#✨│ai-help
viral mason
uh..
viral mason
what version of the voice changer do you have? I would recommend switching to vonovox, there is going to be a new beta release in about 3 hours
any pitch past 12 btw is really too high for any model
lapis kiln
what version of the voice changer do you have? I would recommend switching to v...
i have FORK
lapis kiln
what version of the voice changer do you have? I would recommend switching to v...
what this beta
How much do I put CHUNK
viral mason
this https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/
that one is a little bit outdated compared to both wokada tg fork and vonovox
viral mason
what this beta
it's in the official vonovox server, I could send you an invite if you wish
lapis kiln
it's in the official vonovox server, I could send you an invite if you wish
YES Give me the link pls
lapis kiln
that one is a little bit outdated compared to both wokada tg fork and vonovox
So what's new?
viral mason
YES Give me the link pls
ok! I will send it via dms as if I send the discord link here I will die
lapis kiln
could be helpful for people that already updated to the current latest nvidia dr...
what this
simple ore
could be helpful for people that already updated to the current latest nvidia dr...
that is why i update drivers maybe once in 6 months
at least wait a few weeks for the new release
bold salmon
well it completely stable and fine but when joining roblox i have to set the graphic to the lowest in order for the program to work fine (not yet test in valor)
viral mason
well it completely stable and fine but when joining roblox i have to set the gra...
yea any of the voice changers need pretty much all games to be set at low graphics to work properly
bold salmon
oh is that how it work
hardy yew
(especially with a 1060)
hardy yew
supposedly very significant with today's beta (not available yet but soon), although i don't quite remember all the details
you can find lots of info in vonovox discord
hardy yew
sure! sent
prisma grove
lapis kiln
I want to download the improved version of
Tg Develop's W Okada Fork
how?
for Nividea Rtx 4070
sullen raven
Hey I am looking for a way to use AI to convert/translate a SUB anime episode into a DUB anime episode on my computer locally. What is the best way to do that?
OS: Aurora Linux
GPU: Intel Arc A770 16GB
CPU: Intel Core i5-12400F
Memory: 32GB DDR4
Storage: 1TB
rugged magnet
brisk trail
Hello peeps! I am a newbie model maker and I wanted to share my model of Kyle Crane from Dying Light: The Beast but I don't know how to upload my model on huggingface.com any help would be much appreciated!
viral mason
Hello peeps! I am a newbie model maker and I wanted to share my model of Kyle Cr...
here ya go man
brisk trail
Thank you so much I'll try it now
lofty sierra
ai bros
Is this chill
If i set it to 1 channel
Since my mic only has 1 channel
(it works good i tested in voice just want a confirmation)
viral mason
idk I wouldn't change the default settings on that
zenith ridge
Yo which version of okada should I use with amd gpu
viral mason
Yo which version of okada should I use with amd gpu
the first link is a virtual audio cable like VB cable, it's recommended to use the one I sent over VB
second one is the voice changer, Wokad TG Fork
it's pretty awsome
zenith ridge
it's pretty awsome
Thanks 🙂
zenith ridge
for vac47 run setup64, then for wokada tg fork run mmvcserversio
do yk where i can find a tut on how to shit ts up with the virtual cable and shi
viral mason
there are none for the newer stuff
just
open setup64 for vac 47
and mmvcserversio for tg fork
not that hard 🩶
quaint sundial
Can someone help me fix the audio I cant use ingame]
craggy current
how do i lower my vonovox speaking delay without it losing its quality
block size: 0.70
extra time: 4.0
crossfade: 0.10
pitch smoothing: 10.0
viral mason
Lower block size, also pitch smoothing isn't really needed at all
Change crossfade to 0.15
craggy current
Lower block size, also pitch smoothing isn't really needed at all
0.60 good?
hardy yew
0.3 is default and should be fine
dark orbit
guys why does the start http thing close right after i open it
scarlet meteor
File "torch\utils_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "voice_changer\VoiceChangerV2.py", line 190, in on_request
result, vol = self.process_audio(audio_in)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "voice_changer\VoiceChangerV2.py", line 133, in process_audio
audio, vol = self.vcmodel.inference(audio_in)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "voice_changer\RVC\RVCr2.py", line 226, in inference
raise PipelineNotInitializedException()
Exceptions.PipelineNotInitializedException: 'Pipeline is not initialized.'
2026-02-27 05:21:16,406 ERROR [VoiceChangerManager] 'Pipeline is not initializ
hallow thistle
File "torch\utils\_contextlib.py", line 116, in decorate_context
return func...
What W-Okada is this one supposed to be?
hallow thistle
guys why does the start http thing close right after i open it
You're likely using the original and outdated W-Okada version. There are workarounds.
dark orbit
You're likely using the original and outdated W-Okada version. There are workaro...
Which is the correct one
And is it supposed to be the beta or alpha or whatever
hallow thistle
Which is the correct one
!howtoask
hallow thistle
What is your PC GPU?
gusty bane
Why when i use custom models it doesn't work i have a amd processor
hallow thistle
Y'all be asking questions like if anyone this server knows about the voice changer lmao.
hallow thistle
Why when i use custom models it doesn't work i have a amd processor
What is your AMD Radeon or NVIDIA GeForce GPU? And did you follow any tutorial or guide about something something?
lapis kiln
I want to download the improved version of
Tg Develop's W Okada Fork
how?
for Nividea Rtx 4070
What is the version designed for RTX 4070
hardy yew
cuda.zip 001 and 002. You need to download both files and then extract the 001.
lapis kiln
cuda.zip 001 and 002. You need to download both files and then extract the 001.
How do I download both files? I don't understand.
Where do I put file 002?
this?
hardy yew
Download both 001 and 002, put them in one place, then just extract 001 (using e.g. 7zip)
lapis kiln
Download both 001 and 002, put them in one place, then just extract 001 (using e...
Like this
??
hardy yew
Did you rename it? You shouldn't
hallow thistle
Like this
If you disable "Hide extensions for known file types" in your Explorer, both .zip.001 and .zip.002 files should look like mine. https://cdn.discordapp.com/attachments/1159290139609137264/1456293998196035707/image.png?ex=69a253b9&is=69a10239&hm=89592e408278a60658c281548ec01bfa665d8cabc7bbd735bc078d13612dbda7&
hallow thistle
my gpu wont show
Is your PC GPU Intel HD Graphics?
neat apex
Is your PC GPU Intel HD Graphics?
nah i have a rx 7600
hallow thistle
nah i have a rx 7600
Well, make sure to download and use this voice changer. This voice changer works better for AMD Radeon RX GPU. https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip
neat apex
Well, make sure to download and use this voice changer. This voice changer works...
Ty il try it
vale pasture
how to make an account for weights it won’t let me
pure lantern
i have a rtx 4070 and currently using v.1.5.3.18a onnxgpu-cuda is there a better vers?
viral mason
how to make an account for weights it won’t let me
Weights is shutting down, it says at the top of the site
So you can't
vale pasture
Weights is shutting down, it says at the top of the site
so i can’t use the ai voice models ??
are there alternatives
viral mason
You can, just can't do anything on weights.gg anymore
There are pretty much no alternative that is just as good as weights sadly
viral mason
i have a rtx 4070 and currently using v.1.5.3.18a onnxgpu-cuda is there a better...
Download this version, it's wokada tg fork, the version you're using now is extremely outdated
https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip
vale pasture
You can, just can't do anything on weights.gg anymore
oh ok
viral mason
Did you rename it? You shouldn't
Sometimes if someone can't extract 001 I tell them to remove anything after it says .zip
And it usually works
It's kinda dumb to put anything after .zip in the file name tbh, should have been put before
hardy yew
Sometimes if someone can't extract 001 I tell them to remove anything after it s...
sounds more like "windows doesn't autodetect this as an extractable archive"
(in which case it's understandable why your advice works xD)
pure lantern
Download this version, it's wokada tg fork, the version you're using now is extr...
it keeps playing a loud buzz kind of noise
instead of the actual voice
viral mason
That's weird, do you have either VB cable or vac lite installed?
viral mason
for example if i say hello its like he- BUZZ llo
Ping the helper role, I'm not sure how to fix errors like that, tho it is recommended to use vac lite over VB cable
Here's the download
Just run setup64, not as admin
elfin nebula
Is there a locally hosted trainable llm framework?
normal horizon
Hi, I have a question. I'd like to modify my voice live. Is that possible with RVC? If so, how do I do it? If not, what software can I use?
viral mason
Hi, I have a question. I'd like to modify my voice live. Is that possible with R...
