#✨│ai-help
1 messages · Page 309 of 1
yo i need help it stops on "[Voice Changer] MMVC_SocketIOApp initializing... done."
guys i found the fix just become a voice actor
i have nvidia graphics driver 581. is that the correct one?
By the way, for less problematic at running Applio RVC locally, because your laptop GPU is not GeForce RTX but rather GeForce MX which has just 2GB of VRAM, consider for an online option like Kaggle instead. When you attempt to run Applio RVC like that, you can expect for a potential false hope.

2 GB is too low so it might be choked, 4 GB is bare minimum to go
but hwinfo says i have 4gb of VRAM
Applio RVC does in fact launches because it uses CPU at first, but your laptop GPU is not ready for such task.
time to upgrade I guess, dammit
is there a guide available to help me set up Applio > Kaggle for inferencing?
Your GeForce MX GPU typically has 2GB, the other 2GB might be of the shared memory from your main PC RAM. There's another variant that has VRAM as low as 1024MB (1GB) but it's a very rare one.
how sad, I am underpowered
The GeForce GTX mobile GPU that actually features 4GB VRAM would be those NVIDIA GeForce GTX 1650 or 1050 Laptop, the more step up from any GeForce MX GPU. For much better choice, "desktop" RTX 2060 or newer is more preferred over those laptop/mobile GPUs, although high-performance mobile GPU like RTX 4070 Mobile is a significant step up and can run Applio RVC as well.
hey @hallow thistle what should i put these to? and should i use GPU or CPU, both are really strong (7800x3d and gpu is 9070xt)
im talking about the chunk size and extra processing time
AMD Radeon RX 9070 XT should be GPU. The "AMD Ryzen 7 7800X3D" CPU doesn't always mean it work better than GPU for AI tasks.
and what about the chunk size? and extra processing time
Chunk: around 60 - 70 ms, always check perf number at top right if the number always green 
Extra: 2.7 s
Pitch extraction: rmvpe_onnx
im trying to use it for trolling in games, that'll work in game aswell right
dumb question actually
dont mind it
alright so
Ok so since I got amd do u know like the numbers?
wdym numbers
Like rx 570 or something like that
ohh
let me check them agaib
cuz i dont remember
i think i had a pretty bad one but ima buy a new pc soon anyway
lemme check
fuck
i remember it that bad
I need gpu btw
i have an rx560
Yikes yeahhh
shit fuck
Idk if it'll run
its okay lemme atleast try
But we can still try ✨
i saw vids in yt that peoples runs it without a gpu
Anything on yt is outdated
i got the old rvc downloaded (not installed) and the new wokada downloaded but not installed also yet
yeah ik
New wokada?
i tried to look the vids that were more recent
like i mean the one that runs in browser
i call it the new one
That do that
Just in case u have the wrong one, here's the AMD version
https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip
O
and found that link u sent before
how is it only like 300mb when the old wokada is like 2gb
i needa download vac lite
ill do it in a sec
but yeah keep going
Old wokada wayyyy less optimized
Like really badly
Optimized
oh okay
Ye
why doies this one come straight with an exe
More convenient I guess, I'm not sure about every change it has
hi guys i have downdloaded the ai model but there is a lot of delay also when it comes to recording or playing games the voice model lags or glitches is there a way to fix this?
alr so do i need to open the exe?
and as for the updated vers of rvc
"applio"
Did you get the voice changer off YouTube?
where to download
-rvc
I use applio on Kaggle which is a website so this is the best I can give you
why do yall use applio and not the "old" rvc
I followed a youtube turotial i downloaded the git hub then a virtual cable and then the voice model i downloaded it from this discord
And applio supports all types of pretrains as well as is faster I think
do u think this guy uses applio ? https://www.youtube.com/watch?v=fdKHp0smV0c
ik they prolly do
Yup you got an old voice changer because YouTube
What's your GPU David
ik they use rvc and for calls and interviews wokada
but idk if they use the updated vers
They don't
can i dm you ? or everything should be through here?
Everything should be here
if they use old rvc how is it so smooth then
I dunno
Applio updates all the time
how do u know if they use the old one or not
Because applio is always updating and getting better
Old one doesn't change
And it looks very different
so this song sounds like old one?
that is really bad
GPU 1 is Nvidia Geforce GTX 1050 Ti
My laptop has two gpus
when i use the github i always choose the Nvidia gpu
But like if i have nothing running just the github it has some delay but its not glitchy
when i start recording or playing games it gets pretty bad
You might be able to run it but it won't work good in games
like the delay is 1 min
Damn
and it glitches sometimes
Y
for sum reason
like it doesnt captures every sound i make
What kinda games are u trying to play?
valorant
i havent tried with other games
but isnt this voice changer supposed to work smoothly or it depends on the game and the computer?
local worm
Depends on your gpu and what else you're running with it
if i run it in a webiste will it run better for my gpu then ?
