and then empty the trash

[woeful basin]

ok...

👍

[solemn burrow]

here u go my fellow accomplaice:
2025-07-18 19:10:56,971 INFO [RVCr2] Initialized.
Context leak detected, msgtracer returned -1
Context leak detected, msgtracer returned -1
Context leak detected, msgtracer returned -1
voice_changer/RVC/inferencer/RVCInferencerv2.py:60: UserWarning: MPS: The constant padding of more than 3 dimensions is not currently supported natively. It uses View Ops default implementation to run. This may have performance implications. (Triggered internally at /Users/runner/work/pytorch/pytorch/pytorch/aten/src/ATen/native/mps/operations/Pad.mm:465.)
res = self.model.infer(

[flint vapor]

After you uninstall, install this

[alpine onyx]

ohh how do I add pngs to the models btw ? :O

[flint vapor]

Go on edit, then you have the box on left, you can click there and select an image 😄

[knotty moth]

and so you didn't hear the output at all, not even some quiet noise?

[alpine onyx]

are there any well known models with clear pronounciation ? 😭 every one I used so far is a bit messy

[flint vapor]

You have to find it haha there are a lot with clear pronunciation

You can go on weights.gg

For search models

[alpine onyx]

Can you download them on that site ?

[ocean carbon]

its so delayed i speak after like 20 years

[flint vapor]

Yes only login there

[ocean carbon]

how do i make it speak in real time with me

[flint vapor]

Which gpu do you have? Which voice changer are you using

[ocean carbon]

4080 and MMVCServerSIO_win_onnxdirectML-cuda_v.1.5.3.15

[flint vapor]

Ok, uninstall the voice changer, i will give you a better compatible and updated version, deiteris fork

I will send you now a guide

With direct download link

-wokada

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[flint vapor]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

Guide

[ocean carbon]

and it will speak in real time with me right ?

[flint vapor]

I will send you now direct link

Yes don't worry

Of coruse a bit of delay you will have the same but few

https://huggingface.co/Shadicti/deiteris-Fork/blob/main/voice-changer-windows-nvidia-b2332.zip

Is impossible not have delay

But 4080 is very good you will have very few of delay

Just 1-2 sec max

First read the guide before install 😄

[ocean carbon]

thats perfect

[wary carbon]

Hey, the AI Voice doent work for me, what can I do? The passtrough channel works though.

[flint vapor]

Wait, first of all, which voice changer are you using, and which gpu do you have?

[wary carbon]

https://github.com/w-okada/voice-changer/blob/v.2/docs_i18n/README_en.md this one, i have AMD and I downloaded the AMD File

[flint vapor]

Uninstall this, i will give you update d version

Wait

[flint vapor]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

Which gpu you have?

[wary carbon]

the only voice working is the Beatrice thingg

[alpine onyx]

hmm the voices also all sound a bit raspy no matter how much I change the settings and I don't even have a raspy voice

[wary carbon]

AMD Radeon RX 6600 XT

[ocean carbon]

@flint vapor i luv u

[alpine onyx]

also @flint vapor on the chart for the chunkz it says 90ms for 3060 gpu ??

is that right?

[flint vapor]

Ok is compatible

Wait

[flint vapor]

https://github.com/deiteris/voice-changer/releases/download/b2332/voice-changer-windows-amd64-dml.zip

Download link

[flint vapor]

Yes but i use more higher

[alpine onyx]

what would you recommend for gaming ?

[woeful basin]

how do i install it

[flint vapor]

Just click and it will install

Auto

[flint vapor]

Put 298 chunk

And 2.7 extra

[woeful basin]

but its a .zip

[flint vapor]

Yes, then you have to extract it

[woeful basin]

then

then what

[flint vapor]

Then you have to extract the .zip

[low shard]

paid commissions will never comeback in the server, they are not allowed

[low shard]

Do you still need help?

[solemn burrow]

Not witht he media perms.