As long as you have a PC it's possible yes
lone kindle
hi
neat apex
Uhm my pc keeps spiking to 10 fps when i have the vc onn
any fix?
i have a ryzen 5500 and rx 7600
sharp cliff
anyb know how to fix the 'TypeError Failed to Fetch" think
like if i click on or try to edit literally anything i js get blasted with error popups, tried fully redownloading the files and no difference
hallow thistle
anyb know how to fix the 'TypeError Failed to Fetch" think
like if i click on or...
What is this Python environment supposed to be?
hallow thistle
Hi, I have a question. I'd like to modify my voice live. Is that possible with R...
Use Vonovox if you have NVIDIA GeForce RTX GPU; otherwise use Tg Develop's W-Okada fork the more recent voice changer.
sharp cliff
What is this Python environment supposed to be?
its alright no worries i figured it out i was js being stupid
pale trout
Use Vonovox if you have NVIDIA GeForce RTX GPU; otherwise use Tg Develop's W-Oka...
when i download the voice changer and extract it. it just disappears. im also using windows 10, i tried to search up the file but it also doesnt show up but when i go to redownload it it asks if i want to replace it
hallow thistle
when i download the voice changer and extract it. it just disappears. im also us...
!howtoask
hallow thistle
What is your PC GPU? And did you follow any tutorial or guide before? Answer these question directly.
pale trout
What is your PC GPU? And did you follow any tutorial or guide before? Answer the...
it is a NVIDIA GeForce RTX 3050 Ti Laptop and one of my friends (someone who uses the app) was walking me through it
hallow thistle
it is a NVIDIA GeForce RTX 3050 Ti Laptop and one of my friends (som...
Either use Vonovox or Tg Develop's W-Okada fork. These voice changers are more recent.
-realtime
willow jetty
what is the best voice changer, i have the voice model but i dont have the ai. i have ryzen 5 5600x and rtx 3060ti is that enough?
viral mason
what is the best voice changer, i have the voice model but i dont have the ai. i...
for your gpu, you'd have to use wokada tg fork, I can get the downloads for you ^^
your gpu is good for it btw ye
https://software.muzychenko.net/freeware/vac470lite.zip this is the virtual audio cable, if you downloaded the other thing delete it or smth
hardy yew
for your gpu, you'd have to use wokada tg fork, I can get the downloads for you ...
Whaat but they said RTX3060Ti and you sent AMD version of the app. Looks like a misunderstanding to me
hardy yew
I'm still waking up and wanna stay in bed for a little longer xd but my doggo will "motivate" me to get up soon for sure
viral mason
Thx for letting me know tho or I would of never noticed
hardy yew
But anyway, they can use Vonovox or the CUDA version of tg develop
And both will be nice
hardy yew
Good thing it's his choice, either will let me stay for a little longer, or... Won't xD
low shard
If anyone's looking to submit their model to get the model maker role to post in #1175430844685484042 , The Docs Page has been revamped and updated!
clever dagger
Hi all, how to create RVC model?
viral mason
-rvc
low shard
Hi all, how to create RVC model?
what's your PC OS and GPU?
clever dagger
what's your PC OS and GPU?
Windows 11.
NVIDIA GeForce RTX 4060 Ti
SSD 1.5TB
32 RAM
Intel(R) Core(TM) i5-14600KF
low shard
Windows 11.
NVIDIA GeForce RTX 4060 Ti
SSD 1.5TB
32 RAM
Intel(R) Core(TM) i5-146...
your pc is good enough to train RVC locally
check https://docs.aihub.gg/rvc/resources/dataset-isolation/ and https://docs.aihub.gg/rvc/local/applio/
clever dagger
your pc is good enough to train RVC locally
check https://docs.aihub.gg/rvc/res...
I've already started trying to make my own voice model on Applio.
Thx
clever dagger
your pc is good enough to train RVC locally
check https://docs.aihub.gg/rvc/res...
What is the best choice?
HiFi-GAN
MRF HiFi-GAN
RefineGAN
hallow thistle
## If anyone's looking to submit their model to get the model maker role to post...
What about the "last update" date? 
low shard
What is the best choice?
HiFi-GAN
MRF HiFi-GAN
RefineGAN
hifi gan is the usual one used for most models
refinegan is more of experimental but it might give better results but the issue is lack of pretrains for now
I would suggest going for HifiGan with contentvec for now
low shard
What about the "last update" date? <a:Aruthink:1142367781883494470>
thx for noticing, just updated it
hallow thistle
What is the best choice?
HiFi-GAN
MRF HiFi-GAN
RefineGAN
HiFi-GAN.
low shard
You want to like dub with AI episodes? From what language to what language? You can use smt like https://github.com/R3gm/SoniTranslate maybe
sullen raven
You want to like dub with AI episodes? From what language to what language? You ...
From Japanese to English as that is what anime language normally is.
I will take a look at that though
sullen raven
You want to like dub with AI episodes? From what language to what language? You ...
I struggled with trying to get it to work right but always failed. Is there something easier?
static holly
How to use AI voice
upper furnace
@low shard Can you unblock me so I can write to you? After the hacker attack.
gaunt ridge
can anyone help, my voicechanger is cutting out, not even detecting my voice sometimes and flat out not working 😭
gaunt ridge
my cpu maxes out but my gpu is barley even getting used
fresh cairn
whgats the updated real time voice changer for linux
viral mason
can anyone help, my voicechanger is cutting out, not even detecting my voice som...
Did you get it off a yt tutorial
gaunt ridge
yes?
hazy gust
https://software.muzychenko.net/freeware/vac470lite.zip
thanks
gaunt ridge
😭
viral mason
What's your PC gpu (Nvidia or AMD)
gaunt ridge
AMD
viral mason
Bet 1 sec while I get the download
viral mason
gaunt ridge
whats the vac one?
viral mason
how would i know if its for sure outdated?
Anything that isn't from this server is outdated, the people on YouTube use old stuff either to be annoying or because they're not keeping up with what is actually new
gaunt ridge
Anything that isn't from this server is outdated, the people on YouTube use old ...
okie dokie
viral mason
whats the vac one?
It's a virtual audio cable, like VB cable
It's recommended to use that one over VB
viral mason
thanks for the help
You're welcome!
For vac lite, run setup64 (not as admin)
And for the voice changer run mmvcserversio
viral mason
You ran the wrong file then 💔
gaunt ridge
🥀
viral mason
It's ok u can just do it again
viral mason
The voice changer does launch on web browser yea
If vac lite opened a website just close it
gaunt ridge
viral mason
Yea that's the voice changer 👍
gaunt ridge
seems alot more complicated lmao
viral mason
It looks complicated but it's really not lol
fresh cairn
It looks complicated but it's really not lol
do you know where the linux version is
for AMD
viral mason
where do i change input device?
On the right, should say audio settings
hardy yew
(also change driver to WASAPI)
viral mason
Capy help the humble Linux user out as I cannot
fresh cairn
xD
gaunt ridge
how would i test how it sounds?
viral mason
xD
The download for the Linux version is somewhere in here, I'd wait before downloading anything tho
https://github.com/tg-develop/voice-changer/releases/tag/b2397
hardy yew
do you know where the linux version is
https://github.com/tg-develop/voice-changer/releases/tag/b2397
voice-changer-linux-amd64-rocm.tar.gz
hardy yew
damn, i'm too slow x)
viral mason
We sent the same thing 😭
fresh cairn
why wait?
viral mason
how would i test how it sounds?
Upload a model and then as long as your mic settings are set correctly it should work, make sure to set CPU to your AMD gpu
gaunt ridge
Upload a model and then as long as your mic settings are set correctly it should...
anything look wrong here? i cant hear anything
hardy yew
anything look wrong here? i cant hear anything
did you start it? what does the graph in top right show?