Probably yea bc it won't be using your GPU just running on a website
alr then i need to use kaggle or however is called right
I showed how to use it to someone else yesterday lemme find that so u can follow
k i madea an account already btw
so what would be your suggestion, i want to use the voice changer for in game chat if not at least be able to use it on discord while i play games.
oh shit yeah and also, does this other song also sound like they use old rvc ? https://www.youtube.com/watch?v=Y-wk3Ha0DaM
Starts here yea
alr this how my shit be looking
I dunno what the old one sounds like but if it's bad probably is
i mean it sounds good the song
no sloppery delivery
no lag
no cracks no bad shit so
oh i needa verify my phone ?
ima do that in a sec
So go to settings and phone verify and follow that video I made
Ye
if im not a dumb shit
as the tutorial says
it should look like this right
is this for me?
guys why beatrice works and not rvc i downloaded the latest version
yo @viral mason which one should i use for my shitty gpu
yea that's good
use the default
beatrice is old, ancient
what gpu do u have
gtx 1080 8gb
what do u suggest for my issue local worm?
it's not for wokada so no, unless u also want to make covers
u can maybe run wokada tg fork
ah i closed and started it again and now its downloading this
btw ts doesnt appear
i want real time voice changer to use for games and/or discord calls. not covers
then use wokada instead of rvc
https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002
u need these two since u have nvidia, u can try and get it to to work but it probably won't run any better than what u already have
why you say so?
@viral mason so am i set now?
lol 2007 like thumbnails
what'd u download?
vcc
vcc client win cuda
old
ah i have downloaded that twice doesnt work for me
why do u say it wouldnt work better for me? is it my laptop or i need to find a different voice changer for my purposes?
is there any tutorial for uploading sound to okada?
btw why do u touch some of the code for
i mean i copied as u did
but jst asking
well since we are in the topic is this this the best hub for voice changers ? those who have experience
bc laptops just don't have the power to run ai that good
I change ngrok to gradio tunnel bc it doesn't require extra stuff but I remove the numbers in the first bit because it garuntees it'll be the most recent version that runs
omg so am cooked. so i need an actual pc? any suggestions?
tbh yea ur cooked for laptop, but the best u probably could get that would run good would be one with an nvidia gpu in the 3000 series or up
like 3080ti or anything like that
alr i runned it and its like isntalling now
it keeps saying something about some files or idk whta they are and it says "......is not a symbolic link)
hold on it keeps going now
again like all good i think
why the fuck is it even searching for my dc
alr i got the link
it does that, idk why
how to do this
yo is there any way of instead of putting the link i can just put the files
the pth and index files yk
i wanna use the yunyun anime girl voice model (i found the model on this server) and i have the files downloaded
oh wait holdon it dont matter
the download link is attached also
the links are to get to the software
yeah im dumb like a shit mb
can u help me with em settings now?
knowing i have a trash gpu and shit
any settings recommended?
uhm wtf is this for
which one do i select
the voice model ima use is this one https://huggingface.co/Yuki354/YunYun-LAT/blob/main/YunYun-V2.zip
same dw
those are pretrains, not needed for what u are gonna do
just copy paste that into the download section of applio
so for what im gon do i just leave it like this (default)
well i mean i touched the pitch cuz for what im gon do they pitch the voice to 12 so yeah
if the voice is female and the singing is male yea
if it's both female keep at 0
hey it counts as "ai-help" if i need help or a tutorial to make this github ai project to run it as non-ui in google collab?
its an alternative to uvr5
u could just use this one it's a no ui colab
https://colab.research.google.com/github/Eddycrack864/UVR5-NO-UI/blob/main/UVR5_NO_UI.ipynb
uvr
this is the github https://github.com/0x0funky/audioghost-ai
uses this new model called sam-audio
tbh idk what that even is
uses natural language to isolate sounds in video/audio
👍
that's actually kinda cool
but can it also remove sounds like music? i saw it can remove baby crying kinda
yes
but the output was kinda mushy
it uses a model created by meta
but these project used some tricks to reduce the vram usage
so it can be used in consumer hardware
but i failed to run it in my laptop so i wondered if there is a method to use it in gradio with no ui

I don't trust meta tbh, only models made that are in uvr and such
the model are in huggingface you want the link?
its okay
I wouldn't know for sure but someone like Noobies might know, or Lyery
they're both pretty code smart
ohhh
is there any tutorial for adding a new voice in okada?
did u get wokada off youtube tutorial?
and when someone is speaking on the game, my voice changer followed the voice?
did i ping one of them or just wait to them to appear in this channel?
ohh i see
but idk how to fix this anyway
I'm asking if u got it off of youtube
no?
oh, what one do u have then, there are three
wokada deiteres, wokada tg fork, and vonovox
cant find it on yt
I'd say wait a bit and if not, u can ping the helper role
Is there any option for me to run ai voice changer without upgrading to pc? Like the nvidia ai 3080 costs like 1k and I am broke.
I heard something about clouds
there's online versions but idk how to help with those
you would have to talk to Namari for that
they're pretty smart
is this to fix doubled voice?
Do I have to contact him on this channel or dm him ?
I'd wait for him to get online then contact him here
going to dms without asking would be rude
thank you
Thank you
you're both welcome
-colab
Google Colab is a Cloud (Remote Good PC) Service. While the Free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.
by IA Hispano
Google Colab
by Hina
Google Colab
by Eddy
Google Colab
by Eddy
Google Colab
by Deiteris & Hina
Google Colab
by Shiro & Eddy
Google Colab
by Nick088
Google Colab
by Nick088
Google Colab
by Jarredou & Makidanye
Google Colab
How’s your experience using -collab? I want to use the ai voice changer for games and discord calls but sometimes it gets very delayed and other times it glitches or it doesn’t capture the voice.
did you tried first the voice changer using your pc?
to see if is too laggy?