[woeful basin]

then what

[low shard]

Google colab shouldn't be ur first option, cloud isn't the best if you got a good pc

Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using

[flint vapor]

I said

When you download the zip

You have to extract it

Right click

Extract

[low shard]

That one is old original wokada, don't use yt tuts

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using

[woeful basin]

what do i do after i extract it

[flint vapor]

Open the folder

Open MMVC folder

and open the .exe

[woeful basin]

this?

[flint vapor]

Yes

[low shard]

I remember talking to you, you got wokada deiteris fork

You can't rename models on the web user interface, you have to rename the model files and then reupload in the slot

[woeful basin]

i did that already

[flint vapor]

Yes but this is updated version

The one you was use before was old

[woeful basin]

its been like this for awile

[flint vapor]

It should have open you a page on your browser

With an interface

If not, copy that link and put it on browser

[woeful basin]

oh it didnt tell me or open it

[flint vapor]

Now you are on the interface

[woeful basin]

why is it on the brower anyway

[flint vapor]

Cliclk allow and then start

[flint vapor]

Is a webui, is always using your local components

[woeful basin]

well tehres no more built in voices

[flint vapor]

Click upload

Then put the .pth and the .index

of the model you want put

[simple scaffold]

Guys, hey. I've got a problem here, how can I solve it? I have no idea at all.

Looks like there's a bit of a problem.
unknwon message

If you clear the information being managed by this app, it may be recoverable.

[flint vapor]

With which tool you having problem?

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[flint vapor]

RVC, Voice changer, Comfy, or other?

[simple scaffold]

Voice Changer

[low shard]

Please elaborate all I asked

[flint vapor]

Which version are you using right now, and which is your gpu?

[low shard]

Those are crucial infos

[simple scaffold]

Wait, i'll go check

[low shard]

!give-media-perms 1h @simple scaffold

[alpine onyx]

I'm really getting a bit exhausted... none of the voices I've picked so far really work. All sound off but when I listen to some of the videos or samples they sound amazing... Either I don't get the settings right at all or something else is missing :/

[low shard]

All yt videos are old

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[simple scaffold]

MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a, GPU GTX 1650, If I understood correctly

[low shard]

That is old asf

Delete it

[flint vapor]

Which chunks are you using now?

[low shard]

Uninstall vb audio cable

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

1st link, read wokada deiteris fork

You had an over year old version of original wokada, and vb audio cable creates issues on windows

[flint vapor]

Try with higher chunks, try with the chunk i use for gaming like 298

Extra 2.7

Just for test again

[alpine onyx]

Nividia 3060
Windows
I want the voice to actually just sound like the sample and not robotic/or other issues
No tutorial link

@flint vapor I'm currently using 298.7ms Chunk and Extra 2.7s

[low shard]

That all depends on his GPU nails

[flint vapor]

Formant shift put 0

[flint vapor]

Oh he have my same gpu 😄

[alpine onyx]

oh yeah it was put to 0 before I was just testing before sending the pic

[flint vapor]

Input volume keep 100

[alpine onyx]

done what next

[low shard]

Does it say CUDA b2332 at the top? Seems like wokada deities fork

on wokada deiteris fork, you can **optionally **use more advanced settings for benefits:

• Advanced Settings -> Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Advanced Settings -> Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Reduce the delay on Windows via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

Also, try other models and playing with the pitch, what model are you looking for?

[flint vapor]

Yes @alpine onyx try also one of these settings

[low shard]

@simple scaffold let me know

[flint vapor]

If you go down on the guide there are some models to try

[simple scaffold]

And so I looked through

[wary carbon]

thx works

[simple scaffold]

I downloaded the archives from the first and second links

What next?

[low shard]

Did you setup vac lite?

Uninstall vb audio cable from windows app settings

[simple scaffold]

I'll install it now, wait

[alpine onyx]

It says b2332 NVIDIA-CUDA at the top.
Tried the advanced settings and it's starting to sound way better!

I also tried with server everything on WASAPI but it doesn't sound any better.

I'm honestly just trying all the female models rn