BTW you shouldn't set output and monitor to the same device
gaunt ridge
BTW you shouldn't set output and monitor to the same device
i changed it
viral mason
anything look wrong here? i cant hear anything
Your output should be your virtual audio cable (should say line 1)
hardy yew
output is usually for the actual output (VAC) and monitor is for feedback (your headset/speakers etc so that you can hear it)
viral mason
Green is good
gaunt ridge
but i cant hear anything
gaunt ridge
model issue?
viral mason
Audio settings
gaunt ridge
okay
viral mason
And make sure you have "listen to this device" on for line 1 in the recording section
hardy yew
And make sure you have "listen to this device" on for line 1 in the recording se...
That's redundant for tg develop, that's what monitoring is for
viral mason
Oh yea
hardy yew
(although obviously it can be done that way with monitoring disabled)
viral mason
You can disable that since it didn't work
viral mason
It's a monitor setting then, I usually just set that to my headphones
gaunt ridge
i did that
hardy yew
I wonder why it'd make no sound when it's running and supposedly processing fine judging by the graph
hardy yew
I doubt there's a chance for the model to make no sound at all
viral mason
Your output should be line 1 which would be your virtual audio cable
I dunno if that would help
gaunt ridge
I doubt there's a chance for the model to make no sound at all
inless i imported it wrong
swift thunder
Does Xeus Embedder also work for singers?
gaunt ridge
Your output should be line 1 which would be your virtual audio cable
still no sound
viral mason
What model did you use? If you can link it from the server or smth
fresh cairn
wait, this can be used on discord right
gaunt ridge
yeah sure
viral mason
wait, this can be used on discord right
Yea
gaunt ridge
viral mason
You just need a virtual audio cable, vac lite
viral mason
https://discord.com/channels/1159260121998827560/1308872482358300815/13088724823...
Hm...
Why that one in particular?
fresh cairn
speaking of, i need to find a nice model. woo.
viral mason
We have almost anything
And also there is this site but you would have to test the model with a audio sample
https://voice-models.com
viral mason
That's ok!
fresh cairn
tried last year but the AMD okada build was broken on linux
viral mason
Damn
fresh cairn
yea hopfully it works
hardy yew
tried last year but the AMD okada build was broken on linux
yeah you probably have the least commonly used combo, linux + AMD
so the support is definitely worse xD
fresh cairn
yeah you probably have the least commonly used combo, linux + AMD
yea, it was a huge thing.
gaunt ridge
i can hear it now!!
viral mason
Epic!
fresh cairn
i havent been on windows in a year now
gaunt ridge
it sounds diffrent though..
fresh cairn
besides for bf6
viral mason
it sounds diffrent though..
You gotta up the pitch because it's a female model
Start at like 3 and go up till it sounds good on your voice
hardy yew
speaking a bit higher to compensate is often convenient too
you can play around and test various approaches anyway
see what works the best for you
gaunt ridge
gotcha thanks
viral mason
If you speak in a monotone way it'll never sound good 🙏
gaunt ridge
whats formant shift?
fresh cairn
why is that?
hardy yew
whats formant shift?
slide it around and you'll hear
hardy yew
perhaps it needs to be stopped first
gaunt ridge
false
hardy yew
(don't exactly remember, haven't used it for a while)
gaunt ridge
lmao
hardy yew
Oh, maybe it's in client mode only then?
gaunt ridge
server
hardy yew
I mean, maybe it can only be used in client
hardy yew
More like decreasing it might make it worse
Increasing usually shouldn't make much of an improvement
viral mason
Oh, maybe it's in client mode only then?
Yea the noise suppression can be used only in client mode which I haven't gotten to work since November of 2024
gaunt ridge
is client or server better?
viral mason
Not sure, I think client is quicker but I could be wrong
gaunt ridge
thanks!
viral mason
If you have any questions just lemme know! Or ask here as well
gaunt ridge
okay so.. if i was wanting to record the output
how would i do that?
normal recording software?
hardy yew
yeah, you can do that and record the VAC. But.. if you want recordings, you don't need the realtime app at all
you can download applio and run inference on audio files
fresh cairn
darn, still broke
hardy yew
yeah i want the realtime
aah okay, in that case yes, just record line 1 from the virtual audio cable
gaunt ridge
aah okay, in that case yes, just record line 1 from the virtual audio cable
o7
on the recording it sounds diffrent
eh
nevermind
fresh cairn
anyone know why the app isnt working?
for linux
wont launch at least
i run CachyOS and a RX7900GRE
hardy yew
i'm afraid i can't help much with linux, haven't ever run RVC there
fresh cairn
ah
hardy yew
how about sudo apt-get install python3-pyaudio
fresh cairn
ill try finding that
fresh cairn
how about `sudo apt-get install python3-pyaudio`
checke dit
its a old file in the lib
i did fix it
however, the error after that is a security issue
hardy yew
ooh my bad i didn't scroll right and misinterpreted it
low shard
<@911742715019001897> Can you unblock me so I can write to you? After the hacker...
I didn't block you, I gave you back your roles before you got banned from being hacked
upper furnace
I didn't block you, I gave you back your roles before you got banned from being ...
So what should I do?
to get everything back
viral mason
to get everything back
you've got your roles and stuff back so you can go back to making models and posting them yea :D
upper furnace
you've got your roles and stuff back so you can go back to making models and pos...
thanks
lapis kiln
Did you rename it? You shouldn't
how?
lapis kiln
If you disable "Hide extensions for known file types" in your Explorer, both .zi...
I don't understand. Please explain it to me. How do I do it?
Guys, I don't understand how to download a copy ( Tg Develop's W Okada Fork )
It's complicated to merge the zip001 + zip002 files.
I would appreciate a full explanation or a video clip, please.
hardy yew
did you try just extracting the first file?
viral mason
I have a hard time communicating with them, they may need a video example of what to do
fresh cairn
ok, it has a issue finding the the libs because it is arch
pallid hedge
is Vonovox simple to use?
tame oracle
@pallid hedge https://docs.aihub.gg/realtime-voice-changer/local/vonovox/
download the latest .zip from here and extract it, after run setup.bat then start.bat
also do you have virtual audio cable?
Models you're going to have to look for in #1175430844685484042
pallid hedge
<@1473062720948408566> https://docs.aihub.gg/realtime-voice-changer/local/vonovo...
yeah
tame oracle
I don't know all the models, but this one sounds good https://discord.com/channels/1159260121998827560/1470978983930892308
pallid hedge
<:AES_salute:1091143801843752980>
when I use vonvox I hear nothing lmao ihave the right input and output and a model but nothing
tame oracle
when I use vonvox I hear nothing lmao ihave the right input and output and a mod...
Do you have virtual audio cable
like
VAC Lite?
pallid hedge
yeah I have one
where would I use it?
all I'm seein Starting voice conversion...
[SOLA] Search range set to 10ms (multiplier: 1)
[Debug GUI] Upscaler inactive. Source SR for final resample: 48000
WASAPI Configuration:
Input Device: Microphone (Realtek USB2.0 Audio)
Output Device: Headphones (Realtek USB2.0 Audio)
Configured Sample Rate: 48000
WASAPI Mode: Shared
WASAPI Stream Information:
Input Device: Microphone (Realtek USB2.0 Audio)
Output Device: Headphones (Realtek USB2.0 Audio)
Input Sample Rate: 48000.0
Output Sample Rate: 48000.0
WASAPI Mode: Shared
Initialization Complete
pallid hedge
the ui is weird I'm tryna figure it out
idk where that wouldbe
where is that option?
tame oracle
hang on
pallid hedge
alr
tame oracle
i deleted it off my other hard drive i need to download it again
hardy yew
In vonovox there's no way in GUI to choose GPU
pallid hedge
ah alr
hardy yew
It just uses the primary nvidia GPU by default
pallid hedge
It just uses the primary nvidia GPU by default
Initializing CPU limits: 4 cores, 4 threads [ID:3525]
PyTorch thread limits set to 4
CPU affinity set to 4 cores (0-3) out of 16 available (total/4 = 4)
tame oracle
Sadly the most up to date voice changers have a different ui
so uh
your block size
what is it
tame oracle
try these settings
Block: 0.30
Extra: 2.00
Crossfade: 0.10
Input: [Your Mic]
Output: [Virtual Cable Input]
GPU: Your GPU
hardy yew
no monitoring in vonovox directly either
you can go to windows sound settings, properties of the VAC
and enable "listen to this device"
hardy yew
and then you'll hear what comes through the VAC in your speakers
you can also just do a test recording in e.g. audacity
or use the discord microphone test
or you can set vonovox output to your speakers/headphones/whatever temporarily to test it
pallid hedge
works ty
viral mason
What are you doing with them hmmmmm?
lapis kiln
did you try just extracting the first file?
yes
How do I merge the two zip files, 001 and 002?
hardy yew
You don't need to
lapis kiln
You don't need to
Could you please explain to me what I am carrying or doing?
hardy yew
If you extracted it and there was no error, it's all ready and it used both files while extracting
lapis kiln
Is there an explanatory video?
tame oracle
What are you doing with them hmmmmm?