I have a laptop, it has some delay when I am just using the voice changer but I got it from yt tutorial I haven’t tried yet the one local worm sent. But local worm said it won’t run any better because I have a laptop with nvidia 1050 ti. The issue is that when I try to game and use the voice changer it lags or glitches.
sure
ive used the yunyun rvc voice model but it dont sound nothing like yunyun
u can use applio in kaggle to do it
it sounds like some weird retrded girl
alr can u help me
yea, you first need to get the audio of the character, like get episodes of the anime or whatever she's from
how can i possibly do that
do i need to download the episodes, extract the voice from the other audios such as the background music and shit and allat?
😭
i want it to sound like this
i used musicfy.lol to convert the voice
and i want the yunyun voice to sound like that
maybe if u use another laptop to play
weights.gg is prob a better source than that
yeah i meaan, prolly not
its unusable
it's unusable rn but when the que isn't busy it's fine
oh yeah and one thing
there's a version of weights u can put on ur pc locally that does covers ect for free
called replay
setupX64, run normally not admin
I was thinking the same thing
it asks for admin anyway dont u see it on the logo
i run the one i have selected in blue in the pic i sent right??
cuz theres no file with an x
But i don’t want to invest in another laptop (am broke)
u could sell urs and with that money add a lil bit more of money u got saved and buy a good new one or a pc
yea
forgot that vac lite doesn't have an x
idk what that is
looks like sum random russian guy tf
the name of like the creator kind of
the editor of the program
where is that even at?
when u click on run as admin
on the admin prompt
if the editor is verified it appears their company name or whatever
it's not sketchy I have it on my pc and it's not messed up or anything
qwhat am i supposed to do here
press yes
then set your mic settings back to normal bc it will change the default to line 1 which is kinda dum
I need help
Bruh I can’t send pics
i get some errors for some milliseconds if i open Okada ''start_http''
how do i fix this?
yo i managed to get some vocals from yunyun
from the show
can someone help?
without background music shit
yea
sent
can anyone help me the model thingy is red same for index
nice
yo bro can you help me
whatcha need help with
lemme send you it in dms
I prefer here
it wont let me send me the picture
sorry I don't help people who use egirl stuff
happens the same with adam smasher
what are you having issues with?
im trying to extract okada tg fork and it just says error archive is corrupt
what you're using btw looks old
002.zip doesn't extract just place it in the same folder as 001 and run mmvcserversio.exe
should work
idk what it even does
002
u want me to send em here ? so u can help me with training the voice
got itttt
since u said u dont do it on dms
I'm fine with u sending in dms since I've been talking to u for a bit
and trust u
horayy
if u want I could get u a newer voice changer what gpu do u have
geforce rtx 5060 16gb
so will adam smasher work with this one?
it should, to run this one just run setup first then run start
alright
as of now everything works but when i put the VB Cable output into discord it doesnt work can someone help me with that?
can i send you a ss in dms?
sure
sent
I saw what u sent, what's ur gpu
its kinda working now BUT
4060ti
but
uh wait
its stuck on the screen i sent you
nothing happens
it downloaded the files already
u download the 001 and 002.zip?
yes
hm
it says checking for mandatory models
close and reopen?
but in the models folder there is nothing
doesnt work
u put both 001 and 002 into the same folder?
hm
try restart pc?
it wouldn't hurt
Okay
it does?
It's now opened in my browser
nowww
uhm
how does the model selector work?
wait nvm
ill try my self
-colab
Google Colab is a Cloud (Remote Good PC) Service. While the Free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.
by IA Hispano
Google Colab
by Hina
Google Colab
by Eddy
Google Colab
by Eddy
Google Colab
by Deiteris & Hina
Google Colab
by Shiro & Eddy
Google Colab
by Nick088
Google Colab
by Nick088
Google Colab
by Jarredou & Makidanye
Google Colab
gpu
Okay because standard is cpu
it runs better for gpu
okay
how do i fix "5060 ti gpu with cuda capability sm_120 is not compatible with the current PyTorch installation"
use an updated voice changer
do you know which one i should get or the version or whatever
wath gpu have you?
<@&1159293204038955078> do you know a tutorial about how to make a NO UI of this ai project https://github.com/0x0funky/audioghost-ai to run it in collab like the others projects?
its an alternative to UVR5 to isolate audio but now its using natural language like you write "vocals", "drums", "a dog barking" etc... the model that uses its from here https://huggingface.co/mrfakename/sam-audio-base

Nvidia 5070ti
wath config for chunk size and time you recomend , i has rx 6600
There are models in UVR that can separate into 6 stems (one of them being drums), though most of these models usually trained for fixed purposes. A vocal model (like voc_fv6) can also separate "chicken" sounds from an audio, but the model isn't always made for it. 
Aside from UVR5, I haven't tried or know these tools, so information can be quite limited.
https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/ https://github.com/tg-develop/voice-changer/releases/tag/b2397
what files do I download?