Have fun detective
viral mason
Have fun detective
I wonder what would a dinosaur need with a girl voice changer (Because they are a Yoshi lol)
tame oracle
I wonder what would a dinosaur need with a girl voice changer (Because they are ...
uh
"stuff"
viral mason
What do I download? Which version is specific to ( RTX 4070 )
If you have already downloaded both 001 and 002 that's all you need
lapis kiln
I want to start over, please follow up with me
hardy yew
You already downloaded both and sent a screenshot 😭
No need to start over
And if you do, just read again the advice we gave you before
lapis kiln
I'm downloading these two now
just this
yes?
for RTX 4070
hello
voice-changer-windows-amd64-cuda.zip.001
voice-changer-windows-amd64-cuda.zip.002
I've downloaded this now, what do I do?
pls help me?💔
hardy yew
I'm starting to question my own sanity
lapis kiln
Why is zip002 without a file, just white in color?
lapis kiln
I'm starting to question my own sanity
Please just follow along with me, and I apologize for bothering you.
lapis kiln
Should I put them all in one file?
New Floder
?
Now I'm only extracting 001 from the same file and nothing else?
@hardy yew@viral mason
I can't extract zip002
viral mason
Now I'm only extracting 001 from the same file and nothing else?
<@2141346260655...
extract 001, u don't need to extarct 002 since you cant
put 002 into the folder for 001 after extraction
lapis kiln
put 002 into the folder for 001 after extraction
how i do this?
viral mason
drag the 002 file into the folder
lapis kiln
drag the 002 file into the folder
put 002 in MMVCServerSIO
or what?
Or leave it like that??
viral mason
put 002 in MMVCServerSIO
leave it like this ye
lapis kiln
leave it like this ye
Is this true?
lapis kiln
put 002 in MMVCServerSIO
I don't do that, right?
- voice-changer-windows-amd64-dml.zip
Do I need this file?
viral mason
ai takes a lot of gpu
lapis kiln
ai takes a lot of gpu
I mean ram Memory
37% why?
Also, why can't I upload a picture of my model?
lapis kiln
Is this correct or is there a mistake?
?
white shadow
Lowkenuinly what voice changer does this guy use
https://www.youtube.com/watch?v=9AqkEqNSclI
idk if its a personal one or not
robust blade
Lowkenuinly what voice changer does this guy use
https://www.youtube.com/watch?v...
It's eleven labs
white shadow
It's eleven labs
i thought that too but he also uses it in real time
not only editing it into the vid
robust blade
i thought that too but he also uses it in real time
Then it's a voice changer
viral mason
bro has rolled a 1 in skill
white shadow
lmao
viral mason
I saw you typing mr worm.
elder coral
this is the model output of OMFG talkbox but i can't submit it because it is a robotic model
white shadow
this is the model output of OMFG talkbox but i can't submit it because it is a r...
what on earth is gripping my ear lobes with its toes
elder coral
I'll send one that's more cleaner
I got the model maker last time, then left the server, now it's gone
elder coral
i used noise reduction for baldi's dataset and this turned out great because the dataset is long
baldi pronounced garage wrong
hallow thistle
Then it's a voice changer
RVC (retrieval-based voice conversion) is also called "voice changer", but not always realtime like the way "realtime voice changer" does.
elder coral
this is so bad
plain totem
I am currently trying to upload a model I found online that isn’t uploaded on weights and I am trying to upload the model on weights but for the life of me can’t find the upload button, all I see is train a model or use a model and not upload a model
hardy yew
weights is shutting down on 31.03
sleek ocean
Is anyone here using Kie.ai?
I tried generating an image with Nano Banana 2, but the result shown in the gallery is rotated 90 degrees compared to the original. Has anyone experienced this issue?
somber adder
I want an Ai model that can help with quality standard reviews. I’m thinking of creating a shared Microsoft copilot agent for each type of guidance to check drafting standards and discrepancies. But wonder if there’s a tool specifically made for that already. I.e. checking sheet numbers, outdated dates, discrepancy between descriptions vs specifications.
low shard
So what should I do?
You got your model maker role back, wdym?
low shard
I am currently trying to upload a model I found online that isn’t uploaded on we...
Weights.com is shutting down on March 31st 2026
upper furnace
You got your model maker role back, wdym?
yes
paper ridge
Hi, Quick Question, i dont know if this is the correct Channel
i used outdated software for the AI Voice im using for Videoediting, sadly the Software collabsed and got a Heart Attack and now refuses to work propperly again, what software can i upgrade to so i can continue using the Voice Models from #1175430844685484042 ?
hardy yew
Applio
hardy yew
i can get .wav file outputs from it correct?
yeah, use the "Inference" tab and there you can input audio files and get audio files as output
paper ridge
yeah, use the "Inference" tab and there you can input audio files and get audio ...
That works too? thats cool, and Live output works too?
paper ridge
yeah, use the "Inference" tab and there you can input audio files and get audio ...
does it work offline?
hardy yew
Sure
hardy yew
Works best with an nvidia GPU, by the way
paper ridge
Works best with an nvidia GPU, by the way
i have a few GPU's dw
storm dew
Do you have any alternatives to Weights.gg? I'm not particularly good at training vocal models and Weights.gg was very easy to use and convenient
low shard
Do you have any alternatives to Weights.gg? I'm not particularly good at trainin...
What's your PC OS & GPU?
next sphinx
hey guys, I have problem with "Get unlimited voice features" in Voice Changer App... firstly i could just close it with a X in corner, but not now a I don't know what to do... I have the Starter plan...
storm dew
What's your PC OS & GPU?
Windows 11 and Intel UHD
noble canopy
In client voice is perfect but in discord vc it sucks anyone know how to fix this?
low shard
Windows 11 and Intel UHD
Intel UHD means you have integrated GPU which isn't good for running local AI on your machine
You can try other free limited cloud options
You need to first manually make the dataset https://docs.aihub.gg/rvc/resources/dataset-isolation/ (it isn't automatically cleaned like you did with Weights)
Then you can try Cloud Options like Applio Kaggle https://docs.aihub.gg/rvc/cloud/applio-kaggle/ or anything in the RVC Cloud Tab
low shard
hey guys, I have problem with "Get unlimited voice features" in Voice Changer Ap...
Are you using Voice.AI from a Youtube Tutorial? we don't suggest nor are affiliated with those
it's better if you elaborate
!howtoask
low shard
In client voice is perfect but in discord vc it sucks anyone know how to fix thi...
this is a general ai discord server, there are multiple programs, elaborate
!howtoask
noble canopy
I'm using RVC Client (I don't know the exact name). I did it according to this guide: https://www.youtube.com/watch?v=FKWVJmLwoYo
In client voice is perfect but in discord vc it sucks I also use a virtual cable
low shard
I'm using RVC Client (I don't know the exact name). I did it according to this g...
RVC means Retrieval-based-Voice-Conversion
All youtube tuts are outdated, this is using a plus over year old version of the original wokada (vcclient)
You should uninstall everything you got from that tutorial and elaborate:
- Your PC OS
- Your PC GPU
- What you're trying to do (since you seem kinda confused about RVC)
noble canopy
my os in Windows 11
gpu is Geforce rtx 5080
trying to change my voice in real time so that it sounds completely different (using vc models)
low shard
my os in Windows 11
gpu is Geforce rtx 5080
trying to change my voice in real ti...
so sounding like Goku in realtime on discord right?
Because AI Covers like making goku sing would require a different software from realtime (that's actually RVC)
lapis kiln
@low shardhello
noble canopy
so sounding like Goku in realtime on discord right?
Because AI Covers like makin...
yes
lapis kiln
I have downloaded • Wokada Tg-Develop Fork
low shard
<@911742715019001897>hello
hello, you need smt?
noble canopy
what is more popular in your opinion?
lapis kiln
hello, you need smt?
I have downloaded • Wokada Tg-Develop Fork
and
I placed the zip001 + zip002 files like this and only extracted the zip001 file. Is this correct?
low shard
what is more popular in your opinion?
it isn't really about popularity, but more about some different features that might interest you
low shard
I have downloaded • Wokada Tg-Develop Fork
and
I placed the zip001 + zip002 file...
I see you're on Windows 11, could you elaborate:
- Your PC GPU
- What are you trying to do
lapis kiln
I see you're on Windows 11, could you elaborate:
- Your PC GPU
- What are you tr...
RTX 4070
Is what I did the correct way to download?
• Wokada Tg-Develop Fork
lapis kiln
I have downloaded • Wokada Tg-Develop Fork
and
I placed the zip001 + zip002 file...
..
Because there are two files: zip001 + zip002
noble canopy
it isn't really about popularity, but more about some different features that mi...
Hi, I did it on Vonovox, the sound quality on discord improved but discord cuts off the sound in the middle of the sentence or at the end, do you know what I can change?
hardy yew
https://docs.applio.org/getting-started/inference/
there's some general instructions here
in general applio expects the models to be in the logs directory
logs/example_model/example_model.index and pth and all related files
because that's where it outputs the files from training/preprocessing etc
so you can put your models in such paths or just use full paths to your pth and index file
hallow thistle
Because there are two files: zip001 + zip002
Use WinRAR or 7-Zip to open the .zip.001 one, extract to somewhere like "C:\MMVCServerSIO" or "D:\MMVCServerSIO".
signal isle
Can someone help me how to make a song with same melody, same lyrics but translated language?
low shard
Hi, I did it on Vonovox, the sound quality on discord improved but discord cuts ...
have you tried turnin off discord echo and noise cancellation?
low shard
Can someone help me how to make a song with same melody, same lyrics but transla...
Mmm kinda hard to do
maybe https://github.com/R3gm/SoniTranslate could help
or you could separate the vocals and instrumentals, sing it in what language you want, and then use RVC to make you sound like the original singer
Just letting people know about the latest Nvidia GPU Driver, but be aware that the previous driver removed had issues so you could also wait a lil before updating
noble canopy
have you tried turnin off discord echo and noise cancellation?
yes its off
low shard
yes its off
show a screenshot of your vonovox settings
noble canopy
show a screenshot of your vonovox settings
where should I send ss? I cant send them here
no permissions
hardy yew
Oh, vonovox? If sound is cutting off for you there then check if you have SmartSINE enabled
If you do, try to disable it
it's basically a gate that tries to only pass through what it considers to be speech
it isn't necessarily perfect
low shard
where should I send ss? I cant send them here
!give-media-perms 1h @noble canopy
viral mason
I haven't seen that command in so long
tame oracle
Real
low shard
I haven't seen that command in so long
I forgot about it for a second lmao
tepid bay
Use WinRAR or 7-Zip to open the .zip.001 one, extract to somewhere like "C:\MMVC...
I'm also trying to install it to support the rtx50xx video card. But after unpacking, I launch the application and a window pops up, but nothing happens in it. Do I need to download something else first?
stiff rivet
Yo, is there any way to make w-okada real time voice changer to sound better on other languages, beside English? For example - Russian, idk.
hardy yew
choose appropriate model
trained on that language
an english model will never be able to pronounce the sharp r-s and similar sounds because they just don't exist in english
tepid bay
I'm also trying to install it to support the rtx50xx video card. But after unpac...
or tell me where to download it, because nothing works
hardy yew
I'm also trying to install it to support the rtx50xx video card. But after unpac...
did you wait for a while?
tepid bay
did you wait for a while?
1-2 minutes
shell pulsar
does anyone know where to find more artists ai voice models?
viral mason
does anyone know where to find more artists ai voice models?
check here https://discord.com/channels/1159260121998827560/1175430844685484042 but if you don't find anything u like u can also check here https://voice-models.com
vague nacelle
how to get AI Members
verbal ridge
Is anyone else having trouble getting Applio's TTS to work?
vivid pewter
-colab
vivid pewter
nah
viral mason
what are u lookin for
vivid pewter
nvm
tardy jolt
rvc
tame oracle
cosmic folio
hii
hallow thistle
Assuming you're looking for "a voice changer" especially, which is always expected.
hallow thistle
For W-Okada voice changer, it's possible to make yourself to sound like "Frieren", look up a voice model of her in #1175430844685484042. There's no reason to ask me in direct message lol, it ain't something secret. 
verbal ridge
Is anyone else having trouble getting Applio's TTS to work?
Got it figured out
cosmic dagger
Hi Looking for an OpenClaw expert in Seattle and surrounding areas. Message me or contact me somehow on this platform. Thanks
craggy mist
I need jk ai model link
Where I can get
Ple give me link
safe prawn
i've got the pipeline not initialized error like some others but my gpu is the rtx5070, i have the latest installed from this guide https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/#windows i've read through most of the troubleshooting and responses of others. they're using stuff that's outdated but it seems i have the most up to date version
tepid bay
i've got the pipeline not initialized error like some others but my gpu is the r...
Did you manage it, is it working for you? Or?
hallow thistle
i've got the pipeline not initialized error like some others but my gpu is the r...
Vonovox and Tg Develop's W-Okada fork (which itself is based and a continuation of Deiteris' W-Okada) "have" GeForce RTX 50 series GPU support, while Deiteris' W-Okada is older one and can be tricky to get it to work with RTX 50.
azure sequoia
what the best realtime voicechanger now for NVIDIA 4050 laptop?
hallow thistle
what the best realtime voicechanger now for NVIDIA 4050 laptop?
Vonovox.
lucid isle
What’s the best real-time voice changer for arch Linux using an nvidia rtx 3070 8gb?
hallow thistle
What’s the best real-time voice changer for arch Linux using an nvidia rtx 3070 ...
These files.
noble canopy
what is the best tool for creating/training rvc models
hardy yew
what is the best tool for creating/training rvc models
you can just use applio or some fork of it
noble canopy
you can just use applio or some fork of it
you can send the link/any info??
hardy yew
safe prawn
Did you manage it, is it working for you? Or?
nope, still notworking
safe prawn
Vonovox and Tg Develop's W-Okada fork (which itself is based and a continuation ...
So i messed around with the Tg Develops fork and that one also gave the same pipeline thing, would you suggest i try Vonovox? or would that just be a waste of time
low shard
So i messed around with the Tg Develops fork and that one also gave the same pi...
can you show a screenshot of the wokada tg develop fork settings?
!give-media-perms 3h @safe prawn
tame oracle
-# Crusher ANC nice
viral mason
-# Crusher ANC nice
what's that
viral mason
did u download both 001 and 002 zip? u have nvidia so both would be needed
tame oracle
what's that
Headphones, they can just output A LOT of bass
low shard
Set RMVPE without ONNX
Did you install vac lite and an RVC model?
hardy yew
did u download both 001 and 002 zip? u have nvidia so both would be needed
you couldn't fully extract without errors without all partial files
viral mason
making sure he didn't download the wrong version since there are a lot of downloads
safe prawn
did u download both 001 and 002 zip? u have nvidia so both would be needed
yes, the second is downloaded, its not applicable to extract it
i wasnt able to extract the first one without having the second downloaded
viral mason
oki
safe prawn
Set RMVPE without ONNX
Did you install vac lite and an RVC model?
i did
the rvc model im using is the juice wrld one i found scrolling through voice models and will try setting to without onnx right now
safe prawn
-# Crusher ANC nice
i like bass 😛
viral mason
🐟
safe prawn
So on regular RVMPE still same error pipeline not initialized
viral mason
could try switching from client which for me hasn't worked since last year November
safe prawn
could try switching from client which for me hasn't worked since last year Novem...