.zip.001 and .zip.002
Alr thank you
What do I do after installing them?
can sm1 help me out? im trying to use the rvc client, but i keep getting this error; "An error occured during voice conversion. Check command line window for more details."
no idea what i've done wrong or what im missing :(
hello people, is it worth running applio on ryzen 7 8845hs? I only have integrated graphics. (performance equivalent to a 2060 or 3060 mobile)
u gon use it to convert ur voice ?
or for what are u going to use it
and even with a bad gpu u can run the raeltime
u have amd right ?
i mean for a bad gpu i would recommend running it on a website
i only know how to install it trought a website
itll be better for u also so i recommend it
idk if im explaining myself good , im not really good at teaching
I know what you mean, don't worry, I just prefer using local applio instead of kaggle or collab
oh yeah well
if u preffer it like that then ig
but i dont really know how to do it locally
I was just wondering if it's alright with just the iGPU
I used to do it on collab, it felt easier, but then I learnt how to do it locally and it feels much better
it'd be cpu mode anyway
Ahh, alright, thank you for the information
did you run all the steps before?
wdym
i just selected my voice model , putted the settings for the voice model and when i clicked on start server that error appears
lkike what
to install the voice changer i just clicked on ts
hold on
am i not suposed to use like the crepe tiny as pitch extractron algorithm if i have a bad gpu ?
what gpu do you have?
which voice changer should i get with windows 11 with 64bit and 64 processor?
Guys do they still ban you for using UI in Google Colab?
Is there more websites with AI videos cost free
No rmvpe is fine
Sora ai
Depends on your gpu not what you listed
this the error
"64 bit" likely refers to CPU capable of processing data at 64 bit, so it's the same thing.
I can't help with errors idk what causes them
it said something about "pipeline not initialized"
rn no cuz i closed that cuz it happend a long time ago, but yeah i had the command prompt opened
It should work fine as long as it's open then
But Is there a website I know Mindvideo
Idk what that website is
Assuming that you once said your GPU was trash. When you try to run the program locally, does this mean something?
Is this an ai video website?
yeah but you only get 4 a day there a beta
and it has Sora2
but tell me the free Sora 2
Yea sora is free with no need for a code at least in the usa
Well, some of you still have time to focus on AIs for sure, meanwhile I'm on vacation because new year. 
I am in UK
Oh, I'm assuming getting a VPN wouldn't work there right
Bc of that weird law
That happened
For video generator, Luma Dream Machine is one of only websites I know.
I cant it wants me to pay
You know, not everything on Earth is always free, even those online AI video generators are slightly more expensive than some static AI image generators (including those that use Stable Diffusion). While I don't have any more recommendation about it, some people say there's a free one to try but potentially more complicated to set it up.
For Applio RVC, it can work with only CPU for basic inferencing (AI cover) but with slightly slower performance. And by an AMD Radeon(TM) integrated GPU, it might not be the same level or potentially just below those higher-performace AMD Radeon RX dedicated GPUs or comparable NVIDIA GeForce RTX.
does anybody know how to get rid of lag/glitching of the voice?
like settings that will help please reply to this message
what is the best realtime voice changer?
Hey guys, which software do you recommend for creating voices for a mini-series(example), W-okada Fork or RVC Applio? I'm looking for the one with the best voice quality.
Nvidia Rtx 2060
https://www.youtube.com/watch?v=xhHdKT8NTXA
have anyone watched this and know what voice changer is he using? His screams/cries are really high quality (Especially at the later part of the video where he cries using Rice Shower's voice)
BEFORE this video which he uploaded 3 days ago, it is very audible that he is using something like w-okada's, but this video only(Mambo & Rice) have a very high quality, i was wondering if there is an update that i didn't know that makes it possible to produce that kind of output? (7:50 - 8:15)
how do i get the voice mod
i am looking too i will try voice ai later if it works for me but clownfish and some other github shit dint work
Hey where does one find high quality models?
https://huggingface.co/wok000/vcclient000/tree/main i looked a tutorial but its german and its been good i use the vcclient_win_cuda_2.0.78-beta.zip
anyoen know a really good ai for images
epic ass video mambo
i feel like all the voice models that promise to be good are rly rly rly robotic and I feel like its my settings
but my cpu is to bad i think its slow lagging and slopy if i am in game
anyone know a voice changer that has low cpu rating
i need help everyone plss
is vonovox the best one out there as of now?
or is there something better?
also where can i find force fp32 setting on vonovox?
for using rvc models in realtime? yes
Teach me how to use it
Bro why am i getting delay i tried doing evereything
Me too vroo
I can't find a good voice changer so i can troll my friends as a e girl
You doin same?
whats the issue with the delay?
Twin can you help me set-up?
when i speak and record it takes 1-2 second for the voice to play
what chunk do you have it at
256
dm me il send u the pic of the settings currently
Soo i want a best voice changer to troll my friends as a e girl in Roblox
failed to load file in vonovox when trying to load an mp3 file
why does it do that?
what voice changer are you using?
W okada vroo jus downloading
Vovojeheb is glitchin asf and I can't figure out the settings 😭
search it up on #1175430844685484042
I cannot find any good ones
is that vonovox?