that was actually my first idea
viral mason
same error ig
safe prawn
yeah
viral mason
😔
safe prawn
trust i wouldn't have came on here bothering anyone if i REALLY didn't try my hardest by myself
safe prawn
yes yes, it was set to headphones just because i just connected them via bluetooth
shell pulsar
Hi, do you know how to spend less time processing in the Inference tab of Applio?
safe prawn
i have a VAC set up and have set up a voice changer like this before via this discord, just not sure why its not working this time
tame oracle
i like bass 😛
Real
safe prawn
yes i have done that
low shard
is your windows up to date? Did you update to the latest Nvidia Drivers (the newest came out like yesterda)?
safe prawn
is your windows up to date? Did you update to the latest Nvidia Drivers (the new...
not manually no
im not on yesterdays update
for nvidia
my windows updates have also been paused for a month
due to the latest security update causing screen freezes
low shard
not manually no
mm that might cause some issues
but before you try updating:
- show a screenshot of all your settings
- show a screenshot of the CMD full
- tell where you extracted the wokada tg develop fork (like C: Drive, external drive, etc)
safe prawn
mm that might cause some issues
but before you try updating:
- show a screensho...
c: Drive downloads folder is where it's extracted
low shard
c: Drive downloads folder is where it's extracted
could you try another model, possibly without special characters neither in the model's name nor the path where the wokada tg-develop fork is in, and then select it and click start server?
low shard
its working on a different model
might be that specific model was giving issues because of the special characters in it, do you really need that specific one? if so you could try renaming it but i'm not sure if it will work
btw updating nvidia drivers once in a while might help with performance, don't stay too long on very old versions
safe prawn
might be that specific model was giving issues because of the special characters...
no i dont it was just a test
low shard
no i dont it was just a test
alright, so now that it's working fine on all other models, is this all solved or do you need any other help?
safe prawn
alright, so now that it's working fine on all other models, is this all solved o...
i think its good now
low shard
i think its good now
great, have a nice day
safe prawn
thanks so much
viral mason
enjoy your new power, don't use it for evil
safe prawn
we'll see about that
white shadow
Guys please don't put me in hall of shame 🫣
Mommy voices?
craggy mist
austere cipher
guys what do i pu in my input and output
so i can use it in different apps rvc voice changer
hardy yew
guys what do i pu in my input and output
Input: your microphone.
Output: virtual audio cable (if you don't have it, download VAC lite 4.70 and install it)
In the app (e.g. a game, discord etc) choose virtual audio cable as your input/microphone
daring sail
Can anyone recommend a website where I can edit photos (remove text from photos)? I tried doing this in nanobanana, but it constantly reduces the photo resolution.
tepid bay
Guys, please tell me what and where to download, an up-to-date voice changer like mmvcserver, because it doesn't run on 50xx+ video cards. Maybe someone has managed to install it on modern hardware?
hardy yew
Guys, please tell me what and where to download, an up-to-date voice changer lik...
Vonovox or tg develop w-okada
-rt
tepid bay
Vonovox or tg develop w-okada
Thank you, I'll try it, I hope it works 😌
hallow thistle
Guys, please tell me what and where to download, an up-to-date voice changer lik...
Vonovox and Tg Develop's W-Okada are the only voice changers that can work with GeForce RTX 50 series.
tepid bay
Vonovox and Tg Develop's W-Okada are the only voice changers that can work with ...
Oh, and I'm wondering why it's not working 
low shard
Oh, and I'm wondering why it's not working <:dog_laugh:1344686486095925278>
don't use original wokada from outdated youtube video tutorials
what's your PC GPU and OS btw?
Vonovox works only on Nvidia GPUs on Windows 10/11
hardy yew
He asked for 50x0 RTX cards specifically
low shard
He asked for 50x0 RTX cards specifically
ops didn't see that, then yeah both programs can work for him
low shard
Can anyone recommend a website where I can edit photos (remove text from photos)...
have you tried using nano banana 2 or nano banana pro? are you trying to remove a text watermark or smt? Because it might not guess exactly what's behind it
low shard
https://huggingface.co/xZAFUEx/ZafueRVCMode...
is there something wrong? it's not a full link
low shard
what voice changers can i use ?
it depends, please elaborate:
- Your PC GPU
- Your PC OS
- What type of software you need for changing voice (like making AI Covers of Goku singing, or do goku roleplay with realtime)
limber grotto
can anyone help me improve the reply speed of my AI? i'm currently using llama qwen3.5:4b as the model
the ai is also using a character model for the voice using applio
the delay for the text reply after i message it comes to be around 20-30 seconds and 5-10 more seconds for the voice
my pc specs:
1x16 gb ddr5 6000mhz
r5 7600x
rx 7800xt
please let me know if i need to provide any other information
low shard
can anyone help me improve the reply speed of my AI? i'm currently using llama q...
Could you provide your PC OS and the LLM programs you're using?
btw not using dual channel ram bandwidth does affect your performance
limber grotto
Could you provide your PC OS and the LLM programs you're using?
btw not using d...
Windows 11
For the LLM it is 'qwen 3.5:4b' (though i might be wrong on question you asked for the LLM programs)
And yes I regret not buying another ram stick before the ai bubble T ^ T
raw dragon
i want to get a ai voice changer i can use for calls
gentle sequoia
What is the best ai to generate promps and image2video?, I want to make images and then image2video, im tryna do 3d cartoon pixar style for youtube kids
low shard
Windows 11
For the LLM it is 'qwen 3.5:4b' (though i might be wrong on question ...
with programs I meant what are you using to run the LLMs, like Ollama, Open WebUI, LM Studio, etc
Also you could try using quantization, like the q8 or q4_k_m guff versions
low shard
i want to get a ai voice changer i can use for calls
hello vox, could you elaborate your PC specs?
!howtoask
low shard
What is the best ai to generate promps and image2video?, I want to make images...
what's your PC Specs?
gentle sequoia
what's your PC Specs?
CPU: AMD Ryzen 7 7445HS (6-core / 12-thread)
RAM: 16 GB
GPU: NVIDIA GeForce RTX 4050 Laptop (6 GB VRAM)
Storage: NVMe SSD
OS: Windows 11 Home (64-bit)
limber grotto
with programs I meant what are you using to run the LLMs, like Ollama, Open WebU...
im using Ollama
and looks like my model is already using q4_k_m as the quantization
low shard
im using Ollama
and looks like my model is already using q4_k_m as the quantiza...
I would suggest closing as many programs as possible in backgrounds, trying maybe a smaller model if your tasks don't require much thinking, disable thinking
you could also try llama.cpp instead of ollama for a slight performance improvement at the cost of user friendliness, honestly I use Ollama too
gentle sequoia
what abt me what do you suggest?
limber grotto
I would suggest closing as many programs as possible in backgrounds, trying mayb...
understood
I'm currently trying to just get small responses and tasks to be done so it might really be better for me to switch to a small model
limber grotto
you could also try llama.cpp instead of ollama for a slight performance improvem...
i'll try that too
Thank you~
pearl lark
using tg's fork when i press the gpu option it only says CPU
is this bc im using a 50 series?
viral mason
If you have a Nvidia at all you downloaded the wrong one most likely
It should display your GPU there
low shard
understood
I'm currently trying to just get small responses and tasks to be done...
is it just for roleplay?
you can limit the model response length in Ollama via:
/set parameter num_predict 100
/set parameter num_ctx 2048
pearl lark
If you have a Nvidia at all you downloaded the wrong one most likely
yeah i downloaded "voice-changer-windows-amd64-cpu.zip"
would it be these 3 instead?
or just the 2 that say cuda
low shard
or just the 2 that say cuda
just those
pearl lark
alr thanks
limber grotto
is it just for roleplay?
kind of?
you can say i'm trying for a base neuro sama without the vtuber function
low shard
using tg's fork when i press the gpu option it only says CPU
is this bc im usin...
did you update your nvidia drivers via the nvidia app and also windows?
viral mason
yeah i downloaded "voice-changer-windows-amd64-cpu.zip"
would it be these 3 inst...