Yes
i heard it's better tho
It's not gitchy
But idk settings i uploaded a voice model it sound suttter
Maybe i need to change settings
I deleted it now tho
@hardy crow can you send me a good settings for w okada i have RTX 4060TI i5 12400f 16gb ram
Maybe should i download it again?
i believe it's better so yes
i just use force fp32 mode turned on extra set to 2.7 and crossfade 0.15
where do i turn on that force fp32 mode
Yeah
i only see block size, extra time and crossfade
advanced settings at the bottom
that's okada?
yeah
you stick to okada for other reasons?
got an amd card not to long ago
nvidia card i mean
so couldnt use vonovox for a while
aw i see
Vroo slide me som girl models please
Thanks
@somber bough yk settings for vonovox?
Ew
You ew
This gif you age player playing Roblox
My input device when i record my voice on the app i hear my normal voice not the voice changers, also when im not recording i just hear the voice changer when im speaking but its just my normal voice
Hey local worm
What if I get a laptop with a better graphics card ? Would that give me a chance to run ai voice changer with no glitches?
Or do I really need a pc set up?
.
Ew..
It's honestly better to get a desktop over a laptop
yo how tf do you turn off this hearing ur own voicechanger in this thing
Why 🥲
I don't play fortnite 😭
Also what’s vonovox local worm ?
It’s another voice changer ? And why people say it’s better than okada ?
Yea it's the only one that's still having active development and receiving updates still but it's Nvidia only
nvidia only for now™
didnt he say hes working on amd
or gonna work on amd eventually
When I install Applio or RVC in Windows, the complete installation includes ffmpeg. But I already have ffmpeg installed in a separate folder, and it is in my windows PATH so that other applications can access it. Is it a good idea to configure Applio and RVC to use ffmpeg from windows path? How do I do this?
My gpu is nvidia ☝🏼
Local worm do you think you could help me with choosing the right options for a pc build
which TTS is the best right now?
yo guys, what is the best deepfake (with only 1 photo, not real time) to windows with my rtx 50xx?
anyone who can help me optimize my RVC voice changer so it works well in fortnite?
TTS feature in Applio RVC.
Still not giving up? 
if you look in the forums theres usually settings recomendations in there

I've told them about some working settings, though I doubt if the person still remembers or reviews that discussion.
Click "stop" button, change output device to "Line 1 (Virtual Audio Cable)", click "start" button, and that's about it.
For help about anything, you can use @ helper ping for more reach, and don't always expect a non-helper member like Local_Worm to provide an answer for you. 
For a laptop with a dedicated GPU, any NVIDIA GeForce RTX Mobile of 8GB VRAM minimum will work for W-Okada voice changer. But for full performance otherwise, the desktop (tower PC) one is always better.
No. 
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
what would be the best voice changer if you were on a lower end device
cant seem to get stable results
Even the "original" mainline W-Okada version like v.2.0.78, which is its latest beta, is not always recommended for use either way. It's best to use more recent W-Okada fork as what I replied to you.
!howtoask
- Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
- Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.
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.
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.
- Be Polite & Patient: Our helpers are volunteers. You may ping the
Helpersrole once. - English Only: Please keep all conversations in English.
- Don't Ask To Ask.
AMD Radeon(TM) Graphics on Windows 11
just curious what you'd recommend
If your PC only has integrated GPU, it's better to go for online option like Kaggle instead. When you try to run the voice changer locally, you can only expect to run CPU only which is not always ideal.
Thanks!
What is your PC GPU? If you mean to train (create) a voice model or simple AI cover (inference), Applio RVC fits better, while W-Okada voice changer is a realtime voice changer that uses RVC (retrieval-based voice conversion) voice model.
Last update: September 6, 2025
What is your Lenovo PC GPU? Because you have been asking about "RVC" last few years but never stated one elsewhere.
is there any fix to this? :<
Whenever alt tab into the game and speak the delay occurs but if i alt tab out of the game it keeps working fine
What is even voice changer this? This program has of its own separate standalone window (which is the characteristic of the original W-Okada version) but the GUI inside is Tg Develop's W-Okada voice changer fork?
If you have extracted the W-Okada fork overlapping another W-Okada version inside the same "MMVCServerSIO", because you were confused about these two different versions, better don't do that, and you should instead delete the whole folder and then extract/install W-Okada fork from start.
ohh okk i will do that
hi im having a problem with the voice changer its really laggy for some reason i have a good pc so it cant be that
do me a favor and upload a screenshot of your settings panel
(the vc's panel)
it wont let me send
ah right
DM me the screenshot
.....or upload it somewhere and paste the link lol
catbox, google drive, etc etc
i sent it to you
voice-changer-windows-amd64-cuda.zip.001 or 002 there are 2 which one should i go with for better performance so it doesnt lag and does the voice 3-4s later
its like cutting in and out help
it's weird, even on minimum settings, it's lagging for some people
and the hardware isn't exactly weak
is there a way to fix
i wish i knew
first and foremost, the correct version for your GPU (nvidia vs. amd/intel)
amd
i wish i knew
And then you proceed trying to help them when you ask too little about the issue.
What is your AMD Radeon RX GPU?
Not everythnig have to be in one's private direct message, it's not that personal.
Try this Tg Develop's W-Okada fork, especially its DirectML variant. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/ https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip
Last update: November 22, 2025
which
There's no difference between .zip.001 and .zip.002, these files are split parts of a full single zip, so you must download both .001 and .002, these two files have to be in the same download folder. Do not use built-in Windows archive extractor, use WinRAR or 7-Zip to open the .zip.001 one.

so i extract both and but them into a folder?
dont have winrar
Sure, extract to somewhere like "D:\MMVCServerSIO", anywhere but not on your desktop folder.
can i screen share to u btw ._.
i dont know what im doing </3
There's 7-Zip; if you have neither of these archive program you should download and install either one.
!howtoask
- Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
- Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.
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.
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.
- Be Polite & Patient: Our helpers are volunteers. You may ping the
Helpersrole once. - English Only: Please keep all conversations in English.
- Don't Ask To Ask.
is this a no to the screen share 🥲
Better no.
okay 😭
I'm more better at text, and I don't have time to watch your screen like a Teamviewer.
im slow ._.
I go faster, so you should be careful what you're doing. Click download "voice-changer-windows-amd64-dml.zip" from this GitHub. https://github.com/tg-develop/voice-changer/releases/tag/b2397
advantage of text is, you can read it at your own pace
theres cpu and cuda which one ._.
This "voice-changer-windows-amd64-dml.zip" one is DirectML, which is compiled for AMD Radeon and even Intel Arc GPU especially. The "voice-changer-windows-amd64-cpu.zip" one is simply CPU-only.
@compact panther if you're not being helpful, better don't comment anything nonsense here.
4 hours of download can indicate of slow internet speed, but if it takes under minutes to download it should be fine.

Extract the files to somewhere like "D:\MMVCServerSIO" or "C:\Users\your username\Downloads\MMVCServerSIO", anywhere that's accessible but not on your desktop folder.
The what?
Inside MMVCServerSIO, double click on "MMVCServerSIO.exe" to launch the program.
When you ask for every step-by-step, like how to run the program, it can be very annoying and messy to going. So, while I can focus about some settings in the program, you should have some respect and review about your step by yourself.
nonsense is my forte
but on a serious note, i used the VC when i had a GTX 1060, and it would lag constantly after a while... i couldn't get it to run on any setting, not even minimum

only when i switched to a much stronger GPU, it worked
so it opened microsoft edge for some reason
If your Windows set "Microsoft Edge" as default browser, any program even "W-Okada voice changer fork" will open Microsoft Edge, by default. It's completely normal for this specific W-Okada version to opens your browser.
it runs in the browser not as an electron-based standalone program, it's a feature not a bug (im serious tbh)
For help about W-Okada voice changer, let take thing to here, here, not in someone's DM. What is your PC GPU? There's a better W-Okada version you might like to try.
This is Tg Develop W-Okada voice changer.
Could've ask for settings sooner, here's your settings:
Chunk: around 60 or 70 ms 
Extra: 2.7 s
GPU: AMD Radeon RX 9070 XT
Pitch extraction: rmvpe_onnx
sorry for being aburden i just dont know wtf im doing 🥲
for "chunk" do u mean chunk size or
"Chunk" and "chunk size" are the same thing.

IM SORRY 
This might give you some idea.
i dont use the cable thingy right
If you still don't know how to use the program, it can mean you didn't read the guide I sent to you and think alongside my steps, just expecting me to provide every answer because you say you don't know as if it's an excuse.
No no no i did read it
Try Virtual Audio Cable lite. https://software.muzychenko.net/freeware/vac470lite.zip
i have this already
You should know what you're doing, by the way. 
ur just smart and i wanna make sure im doing it right im reading the guide tho 
On Discord, in settings, Voice & Video, there are input and output devices.
problem
and ur really gonna call me stupid for this
on the left hand side
the input sensitivity
i turn that to the right
correct
This part is noise reduction. I'm not sure how to explain about this.
... can i say something now
Be quiet.
I have like 3 more
If I tell you how the program works, you can expect it to be very confusing for you, so better no.
What do I set the sample rate as what should this monitor stuff be set to and how do I hear myself 🥹
nevermind abt the hearing myself part

Can this be solved? This just putting more pressure on me to provide an answer, when that's really anything you can know about the program.
There was a recent announcement saying he probably won't be making an AMD version but he didn't say he wouldn't at all which means there's a small chance of it
Em..
It really just depends on your income and what you can spend
4070 rtx
W-Okada or Applio RVC?
@hallow thistle my mic is like stuttering still ._.
That's my question.😅 ,I just wanted to know which of the two offers better voice quality and naturalness.
but i can use both
Press "PrtScr (print screen)" or Snipping Tool, screenshot your W-Okada and send the image to here.
6750xt
how do i get that
These two programs have different usages; one is realtime, while another is for simple AI cover (inferencing) and an ability to train a voice model. For simple non-realtime AI cover, Applio RVC is better. For talk and convert your voice into something in realtime, there's W-Okada fork, though audio quality usually depends on chunk and extra values and the voice model itself.
thanks
For a guide doc, there's this link. https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/
Last update: November 22, 2025
oops
ok there u go
wait how do i open it after downloading
@hallow thistle did i download the right one
yo i need some help, say I got a model for an artists voice, and I also got some stuff recorded on my own, how do i use the model to convert my voice into the artists voice? like what's the software/procedure?
how do you get that
use the file as input instead of your comm device
Within "server" audio mode, set sample rate to "48000", all audio devices there should set to use WASAPI.