Just 001 and 002
That last one under those is for AMD
low shard
kind of?
you can say i'm trying for a base neuro sama without the vtuber functi...
i don't know much about vtubers but you could limit the model response after running them in ollama with the commands I sent, you could also use a crafted system prompt to tell it to respond in a certain short way since i'm guessing this is AI VTubing
limber grotto
i did try the second thing but it didn't really work in the past but well i can try it again i guess
low shard
understood
I'm currently trying to just get small responses and tasks to be done...
qwen3.5:2b also isn't that bad
it's just for chatting and creativeness after all
limber grotto
hmm right
there are so many models so it gets a little tiring on which one to choose
low shard
hmm right
there are so many models so it gets a little tiring on which one to ch...
qwen3.5 models are the ones I would suggest currently, they are pretty fresh and good
https://huggingface.co/Qwen/Qwen3.5-2B here's the 2b benchmarks
the 9b gets between gpt oss 20b and 120b while the qwen3.5 4b gets near gpt oss 20b
limber grotto
9b is not suited for my work right now, right? if i'm right higher number (2 vs 9) will be slower for small simple answers than the smaller one
low shard
9b is not suited for my work right now, right? if i'm right higher number (2 vs ...
yeah I was just showing you a comparison that if slightly bigger versions of qwen3.5 reach such big models, the lower ones aren't too dumb
in your case you need faster responses than precision
pearl lark
Ran into another issue, not sure if this is normal but the voice sounds very deep pitched at pitch 0
limber grotto
yeah I was just showing you a comparison that if slightly bigger versions of qwe...
oh i see lol
let me check if it gets better after switching model
low shard
Ran into another issue, not sure if this is normal but the voice sounds very dee...
you need to play around with the settings and try different models
want me to check out your settings?
pearl lark
like the model settings?
pearl lark
this is my first time using tg's fork so where do i find the program settings?
low shard
this is my first time using tg's fork so where do i find the program settings?
you can just take a screenshot of the whole program
low shard
you can higher up the pitch value to make it sound more high
you can set your headphones as monitor device to hear urself
you can higher up a bit the extra for a bit extra of quality
low shard
limber grotto
yeah I was just showing you a comparison that if slightly bigger versions of qwe...
qwen doesnt work for me
it is giving json format errors and is taking too much time for a reply even after chaing the model
i switched to "Llama-3.2-3B-Instruct-abliterated-GGUF:Q4_K_M" and it is giving me responses almost instantly so that's great whew
i have a new problem though
whenever the ai speaks it sometimes feels like the voice is cut out for a few milliseconds or you can say a very low pitched glitchy noise
low shard
qwen doesnt work for me
it is giving json format errors and is taking too much t...
that's pretty older but that also makes it faster, be aware that it won't be as good tho
low shard
i have a new problem though
whenever the ai speaks it sometimes feels like the v...
have you tried different voice models?
limber grotto
that's pretty older but that also makes it faster, be aware that it won't be as ...
its not like i have any other choice for now unfortunately
limber grotto
have you tried different voice models?
oh yeah forgot to use other voices XD
low shard
oh yeah forgot to use other voices XD
also be aware that RVC isn't that good when mixed with TTS
limber grotto
also be aware that RVC isn't that good when mixed with TTS
You mean that it performs poorly when the TTS sends the voice to rvc to convert it?
spare dove
I want Google Cloud for Covers ai
spare dove
also be aware that RVC isn't that good when mixed with TTS
I want Google Cloud for Covers ai
spare dove
I want Google Cloud for Covers ai
viral mason
nova harbor
How can I clone myself with AI for video creation online
verbal wren
voices arent working i just hear my voice coming back to me
barren pumice
Hacked by
viral mason
Hacked by<:cat_doom:1420282253459591240>
Evil cat hacked you? 
hardy yew
Evil cat hacked you? <:y_blue_shocked:1159360273904316467>
Gonna spend all the discord vouchers on treats and catmint 
neon grail
im getting shown a black screen
low shard
You mean that it performs poorly when the TTS sends the voice to rvc to convert ...
yeah, as most TTS isn't that expressive
low shard
im getting shown a black screen
!howtoask
low shard
I want Google Cloud for Covers ai
!howtoask
low shard
voices arent working i just hear my voice coming back to me
elaborate:
- your pc gpu
- your pc os
- what are you trying to do
- the tutorial you used
low shard
How can I clone myself with AI for video creation online
!howtoask
hallow thistle
im getting shown a black screen
Did your PC GPU fail? Can you explain about what was the cause of it?
neon grail
Did your PC GPU fail? Can you explain about what was the cause of it?
was just some extremely long boot time
low shard
elaborate:
- your pc gpu
- your pc os
- what are you trying to do
- the tutorial...
@verbal wren I saw you said 5060 then deleted it, it's better you elaborate all info
limber grotto
yeah, as most TTS isn't that expressive
Which one do you suggest?
I'm currently using edge TTS
low shard
I'm currently using edge TTS
edge tts is high quality and multi language, but made for reading than for expressions
I haven't tried recent tts but I heard things like VibeVoice and https://github.com/kyutai-labs/pocket-tts and qwen3-tts
might be better to search for more and see which gives you the best results for the language you need
limber grotto
edge tts is high quality and multi language, but made for reading than for expre...
Understood I will check other tts and report back 🫡
tired star
Can someone tell me why my mic isn't picking anything up when i use W-okada
low shard
Can someone tell me why my mic isn't picking anything up when i use W-okada
please elaborate teto
!howtoask
wild bobcat
hello i have a question related to w-okada
my gpu is a amd radeon rx 7600
is this good enough for real time
?
low shard
is this good enough for real time
hi spamton g spamton, yeah your gpu is good enough, be sure to use wokada tg develop fork and vac lite
wild bobcat
hi spamton g spamton, yeah your gpu is good enough, be sure to use wokada tg dev...
i
would hav eto look into that
but alright thank you
i cant wait to indulge myself into the metaverse
low shard
would hav eto look into that
wild bobcat
https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/
thank you nick
low shard
thank you nick
you're welcome, let me know for any issues
wild bobcat
you look kinda shocked that I know spamton 😭
btw deltarune is goated
oh no i wasnt shocked because you acknowledged my awesome spamton
i forgot that anything rvc is weirdly complicated to setup
lol
wild bobcat
you're welcome, let me know for any issues
will do
limber grotto
@low shard have you used GPT-SoVITS?
low shard
<@911742715019001897> have you used GPT-SoVITS?
I know about it but I haven't used it really
slim frost
AI for webcam?
cerulean rapids
hi guys, i wanted to install the w okada version for amd, i read that it was the dml 2.1 version, but when I try to open the bat program, it just does not download the things needed, I tested with the version of std, and it downloads and works fine, but it uses my cpu, and i wanted to use the dml to use my graphic card for it, is there any solution? i read people had the same problem with the download
viral mason
hi guys, i wanted to install the w okada version for amd, i read that it was the...
you most likely have an outdated version, I'll get you the download for the most up to date version used now for AMD
if u used any yt tutorial it's outdated btw for this kinda stuff
-rt
viral mason
cerulean rapids
you most likely have an outdated version, I'll get you the download for the most...
i used the link for it
viral mason
here ya go
viral mason
i used the link for it
not sure what link you're talking about
viral mason
yea that's the old version of wokada
cerulean rapids
ah that explains it thanks
viral mason
wokada tg fork, and Vonovox (nvidia only) are the only two used now
there is also wokada deiteris fork but it's just a worse version of tg fork
cerulean rapids
finally it started downloading, thanks
viral mason
no problem!
cerulean rapids
wokada tg fork, and Vonovox (nvidia only) are the only two used now
i see
viral mason
btw in case u have VB cable use this one as it's recommended over VB https://software.muzychenko.net/freeware/vac470lite.zip
does the same job just more consistently
cerulean rapids
i was looking for other options and saw one called applio
viral mason
i was looking for other options and saw one called applio
that is for a different thing, Applio is for training ai voice models
cerulean rapids
that is for a different thing, Applio is for training ai voice models
ah so thats for training and okada is to use those trained voices?
viral mason
Applio has had a realtime part to it before but didn't work so I think it got removed
cerulean rapids
i see, seems like i will have to use okada then, thanks
viral mason
np! if u have questions or anything just ask me here ^^
for setup,
Vac lite: run setup64 (not as admin)
Wokada tg fork: run mmvcserversio.exe
it will launch on web browser
cerulean rapids
it will launch on web browser
yea it just downloaded all
and its open now
gonna try one of the models from the server
viral mason
alrighty
quick tip for tg fork, in advanced settings having it like this makes models just sound better, idk why
viral mason
if you're going to use it in games and stuff like that change output to line 1
and input should be fine as long as that is your regular mic
cerulean rapids
if you're going to use it in games and stuff like that change output to line 1
yeah, for that i usually use the scream bee audio or the voice mod micro
that both are virtual
but something strange its that the models i downloaded do not change my voice by a lot xd
does not sound like the trial audios
hardy yew
is good mme?
i'd try WASAPI instead of MME unless there's a reason not to
viral mason
yeah, for that i usually use the scream bee audio or the voice mod micro
not sre what that first one is but it's recommended for line 1 as output, and if u use voicemod put line 1 as your input to add cool effects to the voice changer ect
viral mason
but something strange its that the models i downloaded do not change my voice by...
if they're female voices or just high pitched voices try upping the pitch some
only if ur voice is deep or u are a guy
same for opposite
vast fulcrum
<@&1159293204038955078>
I want to use AI voices, but which program can I use to create these AI voices?
hardy yew
if you mean training the models, then Applio
swift raven
Trying to download b2397 windows cuda version of the tg-develop fork, apparently the zip files are corrupt?
Which one should i download?
hardy yew
Trying to download b2397 windows cuda version of the tg-develop fork, apparently...
corrupt? how come?
did you download both .001 and .002 files?
they're both one zip actually, just split into two files due to size restrictions
you need to download both and then extract the .001 one
it should succeed without an error
in which case it would also mean it used the .002 file too
swift raven
they're both ***one*** zip actually, just split into two files due to size restr...
I know, yes i downloaded both
it says the file is corrupt, downloaded it twice
asked someone else to download it as well, same issue
hardy yew
it was uploaded over 4 months ago, so i doubt there's a chance of corruption, it gets downloaded many times each day
what do you extract it with? what errors do you get?
hardy yew
i just downloaded those files and extracted with no issue
swift raven
weird
hardy yew
those ones, to be exact
swift raven
yes, those ones.
hardy yew
i think there should be no issue with winrar
but
i use 7zip
worth giving a try i guess?
swift raven
well i am downloading vonovox now, hopefully that works
gaunt otter
Can someone give me the link to Vonovox v11
hallow thistle
Can someone give me the link to Vonovox v11
Vonovox 1.1.0? That's an old version. The 1.6.9 is the latest stable Vonovox. https://github.com/dr87/Vonovox/archive/refs/tags/v1.6.9.zip
hallow thistle
Trying to download b2397 windows cuda version of the tg-develop fork, apparently...
swift raven
yes, does it work if you try extracting it? because it didnt for me
hallow thistle
Are you trying to do voice changer as Hitori Gotoh? If you disable "Hide extensions for known file types" in your Windows Explorer, both .zip.001 and .zip.002 files should look like mine. Try 7-Zip to open the .zip.001 one.
swift raven
I see...
well its alright, ive setup vonovox instead and it seems to be working fine
swift raven
Are you trying to do voice changer as Hitori Gotoh? If you disable "Hide extensi...
i am just trying it out, but now that u mention it
might as well try the bocchi one as well
hallow thistle
This is what combined split zip file should look like in both 7-Zip and WinRAR.
swift raven
This is what combined split zip file should look like in both 7-Zip and WinRAR.
yes the file structure was the same as the ss u sent
hallow thistle
But if you see something like this in either archive program, it indicates that the program couldn't find another part of zip (e.g. .zip.002). https://cdn.discordapp.com/attachments/1159290139609137264/1458488426851733765/image.png?ex=69aaf832&is=69a9a6b2&hm=28d179437ce519d11e0f35b20c31e868ffb0879cd8825cde5ff4f7ab88fa97e3&
swift raven
but when i tried extracting it said the file was corrupt
swift raven
But if you see something like this in either archive program, it indicates that ...
ohhhhh
wait, actually yeah i think a file might've been missing
hallow thistle
This is b2397 W-Okada. https://cdn.discordapp.com/attachments/1159290139609137264/1467527432457424969/image.png?ex=69ab8dad&is=69aa3c2d&hm=dcc93072228402da9317ca5804c0ff3bc1b476e623c1b59fa29de57a9bb64274&
hallow thistle
on the audio things
Sample rate: 48000 Hz
Audio driver: WASAPI
Input: (WASAPI) microphone
Output: (WASAPI) Line 1 (Virtual Audio Cable)
Monitor: (WASAPI) speakers
smoky arrow
https://github.com/IllIlIlIllIl/voice-changer/releases/tag/b2335, hello im using this, and i got these error, any idea how to fix or its better to use other voice changer, im on a 5000rtx series
hallow thistle
https://github.com/IllIlIlIllIl/voice-changer/releases/tag/b2335, hello im using...
Use Tg Develop's W-Okada b2397 or Vonovox instead of b2335 W-Okada. These two voice changers have better GeForce RTX 50 series supports.
smoky arrow
where can i find the link ?
hallow thistle
If you have another voice changer instance running in the background either by accident, check if there any and use "Task Manager" to end task. 
hallow thistle
where can i find the link ?
-realtime
smoky arrow
whats the difference between the b2397 and vonovox ?
hallow thistle
This is b2397 W-Okada. https://cdn.discordapp.com/attachments/1159290139609137264/1467527432457424969/image.png?ex=69ab8dad&is=69aa3c2d&hm=dcc93072228402da9317ca5804c0ff3bc1b476e623c1b59fa29de57a9bb64274&
smoky arrow
which one are better ?
viral mason
if you have Nvidia Vonovox is better
smoky arrow
effects are paid ?
hallow thistle
which one are better ?
I'm sorry, but this question is too stressful for me to answer because I've been saying the same answer for many times.
smoky arrow
I'm sorry, but this question is too stressful for me to answer because I've been...
ahh sorry
smoky arrow
I'm sorry, but this question is too stressful for me to answer because I've been...
thanks for the help
viral mason
effects are paid ?
voice effects aren't important and if u neeed them can be worked around with something like voice mod
smoky arrow
voice effects aren't important and if u neeed them can be worked around with som...
oh ok i got a little confuse there my bad
edgy minnow
5090 too much for it lul ¯_(ツ)_/¯
NVIDIA GeForce RTX 5090 with CUDA capability sm_120 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_50 sm_60 sm_61 sm_70 sm_75 sm_80 sm_86 sm_90.
If you want to use the NVIDIA GeForce RTX 5090 GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/
warnings.warn(
2026-03-06 02:30:49,916 INFO [DeviceManager] Switched to 0: NVIDIA GeForce RTX 5090 (CUDA) (cuda:0). FP16 support: True
2026-03-06 02:30:49,918 INFO [IORecorder] -------------------------- - - - C:\Users\3173\AppData\Local\Temp\tmpzy_uq21z\tmp_dir\in.wav, C:\Users\3173\AppData\Local\Temp\tmpzy_uq21z\tmp_dir\out.wav
Traceback (most recent call last):
File "client.py", line 22, in <module>
File "asyncio\runners.py", line 194, in run
File "asyncio\runners.py", line 118, in run
File "asyncio\base_events.py", line 687, in run_until_complete
File "main.py", line 140, in main
File "main.py", line 81, in runServer
File "uvicorn\server.py", line 69, in serve
File "uvicorn\server.py", line 76, in serve
File "uvicorn\config.py", line 434, in load
File "uvicorn\importer.py", line 19, in import_from_string
File "importlib_init.py", line 90, in import_module
File "<frozen importlib._bootstrap>", line 1387, in _gcd_import
File "<frozen importlib._bootstrap>", line 1360, in _find_and_load
File "<frozen importlib._bootstrap>", line 1331, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 935, in _load_unlocked
File "PyInstaller\loader\pyimod02_importers.py", line 384, in exec_module
File "app.py", line 21, in <module>
File "voice_changer\VoiceChangerManager.py", line 66, in init
File "voice_changer\VoiceChangerV2.py", line 38, in init
File "voice_changer\VoiceChangerV2.py", line 102, in _generate_strength
RuntimeError: CUDA error: no kernel image is available for execution on the device
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.
Press Enter to continue...
apparently the 5090 is too good
whatever.
hardy yew
apparently the 5090 is too good
vonovox is compatible with those in general
which version did you download?