"Virtual Audio Cable" and "VB-Cable" are two different programs, yet many people here still mix them up confusing.
Windows wasapi right
so basically this
Scroll a bit up on voice changer, you'll see a section named "Performance Stats". Look in my screenshot, I circle too many circles on a specific part called "perf". This is there you see the perf number, if the audio stuttering, the perf number would generally be red, which indicates an issue. If perf number is red, navigate to "AI Settings" section, look at "chunk size", try raise chunk value up, look at the perf number again if it's green now.
Use client or server..
Tyty
"Server" audio mode can be better than "Client" mode, although the "noise reduction" part on the left would not be usable if set to server.
nah nah i wanted to just change my voice, not realtime, but i figured it out no worries, thanks for the help anyways! :D
Trivia: WASAPI's "shared mode", the sample rate of all devices and software(s) have to be of the exact sample rate (commonly being 48000 Hz), if one software or device being set to a different sample rate (like 96000 Hz) that one must set to 48000 Hz or otherwise will fail. With WASAPI or ASIO "exclusive mode", one software at the time can use any sample rate, although some other softwares that also use the same audio system will not be listenable at the same time.
MME, a legacy Windows audio API, does not have shared or exclusive mode, the sample rate can be of any value, but despite its premise this API generally has the highest latency than the other two.
W-Okada voice changer or Applio RVC?
!howtoask
- Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
- Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.
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.
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.
- Be Polite & Patient: Our helpers are volunteers. You may ping the
Helpersrole once. - English Only: Please keep all conversations in English.
- Don't Ask To Ask.
yeah i used applio, worked perfectly
although i did have to go through docs to figure it out lmao

||but that was what i was going for either way🙃 anyway, glad i could be of help||
atta boy
My laptop gpu is an nvidia 1050 ti what would you recommend to be able to run a voice changer smoothly? A cloud, a specific voice changer, give up and get a new pc ? I don’t have much money. My issue rn is that it delays or it glitches a lot when I play o games or I record and I want to be able to use it in game.
What do you mean NVIDIA GeForce GTX 1050 Ti is a CPU?
wait how do i open it after downloading
Use WinRAR or 7-Zip to open/extract the zip.
so i got them now and what do i do now i only have 2 big paper there that i cant open
yea but how do i open the acutal vc
you're probably best off using a server-based vc for a smooth experience
Go inside MMVCServerSIO folder, double click on MMVCServerSIO.exe to launch the program. Simple.
it doesnt open tho
What does this mean?
...it doesn't make much sense to name "give up and get a new pc" as an option though if you say right after that "i don't have much [i.e. enough] money"
its normal data not a normal zip file like i know i dont know how to extract after download
i cant send pictures here
so uh i compress it into a 7zip or uhh
Why can't y'all just be smarter someday? I think about my steps before sending anything, almost everything I say is so clear enough to understand, yet you still act like you still don't know how to basics, as if it's some excuse. 
i dont know
so i do select both and press compress and archive into 7zip?
When did I say to compress them into one zip? I never said that. I said to use WinRAR or 7-Zip to open the .zip.001 one. Is this instruction clear enough?
but i cant open with windows it doenst let me
which app do i need to do that
and which version to open it from the app
and ahhh 
Please, please stop being stupid about this.
ah yea winrar x64 i got it now
sorry i am a litte 
sometimes
why does it say the 002 is damaged or unkown format

do i need to reinstall?
I'm not sure what to help you, as much as I got into issues with other members in a row who also struggle on the same fucking steps.
i just reinstall maybe it didnt do proper
Try 7-Zip.
yes 7 zip sayed that
Always make sure both are named "voice-changer-windows-amd64-cuda.zip.001" and "voice-changer-windows-amd64-cuda.zip.002" and they all in the same folder. If one is named "voice-changer-windows-amd64-cuda.zip.001.zip" whether if it has to happen with human error or issue with system, you would rename it into "voice-changer-windows-amd64-cuda.zip.001".
No, stop doing this step.
okay
When you say "okay", do you think you know what you're doing? The "dml (DirectML)" one is for AMD/Intel GPU, not NVIDIA GPU.
but i cant add or open 002 with 7 zip i reinstall it and try again
also gonna eat real quick brb
thank you for your support until now
Seriously? If you know better how to get going, we can go faster up to that "settings" part. But because you keep asking for the same thing, it's just it's to the point of making me so frustrated with such requests.
guys i've been following this conversation for a while now and
i gotta be honest, namari is right about one thing
the questions you've been asking show no effort being made at learning or understanding the situation, and trying to defuse it
they weren't supposed to give you a step-by-step guide... that's what the docs and tutorials are for
and namari, don't be so mean to them
we all started somewhere

i know, this conversation alone has been going on for like six hours straight
😶🌫️
but to be fair, the way you talk to them is like they're dumdums ||regardless of if you may be right👀||
If you know how to help people, would you like to apply for a helper? Because for the past many days there's only one helper, which is me, constantly helping member day by day, sometimes Makidanye does help, while some other helpers either refuse to help or offline, and there are potentially non-helper members like you and Local_Worm helping about the very basics while I'm away from screen.
i've scrolled through the chat, and yeah
i basically only saw you in here... hats off
i am back it installed again but i still cant open with 7zip and sorry for being 
And, please understand those repetitive requests from those members that are always bugging me for past few days, even if I have to repeat my steps again and again, there's no way I can hold those frustration even if I tried.
I didn't mean to send that one, but here we are.
! E:\Voice Changer 234\voice-changer-windows-amd64-cuda.zip.001: The archive is corrupt
and also
heeeeeeeeeeeeeell no.
not after what i've seen
By the way, you can send your screenshot to here now.
but once i find some time, i'll check by again — i gotta tidy the house atm
Sure, the actual problem has been identified, this time. Your file name is "voice-changer-windows-amd64-cuda.zip (2).002" and not "voice-changer-windows-amd64-cuda.zip.002". I've been told you about checking the file names. How was even your effort?
the first one was correctly spelled thats the secound download and after changing it now it still doesnt work
i was checking just not the secound time i downloaded it

Don't say it ain't working. I felt of an obvious issue from the start. When I say to "use WinRAR or 7-Zip to open the '.zip.001' one", it means you must double click on voice-changer-windows-amd64-cuda.zip.001, not voice-changer-windows-amd64-cuda.zip.002.
i right click then open with winrar then open then error
Nah, you ain't making any effort on this one.
I am what the hell am i suppost to do 
I am sorry for existing
If neither step works for you, fuck that one b2397. Try the much older b2332 instead, where the zip file isn't split like the b2397 one. https://huggingface.co/Shadicti/deiteris-Fork/blob/main/voice-changer-windows-nvidia-b2332.zip
If you need urgent help, please checkout our AI Hub Docs or ask for help here following the [Guidelines](#1402790586028789830 message)
If you have time to retry about opening the b2397 one, you can try again later, otherwise just give up about it and go for b2332.
thank you
I'm quite done helping with this so repetitive progress, so I better go for another alternative option.
For a guide doc about b2332 W-Okada, better follow this guide. https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork
Last update: September 6, 2025
if i select my audio cable it crashes in the guid it says it should work why is that?
why do u still use vb cable instead of virtual audio cable recommended by the guide?
wait i thought its the same i have guess i should have looked closer
got it now
i was blind
af
now i have it but still crashes could you help again?
i did everything put my devices in but still the site crashes no idea what to do why is it with me always so
time for depression 
i got it i dont know why i was born with so less braincells sorry everyone 
i have my audio virtual cabel set up and when i start using the voice changer it says/does nothing
yeah but like i am wondering what is he using, how is he able to be that realistic and it also includes hiccups, cries, cough
man, i was gone for two hours
but it honestly felt like a week, guy had a meltdown
It's ok I'm dumb too
i helped people too yo , not today but i did yesterday and some other days and stuff i think
@wide oasis @craggy bough @somber brook whys there no index file in the link
im confused on how to import it
the number your trying to reach is unavailable at the moment please call later
rvc wouldnt be able to make a good voice out of high pitched instruments like flutes right
it sounds expensive?
that's because it is expensive
anything with the letters "RTX" and the word "PRO" in it is expensive
@thick portal which VC are we talking about

are you gonna tell me yet or
do you mean this thing, by any chance
.... lmfao
@hallow thistle now i understand why you're so stressed out sometimes
(not that it was ever a mystery)
Namari what do you think
is it possible that a super large multiple hour dataset might produce a lower quality outcome than one that is under an hour?
assuming they're both the same quality
If i use the RVC voice changer on CPU everything works perfectly fine but pretty high delay
ong GPU delay is low but the voice is very buggy there
If you have good audio of it, it makes good models of fnf characters, they just make noises
Does anyone know why Okada uses my CPU and not my GPU? I have an AMD graphics card and downloaded the corresponding version. I've tried a lot of things, but they just don't use the GPU 
Is there a solution? Or am I just screwed for using AMD?
oh yeah i use some of your fnf voices time to time too lol
they are atleast made from human sounds though (mostly) like boyfriends voice
where I can download russian models?
@white pebble@thick portal please refrain from pinging random persons and none of them are helper staff having enough knowledge about rvc or voice changer either 
VRAM is the more important factor in inference and training, but rvc inference should be viable for at least 3 GB vram, and training recommends 8 GB gpu for batch size 4-8. the latest version of applio is recommended for acieving such that efficiency. regarding the performance, even the difference of RTX 2060 and 5090 in inference shouldn't matter much in RVC/TTS inference.
how much performance do you need for the realtime voice changer? I suppose even I saw someone being fine on 4060 performance for gaming purpose (like marvel rivals)
no matter if the rented gpu is as powerful and expensive as H100, the network latency still contributes the most to the point that the long term rental cost could be instead saved for buying a decent GPU like RTX 3060 12 GB
what is that giant number? what version of the voice changer are you using?
dies
what gpu do u have
why extra 131k? please stop using the old wokada and consider using one of these recommendations
-rt
Guides for Programs that use RVC Models in Realtime for Calls/Games
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
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
Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE
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
These options are not recommended for use.
Not suggested, older versions in youtube tuts are even way worse. GUIDE
The program is worse compared to the ones above, and much less updated. GUIDE
u should get wokada tg fork



